Correlation of Molecular Characteristics, Isotype, and In Vitro Functional Activity of Human Antipneumococcal Monoclonal Antibodies
http://www.100md.com
感染与免疫杂志 2006年第2期
Infectious Diseases, Microbiology, and Immunobiology Units, Institute of Child Health, 30 Guilford Street, London WC1N 1EH, United Kingdom
ABSTRACT
Structure-function correlations of pneumococcal antibodies are important in predicting how changes in the pneumococcus (Pnc)-specific B-cell repertoire will influence humoral immunity against invasive Pnc disease. Using a unique panel of human hybridomas derived from memory B cells after pneumococcal conjugate vaccination, we analyzed the structure-function relationship of nine monoclonal antibodies (MAbs) reactive to Pnc polysaccharides. The avidities of the antibodies correlated with the avidity of donor immune serum (R, 0.7; P < 0.025), and this relationship was particularly strong for immunoglobulin A clones (R, 1; P < 0.0005), suggesting that the MAbs may represent important clones contributing to serological memory. Common heavy-light chain combinations and amino acid replacement mutations were seen for clones with the same serospecificity from different individuals. The two highest-avidity MAbs used Vh3-48, and two MAbs with the same serospecificity, using the same V gene pairings (Vh3-7 and Vk2A17), had similar avidities, suggesting that canonical V gene use makes an important contribution to avidity. Although all clones had mutation levels consistent with their being derived from memory B cells, low levels of replacement mutation were associated with high avidities. This relationship was strongest for Vh genes (R, 0.8; P < 0.01). Opsonophagocytosis was demonstrated for all clones, and there was a trend toward clones using canonical genes with low levels of mutation having high opsonophagocytic activities (R, 0.5). These data suggest that the use of canonical genes in the Pnc antibody response is associated with highly functional antibodies and that most somatic mutations seen in these genes are not antigen selected.
INTRODUCTION
Understanding the molecular diversity of the antibody response to polysaccharide antigens has assumed increasing importance with the development of protein-polysaccharide conjugate vaccines that are immunogenic and protect those who are unable to respond to pure polysaccharide vaccines. Many studies have focused on responses to the capsular polysaccharides derived from Haemophilus influenzae type b (Hib), and these studies have demonstrated an oligoclonal antibody (Ab) response with between one and four immunoglobulin (Ig) variable-region (V) genes dominating the Ab repertoire to a specific polysaccharide (2, 25). Similar heavy-chain variable-region genes have been shown to be used by different individuals, and particular V genes appear to be associated with superior functional activity in vitro (24, 26). A number of factors have been shown to influence antigen-specific V gene dominance, including the age of the individual (3), the dose of antigen (18), and the vaccine formulation (24). It is now becoming clear that the choice of V-region gene during a primary immune response may be critical, as this V-region gene use may dominate subsequent memory responses. A murine study by Sanchez etal. demonstrated that the V gene repertoire induced during the initial priming stage of a T-dependent antigen-specific immune response dictated the V gene dominance of the memory response (32).
For some time, it has been suggested that certain V genes present particular conformations (canonical structures) suitable for certain types of antigen (8, 19, 20), although whether these canonical genes in some way reflect a biologically relevant functional advantage has not been clear. Recent data for immunocompromised patients suggest that such canonical genes are functionally important.
Chang and colleagues have shown that individuals with human immunodeficiency virus who are not able to access the optimal canonical genes for specific pathogens produce less effective vaccination responses and are more susceptible to invasive disease (9). With the arrival of new pneumococcal vaccines and their introduction into the infant immunization schedule throughout the world, it is important to understand the factors that dictate Ig V gene use against pneumococcal antigens and the functional consequence of alteration of the antigen-specific V gene repertoire.
We have developed techniques to investigate the genetic structure-function relationship of human anti-pneumococcal polysaccharide (PS) antibodies induced or boosted through immunization. We previously generated a panel of human monoclonal antibodies (MAbs) from healthy adults vaccinated with pneumococcal conjugate vaccines. Mutation analysis of the Ig V(D)J genes used by the MAbs demonstrated that the B cells studied were memory B cells. The isotype distribution of the MAbs suggested that primary antigen encounter occurs viathe mucosa (6). In this study, we describe the in vitro functional activities and avidities of these MAbs and correlate these with MAb immunogenetics.
MATERIALS AND METHODS
Human subjects and vaccination. Four healthy adult volunteers were recruited and given a single dose of Prevenar (a seven-valent polysaccharide conjugate vaccine containing 2 μg each of 4, 9V, 14, 18C, 19F, and 23F and 4 μg of 6B conjugated to a mutant diphtheria toxin [7V-CRM197]). Venous blood (50 ml) was taken from vaccinees prevaccination (day 0) and on day 7 and day 28 postvaccination. The blood was allowed to clot at room temperature, and serum was separated by centrifugation and stored at –80°C. Monoclonal antibodies were derived from the day 7 bleeds as described in detail previously (22).
Monoclonal antibodies. For this study, nine MAbs were selected on the basis of class-switched isotype and high-titer antibody production. All MAbs studied had been shown to be highly serotype specific by a competitive inhibition enzyme-linked immunosorbent assay (ELISA) and did not cross-react with cell wall polysaccharide. Molecular characteristics were defined as described previously (6) and compared with recently described Fab library analyses of serotype 23F- and 6B-reactive clones (36, 37). Homologous gene use and somatic mutations leading to amino acid replacement within the antigen binding sites were identified and described.
Antibody titer determination. Serotype-specific pneumococcal antibody concentrations of MAb and serum samples were determined by ELISA following adsorption with cell wall polysaccharide as previously described (22).
Briefly, Nunc 96-well Maxisorp plates (Life Technologies, Paisley, United Kingdom) were coated with a single serotype antigen at a concentration of 10 μg/ml (obtained from American Type Tissue Collection, Rockville, MD). Plates were washed five times with phosphate-buffered saline-0.05% Tween (PBS-T), and cell supernatants and serum samples preincubated with cell wall polysaccharide (50 μg/ml) and diluted in PBS-T with 1% bovine serum albumin (BSA-PBS-T) were added directly to the coated plates and incubated for 2 h at room temperature (RT). Plates were then washed, and the secondary antibody, horseradish peroxidase-conjugated goat anti-human IgG or IgA (depending on the MAb isotype) (Sigma, Dorset, United Kingdom), was added. Following a further 2-h incubation at RT, wells were rewashed and developed. After 10 to 40 min at RT, 4 M H2SO4 was added, and the absorbance was read at 490 nm. A standard serum, 89SF (kindly supplied by Carl Frasch, CBER), was used as the reference serum. Immunoglobulin concentrations were calculated using ELISA software (Revelations 2.0; Dynatech, United Kingdom) from the standard curve produced by 89SF. Antibody concentrations are expressed in μg/liter.
Avidity. Serotype-specific antibody avidity indexes (AIs) for sera and MAbs were evaluated using a modification of a standard ELISA in which an incubation step using the chaotropic agent ammonium thiocyanate (NH4SCN) was incorporated into the assay. The method has been described in detail previously (16), standardized for Hib and meningococcus C (15, 17), and adapted for pneumococcal antibody avidity determination. In brief, antibodies were diluted in BSA-PBS-T to a concentration known to give an optical density of 1 in the standard pneumococcal ELISA, plated out onto appropriate serotype-coated plates, and incubated for 2 hours. Plates were then washed and incubated with five doubling dilutions in ammonium thiocyanate (NH4SCN), with a starting concentration of 4 M. After 15 min of incubation, plates were washed, incubated for a further 2 h with the secondary antibody (horseradish peroxidase-conjugated goat anti-human polyvalent IgG [Sigma, Dorset, United Kingdom]), and rewashed, and the substrate was then added. After 10 to 40 min, the reaction was stopped with 4 M H2SO4, and the absorbance (read at 490 nm) was determined. All incubations were performed at RT. All studies were performed three times in duplicate with in-house positive (immune serum) and negative (BSA-PBS-T) controls. Results are expressed as AIs, which correspond to the molar concentrations of NH4SCN required to reduce antibody-induced binding (as measured by the absorbance at 490 nm [A490]) by log 50%. Means (± standard errors of the means) of AIs are given for sera and MAbs.
