The search for a safe and effective vaccine against group A beta-hemolytic Streptococcal (GABHS) infections has been ongoing for over 60 years. There are currently several different approaches being investigated, which can be divided into two broad categories: ( 1 ) vaccines containing virulence determinants that are conserved among the different serotypes of GABHS and (2) vaccines that are based on the type-specific, protective regions of the surface M proteins. Strategies for vaccine development are based on our present understanding of the pathogenesis of GABHS infections and the virulence determinants that contribute to pathogenesis. Group A streptococci are human-specific pathogens that are maintained in the population and disseminated among individuals through symptomatic infections of the mucous membranes and skin or asymptomatic carriage, particularly of the pharynx. The organisms are spread by droplets or by hand-to-mouth contact. Attachment to mucosal surfaces is thought to be an important first step in the pathogenesis of infection. A number of ligands that mediate adherence to specific host cells have been defined, including lipoteichoic acid, several fibronectin binding proteins, and M proteins.1 Firm attachment allows the organisms to colonize the mucosal surface, where production of extracellular toxins and possibly specific invasins promote tissue damage and penetration into deeper tissues, respectively.
Once in contact with blood or exúdate, the organisms must resist opsontzation and phagocytosis in order to establish infection. The surface M protein has been considered the major determinant of resistance to phagocytosis.2 Some M proteins bind plasma fibrinogen, which coats the bacterial surface and blocks activation of the alternate complement pathway and nonspecific deposition of C3b.3 M protein and fibrinogen also bind factor H, a potent regulator of the complement cascade, which prevents the generation of C3b.4 Some M proteins and M-like proteins bind immunoglobulins through nonimmune mechanisms, which may mask the surface of the organism and prevent nonspecific immune recognition. Another important virulence factor is the hyaluronate capsule, which for some serotypes appears to be the major determinant of resistance to phagocytosis.5 Group A streptococci also express C5a peptidase that specifically inactivates the potent chemoattractant and reduces the influx of PMNs into the area of infection.6 Extracelluar products of group A streptococci that have an important impact on the host include the pyrogeníc exotoxins, SPE A, B, C, and F. SPE A, C, and F have both toxic and superantigenic properties and likely play an important role in the pathogenesis of severe infections and streptococcal toxic shock syndrome.7 SPE B is also a precursor for a cysteine protease, which appears to be an important determinant of virulence.8,
COMMON PROTECTIVE ANTIGENS OF GABHS
Several of the virulence determinants referred to above contain common epitopes shared by most or all serotypes of GABHS and are actively being evaluated as potential vaccine candidates. C5a-peptidase is expressed by at least 40 different serotypes of group A streptococci. The peptidase specifically cleaves CSa1 which inactivates its ability to attract PMNs into the site of infection. Thus, it is reasoned that the presence of neutralizing antibodies against this streptococcal enzyme would protect against infection by allowing a brisk initial influx of PMNs that follow a steeper grathent of C5a into infected tissue. Indeed, mice immunized intranasally with recombinant C5a-peptidase were protected from pharvngeal colonization after challenge with several serotypes of GABHS.6
All GABHS also produce an extracellular cysteine protease, which is derived from SPE B.s The protease cleaves many biologically active molecules and appears to play an important role not only in pathogenesis of infection but also indirectly in mediating the intense inflammatory response to these organisms. SPE B mutants of GABHS are less virulent in a mouse model of intraperitoneal infection,8 confirming the important role of the extracellular cysteine protease in pathogenesis. Passive and active immunization with SPE B resulted in some protection from death after challenge infections with virulent streptococci.9
Other surface structures that are common among all or many serotypes of GABHS include the hyaluronate capsule, the group carbohydrate, fibronectin binding proteins, and the carboxy-terminal regions of M proteins.15,11 The hyaluronate capsule is identical in structure to the hyluronic acid in mammalian tissues, and thus is poorly immunogenic. The group carbohydrate evokes antibodies in humans that are opsonic, but the exact role of these antibodies in protection against infection is not yet clear.10 GABHS express several fibronectin binding proteins, and in some cases these have been shown to mediate adherence of the organisms to mucosal surfaces.12 In theory, these common antigens may be useful vaccine components that elicit antibodies that block adhesion and colonization, thereby preventing infection at its earliest stage.
