The study was carried out accordingly with the recommendations of the guidelines and policies of the Canadian Council on Animal Care and the principles set forth in the Guide for the Care and Use of Laboratory Animals. The protocols and procedures were approved by the Animal Welfare Committee of the University of Montreal (protocols numbers Rech-1399 and Rech-1523), including euthanasia to minimize animal suffering, which was applied throughout this study when animals were seriously affected since mortality was not an endpoint measurement. Mice had free access to water and food pellets as well as enrichment such as diverse materials to use as bedding alongside access to a refuge and a rubber toy for gnawing.

The well-characterized and encapsulated strain P1/7 (serotype 2, sequence type 1) was used in this study. This strain was isolated from a pig with meningitis in the United Kingdom and its virulence was previously reported in experimental models of infection (31, 32). S. suis was grown overnight on Columbia agar plate supplemented with 5% sheep blood (Oxoid, Nepean, ON, Canada) at 37°C with 5% of CO₂. Isolated colonies were used to inoculate 5 ml of Todd Hewitt broth (THB; Becton, Dickinson, Mississauga, ON, Canada) for 8 h. To prepare working solutions, 10 µl of a 10^-3 dilution of the suspension were inoculated in 30 ml of THB for 16 h at 37°C with 5% of CO₂. Bacteria were washed twice with pH 7.4 phosphate-buffered saline (PBS) and resuspended in THB at the appropriate dilution. To determine the bacterial concentration in CFU/ml, suspensions were diluted and plated on THB agar plates and colonies counted.

Wild-type C57BL/6 mice (5 to 8 weeks of age) were obtained from Charles River Laboratories (Wilmington, MA, USA). Mice were housed under specific pathogen-free conditions in ventilated cages and had unlimited access to water and rodent chow. Mice were intraperitoneally infected with 1 ml of S. suis suspension containing 1 x 10⁷ CFU. In this model, S. suis quickly reaches the bloodstream and spreads causing a systemic inflammatory response, as observed in pig natural infection (31). Control mice (mock-infected) were injected with the same volume of vehicle solution (THB). Clinical signs were observed every 6 h daily to monitor for severe clinical signs of infection. After different times post-infection, mice were euthanized to collect blood by intracardiac puncture, spleen, liver and/or bone marrow. Blood samples were collected from living mice at different time points (see below) via the caudal vein of the tail (to evaluate bacteremia) or the submandibular vein (to quantify pro-inflammatory mediators).

To neutralize the effect of the G-CSF, we applied an anti-mouse G-CSF receptor (G-CSFR) antibody (Ch5E2-VR81-mIgG1κ; CSL Innovation, Parkville, Australia) as previously described (33). Briefly, mice were intraperitoneally injected with 0.2 ml of the anti-mouse G-CSFR antibody (100 µg per dose) or the corresponding isotype control (mouse IgG1κ; either from Bioxcell, Lebanon, NH, USA or from Biolegend, San Diego, CA, USA). The dose of anti-mouse G-CSFR antibody was selected in pre-trials (data not shown). Mice that were infected for 72 h received four doses of antibodies at day -1, day 0 prior to infection and day +1 and day +2 after infection. Mice that were infected for less than 24 h received two doses of antibodies at day -1 and day 0 prior to infection. Mice were infected with S. suis as described above.

Five microliters of blood were collected at 12 h, 24 h, 48 h and 72 h post-infection through the caudal tail vein of mice. Samples were appropriately diluted in PBS, plated on THB agar and colonies counted to assess the blood bacterial load in infected mice.

The blood of infected mice was collected at different time points post-infection either by intracardiac puncture on euthanized mice or by the submandibular vein collection on living mice. Blood was collected in BD Microtainer™ serum separator tubes (BD Biosciences, Mississauga, ON, Canada) and centrifugated at 10,000 g for 2 min. The serum was collected at the surface of the gel and stored at -80°C until analysis. The spleen and the liver were homogenized using a POLYTRON PT 1200E system bundle (Kinematica, Lucerne, Switzerland) in an extraction buffer containing complete Mini, EDTA-free, protease inhibitor cocktail tablets (Roche Diagnostics GmbH, Mannheim, Germany), following the manufacturer’s instructions. The tibias of mice were flushed in 500 µl of extraction buffer. Homogenate supernatants were collected by centrifugation and stored at -80°C until analysis.

