Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
Intern. Rev. Immunol. Vol. 9, 1992, pp. 25-43 Reprints available directly from the publisher Photocopying permitted by license only @I992Harwood Academic Publishers GmbH Printed in the United States of America
The Repertoire of Human Antibody to the Haemophilus influenzae mpe b Capsular Polysaccharide RICHARD A. INSEL,+ ELISABETH E. ADDERSON,$ and WILLIAM L. CARROLL* +Department of Pediatrics and Microbiology, University of Rochester School of Medicine and Dentistry, Rochester, New York 14642, USA *Program in Human Molecular Biology and Genetics, and Department of Pediatrics, University of Utah, Salt Lake City, Utah 84112, USA
Human antibody to the Haernophilus influenzne capsular polysaccharide (Hib CP) is restricted in diversity in the individual and the population with a limited number of variable region genes encoding antibody. Antibody to the Hib CP shows restricted isoelectric focusing gel patterns and light chain usage with frequent restriction to use of only kappa light chains. Shared cross-reactive idiotypes are expressed on antibody. The heavy chain of antibody to the Hib CP is predominantly encoded by two members of the V,, family-LSG 6.1/M8S-like and VH,,/30P1-like. In V, the CDRI, based on complete identity in LSG 6.liM85-like antibodies, CDR2, based on the suggestion of mutation in this region, and CDR3, based on conserved CDR3 usage in unrelated individuals, may be important for antigen binding. Six or more different V, gene families encode antibody. The predominant antibody of the majority of individuals uses the AZ-VKII gene in germline or near germline configuraion, which encodes an idiotype designated HibId-1. Antibody can also be encoded by VKI,non-A2 VKII,VKIII,VKIV,VXII, and VAVII genes. Although different V, genes can be used, unrelated individuals appear to use the same VKIII(A27), VAII (Vh2.1 and VAVII (4A) genes. The V, diversity accounts for differences in fine binding specificity, with AZ-VKII genes not encoding E . coli KlOO CP cross-reactive antibodies and VAVII genes and some of the non-A2 VK genes encoding cross-reactive antibodies. The arginine in CDR3 of both antibody kappa and lambda light chains and the asparagine in CDR2 of V, sequences and in CDRI of LSG6.1-M85 V, sequences of antibody appear to be important residues for antigen binding. A relatively limited degree of somatic mutation has occurred in the non-A2 V, genes, VAVII, and the V, genes. Further studies comparing the polymorphism of germline V genes to antibody-encoding V genes are needed to clarify this issue. Research comparing this repertoire to repertoires directed to other bacterial CP and to self antigens and defining structure-antigen binding relationships is in progress. KEYWORDS: antibody, variable region, V , genes, V , genes
1. INTRODUCTION It has long been appreciated that the antibody responses in animals to polysaccharides can be restricted and shared with evidence of clonal restriction and dominance, cross-reactive idiotypes, shared V gene sequences, and shared fine-binding specificity [I, 21. Polysaccharides lend themselves to studies of antibody repertoires because they can provide single antigenic determinants that completely fill the antibody combining site. They have an advantage over more structurally complex native or even artificial antigens and over haptens which often fail to engage the combining site completely, resulting in specificity to 25
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
26
R. A . INSEL er
CJ~.
haptens being dictated by both the hapten and a portion of the protein complexed to the hapten. It should be noted, however, that while polysaccharides are restricted in overall complexity compared to a protein antigen, they have potentially multiple unique epitopes that can be recognized by antibodies. Kabat and colleagues have shown using immunochemical approaches and anti-idiotypic antibodies that antibodies to polysaccharides can be directed, or complementary, to either non-reducing termini or to internal core regions of a polysaccharide and these specificities are reflected in structures of the combining sites or complementarity determining regions (CDR) [3]. The antibodies binding to internal linear determinants displayed “groove-type’’ antibody-binding sites while those antibodies with specificities for non-reducing ends of the polysaccharide had “cavity-type’’ antibody binding sites [4]. Thus, even antibody responses to relatively simple antigens can be variable and complex. The diversity of human antibody repertoires to defined antigens has been relatively poorly characterized in comparison to the mouse. One exception has been the human antibody response to the capsular polysaccharide (CP) of Haemophilus injuenzae type b. In this review we will address our current understanding of the diversity of the human antibody response to the Haemophilus injuenzae type b (Hib) capsular polysaccharide.
2. RATIONALE FOR CHARACTERIZING THE DIVERSITY OF HAEMOPHlLUS INFLUENZAE B CAPSULAR POLYSACCHARIDE ANTIBODY We began studies on the diversity of human antibody to this antigen because of the extensive work in inbred strains of mice demonstrating restricted diversity of the variable regions of antibodies to polysaccharides and because studying the diversity of the antibody response to the Haemophilus injuenzae C P had several merits. First, this antibody is the major form of protection against a significant human pathogen. Haemophilus injuenzae b, a gram negative bacteria, is the major etiologic agent of bacterial meningitis and sepsis in the United States with a peak age-incidence of infection at 6-12 months of age. The principal virulence determinant on the bacteria is its CP, which is composed of a repetitive polymer of ribosyl-ribitol-phosphate in which the two sugars are in a 1-1 linkage [ 5 ] . Antibody to the type b capsular polysaccharide is the major protective antibody in man against infection and mediates its action through its opsonic properties directly and indirectly through activation of the complement pathways. Infection is infrequent in the first few months of life because of transplacental passage of maternal antibody to the type b CP and in later childhood following the active acquisition of antibody through exposure either to the type b polysaccharide or to cross-reactive antigens on other microorganisms [6]. One such microorganism is Escherichia coli KlOO that has a CP structure identical to the Hib CP except it uses a (1-2) rather than a (1-1) ribosyl-ribitol linkage [7]. We have found that the naturally occurring antibody in children is more commonly cross-reactive with KlOO than that of the adult [8].This presumably reflects increased exposure of older individuals to Hib or other cross-reactive antigens that more precisely express epitopes found on the Hib CF? Supporting this idea is the finding that immunization with the Hib CP decreases the proportion of KlOO cross-reactive antibody compared to preimmunization antibody [8]. These findings indicated that more than one type of variable region gene (K100 crossreactive and non cross-reactive) may be encoding the response. Conceivably, study of
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
REPERTOIRE OF HUMAN ANTI-HIB ANTIBODY
27
naturally occurring and vaccine-induced antibody may demonstrate immunoglobulin variable region (V) genes that have been conserved in the human population to allow response to this virulent pathogen. A second advantage of the use of the Hib CP is that it has been possible to actively immunize humans with the CP to study high-titered antibody responses. A number of Hib CP vaccines were developed in the last two decades to prevent this serious childhood infection [9, 101. Initially, vaccines composed of the purified type b C P were developed and then licensed for children at ages 18-24 months in the United States in 1985. The efficacy of the type b CP vaccines was limited because infants younger than 18 months of age, the age associated with the highest incidence of infection, failed to generate an antibody response and infants between 18-30 months had inconsistent responses. At age 4-5 years, children responded but with low magnitude responses compared to adults. This delay in response to this CP is not unique but is observed with several other bacterial CP [ 111. The development of the antibody repertoire with age could also be studied using the Hib C P antibody model. The unresponsiveness of young infants to the CP, however, is not fixed. It has been found that even young infants respond to the type b C P antigen if the CP is conjugated to a protein carrier [12, 131. Second generation vaccines were developed in which the type b C P or oligosaccharides prepared from the polysaccharides were covalently conjugated to mutant diphtheria toxins, tetanus toxoid, or meningococcal outer membrane proteins [6, 10, 11, 141. These so-called “conjugate vaccines” have the ability to recruit CD4+ T-lymphocytes to provide cooperative help to CP-specific B cells. In addition, conjugate vaccines have the ability to prime the infant to generate an antibody response to immunization with the unconjugated C P [12, 131. Further, conjugate vaccines can bypass poor antibody responses to the C P in some patients with primary immunodeficiency disease (e.g., IgG subclass deficiency) [15]. Conjugate vaccines induce protection against Hib infection in the infant, and were licensed in the United States in 1989 for use in infants beginning at two months of age. Development of these vaccines has allowed comparison of the antibody and antibody repertoires induced by unconjugated and protein-conjugated forms of the Cl? Lastly, certain subpopulations have a higher incidence of systemic Hib infection [16]. Immunoglobulin allotype markers have been associated with an increased risk of invasive Hib disease and with preferential IgG subclass and light chain type antibody responses [17, 181. The possibility that the high risk in such individuals is due to so-called holes or gaps in the immunoglobulin gene repertoire could be explored by analysis of the Hib C P antibody repertoire.