Another shared antigen of GABHS is the carboxyterminal region of the surface M proteins.11 The M proteins are serologically distinct by virtue of their amino-terminal hypervariable domains (see below), but the C-terminal region that contains the so-called C repeats is highly conserved and exposed on the surface in such a way that allows antibody binding. C repeat peptides have been shown to evoke mucosal antibodies that protect mice from colonization and death after intranasal challenge infections with heterologous serotypes of GABHS,11,1 Presumably, the secretory antibodies block attachment and colonization, thus preventing infection. This reasoning is supported by the finding in some laboratories that the C repeat antibodies are not opsonic and therefore could not mediate protection against systemic or invasive infection.11,3
VACCINES BASED ON TYPE-SPECIHC, AMINO-TERMINAL REGIONS OF M PROTEINS
All of the vaccine candidates described above are common antigens or peptides that are designed to evoke broadly protective immune responses after immunization. Although there are sufficient data in animals to suggest that these virulence determinants may contribute to protective immune responses, the observation remains that most individuals acquire more than one streptococcal infection in a lifetime. GABHS have evolved a complex array of type-specific, protective M protein epitopes that largely accounts for the immunologie diversity of the species.2 We believe that the best data, developed over the longest period of time, support the use of type-specific multivalent M protein-based vaccines that would evoke opsonic (protective) antibodies against the epidemiologically important serotypes of GABHS. These vaccines would of necessity be highly complex, and their development in the past has been hampered by some significant obstacles that we believe we can now overcome.
Early attempts to develop group A streptococcal vaccines were based on the observation that bactericidal antibodies were directed against the M protein2 and that these antibodies could persist for as long as 30 years following natural infection.14 Because of the documented causal relationship between streptococcal pharyngitis and acute rheumatic fever, it was clear that immunization with whole, killed organisms carried an unacceptable risk for this immunologie complication. Thus, some of the earliest subunit vaccines were M proteins extracted from viable streptococci and purified to varying degrees. These preparations were not well-tolerated by human subjects because they were contaminated with streptococcal extracellular toxins. The toxicity limited the total amount of protein that could be administered, and sufficiently immunogenic doses of M protein could not always be achieved.
The problem of toxicity of M protein preparations was largely overcome by Beachey and colleagues, who determined that dilute solutions of pepsin released significant amounts of M protein from the surface of the organism while leaving the cell wall relatively intact. The M protein from type 24 streptococci extracted with pepsin (pep M24) was purified to homogeneity using standard procedures and was shown to be immunogenic in laboratory animals.15 The purified pepsin-extracted fragment of type 24 M protein was weíl tolerated in human volunteers in doses sufficient to evoke bactericidal antibodies.15
For the first time it was conceivable that M proteins extracted from multiple serotypes of streptococci could be mixed together to form multivalent vaccines and administered to humans without the fear of toxic reactions. We next turned our attention to type 5 streptococci, a highly prevalent, rheumatogenic serotype. Initial experiments in laboratory animals revealed that pep M5 was immunogenic and evoked not only bactericidal antibodies but also antibodies that crossreacted with human myocardium.16 This finding was not totally unexpected, but these studies showed conclusively for the first time that autoimmune epitopes were contained within the covalent structure of type 5 M protein itself. Although there is no direct evidence that M proteins are involved in the immunopathogenesis of rheumatic fever or rheumatic carditis, the presence of tissue-crossreactive epitopes in M proteins raised the theoretical possibility that M proteins themselves may trigger rheumatic fever, one of the very diseases the vaccine is supposed to prevent. This observation led to a series of studies to identify the structures of M proteins that contained protective and tissue-crossreactive epitopes, in hopes of separating the two functional activities so that vaccines could be developed that were free of autoimmune epitopes.
Figure 1. Schematic representation of the functional and structural regions of type 5 streptococcal M protein.
