Opsonic antibody activity against Actinobacillus actinomycetemcomitans in patients with rapidly progressive periodontitis

Vaccines against periodontitis: a forward-looking review

Periodontal disease, as a polymicrobial disease, is globally endemic as well as being a global epidemic. It is the leading cause for tooth loss in the adult population and has been positively related to life-threatening systemic diseases such as atherosclerosis and diabetes. As a result, it is clear that more sophisticated therapeutic modalities need to be developed, which may include vaccines. Up to now, however, no periodontal vaccine trial has been successful in satisfying all the requirements; to prevent the colonization of a multiple pathogenic biofilm in the subgingival area, to elicit a high level of effector molecules such as immunoglobulin sufficient to opsonize and phagocytose the invading organisms, to suppress the induced alveolar bone loss, or to stimulate helper T-cell polarization that exerts cytokine functions optimal for protection against bacteria and tissue destruction. This article reviews all the vaccine trials so as to construct a more sophisticated strategy which may be relevant in the future. As an innovative strategy to circumvent these barriers, vaccine trials to stimulate antigen-specific T-cells polarized toward helper T-cells with a regulatory phenotype (Tregs, CD4+, CD25+, FoxP3+) have also been introduced. Targeting not only a single pathogen, but polymicrobial organisms, and targeting not only periodontal disease, but also periodontal disease-triggered systemic disease could be a feasible goal.

Immunization to Porphyromonas gingivalis enhances the local pro-inflammatory response to subcutaneous bacterial challenge

BACKGROUND, AIMS

Human and animal studies have suggested that immunization to P. gingivalis might be beneficial for controlling periodontitis, by the induction of protective antibody response. The present study was designed to examine the effect of immunization on the local cellular, cytokine and antibody response to P. gingivalis in mice.

METHODS

Subcutaneous chambers were implanted in 3 groups of mice. 2 groups were then immunized with P. gingivalis in either incomplete Freund's (IFA) or an Alum-based adjuvant. The 3rd group served as the control. At baseline, all mice were challenged with an intra-chamber injection of P. gingivalis. Chamber exudates were sampled at baseline, 1 and 7 days post-challenge, following by determination of leukocyte counts and the cytokines TNF-alpha, IFNgamma (pro-inflammatory) and IL-10 (anti-inflammatory). IgG levels to P. gingivalis were analyzed in both the exudates and serum.

RESULTS

Leukocyte accumulation increased in the chambers over the study period and was more marked in the immunized groups. P. gingivalis challenge induced the expression of the tested cytokines in all groups. Levels of IFNgamma showed a significantly greater increase in the immunized groups on day 1 post-challenge. By day 7, the levels in the controls had reached those of the immunized groups. IL-10 levels were significantly higher in the control group compared to the immunized groups on day 1 and by day 7 they were reduced significantly in all groups to barely detectable levels. While there were no significant differences in TNF-alpha levels between IFA and control groups, they were significantly higher in the Alum group on day 0 and 7. Both immunization protocols induced anti-P. gingivalis IgG. The Alum group achieved the highest antibody levels, which were due to the increased expression of IgG1, a marker of a Th2-response.

CONCLUSIONS

The results suggest that immunization to P. gingivalis results in enhanced expression of pro-inflammatory, tissue-destructive cytokines in the inflammatory site. The nature of the adjuvant used for immunization allows manipulation of the T-cell response, and alum was more effective in reducing the inflammatory response than IFA.

Prior exposure of mice to Fusobacterium nucleatum modulates host response to Porphyromonas gingivalis

Multiple periodontal pathogens sequentially colonize the subgingival niche during the conversion from gingivitis to destructive periodontal disease. An animal model of sequential immunization with key periodontal pathogens has been developed to determine whether T- and B-lymphocyte effector functions are skewed and fail to protect the host from pathogenic challenge. The present study was performed to evaluate the immunomodulatory effect of exposure to Fusobacterium nucleatum prior to Porphyromonas gingivalis. Group 1 (control) mice were immunized with phosphate-buffered saline, group 2 were immunized with F. nucleatum prior to P. gingivalis and group 3 were immunized with P. gingivalis alone. All the T-cell clones derived from group 2 demonstrated type 2 helper T-cell clone (Th2 subsets), whereas those from group 3 mice demonstrated Th1 subsets. Exposure of mice to F. nucleatum prior to P. gingivalis interfered with the opsonophagocytosis function of sera against P. gingivalis. In adoptive T-cell transfer experiments, in vivo protective capacity of type 2 helper T-cell clones (Th2) from group 2 was significantly lower than type 1 helper T-cell clones (Th1) from group 3 against the lethal dose infection of P. gingivalis. Western blot analysis indicated a different pattern of recognition of P. gingivalis fimbrial proteins between sera from group 2 and group 3. In conclusion, these studies suggest that exposure of a host to F. nucleatum prior to the periodontal pathogen P. gingivalis modulates the host immune responses to P. gingivalis at the humoral, cellular and molecular levels.

