Staphylococcus aureus biofilm elicits the expansion, activation and polarization of myeloid-derived suppressor cells in vivo and in vitro

Staphylococcus aureus (S. aureus) is one of the most common causes of biofilm infections in periprosthetic joint infections (PJIs). Accumulating evidence has shown that the immunosuppressive environment established by S. aureus biofilm infection in PJIs involves the presence of myeloid-derived suppressor cells (MDSCs) and M2-macrophages. Due to the diversity of MDSCs, little is known about whether S. aureus biofilm preferentially expands specific MDSC subsets or whether MDSCs can further differentiate into M2-macrophages during S. aureus biofilm infection. Here, we show that in agreement with the results from an established rat PJI model, S. aureus biofilm cocultured with freshly isolated bone marrow cells (BMCs) in vitro significantly increases the proportions of MDSCs, total macrophages and M2-macrophages. Interestingly, we find that treatment of the BMCs in vitro with S. aureus biofilm preferentially promotes the expansion of monocytic MDSCs but not granulocytic MDSCs. Biofilm treatment also substantially enhances the overall MDSC immunosuppressive activity in addition to the MDSC expansion in vitro. Importantly, we provide evidence that S. aureus biofilm is capable of further stimulating the conversion of monocytic MDSCs into M2-macrophages in vitro and in vivo. Collectively, our studies reveal a direct link between MDSCs and M2-macrophages occurring in S. aureus-associated PJIs.


Introduction
Prosthetic joint infections (PJIs) are a devastating complication after arthroplasty, which greatly affect the quality of a patient's recovery [1][2][3]. The incidence of PJIs is about 1% to 2% for primary total joint arthroplasties, and approximately at 2%-6% for a revision joint PLOS

S. aureus strain and biofilm formation
S. aureus (ATCC43300) was purchased from Bioresource Collection and Research Center (BCRC), Taiwan. To prepare S. aureus biofilm, the bacteria were cultured in BHI media for 4 days as previously described [15,27]. Biofilm communities were collected by centrifugation at 3900 xg for 15 min, and then sterilized by autoclaving. Biofilm pellets were resuspended with RPMI medium and total protein amounts of the lysates were measured using BCA protein assay kit (Pierce). For the in vivo study, S. aureus cells were pelleted, washed 3 times, and resuspended in normal saline. Bacterial concentrations were estimated spectrophotometrically by measuring the absorbance at 600 nm (A 600 ). Colony forming units (CFU) were counted on the plates to verify the optical measurement.

S. aureus biofilm infection in rats and in mice
The rat model of the S. aureus orthopedic implant infection was performed as described previously with minor modifications [28]. Briefly, male LEW/SsNNarl rats were anesthetized with a 1:1 mixture of Zoletil 50 and Rompun 2% (1 ml/kg in rat and mice) and the surgical site was disinfected with povidone-iodine. A skin incision was made near the right knee. A burr hole was made in the femoral intercondylar notch extending into the intramedullary canal using a 26-gauge needle, and then a pre-cut 0.5 cm orthopedic grade Kirschner (K)-wire (0.6-mm diameter, Synthes GmbH, Switzerland) was implanted into the intramedullary canal. Following the insertion of K-wire implant, 2×10 3 CFU of S. aureus were inoculated. Subsequently, the burr hole was sealed with bone wax and then the surgical site was closed with Coated VICRYL Ã 4-0 sutures. Rats that received sterile implants served as the sham-operated group. For pain relief, rats were treated with Ketorolac (0.5~1 mg/kg, IM) immediately after the surgery and 24 hours later. For the PJI model in mice, all the procedures of S. aureus infection were the same as those in rats. Herein, after infection with S. aureus for 7 days in C57BL/6J mice, 1×10 7 CD11b-positive MDSCs isolated from Pgk1-EGFP transgenic mice were injected into the infected mice intravenously. Twenty-four hours later, the mice were sacrificed and the tissue near the S. aureus infected site were taken, embedded and frozen in -80˚C.

