The GAS strain 90-226 is an M1 serotype strain originally isolated from a sepsis patient (37), and is herein referred to as GAS-M1. GAS-2W.M1 is a genetically engineered strain in which the 14 amino acid peptide 2W (EAWGALANWAVDSA) has been inserted in frame after the first five amino acids of the mature M1 protein of strain 90-226 (29). All strains were grown without shaking in Todd-Hewitt broth supplemented with 0.2% yeast extract (THY) in 5% CO₂ at 37^°C.

Female C57Bl/6 (wildtype, [B6]) mice were purchased from Taconic and IFN-γ-KO (B6.129S7-Ifng^tm1Ts/J) mice (generated on 129S7/SvEvBrd-Hprt1⁺ , backcrossed 8 generations on C57BL/6) were originally from The Jackson Laboratory. CD4-Cre (B6.CgTg(Cd4-cre)1Cwi/BfluJ) and Bcl6^fl/fl (B6.129S(FVB)Bcl6 ^tml.1Dent/J) mice, both on B6 background, were crossed to generate CD4-Cre x Bcl6^fl/fl animals lacking the T follicular helper (Tfh) cell master transcription factor Bcl6 specifically in CD4⁺ T-cells. For experiments using genetically modified mice, both females and males were used. Experiments were initiated at animal ages of 8 – 12 weeks. All animals were bred (except B6) and maintained at the animal facility at the Lund University Biomedical Center, and experiments were performed in accordance with protocols approved by the Lund/Malmö Animal Ethics Committee.

Mice were immunized using 10⁸ heat-killed (HK) GAS-M1 or GAS-2W.M1. Bacterial over-night cultures were re-inoculated and grown to OD= 0.8 - 1, after which cultures were washed in PBS and incubated for 2 hours at 60˚C. HK cultures were diluted to appropriate concentration, aliquoted and stored in -20°C. Complete bacterial killing was confirmed by plating an aliquot on blood agar for counting. Mice were injected subcutaneously (sc) three times with three-week intervals and bled 19 days after each injection. For protection experiments, a lethal dose of 10⁸ CFU GAS-M1 was distributed intraperitoneally (ip) three weeks after the last immunization injection. Infected mice were monitored every four hours for the first day and then twice daily for one week. After infection, cages were blinded to avoid bias in determining moribundity.

To determine antibody titers in murine serum, HiBond ELISA plates were coated with whole GAS-M1 bacteria, or recombinant M1 protein or isolated M1 HVR [generated as described in (38)] and incubated with serially diluted serum. Captured antibodies were detected using biotinylated goat anti-mouse IgG (Southern Biotech, CAT 1036-08), IgG1 (BioLegend, RMG1-1), IgG2b (BioLegend, RMG2b-1), and IgG2c (Southern Biotech, CAT 1079-08) and streptavidin-HPR (BioLegend, CAT 405210). BSA-coated wells served as negative controls and presented values represent the average of duplicate experimental values with deducted negative control values. All samples were run in duplicate.

Anti-CD16/CD32 (BD, 2.4G2 [anti-FcγR]) was used as a blocking reagent and PI-PECF594 (Invitrogen, CAT P3566) was used to identify dead cells. The following antibodies and reagents were used for germinal center B-cell staining: anti-IgD-BV-605 (BioLegend,11-26c.2a), anti-B220-APCeFlour780 (eBioscience, RA3-6B2), anti-CD38-Alexa 700 (eBioscience, 90), and anti-CD95-PE-CY7 (BD, Jo2). For T follicular helper cell (Tfh) staining, anti-CXCR5 (BD, 2G8), anti-rat IgG (fab’) Alexa-647 (Jackson Immuno Research, Code:712-606-153). 2W-specific T-cells were identified using a 2W:I-Ab tetramer labeled with PE (50959 I-A(b)) or BV421(50960 I-A(b) (NIH Tetramer Core Facility), and the following antibodies: anti-CD3-Alexa flour 700 (BioLegend, 17A2), anti-CD4-FITC (BioLegend, RM4-5), anti-B220-APCeFlour780 (BioLegend, RA3-6B2), and anti-PD1-PE-CY7 (eBioscience, J43). The Click-iT™ Plus EdU Alexa fluor 647 Flow cytometry kit was used for EdU staining of newly synthesized DNA (Invitrogen, CAT C10634). Porcine cardiac myosin (Sigma-Aldrich, CAT: M0531) and a cardiac myosin monoclonal antibody (ThermoFisher, CAT: MA1-26180) was used for investigation of myosin-reactive antibody responses.