Fimbriae of Porphyromonas gingivalis induce opsonic antibodies that significantly enhance phagocytosis and killing by human polymorphonuclear leukocytes

Porphyromonas gingivalis has been strongly implicated in the pathogenesis of human periodontitis. Fimbriae mediate adherence and colonization of the oral cavity by this organism and may, therefore, have potential for use as antigen in an anti-P. gingivalis vaccine. The purpose of our study was to determine whether P. gingivalis fimbriae have opsonic target sites and whether they are accessible on the cell surfaces and cross-reactive among P. gingivalis fimbrial types and serotypes. Rabbits were immunized with a vaccine. The antiserum reacted with a 43-kDa fimbrillin monomer and a 43-kDa component in whole-cell sonicates of P. gingivalis 33277, but it showed only very weak reactivity in the 43-kDa region of Western blots of a whole-cell sonicate of strain DPG3, a mutant that does not express functional fimbriae. The antibody enhanced chemiluminescence approximately six-fold relative to preimmune serum values and significantly enhanced phagocytosis and killing of P. gingivalis 33277 by human polymorphonuclear leukocytes. Peak opsonic activity was observed at week 6 followed by a plateau that remained until week 16. The fimbria-deficient mutant DPG3 did not bind antifimbrial antibody and was not opsonized, whereas strain 381, the parent of the mutant, was opsonized. The specific antibody bound to and opsonized P. gingivalis strains 33277 and 381 (fimbria type I) but not W50, A7A-1-28, 9-14K-1 or FAY-19M-1 (fimbrial types II-V). Specific antibody bound to strain 2561 (fimbrial type I) but, as assessed by chemiluminescence, did not opsonize it. While fimbriae have opsonic target sites that are accessible on P. gingivalis cell surfaces, the relevant opsonic target sites do not appear to be shared across serotypes or fimbrial types. Thus, a vaccine containing, as antigen, fimbrial protein from a single P. gingivalis strain would likely be ineffective against infections by P. gingivalis strains expressing other fimbrial types.

Functional characteristics of antibodies induced by Arg-gingipain (HRgpA) and Lys-gingipain (Kgp) from Porphyromonas gingivalis

Arginine-specific gingipain (HRgpA) and lysine-specific gingipain (Kgp), enzymes produced by Porphyromonas gingivalis, may be candidates for an anti-P. gingivalis vaccine. The purpose of our study was to determine whether HRgpA and Kgp have opsonic target sites and whether these sites are available and accessible on intact P. gingivalis cells. Rabbits were used to generate polyclonal antibodies to both proteins. Animals were immunized and immunoglobulin G (IgG) fractions were isolated from preimmune and immune sera. Functional characteristics of the antibodies were assessed by determining antibody titers by enzyme-linked immunosorbent assay (ELISA), generating Western immunoblots, and measuring antibody enhancement of P. gingivalis opsonization, phagocytosis and killing by polymorphonuclear leukocytes (PMN) of intact cells of strains of P. gingivalis representative of the four serotypes. Strains studied included 33277 (serotype A), A7A1-28 (serotype B), W50 (serotype C) and 381 (serotype D). Both HRgpA and Kgp induced high titers of IgG antibody. Anti-HRgpA and anti-Kgp bound to both HRgpA and Kgp demonstrating a large proportion of shared antigenic epitopes. The two antibodies bound equally well to all four P. gingivalis serotypes with titers ranging from 77 to 205 ELISA units when compared to preimmune IgG set at 1 ELISA unit. The immunoblot patterns of binding of the two antibodies to HRgpA and Kgp and to sonicates of the four P. gingivalis serotypes were virtually identical. Both antibodies detected components in HRgpA at 27, 35 and 45 kDa and in Kgp at 27, 32, 35, 40 and 55 kDa. The antibodies also detected components at or near these same positions in addition to multiple high molecular mass components in the cell sonicates of P. gingivalis. Both proteins induced antibodies that significantly enhanced opsonization as assessed by chemiluminescence, with values ranging from 130 mV to 375 mV for anti-HRgpA IgG and from 240 mV to 475 mV for anti-Kgp IgG. Both antibodies significantly enhanced PMN-mediated bacterial killing of the four P. gingivalis serotypes, although the percentage of killing varied among the serotypes (24-81% for anti-HRgpA and 37-89% for anti-Kgp). Thus, both HRgpA and Kgp express opsonic target sites and induce high titers of antibodies that opsonize and enhance killing of all four serotypes of P. gingivalis. These two proteins appear to be potential candidate antigens for an anti-P. gingivalis vaccine.

