Concurrent Brain Subregion Microgliosis in an HLA-II Mouse Model of Group A Streptococcal Skin Infection

The broad range of clinical manifestations and life-threatening infections caused by the Gram-positive bacterium, Streptococcus pyogenes or Group A Streptococcus (GAS), remains a significant concern to public health, with a subset of individuals developing neurological complications. Here, we examined the concurrent neuroimmune effects of subcutaneous GAS infections in an HLA-Class II (HLA) transgenic mouse model of subcutaneous GAS infection. To investigate changes in the skin–brain axis, HLA-DQ8 (DQA1*0301/DQB1*0302) mice (DQ8) were randomly divided into three groups: uninfected controls (No Inf), GAS infected and untreated (No Tx), and GAS infected with a resolution by clindamycin (CLN) treatment (CLN Tx) (10 mg/kg/5 days) and were monitored for 16 days post-infection. While the skin GAS burden was significantly reduced by CLN, the cortical and hippocampal GAS burden in the male DQ8 mice was not significantly reduced with CLN. Immunoreactivity to anti-GAS antibody revealed the presence of GAS bacteria in the vicinity of the neuronal nucleus in the neocortex of both No Tx and CLN Tx male DQ8 mice. GAS infection-mediated cortical cytokine changes were modest; however, compared to No Inf or No Tx groups, a significant increase in IL-2, IL-13, IL-22, and IL-10 levels was observed in CLN Tx females despite the lack of GAS burden. Western blot analysis of cortical and hippocampal homogenates showed significantly higher ionized calcium-binding adaptor-1 (Iba-1, microglia marker) protein levels in No Tx females and males and CLN Tx males compared to the No Inf group. Immunohistochemical analysis showed that Iba-1 immunoreactivity in the hippocampal CA3 and CA1 subregions was significantly higher in the CLN Tx males compared to the No Tx group. Our data support the possibility that the subcutaneous GAS infection communicates to the brain and is characterized by intraneuronal GAS sequestration, brain cytokine changes, Iba-1 protein levels, and concurrent CA3 and CA1 subregion-specific microgliosis, even without bacteremia.


Introduction
β-hemolytic Group A Streptococcus (Streptococcus pyogenes, GAS) is an important human pathogen that causes a broad spectrum of clinical presentations ranging from mild self-limiting pharyngitis and skin infections such as cellulitis and impetigo to lifethreatening and often fatal invasive infections including necrotizing fasciitis (NF, aka "flesh-eating disease"), a subset of necrotizing soft tissue infections (NSTI) [1].NF manifests as fascial necrosis and is usually accompanied by systemic signs of streptococcal toxic shock syndrome (STSS), leading to a cytokine storm and increasing the risk for multiple organ failure and mortality in susceptible individuals [1][2][3][4].Significant inter-individual variations exist in the clinical presentation, disease severity, and outcomes of invasive GAS infections [5,6].These depend on the heterogeneity in the virulence potential among the GAS isolates and the host risk factors, specifically variations in HLA Class II (HLA-II) alleles that determine the type of immune response elicited [7][8][9].Despite prompt interventions, recurrent GAS infections are not uncommon.GAS mimics of host proteins can prime individuals with HLA-II predilection for increased risk of severity, outcomes, and post-streptococcal sequelae, which include neurological manifestations and neuropsychiatric disorders like Sydenham chorea and the Pediatric Autoimmune Neuropsychiatric Disorders Associated with Streptococcal infections (PANDAS) [10] subgroup of tic disorders and attention disorders [11][12][13], and risk for seizures in some cases [14].While GAS meningitis [15] and brain abscesses have been associated with significant morbidity and mortality [16,17], the mechanisms underlying GAS neuropathogenesis remain incompletely understood.As the frontline mediators of innate and adaptive immune signaling within the brain, the brain resident immune cells astrocytes and microglia are poised to play central roles in determining infection outcomes [18,19].Elucidating the responses of astrocytes and microglia and the localized brain cytokines they govern is crucial for understanding host defense strategies, pathological processes, and antibiotic therapy-mediated sequelae during GAS infections.
HLA-II-expressing humanized mouse models are well-established and frequently used as clinical translation models for studying mechanisms underlying dysregulated host immune responses to infections [10,20,21], streptococcal and staphylococcal superantigen responses, dermatitis, collagen-induced arthritis [22][23][24], food allergens [25], dust allergens [26], progression of demyelination and neurological deficits in Theiler's murine encephalomyelitis virus infection [27], and brain pathology in experimental autoimmune encephalomyelitis [28,29].Building upon our previous work [30], and given the potential for the differential presentation of immune responses and outcomes by DR and DQ alleles, in the present study, we utilized the HLA-DQ8 (DQA1*0301/DQB1*0302) mice (DQ8) to investigate the skin-brain axis and effects of clindamycin (CLN) following subcutaneous GAS infection.Our findings offer unique insights into brain cytokine changes, GAS dissemination, and subregion-specific Iba-1 immunoreactivity that endured beyond an apparent CLN-mediated resolution of skin infection.