Opsonophagocytic assay. The opsonophagocytic assay used was based on an established assay that has been described in detail previously (21) and is optimized for use with small-volume serum samples for rapid throughput. Briefly, bacterial strains of Streptococcus pneumoniae (ATCC, Rockville, MD) were incubated overnight on blood agar plates at 37°C and 5% CO2 and tested for the presence of polysaccharide capsule by using the quellung reaction (5). Well-encapsulated strains were expanded, grown to log phase, heat killed, and fluorescein isothiocyanate (FITC) labeled prior to storage in 109/ml aliquots at –70°C until use.
Polymorphonuclear leukocytes (PMNs) were isolated by density gradient centrifugation from heparinized blood provided by healthy adult donors. Erythrocytes were lysed, and PMNs were resuspended in RPMI 1680. Cells were counted, adjusted to 5 x 106 cells/ml, and kept on ice prior to use.
MAbs of predefined serotypes and isotype-specific antibody concentrations were serially diluted twofold eight times in Hanks buffered salt solution-1% BSA and divided into aliquots (40 μl in 96-well U-bottomed plates [Nunc]). An equal volume of prediluted FITC-labeled bacteria (approximately 4 x 106) was added to each well, and plates were incubated at 37°C for 30 min on a shaker tray at 100 rpm. Eighty microliters of PMNs (4 x 105) was then added to each well, and the plates were further incubated at 37°C for 15 to 30 min, depending on the serotype, with shaking. Eighty microliters of ice-cold Hanks buffered salt solution-1% BSA was added to each well to stop the reaction, and the plates were spun for 10 min at 2,160 x g. Cells were gently resuspended, fixed, and analyzed by flow cytometry using a FACSCalibur cytometer (Becton Dickinson, California) and Cell Quest software (Becton Dickinson, California).
Neutrophils were gated on forward and side scatter, and the level of opsonophagocytosis was determined by the percentage of cells within the gated population demonstrating a fluorescence level (FL1-H) of >10. Approximately 10,000 cells were counted in each analysis.
All samples were run in duplicate in eight doubling dilutions a minimum of two times. Nonopsonin (neutrophils and bacteria only), non-serotype-specific antibody (10% fetal calf serum or a MAb specific to a different serotype), and cell-only controls were run in each assay. The bacteria-to-PMN ratio used was 10:1, which was selected as the optimal ratio to reduce background activity in serum-negative controls based on previous studies within the laboratory.
Same-serotype MAbs were run in parallel to permit sample comparison. Results are expressed as opsonophagocytic activities (OPAs), the reciprocals of the sample concentrations giving a 50% reduction in opsonic activity.
Correlation coefficients (R) and statistical analysis. A number of correlations were made by using sample means to assess the relationships between the different variables, i.e., MAb AIs and pre- and postvaccination serum AIs, AIs and OPAs, and AIs and ratios of amino acid replacement mutations to silent mutations (R/S ratios) of CDR1 and -2 in heavy and light V genes. Pearson's correlation coefficient was used for AI correlations between samples. Spearman's rank correlation coefficient of tied ranked samples was used for all other correlation analyses. A correlation coefficient (R) of 1 represents 100% correlation of ranking between two variables, an R of 0 represents no correlation, and a negative correlation coefficient represents an inverse correlation. Statistical significance was determined with 95% confidence intervals.
RESULTS
Molecular details. Characteristics of the nine monoclonal antibodies, including serospecificities, isotypes, and molecular characteristics, are shown in Table 1. Amino acid translations for two clonally related isotype switch variants (Db8C11 and Db7D4), a clone sharing isotype specificity and using the same Vh gene (CbB2), and two clones using the unusual Vh3-48 gene (AbA5 and CbE2) are shown in Fig. 1. Amino acid replacements resulting from somatic mutations are identified. IgA MAbs were all in a predominantly dimeric form (data not shown).
Avidity. The avidity index was calculated to provide a measure of the functional affinity of the antigen-antibody interaction, independent of the antibody concentration, in the intact Ig molecule. Serotype-specific antibody AIs for the monoclonal antibodies ranged from 0.125 to 3.9. These AIs were compared to the serotype-specific AIs for serum antibodies (isotype matched to MAbs) from vaccinees preimmunization and at 1 month postimmunization (Table 2 and Fig. 2). The AIs for the MAb panel showed a significant correlation (P < 0.025) with postimmunization serum AIs derived from the same donors matched for isotype and serotype (R, +0.7). Upon subgroup analysis, this was particularly strong for the IgA MAbs, for which the MAb AI was almost identical to the postimmunization serum AI (R, +1; P < 0.0005). There was no significant correlation between the MAb AI and the preimmunization serum AI (R, 0.3) (Table 3).
Despite the IgA clones all being in a predominantly dimeric form with the potential for quadrivalent Fab binding of these MAbs, within serotypes, IgG and IgA isotypes had similar AIs, with the exception of Db8C11 (IgA1). Db8C11—the only IgA1 clone analyzed—had an AI more than fourfold greater than that of its isotypic variant Db7D4 (IgG2), despite having fewer replacement mutations (Fig. 1).
The AI was correlated with overall somatic mutation (percentage of nucleotide homology to the assigned germ line gene) and the ratio of amino acid replacement mutations to silent mutations (R/S ratio) within the CDR1 and -2 domains (Tables 2 and 3). Interestingly, antibodies showing a high degree of germ line homology in the Vh genes showed higher avidities (R, 0.4); however, this was not the case for Vl genes (R, –0.3). The relationship between avidity and conservation of the Vh germ line amino acid sequence was more marked. R/S ratios were calculated based on the number of nucleotide mutations resulting in amino acid replacement mutations relative to that resulting in amino acid conservation (silent mutation) in the CDR1 and -2 regions (8). MAbs using V genes with low R/S ratios were of high avidity relative to more mutated clones (Vh R, –0.8; Vl R, –0.3), and this relationship reached statistical significance for the Vh genes (P < 0.01).
Opsonophagocytosis. All antibodies analyzed induced opsonophagocytosis of serotype-specific pneumococci (Pnc), and opsonophagocytic activity could be titrated against the antibody concentration to give an OPA index (Fig. 3). There was no correlation between AI and OPA either for pooled analysis of all MAbs or isotype-specific analysis (Tables 2 and 3). Five of the clones gave an OPA of 1 to 2, and four of the clones had higher OPAs, suggesting more effective opsonic activities (Fig. 3). The high-avidity IgA1 clone Db8C11 had a relatively low OPA index, in contrast to its IgG2 isotype switch variant Db7D4, which, although of relatively low avidity, had a high OPA index.
When the MAbs were ranked for OPA and the mean R/S ratio within CDR1 and -2 of heavy- and light-chain V genes, there was a trend towards high OPAs and low levels of replacement mutation in heavy- but not light-chain genes (R, –0.5 and –0.1, respectively) with the two Vh3-48-using clones, AbA5 and CbE2, which had high germ line homology and relatively high OPA indices.
DISCUSSION
Through correlating the immunogenetic profiles of Pnc Abs with functional activity, it is anticipated that common molecular patterns of highly functional antibodies may be identified and subsequently targeted to induce optimal protection against invasive Pnc disease.
The monoclonal antibodies we have described permit this correlation, and the overall similarity in avidity between the MAb panel and donor-matched postimmunization sera (R, 0.7; P < 0.025) suggests that most of the clones are derived from dominant clones in the Pnc-specific memory B-cell repertoire. Vh3-48, an unusual gene in the circulating B-cell repertoire, has been shown by us and others to be used by successful clones in the memory response to Pnc PS from a variety of serotypes, including serotypes 9V, 18C, and 23F (6, 27, 36). The high avidities and opsonic activities of the two clones using Vh3-48 that we have described suggest that the use of this gene is associated with very good functional activity and may represent a dominant Pnc clone in the B-cell repertoire in many individuals. The differing serospecificities of antibodies using this gene may be attributed to differences in somatic mutation, CDR3 profiles, and light-chain pairing, which have been demonstrated to alter the fine specificity and affinity of antigen binding to both haptenic and polysaccharide antigens (4, 26, 35). In our MAb panel, these characteristics are similar between clones with the same serospecificity but diverse among non-serotype-specific clones.