Antigenic cross-reactivity among Porphyromonas gingivalis serotypes

The goal of our research program is to develop a Porphyromonas gingivalis vaccine. Vaccine development requires identification of antigenic components shared by the many clonal types of P. gingivalis. The purpose of the present study was to evaluate the extent and nature of antigenic cross-reactivity among serotypes of P. gingivalis and to identify shared antigenic components. Strains selected to represent serotypes A-D were 33277, A7A1-28 W50 and 381, respectively. Using intact cells, antibodies were raised in rabbits. Titers were assessed by enzyme-linked immunosorbent assay (ELISA) using intact cells as antigen, Western blots were prepared and biologic activity was measured as opsonization (chemiluminescence expressed as mV) and enhancement of phagocytosis and killing by polymorphonuclear leukocytes. Extensive cross-reactivity that varied greatly among serotypes was observed by ELISA. The Western blots showed an even greater extent of cross-reactivity, with shared protein components at approximately 140, 130, 37, 32 and 28 kDa and a shared variable molecular mass smear considered to be lipopolysaccharide and other carbohydrate. Additional protein components at 110, 85, 35 and 20 kDa appeared to be shared by some but not all serotypes. In the functional assays, strains 33277 and 381 were equally well opsonized by anti-33277 and anti-381 (500-650 mV) but opsonized to a much lesser extent by anti-A7A1-28 and anti-W50 (roughly 125 mV and 350 mV respectively). A7A1-28 and W50 were opsonized by all four immune sera almost equally but to a much lower extent (roughly 400 mV and 250 mV respectively). Enhancement of phagocytosis and killing in the presence of active complement mirrored opsonization with the exception that 381 was reasonably well opsonized by anti-A7A1-28 (400 mV) and anti-W50 (350 mV), but poorly killed. The protein components at 140, 130, 37 and 28 kDa shared by all of the four serotypes appear to have potential as vaccine candidate antigens.

Recognition of antigenic epitopes in lipopolysaccharide and protein from Actinobacillus actinomycetemcomitans by serum antibodies in untreated rapidly progressive periodontitis patients

Actinobacillus actinomycetemcomitans has been associated with early-onset periodontitis, including the localized juvenile and rapidly progressive forms. The immunodominant antigens of A. actinomycetemcomitans recognized by rapidly progressive periodontitis patients remain unidentified. Sera from 22 patients with rapidly progressive periodontitis and 20 periodontally normal subjects were tested by enzyme-linked immunosorbent assay (ELISA) for immunoglobulin G antibodies to whole-cell sonicate, protein, purified lipopolysaccharide and lipopolysaccharide fractions of A. actinomycetemcomitans. The median titers of rapidly progressive periodontitis patients and control subjects to whole-cell sonicate were 25.0 and 14.5 ELISA units, respectively (not significantly different). Binding of antibody from patient sera occurred to both the lipopolysaccharide and the protein fractions, with greater binding to lipopolysaccharide than to protein. We show for the first time that patient sera contain antibodies that bind specifically to antigenic epitopes in lipid A and in the core carbohydrate of lipopolysaccharide that were previously considered to be inaccessible and unavailable, as well as to epitopes in the O side chains. Sera manifesting antibody titers 2-fold or greater than the median titer for control sera were judged to be seropositive. More patients were seropositive for lipid A than for any of the other antigen preparations studied, and the median titer for patient sera to lipid A but to none of the other purified lipopolysaccharide fractions was significantly elevated relative to control values. Of 22 patients, 10 were seropositive to whole-cell sonicate, 7 to protein, 8 to lipopolysaccharide, 7 to the high-molecular-weight lipopolysaccharide-polysaccharide fraction rich in O side chains, and 16 to lipid A. The core carbohydrate did not adhere to the test plate surface, and this precluded ELISA measurements. However, when the core carbohydrate was used in the ELISA inhibition assay, it reduced antibody binding to lipopolysaccharide-coated plates by up to 45%, thereby demonstrating antibody binding to core carbohydrate. The core carbohydrate fraction from the Re mutant of Salmonella minnesota known to contain no O-side chains also inhibited binding of specific antibody to plates coated with A actinomycetemcomitans lipopolysaccharide. Overall, there was extreme variation in responses among patients to the various antigen preparations, with no single pattern dominating. Lipopolysaccharide and its components appear to be the immunodominant epitopes, since most rapidly progressive periodontitis patients are seropositive for lipopolysaccharide and/or its components and they have titers relative to those for proteins.