Scanning electron microscopy
Rats were sacrificed at day 7 after S. aureus infection. The bone rougeur was used to cut the femur longitudinally and a piece of fragment (2×4 mm 2 ) was picked. Femur samples were fixed in 0.1M cacodylate buffer (pH 7.4) containing 3% glutaraldehyde and 2% paraformaldehyde for 2 hr at 4˚C. Subsequently, fixed specimens were cut longitudinally and washed three times in 0.1 M cacodylate buffer (pH 7.4) for 10 minutes at 4˚C. In the fume hood, specimens were soaked in 0.1 M cacodylate buffer containing 1% Osmium Teteroxide for 1 hr at 4˚C. After rinsing three times using 0.1M cacodylate buffer for 10 min at 4˚C, specimens were dehydrated using a graded series of ethanol washes and critical-pointed-dried using a critical point dryer (CPD). Dried specimens were coated with gold-palladium. Samples were viewed using a HITACHI S-5000 scanning electron microscope.

Isolation of specific immune cell populations from mice bone marrow
BMCs were collected from bone marrow of mouse femur. The CD11b-positive cell population was isolated using anti-CD11b antibody conjugating with biotin in combination along with anti-biotin magnetic beads (STEMCELL™) according to the manufacturer's instructions. The purity of the isolated cell population was verified by flow cytometry. Two MDSC subsets, M-MDSCs (Ly6C high Ly6G -) and G-MDSCs (Ly6C low Ly6G + ), were then sorted from the CD11b-positive population by using BD FACSAria Fusion cell sorter.

Coculture of the isolated cell populations with S. aureus biofilm in vitro
Mouse BMCs or the isolated cell populations were seeded in the 24-well plate and cocultured with different concentrations of S. aureus biofilm loaded in the transwell cell culture insert (0.4 μm pore size). All the cocultures of the indicated cell populations with S. aureus biofilm were maintained in complete medium (RPMI1640 with 10% FBS). After coculture for 48 or 72 hr, the biofilm-treated cells were analyzed by flow cytometry (BD FACSCanto II) using specific antibodies.

T cell proliferation assays
Mouse BMCs were seeded in 24-well plates and then left untreated or treated with S. aureus biofilm for 48 hr. After 48-hr treatment, the CD11b-positive cell population from the untreated or biofilm-treated BMCs was isolated using anti-CD11b antibody conjugating with biotin in combination along with anti-biotin magnetic beads (STEMCELL™). The untreated and biofilm-treated BMCs as well as their isolated CD11b-positive cell populations were subsequently cocultured with CFSE-labeled spleen T cells that were stimulated by Dynabeads Mouse T-Activator CD3/CD28 (Gibco). Different ratios of BMCs (or the isolated CD11b-positive cells) to activated T cells (2×10 5 cells/well in a 96-well round bottom plate in RPMI-1640 with 10% FBS) were examined in the study. For isolation of spleen T cells, mouse spleens were gently pressed through a strainer using the plunger end of a syringe to get singlet cell. These cells were washed with PBS buffer and ACK lysing buffer (Gibco). Cells in warm RPMI (2 ml) were injected to nylon wool column, and washed with RPMI. The column was sealed and incubated at 37˚C and 5% CO2 for 45 minutes. Cells were then eluted with 10 ml of warm RPMI.

Analysis of regulatory T (Treg) cells
Mouse BMCs were seeded in 24-well plates and then left untreated or treated with S. aureus biofilm. After 48-hr treatment, the CD11b-positive cell populations were isolated from the untreated or biofilm-treated BMCs. The isolated CD11b-positive cell populations were then cocultured with spleen T cells that were stimulated by Dynabeads Mouse T-Activator CD3/ CD28 (Gibco). After 24-or 48-hr coculture, the proportions of Foxp3 + CD25 + CD4 + Tregs in the cell mixtures were analyzed by flow cytometry using a True-Nuclear One Step Staining Mouse Treg Flow kit (FOXP3 Alexa Fluor 488/CD25 PE/CD4 PerCP; Biolegend). Results were expressed as the percentages of Foxp3 + CD25 + CD4 + T cells in CD4 + cells.