Single cell suspensions were prepared by mechanical disruption of inguinal lymph nodes, and filtering of cells through a 70 μM strainer. The total number of white blood cells per lymph node was determined in a Sysmex hematology analyzer (Sysmex) and divided into two parts for detection of T cells and germinal center B cells. Cell suspensions were incubated with anti-FcγR in 10% rat serum to minimize Fc receptor-dependent background staining, after which the appropriate antibody mix was added. Dead cells were identified using PI-PECF594. For tetramer staining of 2W-specific T cells, cells positive for both tetramers were considered 2W-specific. Intracellular IFN-γ was detected in T cells stimulated with PMA, ionomycin and brefeldin A.

Data were analyzed with Prism version 8.0 (GraphPad Software). Analysis of statistical significance was performed using one-way ANOVA with Tukey’s multiple comparisons test for three or more groups, or Mann-Whitney U test for two unpaired groups. Comparison of Survival Curves was done by Kaplan-Meier survival test with Log-rank (Mantel-Cox) test to evaluate p value. Differences were considered significant when p ≤ 0.05.

All experiments conducted were approved by the Ethical Committee for Laboratory Animals in Lund/Malmö, permit numbers: 7342/2017 and 07178/2020.

B6 female mice were immunized subcutaneously (sc) with 10⁸ heat-killed (HK) bacteria ( Figure S1 ), using either the strain 90-226 (serotype M1; GAS-M1) or the genetically engineered 90-226 strain variant, expressing a chimeric M1 protein with the immunodominant CD4 Th cell 2W epitope inserted in-frame after the first five amino acids of the mature M1 surface protein (GAS-2W.M1) (29). Control mice received PBS only. All mice were immunized three times with three weeks intervals and then injected intra-peritoneally (ip) with a lethal dose of wt GAS-M1 (10⁸ CFU) three weeks after the last immunization ( Figure 1A ). Thereafter the animals were monitored for seven days. Consistent with the results of Lannergård et al. (15), repeated immunization with wt GAS-M1 did not result in protective immunity, and mice subjected to this treatment exhibited equally poor survival to the lethal challenge as control mice receiving PBS only ( Figure 1B ). In contrast, immunization with GAS-2W.M1 resulted in a significantly enhanced survival following subsequent challenge with GAS-M1 – i.e. the strain not containing the 2W epitope ( Figure 1B ).

To assess if the enhanced survival of mice immunized with the recombinant GAS-2W.M1 strain correlated with enhanced antibody response against the parental GAS-M1 strain, blood was collected 19 days after each immunization, and sera were assayed for reactivity against intact GAS-M1 bacteria, the M1 protein, and the M1 HVR domain. While levels of IgG against intact GAS-M1 did not differ between mice immunized with GAS-M1 or GAS-2W.M1 ( Figure 1C ), levels of anti-M1 and anti-HVR IgG were higher in the GAS-2W.M1 immunized animals ( Figures 1D, E ). The post-GAS-infection autoimmune complication acute rheumatic fever (ARF) is believed to involve cross-reactive immune reactions between the M protein and cardiac myosin (39, 40). The ability of the GAS-2W.M1 strain to enhance IgG responses against the M protein and its HVR was however not associated with increased levels of IgG against cardiac myosin ( Figure 1F ), indicating that incorporation of the 2W Th cell epitope into the M protein does not unleash such autoimmune responses. Finally, we assessed the levels of distinct IgG subclasses against intact GAS-M1. We found no difference in IgG1 or IgG2b between the two immunized groups ( Figures 1G, H ), however the geometric mean IgG2c titer was significantly higher in GAS-2W.M1- than in GAS-M1-immunized animals ( Figure 1I ).