Serum immunoglobulin G responses to various Actinobacillus actinomycetemcomitans serotypes in a young ethnographically heterogeneous periodontitis patient group

Sera from young patients with periodontal diseases have been shown to often contain highly elevated antibody levels to Actinobacillus actinomycetemcomitans, in particular serotype b. Such responses were reportedly predominated by antibodies of the immunoglobulin G2 (IgG2) subclass. The aim of this study was to investigate an ethnically diverse group of 14 early-onset periodontitis and 15 rapidly progressive periodontitis patients for the occurrence of elevated antibody titers against the five known A. actinomycetemcomitans serotypes, and to compare the patient's IgG subclass response profiles. Enzyme-linked immunosorbent assays were used to measure both total IgG and subclass specific IgG titers. Twenty-four subjects had markedly elevate total IgG levels against at least one serotype. The frequencies of high responses against serotypes a, b, c, d and e were 7, 11, 6, 4, and 4, respectively. Elevated antibody responses were predominated by IgG2, regardless of the serotype to which the response was directed. The serotype specificity of the host responses was further investigated by competitive binding studies with serotype-specific monoclonal antibodies. Twelve sera were found to contain antibodies capable of strongly inhibit the binding of monoclonal antibodies against a single serotype; four other sera had antibodies against epitopes of two, and one serum against those of three serotypes. The findings document broad serotype diversity in an ethnically heterogeneous group of patients and indicate that strong antibody responses to A. actinomycetemcomitans are predominated by IgG2 regardless of the serotype of the infective agent.

Shared antigens of Porphyromonas gingivalis and Bacteroides forsythus

Periodontitis in humans is caused by a group of predominantly gram-negative, anaerobic bacteria among which Porphyromonas gingivalis and Bacteroides forsythus are prominent. A similar group is present and presumably plays a similar role in experimental periodontitis in the primate Macaca fascicularis. Nevertheless, immunization using a vaccine containing only killed P. gingivalis suppresses the progress of experimental periodontitis in M. fascicularis. We investigated the hypothesis that gram-negative periodontopathic bacterial may share antigens, and immunization with one species may induce antibodies reactive with other gram-negative species. Using enzyme-linked immunosorbent assay (ELISA), Western and dot immunoblots with nonabsorbed and absorbed and immune and preimmune sera we show that monkeys immunized with P. gingivalis produce antibodies reactive not only with antigens of P. gingivalis but also with those of B. forsythus. Similarly, rabbits immunized with P. gingivalis or with B. forsythus produce antibodies that react with antigens of both bacteria. Cross-reactive antibodies bind to epitopes in lipid A and possibly in core carbohydrate of lipopolysaccharide. Using complexes of lipopolysaccharide with polymyxin B, bovine serum albumin and apolipoprotein A1 specificity of binding was documented. Using sera from monkeys immunized with P. gingivalis, cross-reactivity with Actinobacillus actinomycetemcomitans could not be demonstrated by ELI-SA, although binding to lipopolysaccharide but not to lipid A was demonstrated by Western and dot immunoblots. Antibodies to shared lipopolysaccharide epitopes of periodontopathic bacteria may account, at least in part, for the immune protection observed in immunized monkeys, and shared epitopes may have potential as a vaccine for periodontitis in humans.