Immunofluorescence staining and confocal microscopy
Tissue samples were embedded in optimal cutting temperature compound (OCT) and snap frozen in liquid nitrogen. Tissue sections (4 μm) were fixed in ethanol/acetic acid fixative solution for 2-10 min, and then were stained with anti-F4/80-PE (eBioscience), CD206-PerCP-Cy5.5 (Biolegend) overnight at 4˚C in a humidified chamber. After three times of washes, slides were stained with DAPI and mounted with Prolong Gold mounting reagent (Invitrogen). Confocal imaging was performed using Leica SP5 II confocal microscope.

Statistical analyses
Statistical analyses were performed using Student's t-test for independent samples, with significance determined at P<0.05. All data and standard deviation (SD) were calculated and graphed by Microsoft Excel software (Microsoft, Redmond, WA).

Biofilm formation in an S. aureus-infected PJI model
To mimic acute and chronic S. aureus biofilm infection in vivo, we initially established a PJI model in rats. In the PJI model, S. aureus (2×10 3 CFU) were inoculated at a metallic orthopedic implant (Kirschner wire, 0.6 mm diameter x 0.5 cm) in rat femurs. In addition to the shamoperated and S. aureus-infected groups, the S. aureus-infected group receiving vancomycin (2 mg/kg) was also included in the experiment. At 7 days after surgery, all rats were sacrificed and the treated femurs were taken (Fig 1a). Compared to the sham-operated femurs, severe osteolytic bony destruction and formation of abscesses were observed in the S. aureus-infected femurs; however, vancomycin treatment substantially reduced the bony destruction caused by S. aureus infection (Fig 1a). By using scanning electron microscopy, biofilm formation and S. aureus cocci in the lumen of the femur were evidently detected in both S. aureus-infected groups, despite a lower extent of bacterial dissemination for the vancomycin-treated group (Fig 1b). These results indicated that our established PJI model in rats shows persistent biofilm infection and may allow further investigation of host immune responses to S. aureus biofilm infection in vivo.

Significant expansion of MDSCs, total macrophages and M2-macrophages in the peripheral blood of rats during S. aureus biofilm infection
To investigate the possible immunosuppressive effect caused by S. aureus biofilm infection in vivo, we examined if the proportions of the circulating MDSCs, total macrophages and M2-macrophages were changed during S. aureus infection in our PJI model. Blood samples from the sham-operated group, the untreated infection group, and the vancomycin-treated infection group were then collected at different time points after operation. Compared to blood samples obtained immediately prior to surgery (designated as day 0), blood samples collected from day 1 after surgery (or surgery plus S. aureus infection) exhibited higher proportions of MDSCs (CD11bc + HIS48 + ), macrophages (CD68 + ) and M2-macrophages (CD68 + CD206 + ) in all three operated groups (Fig 2a-2c and S1 Fig), suggesting that surgical injury might contribute to systemic expansion of these immune cells. In the sham-operated group, the increased proportions of these immune cells detected on day 1 after surgery could decline gradually to normal levels on days 2 or 3 after surgery (Fig 2). However, when these immune cells were examined in the untreated infection group or in the vancomycin-treated infection group, we found that the proportions of MDSCs, total macrophages and M2-macrophages in the two S. aureus-infected groups were regularly higher than those in the sham-operated group at all postoperative times (Fig 2a-2c). These results suggested that S. aureus biofilm infection substantially stimulated the expansion of MDSCs, total macrophages, and M2-macrophages in the PJI model. Noteworthily, there were no significant differences in the expansion rates of MDSCs, total macrophages, and M2-macrophages between the two infected groups that received no treatment or vancomycin treatment, implicating that vancomycin treatment could not completely eliminate S. aureus biofilm infections during the experimental period.