Mice
20-24 weeks-old females and males expressing the human HLA-DQ8 (DQA1*0301/DQ B1*0302) allele and deficient in endogenous class-II molecules [31] were used as sublethal mouse models of subcutaneous GAS infection.The genotype of the mice was confirmed by PCR-based genotyping using allele-specific oligonucleotide primers Sense 5 -AGG ATT TGG TGT ACC AGT TTA AGG GCA T-3 and Antisense 5 -TGC AAG GTC GTG CGG AGC TCC AA-3 .The surface expression of the HLA-DQ allele was confirmed by flow cytometry (BD Biosciences, Franklin Lakes, NJ, USA) after surface staining whole blood with phycoerythrin (PE) labeled anti-HLA-DQ (clone DQ1) (BioLegend, San Diego, CA, USA) antibody [9].
In vivo infections were performed as described [9,34].One day before infection, mice were anesthetized, and hair on the dorsal side was depilated using Nair cream (Church and Dwight Co., Inc., Lakewood, NJ, USA).Mice were randomly divided into three groups: (Group 1: No Infection (No Inf), Group 2: GAS Infected Untreated (No Tx), and Group 3: GAS Infected and clindamycin treated (CLN Tx)) and were subcutaneously infected with 0.1 mL of GAS suspension (~1 − 5 × 10 8 CFU).No Inf control mice received 0.1 mL of PBS −/− .GAS-infected mice assigned to clindamycin (CLN, Gold Biotechnology, Olivette, MO, USA) treatment received the first dose of CLN (10 mg/kg/0.1 mL) administered intraperitoneally at 4-6 h after GAS infection.The regimen was then continued every day for an additional five days.Mice were monitored for survival, body weight, and lesions for sixteen days post-infection.
2.3.GAS Burden in the Blood, Brain, Skin, and the Subcutaneous Adipose Tissue 0.5-0.7 mL of terminal blood was drawn through cardiac puncture and collected into heparinized tubes (30U heparin, Akron Biotech, Boca Raton, FL, USA).The head was decapitated after blood collection, and the brain was excised rapidly.The left hemisphere was immersed in 4% paraformaldehyde, pH 7.4 (Thermo Fisher Scientific, Waltham, MA, USA) for fixation.The olfactory bulb and the cerebellum were removed from the right hemisphere.The hippocampi (HC) were isolated aseptically from the rest of the cerebral cortex (CC) and were collected into 1 mL sterile ice-cold PBS −/− .Necrotic skin and subcutaneous adipose tissue (SubQ AT) were recovered aseptically in 1 mL of sterile PBS −/− .Tissues were then weighed and homogenized (Omni International, Marietta, GA, USA), followed by plating of tenfold dilutions on blood agar plates for enumeration of bacterial load (calculated as colony-forming units (CFUs) per mL blood or milligram of tissue).The remaining homogenates were centrifuged for 15 min at 12,000× g at 4 • C, and supernatants were stored at −80 • C and used for cytokine profiling and western blot analysis.

Western Blot Detection of the Glial Fibrillary Acidic Protein (GFAP) and Ionized
Calcium-Binding Adaptor-1 (Iba-1) Protein Lysis buffer (2% w/v SDS, 6.25 mM Tris-HCl-7.5, 5mM EDTA) supplemented with protease (Thermo Fisher Scientific, Waltham, MA, USA) and phosphatase inhibitor cocktail (MilliporeSigma, Burlington, MA, USA) [35] was added to the HC and CC homogenates, and centrifuged at 14,000× g for 30 min at 4 • C. The supernatants were aliquoted and stored at −80 • C. Total protein concentration in the HC and CC supernatants was measured using the Pierce BCA Protein assay kit (Thermo Fisher Scientific).50-100 µg of proteins were solubilized in a sample solubilizing buffer, heated for 10 min at 70 • C, and were resolved in Bis-Tris 4-20% SurePAGE gel (Genscript, Piscataway, NJ, USA).The proteins were electrophoretically transferred to the nitrocellulose membrane (Bio-Rad, Hercules, CA, USA).Membranes were blocked for one hour with Intercept protein-free blocking buffer (LI-COR, Lincoln, NE, USA) to avoid non-specific binding and were incubated overnight at 4 • C with primary antibodies against GFAP-1:1000, Cell Signalling Technology, Danvers, MA, USA; Iba-1-1:500, FujiFilm Wako, Richmond, VA, USA; and GAPDH-1:1000, Thermo Fisher Scientific, Waltham, MA, USA, prepared in Intercept blocking buffer (LI-COR) containing 0.2% Tween-20.The membranes were washed in Tris-buffered saline (TBS) containing 0.05% Tween-20 (TBST) and incubated in the dark for 2 h at room temperature with IRDye 680 goat anti-mouse and IRDye 800CW goat anti-rabbit (prepared 1:15,000 in Intercept blocking buffer containing 0.2% Tween-20).Blotted proteins were detected using the Odyssey CLx Imaging System (LI-COR).Western blots were analyzed using Fiji (a distribution of NIH ImageJ), and the average of the band densities was normalized to the loading control GAPDH.