RATIONALE FOR VACCINES CONTAINING AMINO-TERMINAL M PROTEIN FRAGMENTS
Early studies had shown that the pepsin-derived fragments of type 5 M protein evoked opsonic antibodies as well as heart-crossreactive antibodies.16'17 Some of the heart-crossreactive antibodies also reacted with heterologous M proteins, including types 6, 18, and 19, indicating the presence of shared autoimmune epitopes. Structural analyses confirmed that the ammo-terminal regions of M proteins were hypervariable and could account for type-specific immune responses.18 Thus, we reasoned that the amino-terminal structures of M proteins, which were known to be oriented on the outer most surface of the organism,18 would be most likely to evoke type-specific, opsonic antibodies.
The crossreactive M protein antibodies bind to a variety of human tissues and antigens within those tissues. Many of the antibodies crossreact with other alpha-helical proteins such as tropomyosin, myosin, and vimentin.19 Previous studies have shown that M protein antibodies may crossreact with myocardial proteins, antigens in the basal ganglia of human brain, and also with antigens in articular cartilage.17,20,1 Experiments using a series of synthetic and native peptides of M proteins have identified the locations of many of the tissue-crossreactive epitopes (Fig. 1).
Figure 2. Schematic representation of the tetravaleni (top) and octavafent (bottom) hybrid M proteins.
To demonstrate directly that opsonic M protein epitopes could be separated from autoimmune epitopes, we synthesized peptides copying the amino-terminal regions of several M proteins.19 The antisera against these peptides were bactericidal against the respective serotypes of GABHS. Most importantly, none of the immune sera crossreacted with human myocardium. Thus, we had direct evidence that the protective and tissue-crossreactive epitopes of several M proteins were located within different regions of the molecules.
The above discussion summarizes the data that now serve as the basis for our current strategies for M protein vaccine design. The ammo-terminal regions of M proteins contain epitopes that evoke antibodies with the greatest bactericidal activity and are least likely to evoke tissue-crossreactive antibodies. Most of the tissue-crossreactive epitopes have been localized to the B repeats, the A-B flanking regions, or the B-C flanking regions, which are all some distance from the type-specific, amino-terminal epitopes (Fig. 1). Therefore, our current approach is to incorporate limited amino-terminal fragments of multiple M proteins into multivalent vaccine constructs, either as hybrid proteins or individual peptides linked in tandem to unrelated carrier proteins that could be mixed to form multivalent vaccines. The following sections will summarize our data from these different approaches to M protein vaccine development.
RECOMBINANT, MULTIVALENT M PROTEIN VACCINES
We have used recombinant techniques to produce complex hybrid proteins containing amino-terminal peptides of M proteins from different serotypes of GABHS."·23 We first constructed a tetravalent gene thar encoded defined amino-terminal fragments of M24,M5,M6, and MI9" (Fig. 2). PCR primers were synthesized to amplify specific 5' sequences of each emm gene, and each primer was extended to contain a unique restriction enzyme site used to ligate the individual PCR products in tandem. The tetravalent gene contained 113 codons of emm24, 58 codons of emm5, and 35 each from emm6 and emml9.
Rabbits immunized with the recombinant tetravalent protein developed significant antibody levels against all four serotypes of purified native M proteins (Fig. 3A). On the whole, the antisera had higher titers of antibodies against pep M24 and pep M5 than against M6 and M19. The immune sera also contained opsonic antibodies against all four serotypes of group A streptococci, indicating that the antibodies were, in most cases, directed against protective M protein epitopes (Fig. 3B). None of the antisera crossreacted with human heart tissue. These data indicated that the tetravalent protein evoked opsonic antibodies against all four serotypes of group A streptococci, but not all rabbits responded equally to all subunits of the vaccine. The MI9 fragment, located in the carboxyterminal position, was least imrnunogenic. This may have been related to an unfavorable conformation of the peptide in this position. Alternatively, this end of the protein may be susceptible to proteolytic cleavage so that the M19 fragment becomes haptenic and nonimmunogenic after injection. Rather than reconfigure the tetravalent gene to answer these questions, we chose to extend the gene to include four additional M protein fragments.