Treatment of mouse bone marrow cells with S. aureus biofilm increases the proportions of MDSCs, total macrophages and M2-macrophages in vitro
Since the circulating MDSCs, total macrophages, and M2-macrophages could be derived from bone marrow precursor cells (BMCs), we attempted to investigate whether S. aureus biofilm was able to stimulate the expansion of these immune cell types from BMCs in vitro. Here, we developed an in vitro coculture system in which the freshly isolated mouse BMCs were plated on the bottom of transwell cell culture system and then co-cultured with S. aureus biofilms loaded in the transwell cell culture insert (0.4 μm pore size). Notably, mouse BMCs (C57BL/ 6J) rather than rat BMCs were used in the testing because cell-surface markers for diverse immune cell types are better characterized in mouse. After treatment of mouse BMCs with increasing amounts (0.2, 0.6 and 2.0 mg/ml) of S. aureus biofilm for 48 hr, we found that the proportions of MDSCs (CD11b + Gr-1 + ) were significantly increased from 40% to about 67-  The sham group, the S. aureus-infected group, and the S. aureus-infected group receiving vancomycin were included in the experiments and each group contained 4 rats. Blood samples were collected at different time points after operation. After lysis of the red blood cells, the remaining leukocytes were analyzed for the proportions of CD11bc + His48 + MDSCs (a), CD68 + macrophages (b), and CD68 + CD206 + M2-macrophages 80% along with the doses of S. aureus biofilm (Fig 3a and 3b). The MDSC subtypes, including M-MDSCs (CD11b + Ly6C high Ly6G -) and G-MDSCs (CD11b + Ly6C low Ly6G + ), were then evaluated in the experiments. Interestingly, S. aureus biofilm at the low concentration (0.2 mg/ml) sufficiently elicited a marked increase in the proportion of M-MDSCs (from 9.8% to 26.8%), but not the proportion of G-MDSCs (from 3.6.5% to 37.6%), from the treated BMCs (Fig 3c  and 3d). As compared to the untreated group, a mild increase in the proportion of G-MDSCs was detected only when BMCs were treated with S. aureus biofilm at high concentrations (0.6 and 2.0 mg/ml), resulting in an increase from 36.5% to 55% or 48.3% (Fig 3c and 3d). These results implicated that S. aureus biofilm differentially promotes the expansion of M-MDSCs and G-MDSCs.
In addition to MDSCs, the proportions of total macrophages and M2-macrophages were also measured in the experiments. We consistently found that S. aureus biofilm markedly increased the proportions of total macrophages (F4/80 + ) and M2-macrophages (F4/ 80 + CD206 + ) in a dose-dependent manner (Fig 3e and 3f and S2 Fig). In agreement with the PJI model in rats, our in vitro coculture system reflected that S. aureus biofilm could elicit the expansion of MDSCs, total macrophages, and M2-macrophages from BMCs.

S. aureus biofilm promotes the differentiation of MDSCs into M2-macrophages in vitro
Since MDSCs are an immature population of myeloid cells, we then investigated whether MDSCs could be programmed toward M2-macrophages in response to S. aureus biofilms. To do the experiments, we first isolated the CD11b-positive cells from mouse BMCs using CD11b MicroBeads, which correspond to MDSC population ( S3 Fig and see Discussion). We found that treatment of the isolated CD11b-positive MDSCs with increasing amounts of S. aureus biofilm caused significant increases in the proportions of total macrophages (F4/80 + ) and M2-macrophages (F4/80 + CD206 + ) in a dose-dependent manner (Fig 4a and 4b). These results suggested that the CD11b-positive MDSCs had the potential to convert into M2-macrophages upon exposure to S. aureus biofilm. To further elucidate the conversion of these immune cells, two MDSC subsets (M-MDSCs and G-MDSCs) were sorted from the CD11b-positive MDSCs by cell sorter (Fig 5a). These two isolated MDSC subpopulations, M-MDSCs (CD11b + Ly6-C high Ly6G -) and G-MDSCs (CD11b + Ly6C low Ly6G + ), were then cocultured with S. aureus biofilm. Specially, like the CD11b-positive MDSCs, treatment of the sorted M-MDSC cell population (Fig 5a, P1) with S. aureus biofilm also exhibited increased proportions of total macrophages (F4/80 + ) and M2-macrophages (F4/80 + CD206 + ) along with an increasing amount of biofilm (Fig 5b top, Fig 5c and 5d). In contrast, the coculture of the sorted G-MDSCs (Fig 5a, P2) with S. aureus biofilm did not cause expansion of macrophages (F4/ 80 + ) and M2-macrophages (F4/80 + CD206 + ) (Fig 5b bottom, Fig 5e and 5f). These results indicated that S. aureus biofilm is capable of promoting the conversion of M-MDSCs (CD11b + Ly6-C high Ly6G -) into M2-macrophages (F4/80 + CD206 + ) in vitro.