GAS and Iba-1 Immunoreactivity
Paraformaldehyde fixed-left hemisphere of the brain from No Inf, No Tx, and CLN Tx mice were cryopreserved in 30% sucrose for 72 h and subsequently embedded in a Tissue-Tek ® O.C.T TM compound (Thermo Fisher Scientific, Waltham, MA, USA).The brains were sectioned serially between bregma 2.80 mm and −4.24 mm using a cryostat (CM3050S, Leica Biosystems, Deer Park, IL, USA), and every sixth section of 40 µm thickness was collected into six-well plates containing PBS −/− .The brain sections were rinsed in PBS −/− to remove O.C.T and were stored at 4 • C in a cryoprotectant solution of 30% ethylene glycol, 30% sucrose, and 0.02% sodium azide in PBS −/− .Serial brain sections were subjected to antigen retrieval (10 mM Tris, 1 mM EDTA, pH 9, at 95 • C for 10 min), rinsed in PBS −/− , and blocked with 5% goat serum containing 0.3% TritonX-100 in PBS −/− for 1 h at room temperature.The sections were incubated overnight at 4 • C with the primary antibody anti-Group A Streptococcus (1:100, Thermo Fisher Scientific, Waltham, MA, USA) or anti-Iba-1 (1:1000) [36].The following day, the sections were rinsed and incubated with secondary goat anti-rabbit IgG-AF594 antibody (1:1000, Thermo Fisher Scientific, Waltham, MA, USA) for 2 h at room temperature in the dark.Sections were rinsed and counterstained with DAPI (MilliporeSigma, Burlington, MA, USA), mounted onto gelatin-coated glass slides, dried overnight, and coverslipped with Fluoromount (MilliporeSigma, Burlington, MA, USA).

Image Acquisition
Immunodetection of GAS bacteria in the brain was performed using an Olympus Fluoview 3000 confocal laser scanning microscope coupled with an Olympus IX83 microscope (Olympus, Waltham, MA, USA).Z-stack images with 2.5× zoom were obtained using a 60× oil-immersion objective lens (PlanApo N 60×/1.40oil SC ∞/0.17/FN22).All the images were captured with a constant 4.24% laser intensity, 502 HV sensitivity, 1% gain, and 0 offset for DAPI at 405 nm and with 4.42% laser intensity, 584 HV sensitivity, 1% gain, and 0 offset for anti-GAS at 594 nm, with a setup step size of 1.0 µm.Consecutive cross-section images (XYZ) were acquired from top to bottom; after the acquisition, projections of z-stack images were displayed using the Z mode.
For immunodetection of Iba-1 cells, the images were captured at 20× magnification (×20/0.4HC PL Fluotar, CORR PH1 lens) on a DMi8 microscope (Leica THUNDER Imager, Leica Microsystems, Deer Park, IL, USA) connected to a high-resolution CMOS camera (up to ~1000 fps) using the computational clearing feature in the Leica Application Suite X software 3.7.2.22383 (Leica Microsystems, Wetzlar, Germany).Iba-1 immunoreactivity was quantified using images from two sections per mouse and at least three mice per condition using Fiji (a distribution of NIH ImageJ).Subregions of the images were selected using the marque ROI tool, set to 8-bit, preprocessed by adjusting brightness and contrast to facilitate segmentation, and then segmented and counted using a batch-processing macro that incorporated a machine learning segmentation plugin (Trainable Weka Segmentation) [37] and the Analyze Particles feature in Fiji [38][39][40].Before the analysis, the machine learning algorithm was trained by manually defining labeled and unlabeled regions in selected representative images and assigning them to separate classes.The resulting classifier file was then used in the analysis.

Statistical Analysis
Values are presented as mean ± SEM and were analyzed using the two-tailed Student's t-test and one-way or two-way analysis of variance (ANOVA) followed by uncorrected Fisher's LSD test under a statistical threshold of p < 0.05 using Prism ® 10.0.2 (GraphPad Software Inc.San Diego, CA, USA).

Clindamycin Treatment Significantly Reduced Weight Loss and Lesion Areas following Subcutaneous GAS Infection
Our laboratory has developed mouse models of subcutaneous GAS infections in outbred [33], ARI BXD [34], and clinically relevant humanized mice expressing HLA-II alleles associated with varying susceptibility to GAS NSTI/STSS [9].Here, we studied the effect of CLN treatment on survival, changes in body weight, skin lesions, GAS burden in the skin, SubQ AT, and HC and CC regions of the brain following subcutaneous GAS infection in DQ8 mice.There was no mortality associated with subcutaneous GAS infection in the DQ8 mice.Body weight loss has been established as a significant comorbidity associated with poor recovery from subcutaneous GAS infections [33,34].As shown in Figure 1A, compared to uninfected mice (No Inf), weight loss was significant following subcutaneous GAS infection (**** p < 0.0001) (No Tx).Consistent with the significant reduction in weight loss (No Tx vs. CLN Tx **** p < 0.0001; No Inf vs. CLN Tx: * p = 0.0338), CLN treatment also resulted in significant amelioration of lesions in DQ8 mice (No Tx vs. CLN Tx **** p < 0.0001) (Figure 1B).