An octavalent gene23 was constructed using the same approach outlined above for the tetravalent gene (Fig. 2). The additional fragments encoded 35 amino acids each from Ml, M3, M18 and M2. The purified octavalent M protein evoked significant levels of antibodies in rabbits against each of the pep M proteins represented in the hybrid molecule (Fig. 4A). The octavalent protein also evoked opsonic antibodies against six of the eight serotypes of streptococci (Fig. 4B). None of the antisera cTOssreacted with human tissues. The antisera raised against the octavalent protein did not opsonize type 18 or type 2 streptococci, despite the presence of significant levels of antibodies against the respective M proteins. These studies demonstrate the feasibility of evoking broadly protective immune responses against multiple serotypes of group A streptococci using complex hybrid M proteins. Our finding that two of the eight subunits of the hybrid protein did not evoke opsonic antibodies underscores the need for further studies to determine the optimal size and location of the subunits within the hybrid protein.
Figure 3. Immune responses in rabbits immunized with the tetravalent M protein vaccine. ELISA titers against the pep M proteins (A) and percent opsonization of each different serotype of group A streptococci (B) obtained using immune serum from each of three rabbits (represented by the different bars). Data adapted from Dale and colleagues.22
Figure 4. Immune responses in rabbits immunized with the octavalent M protein vaccine. ELISA titers against the pep M proteins (A) and percent opsonization of each different serotype of group A streptococci (B) using immune serum from each of three rabbits (represented by the different bars). Data adapted from Dale and colleagues.23
MUCOSAL DELIVERY OF RECOMBINANT M PROTEIN VACCINES
The recombinant hybtid proteins described above were designed to evoke opsonic antibodies following parenteral injection. Another approach is to develop M protein vaccines that may be delivered via mucosal routes in order to evoke secretory antibodies as well as serum opsonic antibodies. In initial experiments, we constructed a hybrid gene that encoded the entire B subunit of Escheric/iia coU labile toxin (LT-B) linked to 15 ammo-terminal amino acids of type 5 M protein24 (Fig. 5). In this construct, LT-B serves as a carrier for the haptenic peptide and also as a mucosal adjuvant. We immunized groups of mice intranasally with 30 u-g LT-B-M5 or LT-B. Mice that were immunized with LT-B-M5 developed significant levels of serum antibodies against SM5(1-15) and LT-B as determined by ELISA (data not shown). Many of the mice immunized with LT- B-M 5 also developed salivary IgA against the M5 peptide and LT-B. The presence of serum, antibodies after intranasal immunization suggested that the animals may be protected against systemic challenge infections with group A streptococci. Therefore, we challenged both groups with IO6 virulent type 5 streptococci intraperitonealIy, which is a stringent assay for bactericidal antibodies. Of 18 mice immunized with LT-B-M5, only 2 died of infection. Of 20 mice immunized with LT-B, 15 died after intraperitoneal challenge infection (P <0.001, Fisher's exact test).
Figure 5. Schematic representation of the recombinant fusion protein LT-B-M5.
Figure 6. Schematic representation of the recombinant fusion protein LT-B-tetravalent M protein.
Figure 7. Serum immune responses of mice immunized intranasally with LT-B-tetrava)ent M protein. ELISA ODs (A) were obtained using a 1:100 dilution of immune serum. Opsonization assays (B) were performed using undiluted serum. Preimmune sera resulted in ELISA ODs of less than 0.1 and Opsonization levels ?? less than 10%.
These studies showed for the first time that serum opsonic antibodies could be evoked after intranasal immunization with an M protein fragment. We have since extended this observation and have constructed a fusion protein containing LT-B linked to the tetravalent M protein described above (Fig. 6). Mice immunized intranasally with the purified LT-Btetravalent protein developed significant levels of serum antibodies against pep M24, pep M5 and pep M6, but not pep M19 (Fig. 7A). Most importantly, the mice also developed opsonic antibodies against types 24, 5, and 6 streptococci (Fig. 7B). As discussed above, the finding that the M19 component is not immunogenic may be related to its carboxy-terminal location within the hybrid protein. Despite this disappointment, the ability to evoke serum opsonic antibodies against three of the four M protein fragments after intranasal immunization suggests that this approach may be feasible. Mucosal delivery of M protein fragments has the advantage of evoking mucosal IgA, which may block adherence and colonization, and serum opsonic antibodies, which prevent infection at the level of the mucosa and in deeper tissues.