3. CLONOTYPE, LIGHT CHAIN DIVERSITY, AND IDIOTYPY OF ANTIBODY The first evidence for restriction of the human anti-Hib C P antibody response was from the demonstration of restricted numbers of antibody clones by isoelectric focusing [ 19, 201, the restricted light chain distribution of the antibody [21, 221, and the presence of shared crossreactive idiotypes in serum of unrelated subjects [23-251. Some diversity of the response was readily apparent, however, in both individuals and the general population from the finding of expression of both kappa and lambda light chain types [21,22], and E . coli crossreactive and non cross-reactive antibody in the antibody response [7, 81. Analytical isoelectric focusing (IEF) of postimmunization IgG antibody showed that an
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
28
R. A . INSEL et al.
individual serum displayed a restricted repertoire with only 1 to 4 clones, or so-called clonotypes or spectrotypes, on a gel [19]. This type of analysis, which detects charge differences among antibodies, is affected by both the variable and constant region contribution to overall charge as well as post-translational glycosylation differences on immunoglobulins. Most adults immunized with the polysaccharide or protein-conjugated polysaccharide vaccine generate both IgGl and IgG2 subclass antibody responses [26-281, and this heavy chain constant region difference leads to charge differences and different IEF patterns, even when the same V, gene is used. Thus, the finding of distinct clonal patterns suggested a marked degree of antibody restriction. It was also found, somewhat surprisingly, that antibody induced by the protein-conjugated Hib CP vaccines was as highly restricted as that induced by the Hib CP vaccine. Infants immunized with protein-conjugated Hib CP vaccines generated pauciclonal responses 1131. Repetitive immunization with these so-called conjugate vaccines in infancy induces booster serum antibody responses. Based on IEF patterns, these increased levels of serum antibody occur primarily from pre-existing clones activated by earlier immunizations, thus preserving clonal restriction. As noted above, the conjugate vaccines can accelerate ontogeny of responses to the polysaccharide vaccine so that conjugate-primed young infants can respond to the C P vaccine. Booster immunization with Hib CP vaccine in conjugate-primed infants also tends to restimulate pre-existent clones without new clonal recruitment, with some preferential activation of IgG2 vs IgGl expressing clones in some individuals [ 131. Two other aspects of the isoelectric focusing patterns of antibody proved of interest. First, while most individuals have unique clonotype patterns, some clones appeared to be shared between individuals, suggesting use of similar or highly related V genes [19, 201. Lastly, the human antibody response to CP is known to be a long-lived response and clonal dominance was sustained over time. Clones that dominated the response at one month remained dominant several years later [19]. The normal distribution of kappa and lambda light chain types in serum immunoglobulin is approximately 60% kappa and 40% lambda light chain type and most antibody responses conform to this distribution. The light chain distribution of antibody induced by either the CP or protein-conjugated CP was predominantly or, in some individuals, exclusively, kappa light chain type [21, 221. Some individuals do generate, however, a lambda type predominant antibody response, and we know today, as discussed below, that lambda-expressing antibodies can, not uncommonly, have some unique fine-binding specificities. The finding of kappa light chain restricted, or in rare individuals lambda light chain restricted antibody, also suggested a relatively limited repertoire. Shared cross-reactive idiotypes were demonstrated by Lucas and colleagues on naturallyoccurring and postimmunization antibody using polyclonal and monoclonal anti-idiotypic antibodies that appeared to bind in or close to the antibody combining site [23-251. One monoclonal antibody, which detects an idiotype designated HibId-1, whose structural equivalent is detailed below, reacts specifically with the Hib C P or conjugate postimmunization antibody of approximately 85% of individuals [25]. In these individuals, it detects 60% of the total antibody induced by immunization, including both IgGl and IgG2 antibodies. Immunization increases the proportion of Hib Id-1 antibody, which is in contrast to the diminution in KlOO cross-reactive antibody, described above. These findings demonstrate that the majority of individuals and the majority of anti-Hib CP antibodies share a unique cross-reactive idiotype, also indicating conservation of the repertoire.
REPERTOIRE OF HUMAN ANTI-HIB ANTIBODY
29
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
4. RESTRICTION OF ANTIBODY HEAVY CHAIN VARIABLE REGIONS (V,) Initial evidence of V, region restriction of anti-Hib CP antibody came from studies by Scott and colleagues who found the IgG antibody of adults induced by the Hib CP to be encoded exclusively by the v H 3 family, based on N-terminal amino acid sequences of immunoaffinity-purified serum antibody [29]. Predominant usage of v H 3 genes was confirmed by several other approaches. Using VH3 anti-peptide antibodies directed to framework and hypervariable regions of V, families and elements, Silverman and colleagues demonstrated the use of v H 3 in the IgG, IgA, and IgM antibody induced by CP immunization in all subjects studied [30]. In addition, their studies suggested dominant use of two germline V,, elements-9.1/20Pl-and VH26/30P1-associated subfamilies-in the antibody response. Some individuals appeared to use both v H 3 gene members. The postimmunization IgG antibody of some individuals was also detected with anti-VH1and V,, family-specific antibodies. Definitive proof of use of V,, and identification of the specific V,, gene elements arose from nucleotide sequencing of the immunoglobulin heavy chain genes of human antibodysecreting heterohybridomas generated from the peripheral blood B lymphocytes of humans after immunization with Hib vaccines. These hybridomas secrete antibody that resembles the serum antibody response following Hib CP immunization [31, 321. The V, nucleotide sequences of these human antibodies have shown V, restriction to the use of 9.1/2OPlrelated and VH,,/30P1-related elements [31, 321. Antibodies induced by both the CP and protein-conjugated CP use V,, genes and both 9.1/20P1 and VH,,/30P1-related elements. In total, of 10 V, genes sequenced from antibody secreting hybridomas, six have used 9.1/20Pl-related, three have used VH,,/30P1-related, and one used a germline V, gene with overlapping homology with these two. The 9.1/20Pl-like sequences were members of the v H 3 h subfamily that is characterized by two extra codons in CDR2 [33]. Six of the 9.1/20P1like sequences were most similar to the M85 cDNA sequence [34, 351, which uses a V, sequence that differs from the germline 9.1 gene by only three nucleotides, two being silent replacements [36]. A germline VH gene identical to M85 cDNA, LSG6.1, was, in fact, sequenced from the germline of one of the individuals from whom one of the anti-Hib C P hybridomas was generated. LSG6.1, therefore, represents the best candidate v H , h germline gene identified to date. The CDRl of these six LSG6.1-related V, genes was identical to the germline while some sequence variation was observed in CDR2. Two of the six M85-related V,, genes from unrelated individuals had complete nucleotide sequence identity in both CDRl and CDR2 regions with each other. Two hybridomas had identical V,, DH, and J, genes but used different light chains and heavy chain isotype switching patterns and still had preserved antibody activity. These lines, which were generated from the same individual, may have resulted from the progeny of a single pre-B cell undergoing two different initial light chain rearrangements [37] or a secondary light chain rearrangement [38]. The findings of restricted V, gene usage and shared CDR and V, gene usage in different antibodies demonstrate the importance of the VH gene in conferring antigen binding. The nucleotide sequence identity to germline V, genes of the antibody V, genes, ranged from 93-99% in the FR, and 92-97% in the CDR for the six LSG6.1-M85-related V, genes, and from 91-98% in the FR and 67-99% in the CDR for the three V,,,-related genes. Overall, there was 89-97% amino acid homology for the six LSG6.l-M85-related V, genes, and 73-94% amino acid homology for the three V,,,-related genes, but the CDR
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
30
R. A. INSEL et a / .
regions of some clones had amino acid sequence identity as low as 36%. The amino acid replacement to silent (WS) mutation ratios for the LSG 6. l-MWrelated sequences were 1.5 in the FR, devoid of S mutations in CDRI, and 2.6 in CDR2 and were 1.8 in the FR, 2.5 in CDR1, and 3.3 in CDR2 for the three V,,,-related V, genes. Thus, CDR2 was most frequently altered from the germ line in both types of sequences with some CDR2 RfS ratios reaching levels of 7 for VH2,-related and to levels of 5 for two LSG6.1-Mg5-related sequences. There was a lack of clustering of presumable mutations except in CDR2. These changes are consistent with some, but a relatively limited amount of, antigen-driven somatic hypermutation. Cloning of additional germline V, genes will be required to assess the overall degree of somatic hypermutation because of the highly polymorphic nature of human VH genes, particularly the V,, family. Overall, the V, locus is estimated to contain up to 300 V, genes with the V,, gene family representing the largest of the six-seven V, families [36, 391. The VH3bsubfamily represents approximately 20% of the V,, families. Although multiple V,, genes exist, some V,, genes such as LSG6.1 and 9.1 genes are highly homologous. v,, polymorphism at some alleles is present, while some v H 3 alleles are highly conserved in the population [40-421. The VH3b locus may be more highly polymorphic and larger than previously suggested [43], making definitive assignment of the degree of somatic mutation problematic. For some V,, genes, it appears that variation in the V,, germline repertoire observed in the population is due more to diversity of haplotypes than to allelic diversity of the V,, gene locus [40]. It is of interest, however, that both the 9.1/20P1 and V,26/30P1 germline sequences are highly conserved in the population. 