Statistical Analysis
Values are presented as mean ± SEM and were analyzed using the two-tailed Student's t-test and one-way or two-way analysis of variance (ANOVA) followed by uncorrected Fisher's LSD test under a statistical threshold of p < 0.05 using Prism ® 10.0.2 (GraphPad Software Inc.San Diego, CA, USA).

Clindamycin Treatment Significantly Reduced Weight Loss and Lesion Areas following Subcutaneous GAS Infection
Our laboratory has developed mouse models of subcutaneous GAS infections in outbred [33], ARI BXD [34], and clinically relevant humanized mice expressing HLA-II alleles associated with varying susceptibility to GAS NSTI/STSS [9].Here, we studied the effect of CLN treatment on survival, changes in body weight, skin lesions, GAS burden in the skin, SubQ AT, and HC and CC regions of the brain following subcutaneous GAS infection in DQ8 mice.There was no mortality associated with subcutaneous GAS infection in the DQ8 mice.Body weight loss has been established as a significant comorbidity associated with poor recovery from subcutaneous GAS infections [33,34].As shown in Figure 1A

Heterogeneity in GAS Burden in the Skin, Subcutaneous Adipose Tissue, and the Brain
We next investigated the GAS burden in the blood, skin, SubQ AT, and brain.At sixteen days post-infection, GAS was not detectable in the blood of DQ8 mice.As shown in Figure 2A, No Tx mice showed a significantly higher GAS burden in the skin compared with the CLN Tx group (** p = 0.0018).GAS disseminated into SubQ AT, and CLN treatment reduced SubQ AT GAS burden; however, the difference was insignificant (Figure 2B).Most strikingly, GAS translocated into the CC (Figure 2C) and HC (Figure 2D) regions of the brain.The GAS burden persisted in the brain sub-regions after CLN treatment (Figure 2C,D).To determine whether there was a sex difference in GAS burden, CFUs from female and male mice were compared separately.As shown in Figure 2E-G, CLN effectively reduced the GAS burden in the skin (Figure 2E, * p = 0.0140) of male mice but not in the CC or HC of female or male DQ8 mice.
sixteen days post-infection, GAS was not detectable in the blood of DQ8 mice.As shown in Figure 2A, No Tx mice showed a significantly higher GAS burden in the skin compared with the CLN Tx group (** p = 0.0018).GAS disseminated into SubQ AT, and CLN treatment reduced SubQ AT GAS burden; however, the difference was insignificant (Figure 2B).Most strikingly, GAS translocated into the CC (Figure 2C) and HC (Figure 2D) regions of the brain.The GAS burden persisted in the brain sub-regions after CLN treatment (Figure 2C,D).To determine whether there was a sex difference in GAS burden, CFUs from female and male mice were compared separately.As shown in Figure 2E-G, CLN effectively reduced the GAS burden in the skin (Figure 2E, * p = 0.0140) of male mice but not in the CC or HC of female or male DQ8 mice.
Figure 2. Heterogeneity in GAS burden in the skin, subcutaneous adipose tissue, and the brain.DQ8 mice were infected subcutaneously with 1 − 5 × 10 8 CFU of GAS 5448 and were either untreated or treated with CLN (10 mg/kg) administered intraperitoneally every day for five days.Sixteen days post-infection, necrotic skin lesions, SubQ AT, the right cerebral cortex, and hippocampus were aseptically excised and collected into 1 mL of sterile PBS −/− , weighed, and homogenized to enumerate the GAS burden ((A-D) respectively).GAS burden in female and male mice was compared separately in the skin (E), cerebral cortex (F), and hippocampus (G).Data are presented as mean values + SEM (n = 4-9).Each circle or square represents an individual mouse.Statistical differences were computed by unpaired t-test ((A), ** p = 0.0018) or two-way ANOVA ((E), * p = 0.0140).

GAS Bacteria Localize to the Brain
The dissemination of GAS into the brain following peripheral skin infection presents a novel situation that was unknown previously.Our data demonstrate remarkable sexdependent differences in GAS burden in the brain following subcutaneous GAS infection.Z-stacked images of coronal brain sections stained with anti-GAS antibodies demonstrate the presence of GAS in the vicinity of the neuronal nucleus in the neocortex of both GASinfected untreated and CLN-treated male DQ8 mice (Figure 3).Heterogeneity in GAS burden in the skin, subcutaneous adipose tissue, and the brain.DQ8 mice were infected subcutaneously with 1 − 5 × 10 8 CFU of GAS 5448 and were either untreated or treated with CLN (10 mg/kg) administered intraperitoneally every day for five days.Sixteen days post-infection, necrotic skin lesions, SubQ AT, the right cerebral cortex, and hippocampus were aseptically excised and collected into 1 mL of sterile PBS −/− , weighed, and homogenized to enumerate the GAS burden ((A-D) respectively).GAS burden in female and male mice was compared separately in the skin (E), cerebral cortex (F), and hippocampus (G).Data are presented as mean values + SEM (n = 4-9).Each circle or square represents an individual mouse.Statistical differences were computed by unpaired t-test ((A), ** p = 0.0018) or two-way ANOVA ((E), * p = 0.0140).