COMPLEXITY AND VALENCEY OF M PROTEIN-BASED VACCINES
The approach of using amino-terminal fragments of M proteins in multivalent vaccine constructs carries with it concerns of complexity and valency. There are currently over 90 different serotypes of group A streptococci, all of which by definition have different amino-terminal M protein sequences. Certainly, a vaccine designed to prevent all of these infections would be highly complex. However, epidemiologie data suggest that not all serotypes of group A streptococci have the ability to trigger acute rheumatic fever.25 In addition, the majority of serious life-threatening group A streptococcal infections appear to be caused by a limited number of serotypes.26 In a recent survey of over 1 100 group A streptococcal isolates collected in the United States, serotypes I, 3, and 18 were more frequently isolated from patients with serious invasive infections, and types 3 and 18 were more frequently recovered from patients with rheumatic fever, compared with control isolates from uncomplicated cases of pharyngitis.26 The serotypes most commonly associated with uncomplicated infections were types 1, 2, 4, and 12. Based on these data and assuming 100% serotypespecific efficacy, immunization with a dodecavalent vaccine could potentially prevent 84% of infections causing rheumatic fever, 73% of those that cause serious infections and 69%, of uncomplicated infections.26
More recent data from the US27 and Japan28 provide further evidence that the majority of serious or uncomplicated GABHS infections are caused by a limited number of serotypes. In North Carolina between 1994 and 1995, 70% of invasive infections were caused by only six different serotypes.27 During the same period, 88% of uncomplicated cases of pharyngitis were caused by the same six serotypes. Ml and M3 serotypes were responsible for 50% of invasive infections and 60% of uncomplicated infections. In the same series, only 17% of all isolates were nontypeable.27 Interestingly, during approximately the same period of time, similar data were reported from Japan.28 Serotypes Ml and M3 accounted for the majority of invasive and noninvasive infections and together represented 80% of all isolates.28 Only 7% of the total 112 isolates were nontypeable in this series, and only a total of eight serotypes accounted for the remaining 93% of group A streptococci recovered from invasive or uncomplicated infections.28
Based on these recent epidemiologie data, we believe that multivalent M protein-based vaccines containing protective peptides from a limited number of serotypes could have a significant impact on the overall incidence of streptococcal disease. The longterm goal is to design and test vaccines containing peptides from 12 to 16 different M serotypes based on current and historical epidemiologie data. Future constructs will contain protective peptides from all of the known rheumatogenic serotypes, the serotypes associated with invasive disease (many of which are also rheumatogenic), and those that are frequent causes of uncomplicated pharyngitis. We ultimately envision a safe and efficacious multivalent vaccine that will be administered to pre-school aged children to prevent the majority of cases of GABHS pharyngitis and its complications, as well the rare cases of invasive, lifethreatening disease.
1. Hasty, DL, Ofek I, CouitneyHS, Doyle RJ. Multiple adhesions of streptococci, infect Immun. 1992;60:2147-2152.
2. Lancefield RC. Current knowledge of die type specific M antigens of group ? strcpiococci. J Immunol. 1962;89:307-313.
3. Whitnack E, BeacheyEH. Antiopsonic activity of fibronugen bound to M protein on the surface of group A streptococci.; CIm Invest. 1982:69:1042-1048.
4. Fischeni VA, Horstman RD, Panchoii V Location oí the complement factor H binding site on streptococcal M protein, infect immun. !995;63:149-156.
5. Dale JB, Washbum RG. Marques MB, Wessels MR· Hyaluronate capsule and surface M protein in resistance to phagocytosis of group A streptococci. Infect immun. 1996;64:1495-1501.
6. Ji Y, Caitson B, Kondagunta A, Cleaiy PP. Intranasal immunization with C5a peptidase prevents nasophaiyngeal colonization of mice by the group A streptococcus. Infect Immun. 1 997:65:2080-2087.
7. Kotb M. Bacterial pyiogenic enotoxins as superaniigens. CIm Microfcwl Rev. 1995:8:411-426.
8. LukotnsVi S, Sreevaisan S, Arnfcag C, Reichardt W, WoischniV M, Podbielslti A, Musser JM. Inactivation of Streptococcus pyogenes extracellular cysteine protease significantly decreases mouse lethality of serotype M3 and M49 strains. JCIm Incesi 1997;9:2574-2580.