9.1120P1 CDR2 oligonucleotide probes detect a single, identical 3.1 kb-Taq 1 fragment and vH2,/3OP1 CDR probes also detect single bands by Southern analysis in different individuals [40, 411. Conservation of these genes in the population may prevent so-called “holes in the repertoire” and preserve protective antibody responses in the population in spite of the requirement for the use of a highly restricted group of V, gene segments. Preliminary data suggest that the human antibody response to other bacterial CP may also use anti-Hib CP antibody-associated genes. It has been recently shown that antibodies to pneumococcal CP, which may also be restricted in diversity, can be encoded by vH26/3oP1or 9.1/2OPl-like gene elements [44, 44aJ. It is intriguing to suggest that the antibody repertoire to CP of bacterial human pathogens may be arising from a restricted set of conserved and evolutionarily preserved shared V, genes. There are some precedents for the same V, genes being used for multiple polysaccharides. For example, in the mouse, VH44,, a member of the X24 V, family, encodes for natural antibody responses to galactan, levan, a l - 6 dextran, 3-fucosyllactosamine, and galactosylgloboside [45]. V, segments used in the anti-Hib CP antibody response are also highly homologous to germline V, gene segments that are expressed in the human fetus, and that encode autoantibodies in man. Of five v H 3 genes expressed in fetal cDNA libraries, three were the vH,6/30Pl, M85,and 9.1/20P1 gene elements 134, 351. This finding, and the ability of young infants to respond to immunization when the polysaccharide is conjugated to a protein, suggest that the poor response of infants to the CP is not due to absence of these V genes in the B cell repertoire of the infant. It is of interest that these fetal sequences also have been shown to encode autoantibodies. The 9.1/20P1 germline sequence was shown to encode an anti-Sm antibody, an antibody that is specifically expressed in systemic lupus erythematosus (SLE) [33]. VH2,/30P1 and related gene sequences can encode antibodies that bind DNA, cardiolipin, platelets, insulin, vimentin, thyroglobulin, and IgG (rheumatoid factors) [46-481. VH2,-encoded anti-DNA antibodies express the heavy chain-associ-
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
REPERTOIRE OF HUMAN ANTI-HIB ANTIBODY
31
ated 16/6 idiotype, which is observed in anti-Dna antibodies in SLE and antibody induced by systemic infections with Klebsiellu and Mycobacteriu tuberculosis [49, 501. Human antibodies induced by pneumococcal polysaccharide vaccine or infection with Streptococcus pneumoniae can express a kappa light chain idiotype (31) associated with antiDNA antibodies [51]. The role of microbial antigens in eliciting autoantibody formation is one of active investigation. The close structural similarity between DNA and bacterial polysaccharides, the decreased titers of anti-DNA antibodies in NZB/W mice raised in a germ-free state [52], the protective effect on production of anti-DNA antibody of an Xid mutation (a mutation that prevents responses to polysaccharide antigens [53]), and the ability of bacterial CP to elicit V genes similar to those encoding anti-DNA antibodies all suggest exposure to bacterial antigens may contribute to the production of autoantibodies. Some germline genes elicited by antimicrobial antigens may mutate to give rise to high affinity autoreactive antibodies. While the contribution of V, segments to the specificity of antibody binding appears paramount, the role of the individual VH-CDRSin binding is not known. The single coding difference between germline LSG6.1 and germline 9. I is a serine to asparagine change in CDRl. This change to asparagine suggest this residue is involved in C P binding. The base differences suggestive of mutations in CDR2 of anti-Hib CP V, segments would also implicate this region as being important. V, CDR3 regions also appear important for antibody binding. The D, gene segments of the V, of anti-Hib CP antibodies were variable in size, ranging from 3-34 bases, generating a CDR ranging in size from as short as 3 to as long as 15 amino acids. The D regions were homologous to a variety of germline D elements (DXP-1, DXP-4, DIR-2). No bias of J, gene segment usage was observed and the J, segments expressed had slight differences from published germline J, polymorphisms, suggesting mutation or new polymorphisms. Shared CDR3 sequences of antibodies from unrelated individuals were observed. It was twice observed that antibody from unrelated individuals encoded by highly homologous but different and unique V, genes used identical VDJ joints. This finding suggests an important degree of antigenic selection acting on this region. There were also V, sequences from unrelated individuals that expressed identical CDRl and CDR2, but different CDR3. At this time, studies to correlate CDR3 expression and size with fine-binding specificity have not been performed. The cavity-type and groove type anti-cr. [l, 61 dextran antibodies show differences in size of CDR3 [3,4]. The cavity-type has a long V, and V, CDR3 region to allow formation of the cavity wall of the combining site. We have found that the postimmunization Hib CP antibodies can display differences in the size of oligosaccharide that fits into the binding site [21], but these studies need to be performed with monoclonal antibodies whose sequences have been characterized to assess the relationship between structure and binding specificity. As noted above, heterohybridoma antibodies developed from the same individual used identical V,, D,, and J, gene elements paired with different V, gene segments. This finding suggests that the V, may be making the major contribution to antigen binding in these two antibodies. This type of “promiscuous” V,-V, pairing has been previously observed for other antibodies. In combinatorial joining of V, and V, in-vivo to generate antibody diversity, it appears that identical V, domains can pair with different V, domains in antibody with distinctly different specificities, demonstrating a greater contribution of V, to antigen binding [54]. Different L chains can associate with the same H chain to yield functional murine autoantibody that binds ssDNA, but the combinatorial pairings differ in their avidity for dsDNA, RNA, and cardiolipin and in their nuclear staining patterns [ 5 5 ] . In generation of recombinant combinatorial libraries, promiscuous V, and V, pairing with
32
R. A. INSEL et al.
retention of antibody specificity has also been demonstrated for several murine as well as human antibodies to microbial antigens [56]. As described below, it appears that pairing of different V, with the same V, can alter the fine-binding specificity of antibodies to the Hib CP because in this particular pairing when V,, was paired with a VXVII light chain the antibody was cross-reactive with the E . coli K 100 CP, while pairing with a VXII did not encode a cross-reactive antibody, as detailed below.
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
5. REPERTOIRE OF ANTIBODY LIGHT CHAIN VARIABLE REGIONS (V,) As noted above, both kappa and lambda variable regions can encode anti-Hib CP antibody responses, with most individuals showing kappa type antibody predominantly. Overall, the V, of anti-Hib CP antibody is encoded by a limited number of V, genes, but with greater diversity than observed with V,. At least four VK and two VX gene families encode antibody [29, 57-59]. The majority of the kappa light chain antibody response and the majority of antibody in most individuals is encoded by the VKIIgene-A2 [24,25, 571. The VK locus consists of approximately 75 functional VK genes and retains a large duplicated region 160, 611. The A2 VKII germline gene exists, however, as a single copy member. Greater than 60% of individuals immunized with the CP or protein-conjugated CP generate a major antibody response encoded by A2 and about 85-95% of individuals generate some A2 antibody after immunization [25, 581. Not only is it surprising that a single VK gene is used for the major antibody response in most individuals, but the A2 gene appears to be used primarily in germline form. Amino acid sequencing of VKII antibodies generated in response to Hib CP vaccine has shown an identical partial sequence in 13 of 15 sequences, with these two having a single CDR3 substitution at position 94. Nucleotide sequencing of the V, of one antibody induced by Hib conjugate vaccine, had an unmutated germline pattern [57, 591, as shown (Fig. 1). The idiotype Hibid-1, described above, detected with both polyclonal and monoclonal anti-idiotypic antibodies, is specific for antibody that expresses the A2 VKIIlight chain [25]. The epitope on A2 recognized by this anti-idiotypic antibody is not described at this time. Detection of A2 with these anti-idiotypic antibodies has been useful in assessing A2 distribution in antibody responses. For example, in infants immunized with different conjugate vaccines, HibId-1 expression appears to vary depending on the carrier in the vaccine [62]. VKI, non A2 VKII, VKIII, and VKIV can also encode antibody, representing a minor component expressed along with the predominant V K I I - Aantibody ~ response in about onehalf of adults immunized [%]. By amino acid and nucleotide sequencing, several members of the VKI gene family may encode antibody. At least three different VKI genes-08/018, L11, and clone KC-1 [62A] can encode antibody [ 5 8 , 591. The VKIII gene used by four different individuals was the A27 gene and the partial amino acid sequence was identical in all subjects [58]. An asparagine at residue 53 in the CDR2 apparently arose by somatic mutation in all these A27 gene sequences, suggesting the importance for antigen binding of this residue, which is a relatively conserved asparagine residue in other V, sequences. Only rare non A ~ - V and K ~V K sequences ~ have been characterized. The V K gene ~ family is a single germline gene. Therefore, at minimum, at least seven of the approximately 75 VK genes can be expressed on anti-Hib CP antibody. Of interest, the VKI,non-A2 VKII, VKIII, and VKIV sequence appear to have undergone limited somatic mutation. The VX region of antibody may be encoded by VXII or VXVII [59]. Six lambda type light chains were characterized of which VXVII elements [63] were used in four lines and VXII
RCJ
gwmtiie A 2
RC3
germline A 2
getmlinr A 2 RC3
gemline A2 RC 3
Qermlme A2 RC 3
S
L
~
V
S
M
L
T
L
O
H
T
S
P 10
D
L
G
S
L
K
S
T
V
Y
T
L
I
1
P
Y
G
W
~
Y
P
L
A
Q
S
K
I
P
G
L
L
I
Y
Q
T
F
.............
L
D 70
P R CDR . 3 ATA CAG CTT CCT C
I
G
GA
T
’
L
F
S
G
A
P
V
ACT TTC 133C CCT
T
I
GGG
G
E 80
K
E
V
O
D
V
I
G
K
V
R
Y
Y
ACC AAA GTG GAT ATC AAA CGT germhe J, 3 .-A
1
A
.a .................................
TTC
F
K
........................................................................
~
E V S N R F S G V P D A F S G S G I50 CDR - 2 1 60 CCA CAG CTC CTG ATC TAT G A A GTT TCC AAC CGG TTC TCT GGA GTG CCA GAT AGG TTC AGT GGC AGC GGG TCA
P
........................................................................