GAS Bacteria Localize to the Brain
The dissemination of GAS into the brain following peripheral skin infection presents a novel situation that was unknown previously.Our data demonstrate remarkable sexdependent differences in GAS burden in the brain following subcutaneous GAS infection.Z-stacked images of coronal brain sections stained with anti-GAS antibodies demonstrate the presence of GAS in the vicinity of the neuronal nucleus in the neocortex of both GASinfected untreated and CLN-treated male DQ8 mice (Figure 3).Representative images showing immunoreactivity of the GAS antibody in the cerebral cortex of No Inf, No Tx, or CLN Tx DQ8 male mice.The 40µm thick sections of the brain (left hemisphere) were stained with rabbit polyclonal anti-GAS antibody (1:100) followed by Alexa Fluor 594-conjugated goat anti-rabbit secondary antibody (1:1000) (Red).Nuclei were counterstained with DAPI.Images were captured using an Olympus FV3000 confocal laser scanning microscope coupled with an Olympus IX83 microscope at 60x magnification with 2.5× Zoom (scale = 20 µm).The arrows in No Tx and CLN Tx panels point to positive GAS immunoreactivity localized near the nucleus.Magnified insets in panels No Tx and CLN Tx show GAS bacteria in close proximity to cell nuclei.

Clindamycin Treatment Significantly Altered the Cytokine Profile in the Skin and Cerebral Cortex
We next wanted to investigate whether the significant reduction in lesion size and reduced GAS burden in the skin but not the brain following CLN treatment would correlate with changes in the cytokine profile.We assessed the levels of multiple prototypical Th1 (IFN-γ, IL-2), Th2 (IL-4, IL-5, IL-13), Th17/Th22 (IL-17A, IL-22), pro (TNF-α and IL-6), and anti-inflammatory (IL-10) cytokines as dominant readouts of immune modulation.The levels of IL-4, IL-17A, IL-22, and IL-6 were significantly reduced in the skin homogenates of CLN Tx female mice compared to the No Tx group (No Tx vs. CLN Tx: **** p < 0.0001, IL-4; * p = 0.0442, IL-17A; ** p = 0.0092, IL-22; * p = 0.0192, IL-6; Figure 4B-D).In males, the levels of IFN-γ, IL-4, IL-22, and IL-6 were significantly reduced in the CLN Tx group compared to the No Tx group (No Tx vs. CLN Tx: * p = 0.0405, IFN-γ; **** p < 0.0001, IL-4; * p = 0.0172, IL-22; *** p = 0.0002, IL-6; Figure 4E-H).Next, we investigated whether GAS dissemination into the brain or peripheral inflammatory cytokines communicated to the brain and altered the brain cytokine profile.As shown in Figure 4I-L, there was a significant increase in the levels of IL-2, IL-22, and IL-10 in the cerebral cortex homogenates of CLN Tx female mice compared to the No Tx group, while IL-13 levels were significantly upregulated compared to both No Inf and No Tx groups (No Tx vs. CLN Tx: * p = 0.0280, IL-2; ** p = 0.0040, IL-13; * p = 0.0340, IL-22; * p = 0.0108, IL-10; No Inf vs. CLN Tx: * p = 0.0130, IL-13).In the males, however, cortical IL-2 levels were significantly reduced in the CLN Tx group compared to the No Inf group (Figure 4M).Compared to the No Inf group, IL-22 levels were significantly reduced in the No Tx group, which did not improve with CLN treatment (Figure 4O) (No Inf vs. CLN Tx: ** p = 0.0092, IL-2; and No Inf vs.No Tx: * p = 0.0462, IL-22).

Subcutaneous GAS Infection Increased Cortical and Hippocampal Iba-1 Protein Levels
GFAP is the characteristic intermediate filament protein of astrocytes and is an important neuroinflammation marker [41,42].Iba-1 is widely used as a pan-microglial marker to identify and quantify microglia in the brain.Increased Iba-1 levels have been correlated with inflammatory states, as their levels are dramatically upregulated in response to various neuroinflammatory stimuli [43,44].We assessed the protein levels of GFAP and Iba-1 to quantify glial changes in the cortical and hippocampal lysates.While GFAP protein levels were not altered, there was a significant increase in the cortical Iba-1 levels in the female and male No Tx group compared to No Inf controls (Figure 5B, * p = 0.0477 Female No Inf vs.No Tx); Figure 5E, * p = 0.0487 Male No Inf vs.No Tx), suggestive of a microglial response to subcutaneous GAS infection.Correlating with the persistence of GAS despite CLN treatment in the male mice, Iba-1 levels also remained significantly elevated in CLN Tx male mice compared to No Inf controls (Figure 5E * p = 0.0471, No Inf vs. CLN treatment) in the cortex, and in the hippocampus (Figure 5K, * p = 0.0338 No Inf vs. CLN Tx).