Altogether, these results show that insertion of the CD4 Th cell epitope 2W into the N-terminal part of the M1 protein gives rise to a recombinant strain with an enhanced ability to induce protective immunity against the parental wild type (wt) GAS-M1 strain, and that this protection is associated with increased IgG responses against the M1 protein as well as a more IgG2c-biased response against the bacteria.

To determine if the ability of GAS-2W.M1 to confer protection against the parental GAS-M1 strain involves T cell-dependent antibody responses, we used CD4-Cre.Bcl6^fl/fl mice with a specific deletion of the transcriptional repressor Bcl6 in T cells. In these mice, germinal centers (GC) and associated high-affinity antibody responses fail to develop due to impaired T follicular helper (Tfh) cell development (41, 42). While Bcl6^fl/fl littermate controls (Cre^-) developed robust anti-GAS-M1 IgG responses after GAS-2W.M1 immunization, CD4-Cre.Bcl6^fl/fl mice completely failed to do so ( Figure 2A ), demonstrating that the IgG response elicited by GAS-2W.M1 immunization is T cell-dependent. In subsequent survival experiments, immunized CD4-Cre.Bcl6^fl/fl mice were equally sensitive to the lethal dose of GAS-M1 as unimmunized littermates ( Figure 2B ). Compared to both of these groups, immunized Cre^- control mice displayed increased survival ( Figure 2B ). Although this difference did not quite reach significance (p=0.055), collectively the results indicate that T cell-dependent antibodies contribute to the overall protection induced against GAS-M1 following immunization with GAS-2W.M1.

Children are more susceptible to superficial GAS infections than adults and this correlates with reduced IgG3 and IFN-γ production (20). In mice, IFN-γ is promoting IgG2c responses, and given the increased IgG2c response observed after GAS-2W.M1 immunization (see Figure 1I ), we next set out to compare IFN-γ production by GAS-specific CD4 T cells in mice immunized with GAS-M1 or GAS-2W.M1. B6 mice were again injected three times with PBS or HK bacteria (GAS-M1 or GAS-2W.M1). To track the CD4 T cells specifically responding to the immunizations, all mice received the thymidine analogue ethynyl deoxyuridine (EdU) via ip injection two days after the third immunization, which specifically labels all cells undergoing cell division during the in vivo pulse period (43). Two days after the EdU injection, mice were sacrificed and cells from draining inguinal lymph nodes were analyzed by flow cytometry following a brief ex vivo stimulation with PMA and ionomycin. The percentage of CD4 T cells that had incorporated EdU during the two-days pulse period was ~3-4 fold higher in mice immunized with HK bacteria relative to mice receiving PBS only, demonstrating that labelling with EdU to a high extent reports CD4 T cells proliferating in response to the GAS immunizations ( Figures 3A, B ). There was however no significant difference in EdU labelling between the GAS-M1 and the GAS-2W.M1 groups, indicating that insertion of 2W into the M1 protein does not lead to an enhanced overall expansion of GAS-specific CD4 T cells. Next, we focused on IFN-γ production by the CD4 T cells responding to the bacterial immunizations by gating on CD4⁺CD44⁺ (i.e. activated effector and memory) T cells with detectable EdU labelling. Among the EdU-labelled CD4⁺CD44⁺ T cells, a significantly higher percentage of cells produced IFN-γ in mice immunized with the GAS-2W.M1 strain as compared to mice immunized with GAS-M1, lacking the 2W epitope ( Figure 3C, D ). To determine if the 2W epitope itself contributes to this enhanced IFN-γ response, we used a 2W peptide-MHCII (2W:I-A^b) tetramer to identify 2W-specific CD4 T cells (29). As expected, CD4 T cells binding to the 2W:I-A^b tetramer were present in mice immunized with GAS-2W.M1 but not detected in mice immunized with GAS-M1 or mice receiving PBS only ( Figures 3E, F ). Focusing on the GAS-2W.M1 immunized animals only, we next compared IFN-γ production by 2W:I-A^b positive and negative CD4⁺CD44⁺ T cells, respectively, also labelled by EdU. This approach allowed us to directly compare the 2W:I-A^b-specific Th cells with activated effector and memory Th cells responding to peptide:MHCII complexes containing GAS-derived peptides distinct from the 2W epitope (derived from M1 or other GAS proteins). As shown in Figures 3G and H , the percentage of IFN-γ producing cells was significantly higher among the 2W:I-A^b-positive than the 2W:I-A^b-negative subset. Taken together, these results suggest that incorporation of the 2W Th cell epitope into the M1 protein generates a GAS-M1 strain that drives enhanced IFN-γ responses, as compared to its isogenic parental strain.