CDR- 1 30 40 TCT AGT CAG AGC CTC CTG CAT AGT GAT GGA AAG ACC TAT TTG TAT TGG TAC CTG CAG AAG CCA GGC CAG CCT
S
D
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
S
~
P
W
B
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
34
R. A . INSEL et al.
[64] in two lines. The VXVII sequences had 96-98% homology to one another (Fig. 2). At this time there is only a restricted number of VXVII germline genes sequenced so that it is difficult to judge precisely the degree of somatic mutation of the germline gene. Although up to 10 VX genes may belong to the VXVII family [63,65,66], only one VXVII gene (4A) appeared to encode the antibody from four different individuals based on hybridization of genomic DNA with a CDR oligonucleotide probe. This conservation of sequences suggests that the VXVII genes are arising from a single group of closely related germline species. The translated amino acid sequences of these VXVII sequences had 90-97% homology, with FR showing 94-100% and CDR 72-96% homology with the germline. The two VhII genes had 91% nucleotide sequence identity with each other and had nucleotide sequence identity to germline VX2.1 gene (overall 91%, 99% in the FR and 74% in CDR for JB21, and overall 89%, and 95% in FR and 77% in CDR for 16M3C8)(Fig. 3). When translated, this represented for JB21 87% (99% in FR and 60% in CDR) and for 16M3C8 82% (92% in FR and 60% in CDR) amino acid sequence identity with VX2.1. Each of the VXVII encoded antibodies had cross-reactivity with the E . coli KlOO CP, but neither of the VXII antibodies cross-reacted. It appears that the VXVII gene makes a major contribution to KlOO binding because of the complete concordance of VXVII expression and cross-reactivity and because there were two hybridomas that expressed identical V, regions (LSG6.l-M85-like) with different V, (VXVII vs VXII) that displayed (VXVII) or lacked (VXII) KlOO cross-reactivity. The A,-VKII encoded antibody fails to cross-react with E . coli KlOO CP while cross-reactivity with E . coli KlOO CP is observed not infrequently with VKI, VKIII, and VKIV encoded antibody [58]. The CDR3 region of the VL of antibody to the CP, which arises from VL-JL joining, had some interesting properties. Each of the six JX segments of the heterohybridoma antibodies was encoded by a segment common to JXII and JXIII with substitution of a translationally silent mutation or polymorphism at position 97. JKI, 11, 111, or IV were used in the light chain antibody response [57-591. The JKIII and JKIV gene segments were shown to be quite similar to the germline sequence [59]. The VL-JL junctions showed a characteristic motif. Scott and colleagues first demonstrated by amino acid sequencing an almost invariant arginine at amino acid position 95 in A2-VKII-expressing antibodies [57]. Similarly, an arginine residue was generated at amino acid position 95d96 at the joint of six of seven of eight hybridoma antibody including all of the four VXVII, the two VXII, and the one germline A ~ - V K Iencoded I antibodies. This arginine at the V-JK junction is not encoded by the germline and may arise from either alternative V,J, joining after exonuclease cutting back of V and J or from N region addition at the joint. Alternative joining probably explains the joint with JKI, which has TGG at its 5‘ end that can join with CC at the 3’ end of the A2 V K gene ~ to give rise to the CGG arginine codon. N region addition, however, would best explain arginine at J K ~ - joints 5 and some of the VX joints. N region addition is suggested in the A2-encoded antibody RC3 (Fig. 1) and the VX2 encoded antibodies (Fig. 3), and in one of Scott’s antibodies [57]. N region addition is an event thought to be mediated by the enzyme terminal deoxynucleotidyl transferase and is a relatively uncommon event in V, gene rearrangements [67, 681. These findings suggest that a rare event in the developing B cells-N region addition during L chain rearrangement-is not infrequent and such an event may be antigenically selected. Although N region addition is delayed in lymphocyte ontogeny in the mouse 1691, N region addition in humans is occurring at the time of birth and is presumably occurring in some of the antibody induced in the young infant that displays arginine at this position. The frequent finding of an arginine at this position in both VK and VX antibodies suggests that this residue at this position
REPERTOIRE OF HUMAN ANTI-HIB ANTIBODY
35
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
contributes importantly to binding to the CP Many arginine residues in anti-DNA antibodies are encoded by the use of an unusual D reading frame, D-D fusions, D inversions, or alternative D-J joints [70]. It has been suggested that the negatively charged phosphate residues of DNA may be interacting with positively charged arginine. Binding to the phosphate residues in the Hib CP may occur similarly with this near invariant arginine in CDR3. Arginine at V-J junctions has been shown to be critical for binding to other antigens. The murine light chain in anti-parazophenylarsonateantibodies also contain an invariant arginine at position 96 generated by a VK10-Ars-JK1 joint that is essential for Ars binding [71].
6 . CONCLUSIONS In summary, human antibody to the Hib CP is restricted in diversity with the individual using a limited number of V genes to encode antibody. In the human population, V gene usage is also restricted and shared. The heavy chain of antibody is predominantly encoded by two members of the V,, family-LSG 6.liM85-like and VH,,/30Pl-like. The data suggests that in V, the CDRI, based on complete identity in LSG 6. UM85-like antibodies, CDR2, based on the suggestion of mutation in this region, and CDR3, based on conserved CDR3 usage in unrelated individuals, are inmortant in antigen binding. In contrast to the marked V, conservation, with use of a small number of some 200-300 V, genes, six or more different V, gene families encode antibody. The predominant antibody of the majority of individuals use the A ~ - V K Igene I in germline or near germline configuration. This gene has not been shown to be used for other antibody responses and appears to encode the HibId-1 determinants. Antibody can also be encoded by VKI, non-A2 VKII, VKIII, VKIV, VXII, and VXVII genes. Although different genes can be used, unrelated individuals appear to use the same VKIII(A27), VXII (VA2.1) and VXVII (4A) genes. The arginine in CDR3 of both kappa and lambda light chains suggest this is an important residue for antigen binding, and the asparagine in CDR2 of V, sequences and in CDRl of LSG6.1-rn85 V, sequences of antibody may be important for antigen binding. Evidence would suggest that a relatively limited degree of somatic mutation has occurred in the non-A2 V, genes, VAVII, and the V, genes. Further studies comparing the polymorphism of germline V genes to antibody encoding V genes are needed to clarify this issue. The V, diversity accounts for differences in fine binding specificity, with AZ-VKII genes not encoding E. cali KlOO CP cross-reactive antibodies and VAVII genes and some of the non-A2 VK genes encoding cross-reactive antibodies. Some of the KlOO cross-reactive naturally occurring antibody may be induced in the young infant prior to contact with the Hib organism. Immunization or exposure to the Hib CP appears to activate the VH3-V~II-A2 repertoire at the expense of the KlOO crossreactive repertoire. The reason for this repertoire alteration will require further study of differences in the affinity and protective capacity of antibody. It appears that protective antibody to this human pathogen is encoded by relatively few V regions that are shared in expression in the population. But, as noted above, polysaccharides can express multiple epitopes. Further, even in inbred strains of mice immunized with T dependent forms of polysaccharides or haptens there is an extensive increase in repertoire diversity with time after primary immunization or with secondary immunization. This does not seem to be occurring with responses to the Hib CP, as described. These genes may have been conserved with evolution because of the high protective capacity of their encoded antibodies. The possibility that these same genes also encode a broad group of antibodies to different CP of bacterial pathogens for man, with binding specificity differences based on a
...
-32
LSIZ
CAG
ACT
QQA
CTC
CTT
AGG
...... ...... ...... ......
CCC
QAO
CCC
QGG
I . .
TCA
CTG
ACT
TAT
. . T
TGC
-1.
. .
. . . . . . . .
..
GTQ
TCC
CCA
GGA
COG
ACA
QTC
."
ACT
C l C
ACC
11.1 G C T
GCC
I
w
...
. . Q
...
CTC
. . .
1 7 . . T A
1 . .
CTG
. . . .
C C C C
. . . .
TAC
an4
... ... . .
...
TAT
A . .
AC.
A . A , .
GOT
CTG
...
ACA
. C . . A .
. G .
.G.
. . . . . C
CAG .G. . G .
QCT
..c
...... ......
GGT
CCA
... . . c ... . . C ... . . C ... 0 . C
TAT
............
IGQ ACC CCT
GCC
M
............
TOO T T C C A G C A G 1.. I... . . . . . . . . . . . ............ 7 . . 1..
~
AAC
CGG
. - G . - G
. . G
..G
AAA
GQL
T I C
TCA
- . C - . C . - C
... ..c
... ... ...
40 CCT
TCA
010
GAG
GGG * c c
' . . A
. .G !
. . T
. . T
. . T
. . I
GAC
. . . .
GCT
GAG
. . . . . . . . . . . . . . . . . . . . . . . . . . . c
GAG
. . . . . .
T
K..OI.
I.-01-
m..oZ.
9.'-
K.01-
. Ke.01-
.
L ~ GC T G L C C G T C C T A
. . . . . . . . . ......... . . . . . . . . .
..*
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
00 GGA
.c.
M CCT
.........
CAG
.G#.... . . . . . . . . . . . . .
GGC
...
... .c. ... .c.
... . c .
GQT
............
.............................. . . . . . . . . . . . . . . . . . . . . . . . . .
..G
GTA TTC . . G . . G I . . .
. . T
. . T ..G . . T ..G . . I. . G
CTQ
.........
. . . . . . . . . ... . . A . . . . . . . . . . . . . . . .................. ..................... . . . . . . . . . . . . I . .. . . . . . . . A . . . . . . . . . . . . . . . . Q .
TIC C.T C . 1 C . 1 I . .C . 1
an GGT
......
.........
.........
10
AQT
............
.........
GGC
ACC
............ ............
.........
ATT
..T . . T .AT .GT
a c Q ~TC
ca-l
.........
CTQ
A . . 1.. I..
QCA
............ ............ ............
AQC
............
TCC
. . . . . . .
. . . . . .
TAC
TAT
n w
sa$m
ACT
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
-Vila
msz
LSCI
GIG
.. .. .. .. .. .. .. .. ..... .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ....... .. .. .. .. .. .. .. .. .. .. .. .. .. .. ....... . . . . .. .. .. ... .. .. . G ......................................................... .........................................................
GTQ
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..c . . . . . . . . . 1.. . . . . . . . . . . . . . . . . . . . c . . . . . . . . . . . . . . . . . . . . . . . . . . ..t
I C C
..c
G r n H V 1 V I I GGC
aAr. sa5m
LWZ m.32
..C . - c
GCA . . C
GCT .OC .QC .QC .GC
6..
Q . .
Q . .
ACT 0 . .
... ...
...
P5m8
am.
C A I
... ... ... ...
TQI
.. .. .. ... ...
CAG
O m V l V U
w
LsF1
-qva n w
-
LWZ a 2
1 =,m 1 1 .
-ViW
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
J1ZIJ
?