The concentrations of G-CSF, interleukin (IL)-6 and the C-X-C motif chemokine ligand 1 (CXCL1)/keratinocyte chemoattractant (KC), were determined by sandwich enzyme-linked immunosorbent assay (ELISA) using pair-matched antibodies from R&D Systems (Minneapolis, MN, USA). Cytokine measurements were performed according to the manufacturer’s recommendations, as previously described (34, 35). The concentrations of other cytokines and chemokines were measured using a Bio-Plex Pro assay (Bio-Rad, Mississauga, ON, Canada) according to the manufacturer’s instructions. The assay covered the G-CSF, the interferon-γ (IFN-γ), the IL-12p70, IL-6, C-C motif chemokine ligand 2 (CCL2)/MCP-1, CCL3/MIP-1α, CCL4/MIP-1β, CCL6/RANTES, CXCL9/MIG, and CXCL2/MIP-2. Acquisition was performed on the MAGPIX platform (Luminex, Toronto, ON, Canada), and data were analyzed using Bio-Plex Manager 6.1 software (Bio-Rad).

Blood was collected in tubes coated with lithium heparin (BD Biosciences). Organs freshly collected from euthanized mice were homogenized to obtain single-cell suspensions. In the samples, red cells were removed by suspending the cells in a commercial lysis buffer (Invitrogen, Burlington, ON, Canada). Suspensions were first stained with a Fixable viability Dye eFluor™ 506 (Invitrogen) to exclude the dead cells from the analysis. After washing with PBS containing 2% fetal bovine serum, surface Fc receptors - that non-specifically bind antibodies - were blocked using FcBlock (BD Biosciences) for 15 min before staining. The cells were then stained for the markers of interest (detailed below) for 30 min in PBS containing 2% fetal bovine serum before being fixed by suspending them for 15-30 min in cold paraformaldehyde-PBS (2% for extracellular staining and 4% for intracellular staining) and stored in PBS at 4°C until data acquisition. To count the cells, 10 µL of Countbright™ Absolute Counting Beads (Invitrogen) were added in 400 µL of cell suspension just before acquisition. Data were acquired on a BD Fortessa or a BD FACSAria Fusion using the BD FACSDiva software and analyzed using the Flow Jo software (BD Biosciences).

The different subsets of cells were identified as previously described (36) using the following antibodies: PE-conjugated anti-mouse CD45 (clone 30-F11), APC-conjugated anti-mouse CD11c (clone HL3), BV421-conjugated anti-mouse CD11b (clone M1/70), FITC-conjugated anti-mouse Ly6G (clone 1A8). In addition to neutrophil markers, cells were stained with the following antibodies: PE CF594-conjugated anti-mouse CXCR4 (clone 2B11/CXCR4), BV605-conjugated anti-mouse CD49d (clone R1-2), Alexa700-conjugated anti-mouse CD62L (clone MEL-14), BV711-conjugated anti-mouse CXCR2 (CD182) (clone V48-2310). For BV-conjugated antibodies, the BD Horizon™ Brilliant Stain Buffer was used. To stain intracellular G-CSF, cells were permeabilized by suspension in BD Perm/Wash™ Buffer for 15 min in the dark and stained with AlexaFluor700-conjugated anti-mouse G-CSF antibody (clone 67604; R&D Systems). Unless otherwise indicated, all antibodies and buffers were from BD Biosciences. The gating strategy is depicted in Supplementary Figure S1 .

Experiments were repeated at least three times, and the number of animals varied from three to thirty-six depending on the condition or time point (see figure legends for details). For all data, the normality of distribution was verified using the Shapiro-Wilk test. Accordingly, parametric tests (unpaired t test) or nonparametric tests (Mann-Whitney rank sum test) were performed to evaluate statistical differences between groups. Data are presented as the mean +/- standard error of the mean (SEM). P < 0.05 was considered statistically significant.

S. suis triggers an exaggerated inflammatory response in the host, including the production of multiple chemokines (31). Since those molecules participate in immune cell migration (37), we evaluated the changes in immune cell populations in the blood and the spleen at the early stages of S. suis infection. The markers and gating strategy allowed to distinguish leukocytes, lymphocytes, neutrophils and the pool of dendritic cells (DC), monocytes (Mo) and macrophages (Mph) ( Supplementary Figure S1 ). During S. suis infection, the percentage of neutrophils significantly increased by 10-fold in the blood and by 3-fold in the spleen ( Figures 1A, B ). Other changes occurred in the blood: the percentage of lymphocytes decreased while the percentage of the pool DC-Mo-Mph increased ( Figure 1A ). However, no changes of these two cell populations were observed in the spleen ( Figures 1B, C ).

Since neutrophils are recruited in the blood and in the spleen during the infection, we characterized the kinetics of neutrophil mobilization in these organs as well as in the bone marrow, the headquarter of neutrophil production (38). Flow cytometry analyses revealed that neutrophil rates in the blood and spleen of mice increased as soon as 6 h post-infection (p.i.) and peaked at 12 h p.i. ( Figure 2 ). A return to basal level was observed after 24 h of infection. In parallel, the bone marrow, in which neutrophils represent 20% of all leukocytes during homoeostasis, showed an important drop in percentages of neutrophils from 6 h to 24 h p.i. ( Figure 2 ). From 48 h onwards, bone-marrow replenished its stock in neutrophils, returning to basal levels.