S. aureus biofilm enhances the immunosuppressive activity of BMCs and the CD11b-positive MDSCs in vitro
In addition to the expansion of the immunosuppressive cells induced by S. aureus biofilm, we next determined whether S. aureus biofilm substantially affected their immunosuppressive (c) by flow cytometry. Symbol # indicates significant difference vs. day 0 (p < 0.01) and symbol * indicates significant difference vs. the shame group (p < 0.01). https://doi.org/10.1371/journal.pone.0183271.g002 Expansion, activation and polarization of MDSCs by Staphylococcus aureus biofilm function in vitro. Herein, mouse BMCs were first left untreated or treated with S. aureus biofilm (0.2 mg/ml) for 48 hr. When the untreated or biofilm-treated BMCs were cocultured with activated CFSE-labeled T cells at different ratios (1:1, 0.5:1 and 0.25:1), we found that biofilmtreated BMCs evidently displayed greater suppressive activities on T cell proliferation as compared to the untreated BMCs (Fig 6a). Since biofilm treatment could promote MDSC expansion, equal numbers of the CD11b-positive MDSCs isolated from the untreated or biofilmtreated BMCs were used to compare their immunosuppressive function. In T cell suppression assays, we consistently found that the CD11b-positive MDSCs isolated from biofilm-treated BMCs had a much stronger immunosuppressive activity than those from untreated BMCs (Fig  6b).
To further confirm the elevated immunosuppressive activity in these tested effector cells caused by biofilm, the expression patterns of various anti-inflammatory-related genes including Arginase-1, iNOS, IL-6 and IL-10 were measured in the untreated or biofilm-treated cells (Fig 7). Quantitative RT-PCR analysis showed that BMCs treated with increasing amounts of S. aureus biofilm substantially displayed enhanced expressions of Arginase-1, iNOS and IL-10 mRNAs (Fig 7a-7c), supporting that biofilm-treated BMCs have higher immunosuppressive activities than the untreated BMCs. Moreover, when gene expression patterns were compared between the CD11b-positive MDSCs isolated from the untreated and biofilm-treated BMCs, the levels of Arginase-1, iNOS and IL-10 mRNAs were also particularly higher in biofimtreated MDSC populations (Fig 7e-7g). Notably, we noticed that the levels of IL-6, which has both pro-and anti-inflammatory properties [29], could be abundantly induced in BMCs or in the CD11b-positive MDSCs when exposed to a low concentration of biofilm (0.2 mg/ml). However, treatment of these cell populations with biofilm at high concentrations (0.6 and 2.0 mg/ml) caused only a moderate increase in the levels of IL-6 as compared to the untreated controls (Fig 7d and 7h).
In addition to MDSCs and M2-macrophages, other immunosuppressive cells such as CD4 + CD25 + Foxp3 + regulatory T cells (Tregs) could possibly involve the immune modulation during S. aureus infection [30-32]. Since S. aureus biofilm significantly enhanced the immunosuppressive activity of MDSCs to inhibit T cell responses, we here attempted to determine  whether the biofilm-treated MDSCs could modulate expansion of Tregs from spleen T cells in vitro. To do the experiments, the CD11b-positive cell populations (MDSCs) were isolated from mouse BMCs that were untreated or treated with increasing amounts of biofilm for 48 hr. The isolated MDSC populations were subsequently co-cultured at the 1:1 ratio with spleen T cells that were activated by anti-CD3/CD28 beads. As compared to the coculture of spleen T cells with the untreated MDSCs, we unexpectedly found that the percentages of CD4 + CD25 + Foxp3 + Tregs in CD4 + lymphocytes were significantly elevated in the cocultures of spleen T cells with the biofilm-treated MDSCs after 24-or 48-hr coculture (Fig 8). Furthermore, the isolated MDSCs undergoing treatment for biofilm at higher concentrations had higher capacity to increase the percentages of CD4 + CD25 + Foxp3 + Tregs in CD4 + lymphocytes (Fig 8). These results strongly suggested that biofilm-treated MDSCs could suppress immunity through multiple different mechanisms.