P
W 01
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
REPERTOIRE OF HUMAN ANTI-HIB ANTIBODY
c
-1
37
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
38
R. A. INSEL ef al.
L
I-, I-
Intern. Rev. Immunol. Vol. 9, 1992, pp. 25-43 Reprints available directly from the publisher Photocopying permitted by license only @I992Harwood Academic Publishers GmbH Printed in the United States of America
The Repertoire of Human Antibody to the Haemophilus influenzae mpe b Capsular Polysaccharide RICHARD A. INSEL,+ ELISABETH E. ADDERSON,$ and WILLIAM L. CARROLL* +Department of Pediatrics and Microbiology, University of Rochester School of Medicine and Dentistry, Rochester, New York 14642, USA *Program in Human Molecular Biology and Genetics, and Department of Pediatrics, University of Utah, Salt Lake City, Utah 84112, USA
Human antibody to the Haernophilus influenzne capsular polysaccharide (Hib CP) is restricted in diversity in the individual and the population with a limited number of variable region genes encoding antibody. Antibody to the Hib CP shows restricted isoelectric focusing gel patterns and light chain usage with frequent restriction to use of only kappa light chains. Shared cross-reactive idiotypes are expressed on antibody. The heavy chain of antibody to the Hib CP is predominantly encoded by two members of the V,, family-LSG 6.1/M8S-like and VH,,/30P1-like. In V, the CDRI, based on complete identity in LSG 6.liM85-like antibodies, CDR2, based on the suggestion of mutation in this region, and CDR3, based on conserved CDR3 usage in unrelated individuals, may be important for antigen binding. Six or more different V, gene families encode antibody. The predominant antibody of the majority of individuals uses the AZ-VKII gene in germline or near germline configuraion, which encodes an idiotype designated HibId-1. Antibody can also be encoded by VKI,non-A2 VKII,VKIII,VKIV,VXII, and VAVII genes. Although different V, genes can be used, unrelated individuals appear to use the same VKIII(A27), VAII (Vh2.1 and VAVII (4A) genes. The V, diversity accounts for differences in fine binding specificity, with AZ-VKII genes not encoding E . coli KlOO CP cross-reactive antibodies and VAVII genes and some of the non-A2 VK genes encoding cross-reactive antibodies. The arginine in CDR3 of both antibody kappa and lambda light chains and the asparagine in CDR2 of V, sequences and in CDRI of LSG6.1-M85 V, sequences of antibody appear to be important residues for antigen binding. A relatively limited degree of somatic mutation has occurred in the non-A2 V, genes, VAVII, and the V, genes. Further studies comparing the polymorphism of germline V genes to antibody-encoding V genes are needed to clarify this issue. Research comparing this repertoire to repertoires directed to other bacterial CP and to self antigens and defining structure-antigen binding relationships is in progress. KEYWORDS: antibody, variable region, V , genes, V , genes
1. INTRODUCTION It has long been appreciated that the antibody responses in animals to polysaccharides can be restricted and shared with evidence of clonal restriction and dominance, cross-reactive idiotypes, shared V gene sequences, and shared fine-binding specificity [I, 21. Polysaccharides lend themselves to studies of antibody repertoires because they can provide single antigenic determinants that completely fill the antibody combining site. They have an advantage over more structurally complex native or even artificial antigens and over haptens which often fail to engage the combining site completely, resulting in specificity to 25
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
26
R. A . INSEL er
CJ~.
haptens being dictated by both the hapten and a portion of the protein complexed to the hapten. It should be noted, however, that while polysaccharides are restricted in overall complexity compared to a protein antigen, they have potentially multiple unique epitopes that can be recognized by antibodies. Kabat and colleagues have shown using immunochemical approaches and anti-idiotypic antibodies that antibodies to polysaccharides can be directed, or complementary, to either non-reducing termini or to internal core regions of a polysaccharide and these specificities are reflected in structures of the combining sites or complementarity determining regions (CDR) [3]. The antibodies binding to internal linear determinants displayed “groove-type’’ antibody-binding sites while those antibodies with specificities for non-reducing ends of the polysaccharide had “cavity-type’’ antibody binding sites [4]. Thus, even antibody responses to relatively simple antigens can be variable and complex. The diversity of human antibody repertoires to defined antigens has been relatively poorly characterized in comparison to the mouse. One exception has been the human antibody response to the capsular polysaccharide (CP) of Haemophilus injuenzae type b. In this review we will address our current understanding of the diversity of the human antibody response to the Haemophilus injuenzae type b (Hib) capsular polysaccharide.
2. RATIONALE FOR CHARACTERIZING THE DIVERSITY OF HAEMOPHlLUS INFLUENZAE B CAPSULAR POLYSACCHARIDE ANTIBODY We began studies on the diversity of human antibody to this antigen because of the extensive work in inbred strains of mice demonstrating restricted diversity of the variable regions of antibodies to polysaccharides and because studying the diversity of the antibody response to the Haemophilus injuenzae C P had several merits. First, this antibody is the major form of protection against a significant human pathogen. Haemophilus injuenzae b, a gram negative bacteria, is the major etiologic agent of bacterial meningitis and sepsis in the United States with a peak age-incidence of infection at 6-12 months of age. The principal virulence determinant on the bacteria is its CP, which is composed of a repetitive polymer of ribosyl-ribitol-phosphate in which the two sugars are in a 1-1 linkage [ 5 ] . Antibody to the type b capsular polysaccharide is the major protective antibody in man against infection and mediates its action through its opsonic properties directly and indirectly through activation of the complement pathways. Infection is infrequent in the first few months of life because of transplacental passage of maternal antibody to the type b CP and in later childhood following the active acquisition of antibody through exposure either to the type b polysaccharide or to cross-reactive antigens on other microorganisms [6]. One such microorganism is Escherichia coli KlOO that has a CP structure identical to the Hib CP except it uses a (1-2) rather than a (1-1) ribosyl-ribitol linkage [7]. We have found that the naturally occurring antibody in children is more commonly cross-reactive with KlOO than that of the adult [8].This presumably reflects increased exposure of older individuals to Hib or other cross-reactive antigens that more precisely express epitopes found on the Hib CF? Supporting this idea is the finding that immunization with the Hib CP decreases the proportion of KlOO cross-reactive antibody compared to preimmunization antibody [8]. These findings indicated that more than one type of variable region gene (K100 crossreactive and non cross-reactive) may be encoding the response. Conceivably, study of
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
REPERTOIRE OF HUMAN ANTI-HIB ANTIBODY
27
naturally occurring and vaccine-induced antibody may demonstrate immunoglobulin variable region (V) genes that have been conserved in the human population to allow response to this virulent pathogen. A second advantage of the use of the Hib CP is that it has been possible to actively immunize humans with the CP to study high-titered antibody responses. A number of Hib CP vaccines were developed in the last two decades to prevent this serious childhood infection [9, 101. Initially, vaccines composed of the purified type b C P were developed and then licensed for children at ages 18-24 months in the United States in 1985. The efficacy of the type b CP vaccines was limited because infants younger than 18 months of age, the age associated with the highest incidence of infection, failed to generate an antibody response and infants between 18-30 months had inconsistent responses. At age 4-5 years, children responded but with low magnitude responses compared to adults. This delay in response to this CP is not unique but is observed with several other bacterial CP [ 111. The development of the antibody repertoire with age could also be studied using the Hib C P antibody model. The unresponsiveness of young infants to the CP, however, is not fixed. It has been found that even young infants respond to the type b C P antigen if the CP is conjugated to a protein carrier [12, 131. Second generation vaccines were developed in which the type b C P or oligosaccharides prepared from the polysaccharides were covalently conjugated to mutant diphtheria toxins, tetanus toxoid, or meningococcal outer membrane proteins [6, 10, 11, 141. These so-called “conjugate vaccines” have the ability to recruit CD4+ T-lymphocytes to provide cooperative help to CP-specific B cells. In addition, conjugate vaccines have the ability to prime the infant to generate an antibody response to immunization with the unconjugated C P [12, 131. Further, conjugate vaccines can bypass poor antibody responses to the C P in some patients with primary immunodeficiency disease (e.g., IgG subclass deficiency) [15]. Conjugate vaccines induce protection against Hib infection in the infant, and were licensed in the United States in 1989 for use in infants beginning at two months of age. Development of these vaccines has allowed comparison of the antibody and antibody repertoires induced by unconjugated and protein-conjugated forms of the Cl? Lastly, certain subpopulations have a higher incidence of systemic Hib infection [16]. Immunoglobulin allotype markers have been associated with an increased risk of invasive Hib disease and with preferential IgG subclass and light chain type antibody responses [17, 181. The possibility that the high risk in such individuals is due to so-called holes or gaps in the immunoglobulin gene repertoire could be explored by analysis of the Hib C P antibody repertoire.