Subcutaneous GAS Infection Increased Iba-1 Immunoreactivity in Male Mice
Since peripheral GAS infections selectively induced Iba-1 protein levels in the hippocampal lysates, we assessed Iba-1 immunoreactivity in the HC.Our data show that subcutaneous GAS infections significantly increased Iba-1 reactivity in the CA3 (Figure 6A-D) and CA1 (Figure 6E-H) subregions of CLN Tx male mice compared to No Inf controls (No Inf vs. CLN Tx, ** p = 0.0091, CA3 (Figure 6D)); * p = 0.0491, CA1 (Figure 6H).However, the female mice had no such differences (Figure 6A,B,E,F).There were no differences in the Iba-1 immunoreactivity among the groups in the females or males in the dentate gyrus (Figure 6I-L).campal lysates, we assessed Iba-1 immunoreactivity in the HC.Our data show that subcutaneous GAS infections significantly increased Iba-1 reactivity in the CA3 (Figure 6A-D) and CA1 (Figure 6E-H) subregions of CLN Tx male mice compared to No Inf controls (No Inf vs. CLN Tx, ** p = 0.0091, CA3 (Figure 6D)); * p = 0.0491, CA1 (Figure 6H).However, the female mice had no such differences (Figure 6A,B,E,F).There were no differences in the Iba-1 immunoreactivity among the groups in the females or males in the dentate gyrus (Figure 6I-L).Subcutaneous GAS infection increased Iba-1 immunoreactivity in the male mice.DQ8 mice were infected subcutaneously with 1 − 5 × 10 8 CFU of GAS 5448 and were either untreated or treated with CLN (10 mg/kg) administered intraperitoneally every day for five days.Sixteen days postinfection, the left hemisphere of the brain was fixed and serial sectioned (40 µm thickness).Every sixth section was stained with rabbit polyclonal anti-Iba-1 (1:1000), followed by Alexa Fluor 594-conjugated goat anti-rabbit secondary antibody (1:1000) (Red).The nuclei were stained with DAPI, and images were captured using the Leica THUNDER Imager at 20× magnification.Panels

Discussion
There has been a growing interest in understanding the role of bacterial or bacterial product-induced peripheral immune responses as a causative agent of neuroinflammation, neurodegeneration, and related cognitive dysfunction [45][46][47][48].Systemic exposure to gramnegative bacterial endotoxin, LPS, has been widely used to demonstrate neuroinflammation and associated sickness behavior [49].Liu et al. reported that subcutaneous infections with the GAS M28 serotype in female NF-κB reporter mice cause peripheral TNF-mediated neuroinflammation that can be attenuated by administering a dominant-negative selective inhibitor of TNF [50].
The ability of GAS to cross the blood-brain barrier or the meninges and reach the brain following a skin infection has not been demonstrated.Our results show that female mice harbored less brain GAS burden than males despite comparable GAS burden and CLN-mediated resolution in the skin.The lower initial brain GAS burden was, perhaps, more readily cleared by CLN in the female mice, in contrast to persistence observed in the male mice.The sex-dependent and compartmentalized antibiotic responses are intriguing and remain to be investigated.These data suggest that the brain constitutes a viable niche for GAS despite CLN intervention and might serve as an endogenous source for reactivation or relapse.While we have not been able to detect free bacteria in the brain, we could identify likely neuronal cells that have internalized GAS bacteria.Our findings raise the possibility that CLN, a bacteriostatic macrolide and the standard line of therapy for invasive GAS infections [51,52], may have limited ability to eliminate intracellular GAS reservoirs sequestered within brain cells.It is established that host mimics of GAS proteins result in anti-dopaminergic autoantibodies through molecular mimicry affecting dopamine signaling pathways [53].However, whether GAS produces adhesins or other host mimics that can target neuronal receptors is unknown.We cannot exclude other cell types in the brain as targets, and they need to be identified.Further, the mechanism of GAS invasion, including the indirect "trojan horse mechanism," through riding in GAS-infected peripherally derived macrophages and spreading from infected macrophages to neurons remains to be elucidated.
The therapeutic efficiency of CLN was evident based on the significant suppression of IFN-γ, IL-4, IL-17A, IL-22, and IL-6 levels in the skin homogenates from CLN Tx mice, which correlated with reduced lesion size.Specifically, our data indicate that the IL-6/Th17/Th22 axis is crucial to the pathogenesis of subcutaneous GAS infections.It is wellestablished that IL-6 plays a critical role at multiple levels in polarizing the differentiation of naïve CD4 + T cells into IL-17/IL-22-secreting effector Th17/Th22 subsets through direct and indirect mechanisms [54].We predicted that the exaggerated local inflammatory response, a Th17/Th22 axis driven by IL-6 levels, can transmit signals to the brain that might profoundly affect the cytokine profile and neuroimmune responses.CLN is considered a first-line treatment choice in the clinical management of invasive GAS infections [55,56]; however, the sex-dependent effects of CLN have not been widely examined in invasive GAS infections.Our data demonstrating the coordinated upregulation in the cortical levels of IL-2, IL-13, IL-22, and IL-10 in CLN Tx female mice that lacked GAS burden suggests that CLN triggers a concerted shift toward cytokines recognized to exert immunosuppressive effects [57,58].Interestingly, in the males, where cortical and hippocampal GAS burden was observed, IL-2 levels were significantly reduced in the CLN Tx group compared to the No Inf group.IL-2, traditionally recognized for its ability to support T cell activation and proliferation, conversely tends to dampen inflammation by augmenting T regulatory cell numbers [59].Whether IL-2 signaling appears poised to exert divergent outcomes under conditions of exposure to GAS in the brain remains to be examined.
Our data show that cortical IL-22 levels were significantly reduced in the male No Tx group compared to the No Inf group.Type 3 innate lymphoid cells (ILC3) are the primary source of IL-22, a member of the IL-10 family, and a key cytokine involved in maintaining intestinal barrier integrity [60].Interestingly, ILC3 cells can be stimulated by enteric glial cells to produce more IL-22 and regulate brain and cognitive function [61,62].Further work is needed to establish whether streptococcal skin infections induce intestinal damage or inflammation like colitis and whether increased IL-22 in the brain following CLN treatment is an unintended consequence of CLN therapy, causing gut dysbiosis and alterations through the bidirectional gut-brain axis.In addition to ILC3 cells, the Th22 subset of CD4 + cells is a primary source of IL-22 during infections [63,64].Liang et al. reported that TGF-β initiates the cascade of pro-inflammatory cytokines IL-17A, IL-22, and IL-6, and together, IL-17A and IL-22 are associated with skin innate immune responses [65].Our study suggests the possibility that subcutaneous GAS infection-induced IL-22 assumes a pathogenic role mirroring IL-17A levels in the skin and a protective role in the brain mirroring IL-10 levels.Our future studies will assess the dichotomous role of IL-22 in the skin-brain axis during GAS infections.
Insults in the brain due to trauma, infections, and peripheral inflammatory mediators prime the microglia for activation that correlates with the release of pro-inflammatory mediators, chemokines, and free radicals and can lead to the distinct polarization of a pro-or anti-inflammatory phenotype depending on the type of insult [66].In the case of pro-inflammatory response, mediators released by activated microglia can, in turn, confer an activated state in responding astrocytes [67].Together, these reactive microglia and astrocytes can potentiate neuron and synapse loss, the progression of neurodegeneration, and chronic neuroinflammation [68].Direct evidence for the role of activated microglia and astrocytes in inducing inflammatory mediators leading to neuroinflammation and cognitive impairment has been established due to HIV-1 infections [69][70][71][72].Our western blot data showed that subcutaneous GAS infections significantly increased the protein levels of the microglia marker, Iba-1, in the CC of female and male mice and HC of male mice.The hippocampus is highly susceptible to changes in a redox state [73], age [74], or pathogens [75] and demonstrates a unique selective neuronal vulnerability [76].In the present study, we observed concerning findings regarding the therapeutic failure of CLN to eliminate brain GAS burden in the HC that coincided with CA3 and CA1 subregion selective changes in Iba-1 immunoreactivity in the male mice.We recognize that an important limitation of the present study was the use of Iba-1 as the primary marker to characterize microglia without the use of additional markers associated with an activated phenotype and functional status.Nonetheless, our current findings provide an initial framework for more detailed follow-up investigations.They are designed to include additional markers for robust microglia profiling to elucidate their protective and pathogenic roles during GAS infections.