During infection, neutrophils undergo phenotypical changes to better respond to pathogen invasion (39). Most of the blood neutrophils expressed CD11b in steady state, but S. suis infection significantly increased this percentage from 85% to 96% ( Figures 3A, G ). In addition, neutrophils from infected mice presented higher levels of CD11b expression at their surface, as determined by the median fluorescence intensity ( Figure 3B ). On the contrary, infection reduced the percentage of neutrophils expressing CD62L and CXCR2 ( Figures 3C, E, G ). For CD62L, positive cells exhibited lower levels of molecule surface expression during the infection ( Figure 3D ), while expression levels remained unchanged for CXCR2⁺ cells ( Figure 3F ). An increased expression of CD11b was also observed in spleen neutrophils from infected mice. However, no statistical difference was observed in the percentage or expression levels of the other markers in spleen neutrophils during S. suis infection compared to controls. Nevertheless, the overall pattern of marker expression was similar between blood and spleen ( Supplementary Figure S2 ). To compare the effect of a mild infection and a severe infection on neutrophil phenotype, we infected mice with a low dose (1 x 10⁶ CFU) or a high dose (1 x 10⁷ CFU, the dose used throughout this study) of S. suis. However, the phenotype of neutrophils was overall similar between the two doses ( Supplementary Figure S3 ).

Among the factors produced during infections, the G-CSF represents the key cytokine involved in the control of neutrophils (21). Thus, we investigated if the G-CSF was produced in the serum and organs of S. suis-infected mice at different times post-infection. Important levels of G-CSF were produced in the spleen, the liver, the serum and the bone-marrow of infected mice with a peak of production at 12 h p.i. ( Figure 4 ). Specifically, serum concentrations of G-CSF represented the highest of all evaluated cytokines at this time point ( Supplementary Figure S4 ).

The massive production of G-CSF during S. suis infection indicates that some cells, somehow, produce this factor. Since dendritic cells (DCs) and macrophages produce G-CSF in vitro when stimulated with S. suis (30), we hypothesized that immune cells could take part in G-CSF production in vivo. By quantifying intracellular G-CSF in immune cells, we found that G-CSF increases in blood leukocytes in response to S. suis infection ( Figures 5A, C ). Particularly, this rise occurs among both lymphocytes and the pool of DCs, monocytes, and macrophages. However, neutrophils, which express the highest levels of G-CSF at the basal state, decreased their intracellular levels during infection ( Figure 5A ). In the spleen, none of the hematopoietic cells showed an increase of intracellular G-CSF, but intracellular G-CSF decreased among the neutrophil population at 16 h p.i. ( Figure 5B ).

Since G-CSF controls neutrophil recruitment (20), we neutralized its effect in a mouse model using an anti-G-CSFR antibody (αG-CSFR) prior to S. suis infection ( Figure 6 ). At 12 h p.i., treatment with αG-CSFR prevented neutrophil recruitment in the blood, exhibiting similar levels to those of control mice (mock-infected) ( Figures 6A, C ). Simultaneously, the drop of the neutrophil percentage in the bone marrow caused by the infection did not occur during the G-CSF receptor blockade ( Figure 6B ). The stock of neutrophils was therefore maintained in the bone-marrow under this condition. The analysis of the different leukocyte populations revealed that mainly neutrophil population was affected by G-CSFR blockade ( Supplementary Figure S5 ).

Since G-CSF controls neutrophil egress from the bone marrow to the blood ( Figure 6 ) and neutrophils promote host survival by controlling S. suis bacterial burden (10), we evaluated the role of the G-CSF in S. suis clearance from the blood. Results showed that the treatment with αG-CSFR did not affect the concentration of bacteria in the blood of S. suis-infected mice at any evaluated time ( Figure 7 ).

Neutrophils participate in the production of inflammatory mediators in systemic infection (10). Since G-CSFR blockade induced less neutrophil release from the bone marrow, there might also be an impact on the levels of pro-inflammatory mediators produced during inflammation. We quantified CXCL1 (KC) and IL-6 in the plasma of S. suis-infected mice at 12 h p.i., time at which most cytokine levels peak (40, 41). However, G-CSFR blockade did not significantly change levels induced by S. suis infection ( Figure 8 ). We performed additional cytokine analyses by Luminex and showed that αG-CSFR treatment did not significantly modulate the plasmatic levels of most of the cytokines (despite an upward trend), but increased IFN-γ and MIP-1α production ( Supplementary Figure S6 ).