3. CLONOTYPE, LIGHT CHAIN DIVERSITY, AND IDIOTYPY OF ANTIBODY The first evidence for restriction of the human anti-Hib C P antibody response was from the demonstration of restricted numbers of antibody clones by isoelectric focusing [ 19, 201, the restricted light chain distribution of the antibody [21, 221, and the presence of shared crossreactive idiotypes in serum of unrelated subjects [23-251. Some diversity of the response was readily apparent, however, in both individuals and the general population from the finding of expression of both kappa and lambda light chain types [21,22], and E . coli crossreactive and non cross-reactive antibody in the antibody response [7, 81. Analytical isoelectric focusing (IEF) of postimmunization IgG antibody showed that an
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
28
R. A . INSEL et al.
individual serum displayed a restricted repertoire with only 1 to 4 clones, or so-called clonotypes or spectrotypes, on a gel [19]. This type of analysis, which detects charge differences among antibodies, is affected by both the variable and constant region contribution to overall charge as well as post-translational glycosylation differences on immunoglobulins. Most adults immunized with the polysaccharide or protein-conjugated polysaccharide vaccine generate both IgGl and IgG2 subclass antibody responses [26-281, and this heavy chain constant region difference leads to charge differences and different IEF patterns, even when the same V, gene is used. Thus, the finding of distinct clonal patterns suggested a marked degree of antibody restriction. It was also found, somewhat surprisingly, that antibody induced by the protein-conjugated Hib CP vaccines was as highly restricted as that induced by the Hib CP vaccine. Infants immunized with protein-conjugated Hib CP vaccines generated pauciclonal responses 1131. Repetitive immunization with these so-called conjugate vaccines in infancy induces booster serum antibody responses. Based on IEF patterns, these increased levels of serum antibody occur primarily from pre-existing clones activated by earlier immunizations, thus preserving clonal restriction. As noted above, the conjugate vaccines can accelerate ontogeny of responses to the polysaccharide vaccine so that conjugate-primed young infants can respond to the C P vaccine. Booster immunization with Hib CP vaccine in conjugate-primed infants also tends to restimulate pre-existent clones without new clonal recruitment, with some preferential activation of IgG2 vs IgGl expressing clones in some individuals [ 131. Two other aspects of the isoelectric focusing patterns of antibody proved of interest. First, while most individuals have unique clonotype patterns, some clones appeared to be shared between individuals, suggesting use of similar or highly related V genes [19, 201. Lastly, the human antibody response to CP is known to be a long-lived response and clonal dominance was sustained over time. Clones that dominated the response at one month remained dominant several years later [19]. The normal distribution of kappa and lambda light chain types in serum immunoglobulin is approximately 60% kappa and 40% lambda light chain type and most antibody responses conform to this distribution. The light chain distribution of antibody induced by either the CP or protein-conjugated CP was predominantly or, in some individuals, exclusively, kappa light chain type [21, 221. Some individuals do generate, however, a lambda type predominant antibody response, and we know today, as discussed below, that lambda-expressing antibodies can, not uncommonly, have some unique fine-binding specificities. The finding of kappa light chain restricted, or in rare individuals lambda light chain restricted antibody, also suggested a relatively limited repertoire. Shared cross-reactive idiotypes were demonstrated by Lucas and colleagues on naturallyoccurring and postimmunization antibody using polyclonal and monoclonal anti-idiotypic antibodies that appeared to bind in or close to the antibody combining site [23-251. One monoclonal antibody, which detects an idiotype designated HibId-1, whose structural equivalent is detailed below, reacts specifically with the Hib C P or conjugate postimmunization antibody of approximately 85% of individuals [25]. In these individuals, it detects 60% of the total antibody induced by immunization, including both IgGl and IgG2 antibodies. Immunization increases the proportion of Hib Id-1 antibody, which is in contrast to the diminution in KlOO cross-reactive antibody, described above. These findings demonstrate that the majority of individuals and the majority of anti-Hib CP antibodies share a unique cross-reactive idiotype, also indicating conservation of the repertoire.
REPERTOIRE OF HUMAN ANTI-HIB ANTIBODY
29
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
4. RESTRICTION OF ANTIBODY HEAVY CHAIN VARIABLE REGIONS (V,) Initial evidence of V, region restriction of anti-Hib CP antibody came from studies by Scott and colleagues who found the IgG antibody of adults induced by the Hib CP to be encoded exclusively by the v H 3 family, based on N-terminal amino acid sequences of immunoaffinity-purified serum antibody [29]. Predominant usage of v H 3 genes was confirmed by several other approaches. Using VH3 anti-peptide antibodies directed to framework and hypervariable regions of V, families and elements, Silverman and colleagues demonstrated the use of v H 3 in the IgG, IgA, and IgM antibody induced by CP immunization in all subjects studied [30]. In addition, their studies suggested dominant use of two germline V,, elements-9.1/20Pl-and VH26/30P1-associated subfamilies-in the antibody response. Some individuals appeared to use both v H 3 gene members. The postimmunization IgG antibody of some individuals was also detected with anti-VH1and V,, family-specific antibodies. Definitive proof of use of V,, and identification of the specific V,, gene elements arose from nucleotide sequencing of the immunoglobulin heavy chain genes of human antibodysecreting heterohybridomas generated from the peripheral blood B lymphocytes of humans after immunization with Hib vaccines. These hybridomas secrete antibody that resembles the serum antibody response following Hib CP immunization [31, 321. The V, nucleotide sequences of these human antibodies have shown V, restriction to the use of 9.1/2OPlrelated and VH,,/30P1-related elements [31, 321. Antibodies induced by both the CP and protein-conjugated CP use V,, genes and both 9.1/20P1 and VH,,/30P1-related elements. In total, of 10 V, genes sequenced from antibody secreting hybridomas, six have used 9.1/20Pl-related, three have used VH,,/30P1-related, and one used a germline V, gene with overlapping homology with these two. The 9.1/20Pl-like sequences were members of the v H 3 h subfamily that is characterized by two extra codons in CDR2 [33]. Six of the 9.1/20P1like sequences were most similar to the M85 cDNA sequence [34, 351, which uses a V, sequence that differs from the germline 9.1 gene by only three nucleotides, two being silent replacements [36]. A germline VH gene identical to M85 cDNA, LSG6.1, was, in fact, sequenced from the germline of one of the individuals from whom one of the anti-Hib C P hybridomas was generated. LSG6.1, therefore, represents the best candidate v H , h germline gene identified to date. The CDRl of these six LSG6.1-related V, genes was identical to the germline while some sequence variation was observed in CDR2. Two of the six M85-related V,, genes from unrelated individuals had complete nucleotide sequence identity in both CDRl and CDR2 regions with each other. Two hybridomas had identical V,, DH, and J, genes but used different light chains and heavy chain isotype switching patterns and still had preserved antibody activity. These lines, which were generated from the same individual, may have resulted from the progeny of a single pre-B cell undergoing two different initial light chain rearrangements [37] or a secondary light chain rearrangement [38]. The findings of restricted V, gene usage and shared CDR and V, gene usage in different antibodies demonstrate the importance of the VH gene in conferring antigen binding. The nucleotide sequence identity to germline V, genes of the antibody V, genes, ranged from 93-99% in the FR, and 92-97% in the CDR for the six LSG6.1-M85-related V, genes, and from 91-98% in the FR and 67-99% in the CDR for the three V,,,-related genes. Overall, there was 89-97% amino acid homology for the six LSG6.l-M85-related V, genes, and 73-94% amino acid homology for the three V,,,-related genes, but the CDR
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
30
R. A. INSEL et a / .
regions of some clones had amino acid sequence identity as low as 36%. The amino acid replacement to silent (WS) mutation ratios for the LSG 6. l-MWrelated sequences were 1.5 in the FR, devoid of S mutations in CDRI, and 2.6 in CDR2 and were 1.8 in the FR, 2.5 in CDR1, and 3.3 in CDR2 for the three V,,,-related V, genes. Thus, CDR2 was most frequently altered from the germ line in both types of sequences with some CDR2 RfS ratios reaching levels of 7 for VH2,-related and to levels of 5 for two LSG6.1-Mg5-related sequences. There was a lack of clustering of presumable mutations except in CDR2. These changes are consistent with some, but a relatively limited amount of, antigen-driven somatic hypermutation. Cloning of additional germline V, genes will be required to assess the overall degree of somatic hypermutation because of the highly polymorphic nature of human VH genes, particularly the V,, family. Overall, the V, locus is estimated to contain up to 300 V, genes with the V,, gene family representing the largest of the six-seven V, families [36, 391. The VH3bsubfamily represents approximately 20% of the V,, families. Although multiple V,, genes exist, some V,, genes such as LSG6.1 and 9.1 genes are highly homologous. v,, polymorphism at some alleles is present, while some v H 3 alleles are highly conserved in the population [40-421. The VH3b locus may be more highly polymorphic and larger than previously suggested [43], making definitive assignment of the degree of somatic mutation problematic. For some V,, genes, it appears that variation in the V,, germline repertoire observed in the population is due more to diversity of haplotypes than to allelic diversity of the V,, gene locus [40]. It is of interest, however, that both the 9.1/20P1 and V,26/30P1 germline sequences are highly conserved in the population. 