Recruitment of the transferred EGFP-labeled, CD11b-positive MDSCs into the site of S. aureus infection
To study whether the circulating MDSCs could be recruited into the site of S. aureus infection in femurs and converted into M2-macrophages in vivo, we here employed a mouse model of PJIs to address these questions. In this model, C57BL/6J mice were infected with S. aureus for 7 days, and then 1×10 7 exogenously EGFP-labeled, CD11b-positive MDSCs were intravenously injected into the infected mice. Twenty-four hours later, tissues near the S. aureusinfected site were taken, embedded, frozen in -80˚C, cryosectioned, and then stained with  (Fig 9a and 9b, i and ii). Especially, about 85% of EGFP-positive cells were positive for both F4/80 and CD206 (Fig 9b, iii and iv). These results strongly suggested that S. aureus biofilm infection could promote the conversion of MDSCs to M2-macrophages (EGFP + F4/80 + CD206 + ) in vivo (Fig 9). Additionally, we here noticed that the overall number of M2-macrophages (F4/80 + CD206 + ) was much higher than that of F4/80 + CD206cells (probably pro-inflammatory macrophages) present in S. aureus-infected sites (Fig 9a and 9b, iii and iv). A possible pathway for the development and polarization of MDSCs during S. aureus biofilm stimulation is proposed in Fig 9c.