Conclusions
In conclusion, our data show that despite comparable resolution of skin GAS burden and skin cytokine changes following CLN therapy in GAS-infected female and male DQ8 mice, an altered brain cytokine profile emerged in CLN Tx mice that was sex-dependent.GAS dissemination into the brain and microglial responses were also sex-dependent and subregion-specific.These results open new avenues for detailed investigations, including (a) antibiotic combinations, (b) host-directed therapies as antibiotic adjuncts, (c) mechanisms underlying the sexual dimorphism in GAS invasion and antibiotic responses within the CNS, (d) receptors targeted for intracellular GAS persistence within brain cells, and (e) long-term functional and behavioral consequences of infections caused by GAS, one of the most important human pathogens with the broadest spectrum of clinical manifestations known in medical history.

Figure 1 .
Figure 1.CLN treatment significantly reduced weight loss and lesion areas.DQ8 mice were infected subcutaneously with 1 − 5 × 10 8 CFU of GAS 5448 and were either untreated or treated with CLN (10 mg/kg) administered intraperitoneally every day for five days.Control mice were uninfected (No Inf).Body weight and skin lesion area were measured, and data are presented as mean values + SEM (n = 4-9).Each circle, square, or triangle represents an individual mouse.Statistical differences were computed by one-way ANOVA followed by uncorrected Fisher's LSD test ((A), * p = 0.0338, and **** p < 0.0001) or unpaired t-test ((B), **** p < 0.0001).

Figure 1 .
Figure 1.CLN treatment significantly reduced weight loss and lesion areas.DQ8 mice were infected subcutaneously with 1 − 5 × 10 8 CFU of GAS 5448 and were either untreated or treated with CLN (10 mg/kg) administered intraperitoneally every day for five days.Control mice were uninfected (No Inf).Body weight and skin lesion area were measured, and data are presented as mean values + SEM (n = 4-9).Each circle, square, or triangle represents an individual mouse.Statistical differences were computed by one-way ANOVA followed by uncorrected Fisher's LSD test ((A), * p = 0.0338, and **** p < 0.0001) or unpaired t-test ((B), **** p < 0.0001).