9.1120P1 CDR2 oligonucleotide probes detect a single, identical 3.1 kb-Taq 1 fragment and vH2,/3OP1 CDR probes also detect single bands by Southern analysis in different individuals [40, 411. Conservation of these genes in the population may prevent so-called “holes in the repertoire” and preserve protective antibody responses in the population in spite of the requirement for the use of a highly restricted group of V, gene segments. Preliminary data suggest that the human antibody response to other bacterial CP may also use anti-Hib CP antibody-associated genes. It has been recently shown that antibodies to pneumococcal CP, which may also be restricted in diversity, can be encoded by vH26/3oP1or 9.1/2OPl-like gene elements [44, 44aJ. It is intriguing to suggest that the antibody repertoire to CP of bacterial human pathogens may be arising from a restricted set of conserved and evolutionarily preserved shared V, genes. There are some precedents for the same V, genes being used for multiple polysaccharides. For example, in the mouse, VH44,, a member of the X24 V, family, encodes for natural antibody responses to galactan, levan, a l - 6 dextran, 3-fucosyllactosamine, and galactosylgloboside [45]. V, segments used in the anti-Hib CP antibody response are also highly homologous to germline V, gene segments that are expressed in the human fetus, and that encode autoantibodies in man. Of five v H 3 genes expressed in fetal cDNA libraries, three were the vH,6/30Pl, M85,and 9.1/20P1 gene elements 134, 351. This finding, and the ability of young infants to respond to immunization when the polysaccharide is conjugated to a protein, suggest that the poor response of infants to the CP is not due to absence of these V genes in the B cell repertoire of the infant. It is of interest that these fetal sequences also have been shown to encode autoantibodies. The 9.1/20P1 germline sequence was shown to encode an anti-Sm antibody, an antibody that is specifically expressed in systemic lupus erythematosus (SLE) [33]. VH2,/30P1 and related gene sequences can encode antibodies that bind DNA, cardiolipin, platelets, insulin, vimentin, thyroglobulin, and IgG (rheumatoid factors) [46-481. VH2,-encoded anti-DNA antibodies express the heavy chain-associ-
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
REPERTOIRE OF HUMAN ANTI-HIB ANTIBODY
31
ated 16/6 idiotype, which is observed in anti-Dna antibodies in SLE and antibody induced by systemic infections with Klebsiellu and Mycobacteriu tuberculosis [49, 501. Human antibodies induced by pneumococcal polysaccharide vaccine or infection with Streptococcus pneumoniae can express a kappa light chain idiotype (31) associated with antiDNA antibodies [51]. The role of microbial antigens in eliciting autoantibody formation is one of active investigation. The close structural similarity between DNA and bacterial polysaccharides, the decreased titers of anti-DNA antibodies in NZB/W mice raised in a germ-free state [52], the protective effect on production of anti-DNA antibody of an Xid mutation (a mutation that prevents responses to polysaccharide antigens [53]), and the ability of bacterial CP to elicit V genes similar to those encoding anti-DNA antibodies all suggest exposure to bacterial antigens may contribute to the production of autoantibodies. Some germline genes elicited by antimicrobial antigens may mutate to give rise to high affinity autoreactive antibodies. While the contribution of V, segments to the specificity of antibody binding appears paramount, the role of the individual VH-CDRSin binding is not known. The single coding difference between germline LSG6.1 and germline 9. I is a serine to asparagine change in CDRl. This change to asparagine suggest this residue is involved in C P binding. The base differences suggestive of mutations in CDR2 of anti-Hib CP V, segments would also implicate this region as being important. V, CDR3 regions also appear important for antibody binding. The D, gene segments of the V, of anti-Hib CP antibodies were variable in size, ranging from 3-34 bases, generating a CDR ranging in size from as short as 3 to as long as 15 amino acids. The D regions were homologous to a variety of germline D elements (DXP-1, DXP-4, DIR-2). No bias of J, gene segment usage was observed and the J, segments expressed had slight differences from published germline J, polymorphisms, suggesting mutation or new polymorphisms. Shared CDR3 sequences of antibodies from unrelated individuals were observed. It was twice observed that antibody from unrelated individuals encoded by highly homologous but different and unique V, genes used identical VDJ joints. This finding suggests an important degree of antigenic selection acting on this region. There were also V, sequences from unrelated individuals that expressed identical CDRl and CDR2, but different CDR3. At this time, studies to correlate CDR3 expression and size with fine-binding specificity have not been performed. The cavity-type and groove type anti-cr. [l, 61 dextran antibodies show differences in size of CDR3 [3,4]. The cavity-type has a long V, and V, CDR3 region to allow formation of the cavity wall of the combining site. We have found that the postimmunization Hib CP antibodies can display differences in the size of oligosaccharide that fits into the binding site [21], but these studies need to be performed with monoclonal antibodies whose sequences have been characterized to assess the relationship between structure and binding specificity. As noted above, heterohybridoma antibodies developed from the same individual used identical V,, D,, and J, gene elements paired with different V, gene segments. This finding suggests that the V, may be making the major contribution to antigen binding in these two antibodies. This type of “promiscuous” V,-V, pairing has been previously observed for other antibodies. In combinatorial joining of V, and V, in-vivo to generate antibody diversity, it appears that identical V, domains can pair with different V, domains in antibody with distinctly different specificities, demonstrating a greater contribution of V, to antigen binding [54]. Different L chains can associate with the same H chain to yield functional murine autoantibody that binds ssDNA, but the combinatorial pairings differ in their avidity for dsDNA, RNA, and cardiolipin and in their nuclear staining patterns [ 5 5 ] . In generation of recombinant combinatorial libraries, promiscuous V, and V, pairing with
32
R. A. INSEL et al.
retention of antibody specificity has also been demonstrated for several murine as well as human antibodies to microbial antigens [56]. As described below, it appears that pairing of different V, with the same V, can alter the fine-binding specificity of antibodies to the Hib CP because in this particular pairing when V,, was paired with a VXVII light chain the antibody was cross-reactive with the E . coli K 100 CP, while pairing with a VXII did not encode a cross-reactive antibody, as detailed below.
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
5. REPERTOIRE OF ANTIBODY LIGHT CHAIN VARIABLE REGIONS (V,) As noted above, both kappa and lambda variable regions can encode anti-Hib CP antibody responses, with most individuals showing kappa type antibody predominantly. Overall, the V, of anti-Hib CP antibody is encoded by a limited number of V, genes, but with greater diversity than observed with V,. At least four VK and two VX gene families encode antibody [29, 57-59]. The majority of the kappa light chain antibody response and the majority of antibody in most individuals is encoded by the VKIIgene-A2 [24,25, 571. The VK locus consists of approximately 75 functional VK genes and retains a large duplicated region 160, 611. The A2 VKII germline gene exists, however, as a single copy member. Greater than 60% of individuals immunized with the CP or protein-conjugated CP generate a major antibody response encoded by A2 and about 85-95% of individuals generate some A2 antibody after immunization [25, 581. Not only is it surprising that a single VK gene is used for the major antibody response in most individuals, but the A2 gene appears to be used primarily in germline form. Amino acid sequencing of VKII antibodies generated in response to Hib CP vaccine has shown an identical partial sequence in 13 of 15 sequences, with these two having a single CDR3 substitution at position 94. Nucleotide sequencing of the V, of one antibody induced by Hib conjugate vaccine, had an unmutated germline pattern [57, 591, as shown (Fig. 1). The idiotype Hibid-1, described above, detected with both polyclonal and monoclonal anti-idiotypic antibodies, is specific for antibody that expresses the A2 VKIIlight chain [25]. The epitope on A2 recognized by this anti-idiotypic antibody is not described at this time. Detection of A2 with these anti-idiotypic antibodies has been useful in assessing A2 distribution in antibody responses. For example, in infants immunized with different conjugate vaccines, HibId-1 expression appears to vary depending on the carrier in the vaccine [62]. VKI, non A2 VKII, VKIII, and VKIV can also encode antibody, representing a minor component expressed along with the predominant V K I I - Aantibody ~ response in about onehalf of adults immunized [%]. By amino acid and nucleotide sequencing, several members of the VKI gene family may encode antibody. At least three different VKI genes-08/018, L11, and clone KC-1 [62A] can encode antibody [ 5 8 , 591. The VKIII gene used by four different individuals was the A27 gene and the partial amino acid sequence was identical in all subjects [58]. An asparagine at residue 53 in the CDR2 apparently arose by somatic mutation in all these A27 gene sequences, suggesting the importance for antigen binding of this residue, which is a relatively conserved asparagine residue in other V, sequences. Only rare non A ~ - V and K ~V K sequences ~ have been characterized. The V K gene ~ family is a single germline gene. Therefore, at minimum, at least seven of the approximately 75 VK genes can be expressed on anti-Hib CP antibody. Of interest, the VKI,non-A2 VKII, VKIII, and VKIV sequence appear to have undergone limited somatic mutation. The VX region of antibody may be encoded by VXII or VXVII [59]. Six lambda type light chains were characterized of which VXVII elements [63] were used in four lines and VXII
RCJ
gwmtiie A 2
RC3
germline A 2
getmlinr A 2 RC3
gemline A2 RC 3
Qermlme A2 RC 3
S
L
~
V
S
M
L
T
L
O
H
T
S
P 10
D
L
G
S
L
K
S
T
V
Y
T
L
I
1
P
Y
G
W
~
Y
P
L
A
Q
S
K
I
P
G
L
L
I
Y
Q
T
F
.............
L
D 70
P R CDR . 3 ATA CAG CTT CCT C
I
G
GA
T
’
L
F
S
G
A
P
V
ACT TTC 133C CCT
T
I
GGG
G
E 80
K
E
V
O
D
V
I
G
K
V
R
Y
Y
ACC AAA GTG GAT ATC AAA CGT germhe J, 3 .-A
1
A
.a .................................
TTC
F
K
........................................................................
~
E V S N R F S G V P D A F S G S G I50 CDR - 2 1 60 CCA CAG CTC CTG ATC TAT G A A GTT TCC AAC CGG TTC TCT GGA GTG CCA GAT AGG TTC AGT GGC AGC GGG TCA
P
........................................................................