Next, to assess if IFN-γ is important for the protection conferred by GAS-2W.M1 immunization, we used IFN-γ deficient mice. Mice were again immunized sc three times with HK GAS-2W.M1. Following lethal challenge with GAS-M1, IFN-γ deficient mice displayed a significantly reduced survival as compared to immunized IFN-γ sufficient mice ( Figure 4A ). Despite reduced survival in the absence of IFN-γ, GAS-2W.M1 immunization resulted in equally robust total IgG responses in the IFN-γ deficient and sufficient animals; evident when measuring IgG binding to whole bacteria, the M1 protein, or to the M1 HVR domain ( Figures 4B–D ). In marked contrast, IgG2c responses against GAS-M1, including those targeting the M1 protein and its HVR domain, were strongly reduced in the IFN-γ deficient mice ( Figures 4E–G ). Altogether, these results show that GAS-2W.M1 immunization confers protection against the parental GAS-M1 strain through an IFN-γ-dependent mechanism, and that reduced survival in the absence of IFN-γ production is associated with a specific loss in anti-bacterial IgG2c responses.

As GC B cells and Tfh cells appear to be important for the protection conferred by GAS-2W.M1 immunization (see Figure 2B ), we next determined how these cellular subsets develop in GAS-2W.M1 immunized IFN-γ deficient and sufficient animals, respectively ( Figure 5A ). In agreement with the requirement to immunize multiple times to induce antibody responses and protective immunity against GAS-M1, there was no expansion of CD95⁺ CD38^- B220⁺ GC B cells in any of the mouse strains after primary immunization with either GAS-M1 or GAS-2W.M1 ( Figure S2 ). For all three immunized groups, GC B cells became detectable after secondary immunization and increased further after the third dose ( Figure S2 ). After three doses of HK GAS-2W.M1, GC B cell expansion occurred to the same extent in IFN-γ sufficient and deficient mice, also comparable to the percentage of GC B cells observed in B6 mice immunized with HK GAS-M1 ( Figures 5B, C ). Likewise, Tfh cells, identified as CXCR5⁺PD-1⁺ Th cells, had developed to the same extent in mice immunized with GAS-M1 and GAS-2W.M1 and were not affected by the absence of IFN-γ following GAS-2W.M1 immunization ( Figures 5D, E ). The percentage of CXCR5⁺PD-1⁺ Tfh cells among 2W:I-A^b-binding CD4 T cells did also not differ between IFN-γ deficient and sufficient mice ( Figures 5F, G ). As expected, CD4 T cells binding to the 2W:I-A^b tetramer were not detected in unimmunized or GAS-M1-immunized wt mice (data not shown). Therefore, and in line with our results on the total GAS-specific serum IgG responses (see Figure 1C ), the insertion of the immunodominant CD4 Th cell epitope 2W in the N-terminal part of the M1 protein does not lead to expanded GCs, and neither is the reduced protection observed in IFN-γ deficient animals after GAS-2W.M1 immunization associated with a reduced expansion of GC B cells.