The two Pnc serotype 4 IgG MAbs we describe here (Db7D4 and CbB2), derived from different individuals, are the first Pnc serotype 4 MAbs to be described in the literature. They showed similar avidity indices and used identical heavy- and light-chain variable genes (Vh3-7 and Vk2A17), with a common replacement mutation identified in the framework 3 region. Nonhomologous somatic mutation leading to amino acid replacement was seen in all CDR regions, but in view of the similar AIs shown, it may be assumed that the replacement mutations observed were not associated with significant alterations either in avidity or in specificity, and as such, may not have been selected on the basis of improved binding to Pnc polysaccharide.
In contrast, the two IgG serotype 6B clones DM5 and Db3G9 showed very different avidities and were diverse in molecular structure. Db3G9, the higher-avidity clone, showed remarkable molecular similarity both in terms of V gene use and amino acid replacement mutations within the CDRs, with previously reported Fab panels also being reactive to 6B PS (37). The high avidity of this clone suggests that it may also display the molecular characteristics of human antibodies that dominate in the Pnc PS-specific type 6B memory B-cell repertoire. DM5, which had a low avidity, used an unusual Vh1 gene which has previously been described by limited amino sequence analysis of a 6B antibody (33). Unlike most of the MAbs analyzed, the AI of DM5 was far lower than that of postimmunization serum, which suggests that in contrast to the other MAbs analyzed, DM5 may not represent the molecular characteristics of a dominant Pnc PS-specific type 6B-reactive clone in the repertoire of the vaccinee. Repertoire data are not extensive enough to predict whether the low avidity is likely to be due to V gene use/pairings or to the effects of amino acid replacement mutations, whose occurrence is particularly high in this clone.
It is accepted that the selective use of canonical V genes for polysaccharide antigens may determine the protective efficacy of the antibody; however, the contribution of somatic mutation to affinity maturation and antibody function for polysaccharide antibodies has not been investigated in detail. Affinity maturation does occur in the response to pneumococcal conjugate vaccines (14); however, it is not known whether this is simply due to an increasing clonal dominance of B cells expressing canonical genes optimal for polysaccharide binding or whether somatic hypermutation contributes to affinity maturation.
All clones analyzed in this study had a background mutation level consistent with these clones being memory B cells primed in a T-cell-dependent immune response to protein-complexed polysaccharide antigens, either through natural exposure to pneumococcus or through cross-reactive antigens.
The inverse relationship between R/S ratios within CDR1 and -2 and avidity indexes for the Vh genes suggests that these mutations were not selected on the basis of improved antigen binding for the majority of the clones. The amino acid replacement mutations observed in the MAbs may in fact represent mutational hot spots in which the biased mutation of specific nucleotide triplets has occurred in the absence of improved affinity. In this situation, a high level of amino acid replacement may be observed in the absence of antigenic selection (7, 12, 13), and as we have observed, high R/S ratios may not reflect high-avidity binding. Previous observations both for Pnc type 14-specific Fabs and for Hib polysaccharides support this hypothesis, showing that somatic mutation of canonical heavy-chain genes is associated either with a reduction or no change in the affinity of antigen binding (1, 27). There may, however, be occasional replacement mutations that, even in germ line canonical antibodies, result in increased binding affinities. The identification of two Vh replacement mutations within both CDR1 and -2 in the high-avidity MAb CbE2 which result in a change in polarity within the CDR1 has also been shown with a serotype 23F Fab panel. These mutations may be affinity selected. Mutational bias analysis of Vh3-48 gene nucleotide triplets is required to clarify this issue.
Three of the nine MAbs analyzed were of the IgA isotype. The marked differences in AIs between the IgA1 clone Db8C11 and all other Pnc PS serotype 4 clones, including the isotype switch variant Db7D4, which shared an almost identical profile with Db8C11, are interesting. Unlike affinity measurement, the AI reflects the polyvalent binding affinity of the intact Ig molecule. In addition to molecular characteristics at the antigen binding site, AI may be influenced by the Ig valency, isotype, and subclass, possibly through hinge region dynamics. Since all the IgA clones analyzed were in a dimeric form, dimerism alone may account for the differences in avidity between these isotype switch variants. However, two other explanations are also possible. The difference in Fc characteristics between the IgA1 and -G2 clones may influence the affinity of antigen binding. Affinity assessments using IgG, -A, and -M or IgG subclass constructs with constant variable-region domains have demonstrated that the isotype and subclass can influence the affinity of antigen binding which is independent of Vh and Vl domain-mediated activity and involves the Ch1 domain (28-30). The molecular differences in antigen binding sites between these MAbs must also be considered. Db8C11, although clonally related to Db7D4, does exhibit distinct replacement mutations in both the CDR and FWK regions, altering the charge and/or polarity in these regions. Since it is recognized that even a single point mutation within an antigen binding site may alter the affinity of antigen binding (10, 11), these few unshared mutations between these clones alone could account for the differences in avidity that we observed.
Site-directed mutagenesis and Fab analysis of AIs are now required to determine whether the differences in avidity between clones of different isotypes or subclasses are due to nonhomologous somatic mutations resulting in amino acid charge/polarity changes at the antigen binding site (Fig. 1), IgA dimerism, and/or intrinsic differences in antigen-Ab interactions associated with the Fc region.
The opsonic activity of the antibody panel was evaluated by using an opsonophagocytic flow cytometric assay with fresh human neutrophils and FITC-labeled killed pneumococci. Opsonophagocytic activities of sera against live pneumococci have been shown to be good surrogates for protective efficacy in murine models of invasive disease (23), although it not clear whether the small differences in OPAs detected by OPA assays translate into significant differences in protective immunity.
All of the antibodies we analyzed showed opsonic activities toward well-capsulated pneumococci of homologous serotypes; however, there was no correlation between OPAs and AIs or between OPAs and the molecular characteristics of the Abs. This is in contrast to the strong correlation described for serum between AIs and OPAs, as measured by a bacterial killing assay (31, 34), which may simply reflect differences in the assays used.
In conclusion, our data suggest that antibody responses to a number of pneumococcal serotypes are dominated by antibodies expressing high-affinity canonical Vh genes in near-germ line configuration and that somatic mutation of these dominant genes is unlikely to improve the antigen binding affinity. It is important to understand whether new conjugate vaccines used in previously unresponsive populations such as young infants induce antibodies to polysaccharides that use these important genes, particularly since the V regions used in the primary response are likely to dominate lifelong memory responses.
ACKNOWLEDGMENTS
This work was supported by the Wellcome Trust (Clinical Training Fellowship to H.E.B.) and the special trustees of Great Ormond St. NHS Trust.
We thank F. Stevenson for help in developing the human heterohybridoma technology, J. Reinholdt for IgA subclass assignment, and Nomdo Wessterdaal for the human FcR allotyping.
REFERENCES
1. Adderson, E. E., P. G. Shackelford, and W. L. Carroll. 1998. Somatic hypermutation in T-independent and T-dependent immune responses to Haemophilus influenzae type b polysaccharide. Clin. Immunol. Immunopathol. 89:240-246.
2. Adderson, E. E., P. G. Shackelford, A. Quinn, P. M. Wilson, M. W. Cunningham, R. A. Insel, and W. L. Carroll. 1993. Restricted immunoglobulin VH usage and VDJ combinations in the human response to Haemophilus influenzae type b capsular polysaccharide. Nucleotide sequences of monospecific anti-Haemophilus antibodies and polyspecific antibodies cross-reacting with self antigens. J. Clin. Investig. 91:2734-2743.
3. Adderson, E. E., P. M. Wilson, M. W. Cunningham, and P. G. Shackelford. 1998. Haemophilus influenzae type b polysaccharide-protein conjugate vaccine elicits a more diverse antibody repertoire in infants than in adults. J. Immunol. 161:4177-4182.
4. Allen, D., T. Simon, F. Sablitzky, K. Rajewsky, and A. Cumano. 1988. Antibody engineering for the analysis of affinity maturation of an anti-hapten response. EMBO J. 7:1995-2001.