CDR- 1 30 40 TCT AGT CAG AGC CTC CTG CAT AGT GAT GGA AAG ACC TAT TTG TAT TGG TAC CTG CAG AAG CCA GGC CAG CCT
S
D
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
S
~
P
W
B
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
34
R. A . INSEL et al.
[64] in two lines. The VXVII sequences had 96-98% homology to one another (Fig. 2). At this time there is only a restricted number of VXVII germline genes sequenced so that it is difficult to judge precisely the degree of somatic mutation of the germline gene. Although up to 10 VX genes may belong to the VXVII family [63,65,66], only one VXVII gene (4A) appeared to encode the antibody from four different individuals based on hybridization of genomic DNA with a CDR oligonucleotide probe. This conservation of sequences suggests that the VXVII genes are arising from a single group of closely related germline species. The translated amino acid sequences of these VXVII sequences had 90-97% homology, with FR showing 94-100% and CDR 72-96% homology with the germline. The two VhII genes had 91% nucleotide sequence identity with each other and had nucleotide sequence identity to germline VX2.1 gene (overall 91%, 99% in the FR and 74% in CDR for JB21, and overall 89%, and 95% in FR and 77% in CDR for 16M3C8)(Fig. 3). When translated, this represented for JB21 87% (99% in FR and 60% in CDR) and for 16M3C8 82% (92% in FR and 60% in CDR) amino acid sequence identity with VX2.1. Each of the VXVII encoded antibodies had cross-reactivity with the E . coli KlOO CP, but neither of the VXII antibodies cross-reacted. It appears that the VXVII gene makes a major contribution to KlOO binding because of the complete concordance of VXVII expression and cross-reactivity and because there were two hybridomas that expressed identical V, regions (LSG6.l-M85-like) with different V, (VXVII vs VXII) that displayed (VXVII) or lacked (VXII) KlOO cross-reactivity. The A,-VKII encoded antibody fails to cross-react with E . coli KlOO CP while cross-reactivity with E . coli KlOO CP is observed not infrequently with VKI, VKIII, and VKIV encoded antibody [58]. The CDR3 region of the VL of antibody to the CP, which arises from VL-JL joining, had some interesting properties. Each of the six JX segments of the heterohybridoma antibodies was encoded by a segment common to JXII and JXIII with substitution of a translationally silent mutation or polymorphism at position 97. JKI, 11, 111, or IV were used in the light chain antibody response [57-591. The JKIII and JKIV gene segments were shown to be quite similar to the germline sequence [59]. The VL-JL junctions showed a characteristic motif. Scott and colleagues first demonstrated by amino acid sequencing an almost invariant arginine at amino acid position 95 in A2-VKII-expressing antibodies [57]. Similarly, an arginine residue was generated at amino acid position 95d96 at the joint of six of seven of eight hybridoma antibody including all of the four VXVII, the two VXII, and the one germline A ~ - V K Iencoded I antibodies. This arginine at the V-JK junction is not encoded by the germline and may arise from either alternative V,J, joining after exonuclease cutting back of V and J or from N region addition at the joint. Alternative joining probably explains the joint with JKI, which has TGG at its 5‘ end that can join with CC at the 3’ end of the A2 V K gene ~ to give rise to the CGG arginine codon. N region addition, however, would best explain arginine at J K ~ - joints 5 and some of the VX joints. N region addition is suggested in the A2-encoded antibody RC3 (Fig. 1) and the VX2 encoded antibodies (Fig. 3), and in one of Scott’s antibodies [57]. N region addition is an event thought to be mediated by the enzyme terminal deoxynucleotidyl transferase and is a relatively uncommon event in V, gene rearrangements [67, 681. These findings suggest that a rare event in the developing B cells-N region addition during L chain rearrangement-is not infrequent and such an event may be antigenically selected. Although N region addition is delayed in lymphocyte ontogeny in the mouse 1691, N region addition in humans is occurring at the time of birth and is presumably occurring in some of the antibody induced in the young infant that displays arginine at this position. The frequent finding of an arginine at this position in both VK and VX antibodies suggests that this residue at this position
REPERTOIRE OF HUMAN ANTI-HIB ANTIBODY
35
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
contributes importantly to binding to the CP Many arginine residues in anti-DNA antibodies are encoded by the use of an unusual D reading frame, D-D fusions, D inversions, or alternative D-J joints [70]. It has been suggested that the negatively charged phosphate residues of DNA may be interacting with positively charged arginine. Binding to the phosphate residues in the Hib CP may occur similarly with this near invariant arginine in CDR3. Arginine at V-J junctions has been shown to be critical for binding to other antigens. The murine light chain in anti-parazophenylarsonateantibodies also contain an invariant arginine at position 96 generated by a VK10-Ars-JK1 joint that is essential for Ars binding [71].
6 . CONCLUSIONS In summary, human antibody to the Hib CP is restricted in diversity with the individual using a limited number of V genes to encode antibody. In the human population, V gene usage is also restricted and shared. The heavy chain of antibody is predominantly encoded by two members of the V,, family-LSG 6.liM85-like and VH,,/30Pl-like. The data suggests that in V, the CDRI, based on complete identity in LSG 6. UM85-like antibodies, CDR2, based on the suggestion of mutation in this region, and CDR3, based on conserved CDR3 usage in unrelated individuals, are inmortant in antigen binding. In contrast to the marked V, conservation, with use of a small number of some 200-300 V, genes, six or more different V, gene families encode antibody. The predominant antibody of the majority of individuals use the A ~ - V K Igene I in germline or near germline configuration. This gene has not been shown to be used for other antibody responses and appears to encode the HibId-1 determinants. Antibody can also be encoded by VKI, non-A2 VKII, VKIII, VKIV, VXII, and VXVII genes. Although different genes can be used, unrelated individuals appear to use the same VKIII(A27), VXII (VA2.1) and VXVII (4A) genes. The arginine in CDR3 of both kappa and lambda light chains suggest this is an important residue for antigen binding, and the asparagine in CDR2 of V, sequences and in CDRl of LSG6.1-rn85 V, sequences of antibody may be important for antigen binding. Evidence would suggest that a relatively limited degree of somatic mutation has occurred in the non-A2 V, genes, VAVII, and the V, genes. Further studies comparing the polymorphism of germline V genes to antibody encoding V genes are needed to clarify this issue. The V, diversity accounts for differences in fine binding specificity, with AZ-VKII genes not encoding E. cali KlOO CP cross-reactive antibodies and VAVII genes and some of the non-A2 VK genes encoding cross-reactive antibodies. Some of the KlOO cross-reactive naturally occurring antibody may be induced in the young infant prior to contact with the Hib organism. Immunization or exposure to the Hib CP appears to activate the VH3-V~II-A2 repertoire at the expense of the KlOO crossreactive repertoire. The reason for this repertoire alteration will require further study of differences in the affinity and protective capacity of antibody. It appears that protective antibody to this human pathogen is encoded by relatively few V regions that are shared in expression in the population. But, as noted above, polysaccharides can express multiple epitopes. Further, even in inbred strains of mice immunized with T dependent forms of polysaccharides or haptens there is an extensive increase in repertoire diversity with time after primary immunization or with secondary immunization. This does not seem to be occurring with responses to the Hib CP, as described. These genes may have been conserved with evolution because of the high protective capacity of their encoded antibodies. The possibility that these same genes also encode a broad group of antibodies to different CP of bacterial pathogens for man, with binding specificity differences based on a
...
-32
LSIZ
CAG
ACT
QQA
CTC
CTT
AGG
...... ...... ...... ......
CCC
QAO
CCC
QGG
I . .
TCA
CTG
ACT
TAT
. . T
TGC
-1.
. .
. . . . . . . .
..
GTQ
TCC
CCA
GGA
COG
ACA
QTC
."
ACT
C l C
ACC
11.1 G C T
GCC
I
w
...
. . Q
...
CTC
. . .
1 7 . . T A
1 . .
CTG
. . . .
C C C C
. . . .
TAC
an4
... ... . .
...
TAT
A . .
AC.
A . A , .
GOT
CTG
...
ACA
. C . . A .
. G .
.G.
. . . . . C
CAG .G. . G .
QCT
..c
...... ......
GGT
CCA
... . . c ... . . C ... . . C ... 0 . C
TAT
............
IGQ ACC CCT
GCC
M
............
TOO T T C C A G C A G 1.. I... . . . . . . . . . . . ............ 7 . . 1..
~
AAC
CGG
. - G . - G
. . G
..G
AAA
GQL
T I C
TCA
- . C - . C . - C
... ..c
... ... ...
40 CCT
TCA
010
GAG
GGG * c c
' . . A
. .G !
. . T
. . T
. . T
. . I
GAC
. . . .
GCT
GAG
. . . . . . . . . . . . . . . . . . . . . . . . . . . c
GAG
. . . . . .
T
K..OI.
I.-01-
m..oZ.
9.'-
K.01-
. Ke.01-
.
L ~ GC T G L C C G T C C T A
. . . . . . . . . ......... . . . . . . . . .
..*
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
00 GGA
.c.
M CCT
.........
CAG
.G#.... . . . . . . . . . . . . .
GGC
...
... .c. ... .c.
... . c .
GQT
............
.............................. . . . . . . . . . . . . . . . . . . . . . . . . .
..G
GTA TTC . . G . . G I . . .
. . T
. . T ..G . . T ..G . . I. . G
CTQ
.........
. . . . . . . . . ... . . A . . . . . . . . . . . . . . . .................. ..................... . . . . . . . . . . . . I . .. . . . . . . . A . . . . . . . . . . . . . . . . Q .
TIC C.T C . 1 C . 1 I . .C . 1
an GGT
......
.........
.........
10
AQT
............
.........
GGC
ACC
............ ............
.........
ATT
..T . . T .AT .GT
a c Q ~TC
ca-l
.........
CTQ
A . . 1.. I..
QCA
............ ............ ............
AQC
............
TCC
. . . . . . .
. . . . . .
TAC
TAT
n w
sa$m
ACT
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
-Vila
msz
LSCI
GIG
.. .. .. .. .. .. .. .. ..... .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ....... .. .. .. .. .. .. .. .. .. .. .. .. .. .. ....... . . . . .. .. .. ... .. .. . G ......................................................... .........................................................
GTQ
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..c . . . . . . . . . 1.. . . . . . . . . . . . . . . . . . . . c . . . . . . . . . . . . . . . . . . . . . . . . . . ..t
I C C
..c
G r n H V 1 V I I GGC
aAr. sa5m
LWZ m.32
..C . - c
GCA . . C
GCT .OC .QC .QC .GC
6..
Q . .
Q . .
ACT 0 . .
... ...
...
P5m8
am.
C A I
... ... ... ...
TQI
.. .. .. ... ...
CAG
O m V l V U
w
LsF1
-qva n w
-
LWZ a 2
1 =,m 1 1 .
-ViW
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
J1ZIJ
?
P
W 01
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
REPERTOIRE OF HUMAN ANTI-HIB ANTIBODY
c
-1
37
Int Rev Immunol Downloaded from informahealthcare.com by McMaster University on 11/03/14 For personal use only.
38
R. A. INSEL ef al.
L
I-, I-