Immunogenicity of Actinobacillus actinomycetemcomitans serotype b-specific polysaccharide-protein conjugate

To enhance the immunogenicity of capsular-like serotype b-specific polysaccharide antigen (SPA) of Actinobacillus actinomycetemcomitans, the purified antigen was coupled with bovine serum albumin via an adipic acid hydrazide functional group. The conjugate (SPA-bovine serum albumin) or the native SPA was administered subcutaneously or intranasally to BALB/c mice. Neither subcutaneous immunization with SPA emulsified in Freund adjuvant nor intranasal immunization with SPA and cholera toxin B subunit elicited any antibody responses to the polysaccharide antigen. High serum immunoglobulin M (IgM) and IgG responses to SPA were induced by subcutaneous immunization with SPA-bovine serum albumin in Freund adjuvant. Serum and salivary IgA responses to SPA in addition to IgM and IgG responses were induced by intranasal immunization with the conjugate and cholera toxin B subunit. To investigate the functional activity of SPA-specific antibodies, the opsonic activity of sera from BALB/c mice immunized with the conjugate was assessed by chemiluminescence assay using human polymorphonuclear leukocytes. Murine antisera efficiently opsonized A. actinomycetemcomitans serotype b in the assay, suggesting that antibodies to SPA of the organism might serve as a protective role.

Humoral immunity to commensal oral bacteria: quantitation, specificity and avidity of serum IgG and IgM antibodies reactive with Actinobacillus actinomycetemcomitans in children

The levels, specificity and avidities of serum IgM and IgG antibodies reactive with Actinobacillus actinomycetemcomitans (Aa) serotypes a, b and c were determined in periodontally healthy (PH) children and compared with subjects with localized juvenile periodontitis (LJP). All PH children exhibited IgM and IgG Aa-reactive antibodies whether or not Aa was detected subgingivally but the antibodies were not specific for Aa. In contrast, LJP sera contained high concentrations of IgM and IgG antibodies reactive with Aa that were largely specific for this bacterium. IgM and IgG antibodies in both PH and LJP subjects were of low avidity. With one exception, the avidities of IgG anti-Aa antibodies were significantly greater than those of IgM antibodies in both PH and LJP subjects. However, although the LJP subjects had as much as 115-fold more Aa-reactive IgG antibody than did the PH subjects the avidities of their IgG antibodies were no greater than those of the PH group. The induction by the host of low-avidity antibodies, that are ineffective in immune elimination, may be a reason why commensal bacteria persist at mucosal surfaces and why persons with LJP fail to eliminate Aa from their periodontal pockets.

Critical issues in periodontal research

The purpose of this paper is to highlight briefly the major achievements and the remaining critical issues in the areas of epidemiology, microbiology, pathogenesis, diagnosis, and therapy. Periodontitis affects a relatively small proportion of study populations in the United States and other countries. Prevalence may be decreasing, but that remains to be seen. The identity and characteristics of susceptible individuals and groups are not known, and risk indicators for severe disease are only beginning to be identified. A very large number of different microbial species has been implicated in the etiology. It seems unlikely that all of these are essential participants. Essential participants need to be identified and better characterized. Whether putative pathogens are members of the commensal flora or exogenous species that must be transmitted is unclear. The relationship between the presence of a pathogenic flora and disease status is obscure. Pathogenic bacterial species are essential, but insufficient to cause disease. A susceptible host and local environmental factors--for example, elevated iron concentration--may be necessary for disease to occur. Many clonal types may not be virulent, and numbers greater than certain threshold levels appear to be necessary. The pathways by which bone and connective tissues of the periodontium are destroyed are sufficiently understood to permit development of therapies aimed at their modification. Examples are the use of vaccines, topical application of anti-inflammatory drugs, and use of chemically modified tetracyclines.

Serum immunoglobulin G subclass concentrations in periodontally healthy and diseased individuals

Patients with localized juvenile periodontitis (LJP) often have high titers of antibody reactive with the serotype-specific immunodominant carbohydrate antigen of Actinobacillus actinomycetemcomitans serotype b. The vast majority of this A. actinomycetemcomitans serotype b-specific antibody is immunoglobulin G2 (IgG2). The present study was undertaken to determine whether the overall total levels of IgG2 in the sera of LJP patients are elevated. LJP patients and nonperiodontitis (NP) controls matched for age, race (black and white), and gender were studied. Additional controls included patients with adult periodontitis (AP) and patients similar in age to LJP patients but with the more-severe, generalized form of early-onset periodontitis (SP). Sera from over 700 periodontally characterized subjects were examined by using radial immunodiffusion to quantitate IgG2 as well as IgG1, -3, and -4, which were included for comparison. Serum IgG2 levels increased with age, and this was most dramatic around puberty. Black subjects in all periodontal groups had nearly 1 mg more IgG2 per ml than their white counterparts. Serum IgG2 levels were elevated (about 30 to 40%) in LJP patients of both races compared with their age- and race-matched NP controls (P < 0.01). In contrast, SP patients and AP patients had IgG2 levels comparable to their age- and race-matched NP controls. No other IgG subclass concentration correlated with periodontal diagnosis except for IgG3, which was elevated in white LJP patients. We reason that the high levels of serum IgG2 in LJP may be helpful in localizing periodontal destruction.