5. Austrian, R. 1976. The quellung reaction, a neglected microbiologic technique. Mt. Sinai J. Med. 43:699-709.
6. Baxendale, H. E., Z. Davis, H. N. White, M. B. Spellerberg, F. K. Stevenson, and D. Goldblatt. 2000. Immunogenetic analysis of the immune response to pneumococcal polysaccharide. Eur. J. Immunol. 30:1214-1223.
7. Bose, B., and S. Sinha. 2005. Problems in using statistical analysis of replacement and silent mutations in antibody genes for determining antigen-driven affinity selection. Immunology 116:172-183.
8. Chang, B., and P. Casali. 1994. The CDR1 sequences of a major proportion of human germline Ig VH genes are inherently susceptible to amino acid replacement. Immunol. Today 15:367-373.
9. Chang, Q., J. Abadi, P. Alpert, and L. Pirofski. 2000. A pneumococcal capsular polysaccharide vaccine induces a repertoire shift with increased VH3 expression in peripheral B cells from human immunodeficiency virus (HIV)-uninfected but not HIV-infected persons. J. Infect. Dis. 181:1313-1321.
10. Chen, C., V. A. Roberts, and M. B. Rittenberg. 1992. Generation and analysis of random point mutations in an antibody CDR2 sequence: many mutated antibodies lose their ability to bind antigen. J. Exp. Med. 176:855-866.
11. Chen, C., V. A. Roberts, S. Stevens, M. Brown, P. M. Stenzel, and M. B. Rittenberg. 1995. Enhancement and destruction of antibody function by somatic mutation: unequal occurrence is controlled by V gene combinatorial associations. EMBO J. 14:2784-2794.
12. Dunn, W. D., A. Dogan, L. Boursier, C. M. MacDonald, and J. Spencer. 1998. Base-specific sequences that bias somatic hypermutation deduced by analysis of out-of-frame human IgVH genes. J. Immunol. 160:2360-2364.
13. Dunn, W. D., and J. Spencer. 1998. Strong intrinsic biases towards mutation and conservation of bases in human IgVH genes during somatic hypermutation prevent statistical analysis of antigen selection. Immunology 95:339-345.
14. Eskola, J. 2000. Immunogenicity of pneumococcal conjugate vaccines. Pediatr. Infect. Dis. J. 19:388-393.
15. Goldblatt, D., R. Borrow, and E. Miller. 2002. Natural and vaccine-induced immunity and immunologic memory to Neisseria meningitidis serogroup C in young adults. J. Infect. Dis. 185:397-400.
16. Goldblatt, D., L. van Etten, F. J. van Milligen, R. C. Aalberse, and M. W. Turner. 1993. The role of pH in modified ELISA procedures used for the estimation of functional antibody affinity. J. Immunol. Methods 166:281-285.
17. Goldblatt, D., A. R. Vaz, and E. Miller. 1998. Antibody avidity as a surrogate marker of successful priming by Haemophilus influenzae type b conjugate vaccines following infant immunization. J. Infect. Dis. 177:1112-1115.
18. Gonzalez-Fernandez, A., and C. Milstein. 1998. Low antigen dose favours selection of somatic mutants with hallmarks of antibody affinity maturation. Immunology 93:149-153.
19. Hougs, L., L. Juul, H. J. Ditzel, C. Heilmann, A. Svejgaard, and T. Barington. 1999. The first dose of a Haemophilus influenzae type b conjugate vaccine reactivates memory B cells: evidence for extensive clonal selection, intraclonal affinity maturation, and multiple isotype switches to IgA2. J. Immunol. 162:224-237.
20. Hougs, L., L. Juul, A. Svejgaard, and T. Barington. 1999. Structural requirements of the major protective antibody to Haemophilus influenzae type b. Infect. Immun. 67:2503-2514.
21. Jansen, W. T., J. Gootjes, M. Zelle, D. V. Madore, J. Verhoef, H. Snippe, and A. F. Verheul. 1998. Use of highly encapsulated Streptococcus pneumoniae strains in a flow-cytometric assay for assessment of the phagocytic capacity of serotype-specific antibodies. Clin. Diagn. Lab. Immunol. 5:703-710.
22. Jessup, C. F., H. Baxendale, D. Goldblatt, and H. Zola. 2000. Preparation of human-mouse heterohybridomas against an immunising antigen. J. Immunol. Methods 246:187-202.
23. Johnson, S. E., L. Rubin, S. S. Romero, J. K. Dykes, L. B. Pais, A. Rizvi, E. Ades, and G. M. Carlone. 1999. Correlation of opsonophagocytosis and passive protection assays using human anticapsular antibodies in an infant mouse model of bacteremia for Streptococcus pneumoniae. J. Infect. Dis. 180:133-140.
24. Lucas, A. H., and D. M. Granoff. 1995. Functional differences in idiotypically defined IgG1 anti-polysaccharide antibodies elicited by vaccination with Haemophilus influenzae type B polysaccharide-protein conjugates. J. Immunol. 154:4195-4202.
25. Lucas, A. H., D. M. Granoff, R. E. Mandrell, C. C. Connolly, A. S. Shan, and D. C. Powers. 1997. Oligoclonality of serum immunoglobulin G antibody responses to Streptococcus pneumoniae capsular polysaccharide serotypes 6B and 23F. Infect. Immun. 65:5103-5109.
26. Lucas, A. H., K. D. Moulton, and D. C. Reason. 1998. Role of kappa II-A2 light chain CDR-3 junctional residues in human antibody binding to the Haemophilus influenzae type b polysaccharide. J. Immunol. 161:3776-3780.
27. Lucas, A. H., K. D. Moulton, V. R. Tang, and D. C. Reason. 2001. Combinatorial library cloning of human antibodies to Streptococcus pneumoniae capsular polysaccharides: variable region primary structures and evidence for somatic mutation of Fab fragments specific for capsular serotypes 6B and 23F. Infect. Immun. 69:853-864.
28. McCloskey, N., M. W. Turner, and T. D. Goldblatt. 1997. Correlation between the avidity of mouse-human chimeric IgG subclass monoclonal antibodies measured by solid-phase elution ELISA and biospecific interaction analysis (BIA). J. Immunol. Methods 205:67-72.
29. McCloskey, N., M. W. Turner, P. Steffner, R. Owens, and D. Goldblatt. 1996. Human constant regions influence the antibody binding characteristics of mouse-human chimeric IgG subclasses. Immunology 88:169-173.
30. Pritsch, O., C. Magnac, G. Dumas, J. P. Bouvet, P. Alzari, and G. Dighiero. 2000. Can isotype switch modulate antigen-binding affinity and influence clonal selection Eur. J. Immunol. 30:3387-3395.
31. Romero, S. S., D. M. Musher, M. S. Cetron, L. B. Pais, J. E. Groover, A. E. Fiore, B. D. Plikaytis, and G. M. Carlone. 1999. Reduction in functional antibody activity against Streptococcus pneumoniae in vaccinated elderly individuals highly correlates with decreased IgG antibody avidity. Clin. Infect. Dis. 29:281-288.
32. Sanchez, M., K. Lindroth, E. Sverremark, F. A. Gonzalez, and C. Fernandez. 2001. The response in old mice: positive and negative immune memory after priming in early age. Int. Immunol. 13:1213-1221.
33. Sun, Y., M. K. Park, J. Kim, B. Diamond, A. Solomon, and M. H. Nahm. 1999. Repertoire of human antibodies against the polysaccharide capsule of Streptococcus pneumoniae serotype 6B. Infect. Immun. 67:1172-1179.
34. Usinger, W. R., and A. H. Lucas. 1999. Avidity as a determinant of the protective efficacy of human antibodies to pneumococcal capsular polysaccharides. Infect. Immun. 67:2366-2370.
35. Xu, J. L., and M. M. Davis. 2000. Diversity in the CDR3 region of V(H) is sufficient for most antibody specificities. Immunity 13:37-45.
36. Zhou, J., K. R. Lottenbach, S. J. Barenkamp, A. H. Lucas, and D. C. Reason. 2002. Recurrent variable region gene usage and somatic mutation in the human antibody response to the capsular polysaccharide of Streptococcus pneumoniae type 23F. Infect. Immun. 70:4083-4091.
