Fig 2: IL-17A inhibition partially restores endothelial transcriptional programs after GAS infections but worsens BBB dysfunction in vivo.a, Timeline of recurrent intranasal GAS infections and administration of an α–IL-17A–neutralizing monoclonal antibody (mAb) or isotype control. b, Heat maps of differentially expressed genes (DEGs) related to BBB function, LPS response, interferon signaling, and antigen presentation in olfactory bulb (OB) brain endothelial cells (BECs) from GAS-infected mice treated with isotype control or α–IL-17A mAb. Values are shown as log(z-score); significantly altered genes are indicated in black (adjusted p < 0.05). c, Gene ontology (GO) pathway enrichment analysis of transcripts upregulated and downregulated in α–IL-17A mAb–treated versus isotype-treated BECs following GAS infections. d, Representative immunofluorescence (IF) images of endogenous IgG leakage (green) in the granular layer of the OB. Blood vessels are labeled with Glut1 (cyan). e, f, Quantification of IgG extravasation (relative fluorescence intensity) in the glomerular (e) and granular (f) OB layers from PBS- or GAS-infected mice treated with α–IL-17A mAb or isotype control. Comparisons were performed using two-way ANOVA with Šídák’s multiple comparisons test (p < 0.05; **p < 0.001; n = 3–7 mice per group). g–i, Representative IF images for IFITM3 (pink), Iba1 (yellow), and CD31 (blue) in the OB of GAS-infected mice treated with isotype or α–IL-17A mAb (g,h), and quantification of Ifitm3-positive area within CD31+ vasculature (i). j,k, Representative fluorescence in situ hybridization (FISH) images of Itm2a mRNA (pink) combined with Glut1 (blue) in the glomerular layer (j) and quantification of vascular Itm2a signal (k). l, Representative IF images of the tight junction proteins Claudin-5 (green) and ZO-1 (red) in the OB vasculature of GAS-infected mice treated with isotype or α-IL-17A mAb. m, Serum cytokine concentrations measured by multiplex immunoassay in PBS- or GAS-infected mice treated with isotype or α–IL-17A mAb. Data are mean ± SEM. Comparisons were performed using one-way ANOVA with Tukey’s multiple comparisons test (ns, p > 0.05; p < 0.05; *p < 0.01; **p < 0.001; ***p < 0.0001; n = 3 - 6 mice per group). n, Survival curves of GAS-infected mice treated with isotype control or α–IL-17A mAb (n = 13–48 mice per group). Statistical significance was assessed using the Mantel–Cox log-rank test.

Fig 3: Model summarizing the putative effects of IFNγ, IL-17A, and GM-CSF on brain endothelial and microglial responses to recurrent GAS infections.a–c Proposed model illustrating how effector cytokines produced by Th1 and Th17 lymphocytes that infiltrate the brain after repeated GAS infections may differentially shape transcriptional responses in brain endothelial cells (BECs) and microglia. a IFNγ is predicted to induce interferon-response programs and antigen-presentation pathways in both microglia and BECs. b, c IL-17A and GM-CSF exert partially overlapping effects on microglial activation, including regulation of proliferative responses and induction of disease-associated microglial (DAM) genes, cytokines, and chemokines, but also have cytokine-specific effects. IL-17A signaling may affect more BEC transcriptional alterations following GAS infections. In addition, antigen-presentation signatures remain elevated in conditions of Th17 deficiency or systemic IL-17A neutralization, suggesting that IL-17A may partially suppress these pathways in both microglia and BECs. The model highlights shared and distinct roles for Th17-associated cytokines in shaping neuroimmune and vascular responses in the brain after recurrent GAS exposure in the periphery.

Fig 4: IL-17A inhibition partially restores endothelial transcriptional programs after GAS infections but worsens BBB dysfunction.a Experimental timeline. b, c Heatmaps and GO pathway enrichments of DEGs in OB BECs from GAS mice with two treatments (n = 5 GAS Isotype and n = 7 GAS α–IL-17A mAb mice). Significant genes (adjusted p < 0.05; MAST with Benjamini–Hochberg correction) are in black. d–f IF images of IgG leakage (green) outside OB vessels (GLUT1; cyan) and quantification. Significance by two-way ANOVA with Šídák’s test [(e) GAS Isotype vs α–IL-17A mAb **p = 0.0153; PBS α–IL-17A vs GAS α–IL-17A mAb ***p = 0.0012; f GAS Isotype vs α–IL-17A mAb **p = 0.0388; PBS α–IL-17A vs GAS α–IL-17A mAb ***p = 0.0034; 4–6 mice/group). g–i IF for IFITM3 (pink), IBA1 (yellow), and CD31 (blue) and quantification of IFITM3+ area within OB vessels. Significance assessed by two-way ANOVA with Sidak’s test. h PBS vs. GAS Isotype (*p = 0.0280); PBS vs. GAS α–IL-17A mAb (*p = 0.0470). i PBS vs. GAS Isotype (*p = 0.0455); PBS vs. GAS α–IL-17A mAb (**p = 0.013); 4–6 mice/group. Scale bars are 50 µm (main) and 10 µm (inset). j Survival curves (13–48 mice per group; Mantel–Cox log-rank test; ****p < 0.0001). k, l FISH images for Itm2a mRNA (pink) and GLUT1 (blue) and quantification. Scale bars are 50 µm (main) and 10 µm (inset). Two-way ANOVA with Sidak’s test p values: GAS Isotype vs. α–IL-17A mAb (*p = 0.0333); PBS vs. GAS α–IL-17A mAb (*p = 0.0199); 4–6 mice/group. m Serum cytokine concentrations. Significance by two-way ANOVA with Šídák’s test (GAS Isotype vs. GAS α–IL-17AmAb; and PBS vs. GAS α–IL-17A mAb, respectively are: IFNγ **p = 0.0088; **p = 0.0097), CCL4 (***p = 0.0002; ****p < 0.0001), CCL5 (***p < 0.0001; ***p = 0.0002), CXCL2 (*p = 0.0127; **p = 0.0029), CXCL10 (***p = 0.0073; **p = 0.0009), GM-CSF (***p = 0.0012; **p = 0.0008); n = 4 for PBS, 6 GAS mice). Data are mean ± SEM.