Figure 2 .
Figure2.Heterogeneity in GAS burden in the skin, subcutaneous adipose tissue, and the brain.DQ8 mice were infected subcutaneously with 1 − 5 × 10 8 CFU of GAS 5448 and were either untreated or treated with CLN (10 mg/kg) administered intraperitoneally every day for five days.Sixteen days post-infection, necrotic skin lesions, SubQ AT, the right cerebral cortex, and hippocampus were aseptically excised and collected into 1 mL of sterile PBS −/− , weighed, and homogenized to enumerate the GAS burden ((A-D) respectively).GAS burden in female and male mice was compared separately in the skin (E), cerebral cortex (F), and hippocampus (G).Data are presented as mean values + SEM (n = 4-9).Each circle or square represents an individual mouse.Statistical differences were computed by unpaired t-test ((A), ** p = 0.0018) or two-way ANOVA ((E), * p = 0.0140).

Figure 3 .
Figure 3. GAS bacteria localize to the brain.Representative images showing immunoreactivity of the GAS antibody in the cerebral cortex of No Inf, No Tx, or CLN Tx DQ8 male mice.The 40µm thick sections of the brain (left hemisphere) were stained with rabbit polyclonal anti-GAS antibody (1:100) followed by Alexa Fluor 594-conjugated goat anti-rabbit secondary antibody (1:1000) (Red).Nuclei were counterstained with DAPI.

Figure 3 .
Figure 3. GAS bacteria localize to the brain.

0. 0477
Female No Inf vs.No Tx); Figure 5E, * p = 0.0487 Male No Inf vs.No Tx), suggestive of a microglial response to subcutaneous GAS infection.Correlating with the persistence of GAS despite CLN treatment in the male mice, Iba-1 levels also remained significantly elevated in CLN Tx male mice compared to No Inf controls (Figure 5E * p = 0.0471, No Inf vs. CLN treatment) in the cortex, and in the hippocampus (Figure 5K, * p = 0.0338 No Inf vs. CLN Tx).

Figure 6 .Figure 6 .
Figure 6.Subcutaneous GAS infection increased Iba-1 immunoreactivity in the male mice.DQ8 mice were infected subcutaneously with 1 − 5 × 10 8 CFU of GAS 5448 and were either untreated or treated with CLN (10 mg/kg) administered intraperitoneally every day for five days.Sixteen days postinfection, the left hemisphere of the brain was fixed and serial sectioned (40 µm thickness).Every sixth section was stained with rabbit polyclonal anti-Iba-1 (1:1000), followed by Alexa Fluor 594conjugated goat anti-rabbit secondary antibody (1:1000) (Red).The nuclei were stained with DAPI, and images were captured using the Leica THUNDER Imager at 20× magnification.Panels A and C (CA3), E and G (CA1), and I and K (Dentate gyrus) show representative images (scale bar = 25µm) of Iba-1 immunoreactivity in the brain tissue from No Inf, No Tx, or CLN Tx mice.Bars represent mean values ± SEM of segmented counts of Iba-1 positive cells (n = 3-4 per group).Each circle, Figure 6.Subcutaneous GAS infection increased Iba-1 immunoreactivity in the male mice.DQ8 mice were infected subcutaneously with 1 − 5 × 10 8 CFU of GAS 5448 and were either untreated or treated (A) and (C) (CA3), (E) and (G) (CA1), and (I) and (K) (Dentate gyrus) show representative images (scale bar = 25 µm) of Iba-1 immunoreactivity in the brain tissue from No Inf, No Tx, or CLN Tx mice.Bars represent mean values ± SEM of segmented counts of Iba-1 positive cells (n = 3-4 per group Author Contributions: S.N.: conceptualization, methodology, formal analysis, investigation, writing-review and editing, project administration, funding; S.M.: conceptualization, methodology, writing-review and editing; B.G.: image analysis and quantification, review, and editing; C.C.: conceptualization, writing-review and editing.All authors have read and agreed to the published version of the manuscript.Funding: This work was supported by grants from the European Union (FP7/2012-2017) under grant agreement 305340, grants from the Swedish Research Council under grant number 20150338, UND CoBRE Host-Pathogen Interactions Pilot Award supported by the NIH/NIGMS award P20GM113123, and UND VPRED Post-Doctoral support.Flow Cytometry and Microscopy services were provided by the North Dakota Flow Cytometry Cell Sorting Core and Microscopy and Imaging Core, respectively, supported by the Institutional Development Award (IDeA) from the National Institute of General Medical Sciences of the National Institutes of Health under grant number P20GM103442 and P20GM113123, DaCCoTA CTR NIH grant U54GM128729, and UNDSMHS funds.The funders had no role in study design, data collection, interpretation, or the decision to submit the work for publication.Institutional Review Board Statement: The mice were bred and maintained in compliance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and with the prior approval of the Institutional Animal Care and Use Committees (IACUC) of the University of North Dakota, protocol numbers 1704-1 and 2002-5C.Data Availability Statement: All data are included in the published article.