37. Zhou, J., K. R. Lottenbach, S. J. Barenkamp, and D. C. Reason. 2004. Somatic hypermutation and diverse immunoglobulin gene usage in the human antibody response to the capsular polysaccharide of Streptococcus pneumoniae type 6B. Infect. Immun. 72:3505-3514.(H. E. Baxendale and D. Go)
ABSTRACT
Structure-function correlations of pneumococcal antibodies are important in predicting how changes in the pneumococcus (Pnc)-specific B-cell repertoire will influence humoral immunity against invasive Pnc disease. Using a unique panel of human hybridomas derived from memory B cells after pneumococcal conjugate vaccination, we analyzed the structure-function relationship of nine monoclonal antibodies (MAbs) reactive to Pnc polysaccharides. The avidities of the antibodies correlated with the avidity of donor immune serum (R, 0.7; P < 0.025), and this relationship was particularly strong for immunoglobulin A clones (R, 1; P < 0.0005), suggesting that the MAbs may represent important clones contributing to serological memory. Common heavy-light chain combinations and amino acid replacement mutations were seen for clones with the same serospecificity from different individuals. The two highest-avidity MAbs used Vh3-48, and two MAbs with the same serospecificity, using the same V gene pairings (Vh3-7 and Vk2A17), had similar avidities, suggesting that canonical V gene use makes an important contribution to avidity. Although all clones had mutation levels consistent with their being derived from memory B cells, low levels of replacement mutation were associated with high avidities. This relationship was strongest for Vh genes (R, 0.8; P < 0.01). Opsonophagocytosis was demonstrated for all clones, and there was a trend toward clones using canonical genes with low levels of mutation having high opsonophagocytic activities (R, 0.5). These data suggest that the use of canonical genes in the Pnc antibody response is associated with highly functional antibodies and that most somatic mutations seen in these genes are not antigen selected.
INTRODUCTION
Understanding the molecular diversity of the antibody response to polysaccharide antigens has assumed increasing importance with the development of protein-polysaccharide conjugate vaccines that are immunogenic and protect those who are unable to respond to pure polysaccharide vaccines. Many studies have focused on responses to the capsular polysaccharides derived from Haemophilus influenzae type b (Hib), and these studies have demonstrated an oligoclonal antibody (Ab) response with between one and four immunoglobulin (Ig) variable-region (V) genes dominating the Ab repertoire to a specific polysaccharide (2, 25). Similar heavy-chain variable-region genes have been shown to be used by different individuals, and particular V genes appear to be associated with superior functional activity in vitro (24, 26). A number of factors have been shown to influence antigen-specific V gene dominance, including the age of the individual (3), the dose of antigen (18), and the vaccine formulation (24). It is now becoming clear that the choice of V-region gene during a primary immune response may be critical, as this V-region gene use may dominate subsequent memory responses. A murine study by Sanchez etal. demonstrated that the V gene repertoire induced during the initial priming stage of a T-dependent antigen-specific immune response dictated the V gene dominance of the memory response (32).
For some time, it has been suggested that certain V genes present particular conformations (canonical structures) suitable for certain types of antigen (8, 19, 20), although whether these canonical genes in some way reflect a biologically relevant functional advantage has not been clear. Recent data for immunocompromised patients suggest that such canonical genes are functionally important.
Chang and colleagues have shown that individuals with human immunodeficiency virus who are not able to access the optimal canonical genes for specific pathogens produce less effective vaccination responses and are more susceptible to invasive disease (9). With the arrival of new pneumococcal vaccines and their introduction into the infant immunization schedule throughout the world, it is important to understand the factors that dictate Ig V gene use against pneumococcal antigens and the functional consequence of alteration of the antigen-specific V gene repertoire.
We have developed techniques to investigate the genetic structure-function relationship of human anti-pneumococcal polysaccharide (PS) antibodies induced or boosted through immunization. We previously generated a panel of human monoclonal antibodies (MAbs) from healthy adults vaccinated with pneumococcal conjugate vaccines. Mutation analysis of the Ig V(D)J genes used by the MAbs demonstrated that the B cells studied were memory B cells. The isotype distribution of the MAbs suggested that primary antigen encounter occurs viathe mucosa (6). In this study, we describe the in vitro functional activities and avidities of these MAbs and correlate these with MAb immunogenetics.
MATERIALS AND METHODS
Human subjects and vaccination. Four healthy adult volunteers were recruited and given a single dose of Prevenar (a seven-valent polysaccharide conjugate vaccine containing 2 μg each of 4, 9V, 14, 18C, 19F, and 23F and 4 μg of 6B conjugated to a mutant diphtheria toxin [7V-CRM197]). Venous blood (50 ml) was taken from vaccinees prevaccination (day 0) and on day 7 and day 28 postvaccination. The blood was allowed to clot at room temperature, and serum was separated by centrifugation and stored at –80°C. Monoclonal antibodies were derived from the day 7 bleeds as described in detail previously (22).
Monoclonal antibodies. For this study, nine MAbs were selected on the basis of class-switched isotype and high-titer antibody production. All MAbs studied had been shown to be highly serotype specific by a competitive inhibition enzyme-linked immunosorbent assay (ELISA) and did not cross-react with cell wall polysaccharide. Molecular characteristics were defined as described previously (6) and compared with recently described Fab library analyses of serotype 23F- and 6B-reactive clones (36, 37). Homologous gene use and somatic mutations leading to amino acid replacement within the antigen binding sites were identified and described.
Antibody titer determination. Serotype-specific pneumococcal antibody concentrations of MAb and serum samples were determined by ELISA following adsorption with cell wall polysaccharide as previously described (22).
Briefly, Nunc 96-well Maxisorp plates (Life Technologies, Paisley, United Kingdom) were coated with a single serotype antigen at a concentration of 10 μg/ml (obtained from American Type Tissue Collection, Rockville, MD). Plates were washed five times with phosphate-buffered saline-0.05% Tween (PBS-T), and cell supernatants and serum samples preincubated with cell wall polysaccharide (50 μg/ml) and diluted in PBS-T with 1% bovine serum albumin (BSA-PBS-T) were added directly to the coated plates and incubated for 2 h at room temperature (RT). Plates were then washed, and the secondary antibody, horseradish peroxidase-conjugated goat anti-human IgG or IgA (depending on the MAb isotype) (Sigma, Dorset, United Kingdom), was added. Following a further 2-h incubation at RT, wells were rewashed and developed. After 10 to 40 min at RT, 4 M H2SO4 was added, and the absorbance was read at 490 nm. A standard serum, 89SF (kindly supplied by Carl Frasch, CBER), was used as the reference serum. Immunoglobulin concentrations were calculated using ELISA software (Revelations 2.0; Dynatech, United Kingdom) from the standard curve produced by 89SF. Antibody concentrations are expressed in μg/liter.
Avidity. Serotype-specific antibody avidity indexes (AIs) for sera and MAbs were evaluated using a modification of a standard ELISA in which an incubation step using the chaotropic agent ammonium thiocyanate (NH4SCN) was incorporated into the assay. The method has been described in detail previously (16), standardized for Hib and meningococcus C (15, 17), and adapted for pneumococcal antibody avidity determination. In brief, antibodies were diluted in BSA-PBS-T to a concentration known to give an optical density of 1 in the standard pneumococcal ELISA, plated out onto appropriate serotype-coated plates, and incubated for 2 hours. Plates were then washed and incubated with five doubling dilutions in ammonium thiocyanate (NH4SCN), with a starting concentration of 4 M. After 15 min of incubation, plates were washed, incubated for a further 2 h with the secondary antibody (horseradish peroxidase-conjugated goat anti-human polyvalent IgG [Sigma, Dorset, United Kingdom]), and rewashed, and the substrate was then added. After 10 to 40 min, the reaction was stopped with 4 M H2SO4, and the absorbance (read at 490 nm) was determined. All incubations were performed at RT. All studies were performed three times in duplicate with in-house positive (immune serum) and negative (BSA-PBS-T) controls. Results are expressed as AIs, which correspond to the molar concentrations of NH4SCN required to reduce antibody-induced binding (as measured by the absorbance at 490 nm [A490]) by log 50%. Means (± standard errors of the means) of AIs are given for sera and MAbs.