Immunization against Porphyromonas gingivalis inhibits progression of experimental periodontitis in nonhuman primates

Periodontitis is a common infectious disease in which the attachment tissues of the teeth and their alveolar bone housing are destroyed, resulting in tooth loss. The gram-negative anaerobic microorganism Porphyromonas gingivalis has been closely linked to severe forms of the disease. We show for the first time that immunization of the primate Macaca fascicularis with killed P. gingivalis in Syntex Adjuvant Formulation-M inhibits progression of periodontal tissue destruction.

Effect of treatment on titer, function, and antigen recognition of serum antibodies to Actinobacillus actinomycetemcomitans in patients with rapidly progressive periodontitis

Although periodontal treatment by scaling and root planing (SCRP) is known to induce bacteremia, the effect of this procedure on the host immune response is not known. We have determined pre- and post-SCRP immunoglobulin G antibody titers to antigens of Actinobacillus actinomycetemcomitans in the sera of 22 patients with rapidly progressive periodontitis. We also assessed the ability of these sera to enhance phagocytosis and killing of A. actinomycetemcomitans by human polymorphonuclear leukocytes by using a polymorphonuclear leukocyte chemiluminescence (CL) assay. Specific anti-A. actinomycetemcomitans antibody titers were significantly increased at 6 and 12 months after beginning treatment, and CL values were significantly increased at 12 months, whereas mean interproximal pocket depths were significantly decreased at 12 months after beginning treatment. When patients were classified as either seropositive (twice the median titer of control subjects; n = 10) or seronegative (n = 12), both median titers and CL values were significantly increased for the seronegative group at 6 and 12 months after treatment. In the seropositive group, only the median titer was significantly increased at 12 months. Western blot (immunoblot) patterns for six seronegative and six seropositive patients differed remarkably at the baseline. Before treatment, all of the seropositive patients recognized high-molecular-mass lipopolysaccharide (LPS) and a large number of protein components. Patterns were virtually unaffected by therapy. Before treatment, only one of the seronegative patients recognized the LPS smear and none reacted strongly with protein components. Following treatment, slight LPS staining was observed for five of six seronegative patients and detection of protein bands was enhanced in all cases. We conclude that treatment by SCRP induces a humoral immune response, especially in seronegative patients, and that response may play a role in the observed beneficial effects of periodontal treatment.

Titer and subclass distribution of serum IgG antibody reactive with Actinobacillus actinomycetemcomitans in localized juvenile periodontitis

Most patients with localized juvenile periodontitis (LJP) manifest serum IgG antibodies specifically reactive with antigens of Actinobacillus actinomycetemcomitans serotype b (Aa-b). Whether these antibodies are protective, destructive, or irrelevant to the progress of the disease remains unclear. We report results of studies aimed at assessing the subclass IgG responses in 35 LJP patients and 35 periodontally normal control subjects using well-characterized monoclonal antibody subclass reagents in an enzyme-linked immunosorbent assay. Our data show that the mean value for total IgG reactive with antigens of Aa-b was more than sevenfold higher for patients than for normal control sera (2349.6 micrograms/ml for patients vs 332.2 micrograms/ml for controls). Individual patients and control subjects were classified as high- or low-titer, using twice the median value for total anti-Aa-b IgG in control sera as the cutoff. Of 35 patients, 26 (74%) were high-titer, and 9 (26%) were low-titer. This compares to 5 normal control subjects (14%) high-titer and 30 (86%) low-titer. IgG2 accounted for the major quantitative response in both patients and control subjects. Indeed, the mean IgG2 values for both concentration and percentage of total specific IgG were greater than the combined values for specific anti-Aa-b IgG1, IgG3, and IgG4. Of the 26 high-titer sera, IgG2 predominated in 24, with IgG1 and IgG3 predominating in 1 each; IgG2 predominated in only 2 of the low-titer sera.