Naturally acquired functional antibody responses to group A Streptococcus differ between major strain types

Streptococcus pyogenes (group A Streptococcus [GAS]) is associated with a significant global burden of disease, across a diverse spectrum of clinical syndromes ranging from pharyngitis and impetigo to severe invasive infections and serious post-infectious sequelae such as acute rheumatic fever (1). A surge of invasive GAS infections in multiple countries has occurred contemporaneously with the relaxation of COVID-19 pandemic restrictions (2, 3). Immune protection against GAS is poorly understood. However, the epidemiological observation that GAS disease is most common among children and the elderly suggests that immunity accumulates through exposure until immune senescence in older age (1).

The GAS M-protein was first described as an immunodominant antigen and postulated as the major protective antigen by Rebecca Lancefield over 70 years ago. The hypervariable N-terminal region (HVR) of the protein is the basis of strain typing (emm-typing). Functional, emm-type-specific antibodies (where function is defined as the induction of phagocytosis) have been detected in humans decades after infection (4), and vaccination of animals with M-proteins (and related HVR peptide vaccines) can induce strong, protective immune responses (5). Structurally, M-proteins are elongated coiled coils with a gradient of sequence diversity from highly strain specific at the N-terminus (HVR) to very well conserved at the C-terminus (6). M-proteins vary in length depending on the extent of variable- and repeat-regions and have been categorized into a “three representative protein model” based on emm pattern type, with A–C patterns expressing the longest M-proteins, D-patterns intermediate, and E patterns the shortest (6). Most knowledge on emm-type specific immunity in humans comes from historical studies of A–C pattern strains. However, limited data for other strain types suggest that the relationship between emm-type and immunity is more complex (7).

To explore the contribution of emm-type-specific antibodies to GAS immunity across patterns, we utilized optimized opsonophagocytic killing assays (OPKAs) (8) (Supplemental methods) in combination with intravenous immunoglobulin (IVIG) as a surrogate for population-level immunity (Fig. 1A). IVIG is a clinical blood product containing antibodies purified from thousands of healthy donors and has been similarly used to gain insight into Streptococcus pneumoniae immunity (9). The presence of GAS-specific antibodies in IVIG preparations is well documented, and it is used in the treatment of patients with invasive GAS infections (10). We investigated the functional antibody response for three different emm-types, each a prototype for the three major emm-patterns (emm12 for A-C pattern, emm53 for D-pattern, and emm75 for E-pattern) and included two clinical isolates for each (Table S1). The presence of emm-type-specific antibodies in IVIG (Privigen, CSL Behring) was confirmed using ELISA to 50-mer peptides derived from the HVR as well as the matched full-length recombinant M-proteins (Fig. S1) following published protocols (7) (Supplemental methods), with high titers observed for each of the three strain types (Fig. 1B; Fig. S2). The full-length recombinant M-proteins were shown to be folded with a high proportion of α-helical content by circular dichroism as expected for coiled-coil proteins (Fig. S1). Sera from rabbits vaccinated with the M-proteins were used as controls for type specific anti-M-antibodies as described (8) (Supplemental methods) and similarly showed high titers to homologous HVR peptides and M-proteins, whereas pre-vaccination showed no reactivity (Fig. 1B; Fig. S2).

IVIG enabled opsonophagocytic killing of every GAS isolate across the three emm-patterns (Fig. 1C). Homologous anti-M-protein rabbit sera also induced killing of both emm12 and emm53 isolates, yet neither of the emm75 isolates was killed by homologous sera in an OPKA (Fig. 1C). This suggests that the functional opsonophagocytic activity of M-protein targeted antibodies differ between strains. To dissect this further, specificity assays were undertaken by pre-adsorbing IVIG and rabbit sera with M-protein to deplete emm-type specific antibodies prior to the OPKA (Fig. S3). Complete (100%) inhibition of killing was achieved for emm12 and emm53 when rabbit sera were pre-adsorbed with homologous M-protein and very little or no inhibition was observed with heterologous M-protein pre-absorption (Fig. 1D; Fig. S4). This confirms that killing of emm12 and emm53 GAS by rabbit anti-sera was entirely due to M-protein specific antibodies, as expected. As killing of emm75 isolates was not observed with matched rabbit sera, specificity assays could not be conducted. High inhibition of killing was also observed for both emm12 (80.5%) and emm53 (79.5%) when IVIG was pre-adsorbed with homologous M-proteins, and little or no inhibition was observed with heterologous M-protein pre-adsorption (Fig. 1D; Fig. S4). Thus, functional antibodies in IVIG, which induce emm12 and emm53 killing, were predominantly M-protein specific. In contrast, IVIG induced-killing of emm75 strains was not inhibited by M75 protein pre-adsorption (Fig. 1D; Fig. S4), suggesting that killing of the emm75 strains was not M-protein mediated.

The absence of functional, emm-type-specific antibody mediated killing in both matched rabbit sera and IVIG against emm75 compared with emm12 and emm53 highlights distinct differences between the pattern E and the pattern A–C and D strains investigated. This may be driven by differences in the size as well as the functional and structural characteristics of the M-proteins on pattern A–C and D strains, compared with M-proteins expressed by pattern E strains (6, 11). Flow cytometry was performed with anti-M-protein rabbit serum and one representative clinical GAS isolate per emm-pattern-type. This included an Fc blocking step to reduce non-specific interactions (Supplemental methods) and showed markedly reduced binding of matched anti-M-protein sera to emm75 compared with emm12 and emm53 (Fig. 1E). This suggests reduced accessibility of anti-M antibodies to the shorter, pattern E M-proteins on the surface of the GAS bacterium, or alternatively, reduced M-protein expression in these strain types. Either explanation has the same functional consequence; population-level immunity, as represented by functional antibody responses in IVIG, does not appear to be M-protein mediated for emm75 strains in contrast to strains from the other major emm-pattern types examined (A–C and D). The antigenic basis of emm75 killing remains to be determined, but the recently developed emm75 human challenge model provides a novel means for further investigation (12, 13).

The lack of M-protein mediated killing of the emm75 E pattern strains in this study challenges the dogma that M-protein is the major protective antigen across all GAS strain types. This study was limited by the number of strains investigated. The findings may not be reflective of all pattern E strains, but the findings nevertheless suggest that naturally acquired, protective antibodies found in IVIG target alternative antigens on some GAS strains. This does not preclude the possibility that future vaccines, based on either type-specific or conserved vaccine antigens, will be broadly protective across the spectrum of GAS strains. There are likely to be significant differences between infection- and vaccine-acquired immunity, as has been shown for other vaccine preventable diseases, including S. pneumoniae(9). Overall, these findings provide important insight into the biological basis of naturally acquired immunity for GAS and highlight the importance of strain selection when investigating correlates of protection.

This study was supported by funding from the School of Medicine Foundation at the University of Auckland (#3714694) and the Heart Foundation of New Zealand (#3718052).