Opsonophagocytic assay. The opsonophagocytic assay used was based on an established assay that has been described in detail previously (21) and is optimized for use with small-volume serum samples for rapid throughput. Briefly, bacterial strains of Streptococcus pneumoniae (ATCC, Rockville, MD) were incubated overnight on blood agar plates at 37°C and 5% CO2 and tested for the presence of polysaccharide capsule by using the quellung reaction (5). Well-encapsulated strains were expanded, grown to log phase, heat killed, and fluorescein isothiocyanate (FITC) labeled prior to storage in 109/ml aliquots at –70°C until use.
Polymorphonuclear leukocytes (PMNs) were isolated by density gradient centrifugation from heparinized blood provided by healthy adult donors. Erythrocytes were lysed, and PMNs were resuspended in RPMI 1680. Cells were counted, adjusted to 5 x 106 cells/ml, and kept on ice prior to use.
MAbs of predefined serotypes and isotype-specific antibody concentrations were serially diluted twofold eight times in Hanks buffered salt solution-1% BSA and divided into aliquots (40 μl in 96-well U-bottomed plates [Nunc]). An equal volume of prediluted FITC-labeled bacteria (approximately 4 x 106) was added to each well, and plates were incubated at 37°C for 30 min on a shaker tray at 100 rpm. Eighty microliters of PMNs (4 x 105) was then added to each well, and the plates were further incubated at 37°C for 15 to 30 min, depending on the serotype, with shaking. Eighty microliters of ice-cold Hanks buffered salt solution-1% BSA was added to each well to stop the reaction, and the plates were spun for 10 min at 2,160 x g. Cells were gently resuspended, fixed, and analyzed by flow cytometry using a FACSCalibur cytometer (Becton Dickinson, California) and Cell Quest software (Becton Dickinson, California).
Neutrophils were gated on forward and side scatter, and the level of opsonophagocytosis was determined by the percentage of cells within the gated population demonstrating a fluorescence level (FL1-H) of >10. Approximately 10,000 cells were counted in each analysis.
All samples were run in duplicate in eight doubling dilutions a minimum of two times. Nonopsonin (neutrophils and bacteria only), non-serotype-specific antibody (10% fetal calf serum or a MAb specific to a different serotype), and cell-only controls were run in each assay. The bacteria-to-PMN ratio used was 10:1, which was selected as the optimal ratio to reduce background activity in serum-negative controls based on previous studies within the laboratory.
Same-serotype MAbs were run in parallel to permit sample comparison. Results are expressed as opsonophagocytic activities (OPAs), the reciprocals of the sample concentrations giving a 50% reduction in opsonic activity.
Correlation coefficients (R) and statistical analysis. A number of correlations were made by using sample means to assess the relationships between the different variables, i.e., MAb AIs and pre- and postvaccination serum AIs, AIs and OPAs, and AIs and ratios of amino acid replacement mutations to silent mutations (R/S ratios) of CDR1 and -2 in heavy and light V genes. Pearson's correlation coefficient was used for AI correlations between samples. Spearman's rank correlation coefficient of tied ranked samples was used for all other correlation analyses. A correlation coefficient (R) of 1 represents 100% correlation of ranking between two variables, an R of 0 represents no correlation, and a negative correlation coefficient represents an inverse correlation. Statistical significance was determined with 95% confidence intervals.
RESULTS
Molecular details. Characteristics of the nine monoclonal antibodies, including serospecificities, isotypes, and molecular characteristics, are shown in Table 1. Amino acid translations for two clonally related isotype switch variants (Db8C11 and Db7D4), a clone sharing isotype specificity and using the same Vh gene (CbB2), and two clones using the unusual Vh3-48 gene (AbA5 and CbE2) are shown in Fig. 1. Amino acid replacements resulting from somatic mutations are identified. IgA MAbs were all in a predominantly dimeric form (data not shown).
Avidity. The avidity index was calculated to provide a measure of the functional affinity of the antigen-antibody interaction, independent of the antibody concentration, in the intact Ig molecule. Serotype-specific antibody AIs for the monoclonal antibodies ranged from 0.125 to 3.9. These AIs were compared to the serotype-specific AIs for serum antibodies (isotype matched to MAbs) from vaccinees preimmunization and at 1 month postimmunization (Table 2 and Fig. 2). The AIs for the MAb panel showed a significant correlation (P < 0.025) with postimmunization serum AIs derived from the same donors matched for isotype and serotype (R, +0.7). Upon subgroup analysis, this was particularly strong for the IgA MAbs, for which the MAb AI was almost identical to the postimmunization serum AI (R, +1; P < 0.0005). There was no significant correlation between the MAb AI and the preimmunization serum AI (R, 0.3) (Table 3).
Despite the IgA clones all being in a predominantly dimeric form with the potential for quadrivalent Fab binding of these MAbs, within serotypes, IgG and IgA isotypes had similar AIs, with the exception of Db8C11 (IgA1). Db8C11—the only IgA1 clone analyzed—had an AI more than fourfold greater than that of its isotypic variant Db7D4 (IgG2), despite having fewer replacement mutations (Fig. 1).
The AI was correlated with overall somatic mutation (percentage of nucleotide homology to the assigned germ line gene) and the ratio of amino acid replacement mutations to silent mutations (R/S ratio) within the CDR1 and -2 domains (Tables 2 and 3). Interestingly, antibodies showing a high degree of germ line homology in the Vh genes showed higher avidities (R, 0.4); however, this was not the case for Vl genes (R, –0.3). The relationship between avidity and conservation of the Vh germ line amino acid sequence was more marked. R/S ratios were calculated based on the number of nucleotide mutations resulting in amino acid replacement mutations relative to that resulting in amino acid conservation (silent mutation) in the CDR1 and -2 regions (8). MAbs using V genes with low R/S ratios were of high avidity relative to more mutated clones (Vh R, –0.8; Vl R, –0.3), and this relationship reached statistical significance for the Vh genes (P < 0.01).
Opsonophagocytosis. All antibodies analyzed induced opsonophagocytosis of serotype-specific pneumococci (Pnc), and opsonophagocytic activity could be titrated against the antibody concentration to give an OPA index (Fig. 3). There was no correlation between AI and OPA either for pooled analysis of all MAbs or isotype-specific analysis (Tables 2 and 3). Five of the clones gave an OPA of 1 to 2, and four of the clones had higher OPAs, suggesting more effective opsonic activities (Fig. 3). The high-avidity IgA1 clone Db8C11 had a relatively low OPA index, in contrast to its IgG2 isotype switch variant Db7D4, which, although of relatively low avidity, had a high OPA index.
When the MAbs were ranked for OPA and the mean R/S ratio within CDR1 and -2 of heavy- and light-chain V genes, there was a trend towards high OPAs and low levels of replacement mutation in heavy- but not light-chain genes (R, –0.5 and –0.1, respectively) with the two Vh3-48-using clones, AbA5 and CbE2, which had high germ line homology and relatively high OPA indices.
DISCUSSION
Through correlating the immunogenetic profiles of Pnc Abs with functional activity, it is anticipated that common molecular patterns of highly functional antibodies may be identified and subsequently targeted to induce optimal protection against invasive Pnc disease.
The monoclonal antibodies we have described permit this correlation, and the overall similarity in avidity between the MAb panel and donor-matched postimmunization sera (R, 0.7; P < 0.025) suggests that most of the clones are derived from dominant clones in the Pnc-specific memory B-cell repertoire. Vh3-48, an unusual gene in the circulating B-cell repertoire, has been shown by us and others to be used by successful clones in the memory response to Pnc PS from a variety of serotypes, including serotypes 9V, 18C, and 23F (6, 27, 36). The high avidities and opsonic activities of the two clones using Vh3-48 that we have described suggest that the use of this gene is associated with very good functional activity and may represent a dominant Pnc clone in the B-cell repertoire in many individuals. The differing serospecificities of antibodies using this gene may be attributed to differences in somatic mutation, CDR3 profiles, and light-chain pairing, which have been demonstrated to alter the fine specificity and affinity of antigen binding to both haptenic and polysaccharide antigens (4, 26, 35). In our MAb panel, these characteristics are similar between clones with the same serospecificity but diverse among non-serotype-specific clones.