9. Kapur V, Maffei JT. Gréer RS, Li LL, Adams GJ, Musser JM. Vaccination with sireptococcal extracellular cysteine processe (interleukin-1 beta converrà«) protects mice against challenge with heterologous group A streptococci. Microo Pathogen. 1994; 16:443-450.
10. Salvadori LG, Blake MS. McCarty M, Tai JV. Zabriskie JR Group A streptococcus. liposome EUSA antibody titers to group A polyiaccharide and opsonophagocytic capabilities of the antibodies. } infect Dis. 1995;171:593-600.
11. BessenD, Fischeni VA. Influence of intranasal immunization with synthetic peptides corresponding to conserved epitupes of M protein on mucosal colonization by group A streptococci. Inject Immun. 1988; 56: 2666- 2672.
12. Courtney HS, Dale JB, Hasty DL. Differential effects of the streptococcal nbronecrinbinding protein, FBP54. on adhesion of group A streptococci to human buccal cell: and Hep-2 tissue culture cells, In/id Immun. 199664:2415-2419.
13. BfonicMS, Courtney HS, DaIeJB. Epitopes of group A streptococcat M protein that evoke cross-protective local immune responses .J Immunol. 1992; 146:888-893.
14. Lancefield RC. Persistence of type-specific antibodies in man following infection with group A streptococci.; ?f Mid. 1959:110:271-282.
15. Beachey EH, Stolkrman GH1 Johnson RH, Ofek 1, Bisno AL Human immune response tu immunization with a structurally defined polypeptide fragment of streptococcal M protein. } Exp Mid. 1979:150:862-877.
16. Dale JB, Beachey EH. Protective antigenic determinant of streptococcal M protein shared with sarcoletnmal membrane ptotein of human heart / Exp Med. 1982;156:1165-1176.
17. Dale JB, Beathey EH. Multiple Keart-cioas-reBCtwe epitopes of streptococeal M proteins. J Exp Med. 1985;161:113-122.
18. Rschetti VA. Streptococcal M protein: Molecular design and biological behavior. Clim Microbiol. Rev. 1989;2:285-314.
19. Beachev EH. Brorue MS, Dale JB, Kraus W, Poirier TP, Sargent S]. Protective and autoimmune epitopes of streptococcal M protein vaccine. Vaccini 19S8;6:192-196.
20. Baird RW, Brome MS, Kraus W, Hill HR. Veasey LG, Dale JB. Epitopes of group A sireptococcal M protein shared with antigens of articular cartilage and synovium. J lmmunol. 1991;146:3132-3137.
21. Bronze MS, Dale JB. Epitopesofstreprococcal M proteins that evoke antibodies that cross-read with human brain. J Immunol. 1993;151:2820-2828.
22. Dale JB, Chiang KY, Lederer J W. Recombmani tetravalent group A streptococcal M ptotein vaccine. J Immunol. 1993;151:2188-2194.
23. Dale JB, Simmons M, Chiang EC. Chiang KY. Recombinant, octavalenl group A streptococcal M protein vaccine. Vaccine. 1996; 14:944-948.
24. Dale JB, Chiang EC. lnuanasal immunization with tecombinant group a streptococcal M protein fragment fused ro the B subunit of Eicfiericlua coli labile toxin protects mice against systemic challenge infections. J Infect Dis. 1995:;71:1038-1041.
25. Bisno AL. The concept of theumatogenic and nonrheumalogenic group A streptococci. In: Reed SE, Zabriskie JB, eds. Sireptococcol Diseases and aie immune Response. New York: Academic Press; 1980:789-796.
26. Johnson DR, Stevens DL, Kaplan EL. Epidemiologie analysis of group A stteptoccal seiotypes associated with severe systemic infections, rheumatic fever, or uncomplicated pharyngitis. J Infecí Dis . 1992:166:374-382.
27. Kiska DL, Thiede B, Caraccioln J, et al. Invasive group A streptococcal infections in North Carolina: Epidemiology, clinical features, and genetic and serotypc analysis of causative organisms. J In/m DÍ. 1997:176:992-1000.
28. Nakashima K, lchiyama S. linuma Y, et al. A clinical and bactériologie investigation of invasive streptococcal infections in Japan on the basis of serotypes. toxin production, and genomic DNA fingerprints. Clmln/eciDu. 1 997; 2 5-260- 266.