The two Pnc serotype 4 IgG MAbs we describe here (Db7D4 and CbB2), derived from different individuals, are the first Pnc serotype 4 MAbs to be described in the literature. They showed similar avidity indices and used identical heavy- and light-chain variable genes (Vh3-7 and Vk2A17), with a common replacement mutation identified in the framework 3 region. Nonhomologous somatic mutation leading to amino acid replacement was seen in all CDR regions, but in view of the similar AIs shown, it may be assumed that the replacement mutations observed were not associated with significant alterations either in avidity or in specificity, and as such, may not have been selected on the basis of improved binding to Pnc polysaccharide.
In contrast, the two IgG serotype 6B clones DM5 and Db3G9 showed very different avidities and were diverse in molecular structure. Db3G9, the higher-avidity clone, showed remarkable molecular similarity both in terms of V gene use and amino acid replacement mutations within the CDRs, with previously reported Fab panels also being reactive to 6B PS (37). The high avidity of this clone suggests that it may also display the molecular characteristics of human antibodies that dominate in the Pnc PS-specific type 6B memory B-cell repertoire. DM5, which had a low avidity, used an unusual Vh1 gene which has previously been described by limited amino sequence analysis of a 6B antibody (33). Unlike most of the MAbs analyzed, the AI of DM5 was far lower than that of postimmunization serum, which suggests that in contrast to the other MAbs analyzed, DM5 may not represent the molecular characteristics of a dominant Pnc PS-specific type 6B-reactive clone in the repertoire of the vaccinee. Repertoire data are not extensive enough to predict whether the low avidity is likely to be due to V gene use/pairings or to the effects of amino acid replacement mutations, whose occurrence is particularly high in this clone.
It is accepted that the selective use of canonical V genes for polysaccharide antigens may determine the protective efficacy of the antibody; however, the contribution of somatic mutation to affinity maturation and antibody function for polysaccharide antibodies has not been investigated in detail. Affinity maturation does occur in the response to pneumococcal conjugate vaccines (14); however, it is not known whether this is simply due to an increasing clonal dominance of B cells expressing canonical genes optimal for polysaccharide binding or whether somatic hypermutation contributes to affinity maturation.
All clones analyzed in this study had a background mutation level consistent with these clones being memory B cells primed in a T-cell-dependent immune response to protein-complexed polysaccharide antigens, either through natural exposure to pneumococcus or through cross-reactive antigens.
The inverse relationship between R/S ratios within CDR1 and -2 and avidity indexes for the Vh genes suggests that these mutations were not selected on the basis of improved antigen binding for the majority of the clones. The amino acid replacement mutations observed in the MAbs may in fact represent mutational hot spots in which the biased mutation of specific nucleotide triplets has occurred in the absence of improved affinity. In this situation, a high level of amino acid replacement may be observed in the absence of antigenic selection (7, 12, 13), and as we have observed, high R/S ratios may not reflect high-avidity binding. Previous observations both for Pnc type 14-specific Fabs and for Hib polysaccharides support this hypothesis, showing that somatic mutation of canonical heavy-chain genes is associated either with a reduction or no change in the affinity of antigen binding (1, 27). There may, however, be occasional replacement mutations that, even in germ line canonical antibodies, result in increased binding affinities. The identification of two Vh replacement mutations within both CDR1 and -2 in the high-avidity MAb CbE2 which result in a change in polarity within the CDR1 has also been shown with a serotype 23F Fab panel. These mutations may be affinity selected. Mutational bias analysis of Vh3-48 gene nucleotide triplets is required to clarify this issue.
Three of the nine MAbs analyzed were of the IgA isotype. The marked differences in AIs between the IgA1 clone Db8C11 and all other Pnc PS serotype 4 clones, including the isotype switch variant Db7D4, which shared an almost identical profile with Db8C11, are interesting. Unlike affinity measurement, the AI reflects the polyvalent binding affinity of the intact Ig molecule. In addition to molecular characteristics at the antigen binding site, AI may be influenced by the Ig valency, isotype, and subclass, possibly through hinge region dynamics. Since all the IgA clones analyzed were in a dimeric form, dimerism alone may account for the differences in avidity between these isotype switch variants. However, two other explanations are also possible. The difference in Fc characteristics between the IgA1 and -G2 clones may influence the affinity of antigen binding. Affinity assessments using IgG, -A, and -M or IgG subclass constructs with constant variable-region domains have demonstrated that the isotype and subclass can influence the affinity of antigen binding which is independent of Vh and Vl domain-mediated activity and involves the Ch1 domain (28-30). The molecular differences in antigen binding sites between these MAbs must also be considered. Db8C11, although clonally related to Db7D4, does exhibit distinct replacement mutations in both the CDR and FWK regions, altering the charge and/or polarity in these regions. Since it is recognized that even a single point mutation within an antigen binding site may alter the affinity of antigen binding (10, 11), these few unshared mutations between these clones alone could account for the differences in avidity that we observed.
Site-directed mutagenesis and Fab analysis of AIs are now required to determine whether the differences in avidity between clones of different isotypes or subclasses are due to nonhomologous somatic mutations resulting in amino acid charge/polarity changes at the antigen binding site (Fig. 1), IgA dimerism, and/or intrinsic differences in antigen-Ab interactions associated with the Fc region.
The opsonic activity of the antibody panel was evaluated by using an opsonophagocytic flow cytometric assay with fresh human neutrophils and FITC-labeled killed pneumococci. Opsonophagocytic activities of sera against live pneumococci have been shown to be good surrogates for protective efficacy in murine models of invasive disease (23), although it not clear whether the small differences in OPAs detected by OPA assays translate into significant differences in protective immunity.
All of the antibodies we analyzed showed opsonic activities toward well-capsulated pneumococci of homologous serotypes; however, there was no correlation between OPAs and AIs or between OPAs and the molecular characteristics of the Abs. This is in contrast to the strong correlation described for serum between AIs and OPAs, as measured by a bacterial killing assay (31, 34), which may simply reflect differences in the assays used.
In conclusion, our data suggest that antibody responses to a number of pneumococcal serotypes are dominated by antibodies expressing high-affinity canonical Vh genes in near-germ line configuration and that somatic mutation of these dominant genes is unlikely to improve the antigen binding affinity. It is important to understand whether new conjugate vaccines used in previously unresponsive populations such as young infants induce antibodies to polysaccharides that use these important genes, particularly since the V regions used in the primary response are likely to dominate lifelong memory responses.
ACKNOWLEDGMENTS
This work was supported by the Wellcome Trust (Clinical Training Fellowship to H.E.B.) and the special trustees of Great Ormond St. NHS Trust.
We thank F. Stevenson for help in developing the human heterohybridoma technology, J. Reinholdt for IgA subclass assignment, and Nomdo Wessterdaal for the human FcR allotyping.
REFERENCES
1. Adderson, E. E., P. G. Shackelford, and W. L. Carroll. 1998. Somatic hypermutation in T-independent and T-dependent immune responses to Haemophilus influenzae type b polysaccharide. Clin. Immunol. Immunopathol. 89:240-246.
2. Adderson, E. E., P. G. Shackelford, A. Quinn, P. M. Wilson, M. W. Cunningham, R. A. Insel, and W. L. Carroll. 1993. Restricted immunoglobulin VH usage and VDJ combinations in the human response to Haemophilus influenzae type b capsular polysaccharide. Nucleotide sequences of monospecific anti-Haemophilus antibodies and polyspecific antibodies cross-reacting with self antigens. J. Clin. Investig. 91:2734-2743.
3. Adderson, E. E., P. M. Wilson, M. W. Cunningham, and P. G. Shackelford. 1998. Haemophilus influenzae type b polysaccharide-protein conjugate vaccine elicits a more diverse antibody repertoire in infants than in adults. J. Immunol. 161:4177-4182.
4. Allen, D., T. Simon, F. Sablitzky, K. Rajewsky, and A. Cumano. 1988. Antibody engineering for the analysis of affinity maturation of an anti-hapten response. EMBO J. 7:1995-2001.
5. Austrian, R. 1976. The quellung reaction, a neglected microbiologic technique. Mt. Sinai J. Med. 43:699-709.
6. Baxendale, H. E., Z. Davis, H. N. White, M. B. Spellerberg, F. K. Stevenson, and D. Goldblatt. 2000. Immunogenetic analysis of the immune response to pneumococcal polysaccharide. Eur. J. Immunol. 30:1214-1223.
7. Bose, B., and S. Sinha. 2005. Problems in using statistical analysis of replacement and silent mutations in antibody genes for determining antigen-driven affinity selection. Immunology 116:172-183.
8. Chang, B., and P. Casali. 1994. The CDR1 sequences of a major proportion of human germline Ig VH genes are inherently susceptible to amino acid replacement. Immunol. Today 15:367-373.
9. Chang, Q., J. Abadi, P. Alpert, and L. Pirofski. 2000. A pneumococcal capsular polysaccharide vaccine induces a repertoire shift with increased VH3 expression in peripheral B cells from human immunodeficiency virus (HIV)-uninfected but not HIV-infected persons. J. Infect. Dis. 181:1313-1321.
10. Chen, C., V. A. Roberts, and M. B. Rittenberg. 1992. Generation and analysis of random point mutations in an antibody CDR2 sequence: many mutated antibodies lose their ability to bind antigen. J. Exp. Med. 176:855-866.
11. Chen, C., V. A. Roberts, S. Stevens, M. Brown, P. M. Stenzel, and M. B. Rittenberg. 1995. Enhancement and destruction of antibody function by somatic mutation: unequal occurrence is controlled by V gene combinatorial associations. EMBO J. 14:2784-2794.
12. Dunn, W. D., A. Dogan, L. Boursier, C. M. MacDonald, and J. Spencer. 1998. Base-specific sequences that bias somatic hypermutation deduced by analysis of out-of-frame human IgVH genes. J. Immunol. 160:2360-2364.
13. Dunn, W. D., and J. Spencer. 1998. Strong intrinsic biases towards mutation and conservation of bases in human IgVH genes during somatic hypermutation prevent statistical analysis of antigen selection. Immunology 95:339-345.
14. Eskola, J. 2000. Immunogenicity of pneumococcal conjugate vaccines. Pediatr. Infect. Dis. J. 19:388-393.
15. Goldblatt, D., R. Borrow, and E. Miller. 2002. Natural and vaccine-induced immunity and immunologic memory to Neisseria meningitidis serogroup C in young adults. J. Infect. Dis. 185:397-400.
16. Goldblatt, D., L. van Etten, F. J. van Milligen, R. C. Aalberse, and M. W. Turner. 1993. The role of pH in modified ELISA procedures used for the estimation of functional antibody affinity. J. Immunol. Methods 166:281-285.
17. Goldblatt, D., A. R. Vaz, and E. Miller. 1998. Antibody avidity as a surrogate marker of successful priming by Haemophilus influenzae type b conjugate vaccines following infant immunization. J. Infect. Dis. 177:1112-1115.
18. Gonzalez-Fernandez, A., and C. Milstein. 1998. Low antigen dose favours selection of somatic mutants with hallmarks of antibody affinity maturation. Immunology 93:149-153.
19. Hougs, L., L. Juul, H. J. Ditzel, C. Heilmann, A. Svejgaard, and T. Barington. 1999. The first dose of a Haemophilus influenzae type b conjugate vaccine reactivates memory B cells: evidence for extensive clonal selection, intraclonal affinity maturation, and multiple isotype switches to IgA2. J. Immunol. 162:224-237.
20. Hougs, L., L. Juul, A. Svejgaard, and T. Barington. 1999. Structural requirements of the major protective antibody to Haemophilus influenzae type b. Infect. Immun. 67:2503-2514.
21. Jansen, W. T., J. Gootjes, M. Zelle, D. V. Madore, J. Verhoef, H. Snippe, and A. F. Verheul. 1998. Use of highly encapsulated Streptococcus pneumoniae strains in a flow-cytometric assay for assessment of the phagocytic capacity of serotype-specific antibodies. Clin. Diagn. Lab. Immunol. 5:703-710.
22. Jessup, C. F., H. Baxendale, D. Goldblatt, and H. Zola. 2000. Preparation of human-mouse heterohybridomas against an immunising antigen. J. Immunol. Methods 246:187-202.
23. Johnson, S. E., L. Rubin, S. S. Romero, J. K. Dykes, L. B. Pais, A. Rizvi, E. Ades, and G. M. Carlone. 1999. Correlation of opsonophagocytosis and passive protection assays using human anticapsular antibodies in an infant mouse model of bacteremia for Streptococcus pneumoniae. J. Infect. Dis. 180:133-140.
24. Lucas, A. H., and D. M. Granoff. 1995. Functional differences in idiotypically defined IgG1 anti-polysaccharide antibodies elicited by vaccination with Haemophilus influenzae type B polysaccharide-protein conjugates. J. Immunol. 154:4195-4202.
25. Lucas, A. H., D. M. Granoff, R. E. Mandrell, C. C. Connolly, A. S. Shan, and D. C. Powers. 1997. Oligoclonality of serum immunoglobulin G antibody responses to Streptococcus pneumoniae capsular polysaccharide serotypes 6B and 23F. Infect. Immun. 65:5103-5109.
26. Lucas, A. H., K. D. Moulton, and D. C. Reason. 1998. Role of kappa II-A2 light chain CDR-3 junctional residues in human antibody binding to the Haemophilus influenzae type b polysaccharide. J. Immunol. 161:3776-3780.
27. Lucas, A. H., K. D. Moulton, V. R. Tang, and D. C. Reason. 2001. Combinatorial library cloning of human antibodies to Streptococcus pneumoniae capsular polysaccharides: variable region primary structures and evidence for somatic mutation of Fab fragments specific for capsular serotypes 6B and 23F. Infect. Immun. 69:853-864.
28. McCloskey, N., M. W. Turner, and T. D. Goldblatt. 1997. Correlation between the avidity of mouse-human chimeric IgG subclass monoclonal antibodies measured by solid-phase elution ELISA and biospecific interaction analysis (BIA). J. Immunol. Methods 205:67-72.
29. McCloskey, N., M. W. Turner, P. Steffner, R. Owens, and D. Goldblatt. 1996. Human constant regions influence the antibody binding characteristics of mouse-human chimeric IgG subclasses. Immunology 88:169-173.
30. Pritsch, O., C. Magnac, G. Dumas, J. P. Bouvet, P. Alzari, and G. Dighiero. 2000. Can isotype switch modulate antigen-binding affinity and influence clonal selection Eur. J. Immunol. 30:3387-3395.
31. Romero, S. S., D. M. Musher, M. S. Cetron, L. B. Pais, J. E. Groover, A. E. Fiore, B. D. Plikaytis, and G. M. Carlone. 1999. Reduction in functional antibody activity against Streptococcus pneumoniae in vaccinated elderly individuals highly correlates with decreased IgG antibody avidity. Clin. Infect. Dis. 29:281-288.
32. Sanchez, M., K. Lindroth, E. Sverremark, F. A. Gonzalez, and C. Fernandez. 2001. The response in old mice: positive and negative immune memory after priming in early age. Int. Immunol. 13:1213-1221.
33. Sun, Y., M. K. Park, J. Kim, B. Diamond, A. Solomon, and M. H. Nahm. 1999. Repertoire of human antibodies against the polysaccharide capsule of Streptococcus pneumoniae serotype 6B. Infect. Immun. 67:1172-1179.
34. Usinger, W. R., and A. H. Lucas. 1999. Avidity as a determinant of the protective efficacy of human antibodies to pneumococcal capsular polysaccharides. Infect. Immun. 67:2366-2370.
35. Xu, J. L., and M. M. Davis. 2000. Diversity in the CDR3 region of V(H) is sufficient for most antibody specificities. Immunity 13:37-45.
36. Zhou, J., K. R. Lottenbach, S. J. Barenkamp, A. H. Lucas, and D. C. Reason. 2002. Recurrent variable region gene usage and somatic mutation in the human antibody response to the capsular polysaccharide of Streptococcus pneumoniae type 23F. Infect. Immun. 70:4083-4091.
37. Zhou, J., K. R. Lottenbach, S. J. Barenkamp, and D. C. Reason. 2004. Somatic hypermutation and diverse immunoglobulin gene usage in the human antibody response to the capsular polysaccharide of Streptococcus pneumoniae type 6B. Infect. Immun. 72:3505-3514.(H. E. Baxendale and D. Go)