Discussion
S. aureus biofilm infection in PJIs is the major complication in joint reconstruction. The formation of bacterial biofilm is important not only for protection against antibiotic treatment but also for evasion of host immune responses. In the latter case, S. aureus biofilm that interferes with the immune responses involves accumulation of both MDSCs and M2-macrophages in infected sites [12,[33][34][35]. In this report, we confirmed that S. aureus biofilm triggers the expansion and activation of MDSCs and M2-macrophages in our established in vitro and in vivo models. For MDSC expansion, we found that S. aureus biofilm preferentially induces the expansion of M-MDSCs but not G-MDSCs. Importantly, our results support that M-MDSCs have the potential to further differentiate into M2-macrophages in the presence of S. aureus biofilm. To our knowledge, this is the first study showing the relationship between MDSCs and M2-macrophages in S. aureus biofilm infection. Overall, our findings provide a better  Using our established PJI model in rats, two S. aureus-infected groups that received no treatment or vancomycin treatment were evaluated. Although vancomycin treatment greatly reduced bone infection caused by S. aureus (Fig 1a), S. aureus infection might persistently damage the infected femur even standard treatment with antibiotics. Particularly, we found that aggregates of S. aureus embedded with a self-produced extracellular matrix, which adhere to bone or prosthesis surface, were still observed in the vancomycin-treated group by scanning electron microscopy (Fig 1b). On the other hand, we did not detect significant differences in the expansion rates of the circulating MDSCs, macrophages, and M2-macrophages in the peripheral blood between the two S. aureus-infected groups that received no treatment or vancomycin treatment (Fig 2). Our results emphasize that S. aureus biofilm infection in PJIs is still a great challenge for orthopedic surgeons, and antibiotic treatment may be not so effective to eliminate the persistent bacterial infection.
In addition to a rat PJI model, we here successfully established an in vitro coculture system to study effects of S. aureus biofilm on the expansion of MDSCs, total macrophages and M2-macrophages from freshly isolated BMCs (Fig 3). Originally, we attempted to coculture BMCs with S. aureus biofilm directly. However, the direct contact between BMCs and S. aureus biofilm led to rapid cell death. To avoid the over-reaction of immune cells, S. aureus biofilms in our in vitro system were loaded in transwell cell culture inserts (0.4 μm pore size) and cocultured with BMCs. Consistent with the results found in our rat PJI model, we showed that treatment of BMCs with S. aureus biofilm in our in vitro system stimulated the expansions of MDSCs (CD11b + Gr1 + ), total macrophages (F4/80 + ) and M2-macrophages (F4/80 + CD206 + ) (Fig 3). Most importantly, a dose-response relationship between S. aureus biofilm and the expansions of these immune cells could be established in the in vitro system.
MDSCs are composed of two subsets, M-MDSCs and G-MDSCs. Evidence has shown that these two subsets may manifest different functions under different circumstances [11]. Although earlier studies by Hem et al. [14] have reported that MDSCs function as a central contributor to mediate immunosuppression during S. aureus biofilm infection, the expansion and polarization of MDSC subsets in response to S. aureus biofilm remained poorly understood. Using our in vitro coculture system, we clearly showed that although both M-MDSCs and G-MDSCs could expand upon exposure to S. aureus biofilm, the expansion of M-MDSCs was much greater than G-MDSCs (Fig 3c and 3d). The maximal expansion of M-MDSC population in BMCs could be easily achieved after exposure to a low concentration (0.2 mg/ml) of S. aureus biofilm (Fig 3c and 3d; from 9.8% to 26.8%); however, higher concentrations (0.6 and 2.0 mg/ml) of S. aureus biofilm did not further elicit M-MDSC expansion. There are two possibilities that may explain these results. One possibility is that a low concentration of S. aureus biofilm may be already sufficient to confer the full induction of M-MDSC expansion from BMCs. The other possibility is that, in response to high concentrations of biofilm, the increased proportions of M-MDSCs may be able to further differentiate into another cell types, consequently maintaining a fixed level of M-MDSCs. In the study, we did find that M-MDSCs have the capacity to differentiate into M2-macrophages during S. aureus biofilm stimulation (Figs 5 and 9).
To support the immunosuppressive function of these immune cells, both BMCs and MDSCs (CD11b + ) untreated or treated with biofilm were co-cultured with activated T cells (Fig 6). Our results consistently showed that biofilm treatment significantly enhanced the suppressive activity of BMCs or the CD11b-positive MDSCs on T cell proliferation. Furthermore, quantitative RT-PCR analysis also revealed that the biofilm-treated cell populations abundantly expressed anti-inflammatory mediators, including Arginase-1, iNOS and IL-10 (Fig 7). In addition to these anti-inflammatory mediators, we noticed that IL-6 was markedly stimulated in BMCs or the MDSC (CD11b + ) population after exposure to a low concentration of biofilm (0.2 mg/mL); however, at higher concentrations of biofilm the induction of IL-6 trended toward a modest increase (Fig 7d and 7h). It may be not so surprising because IL-6 is previously known to have both pro-and anti-inflammatory properties [29]. Since S. aureus biofilm enhanced the immunosuppressive activity of MDSCs to inhibit T cell responses, we additionally investigated whether the biofilm-treated MDSCs could modulate Treg cells, another cell type with immunosuppressive function. When compared to the coculture of spleen T cells with the untreated MDSCs, we unexpectedly found that the proportions of CD4 + CD25 + Foxp3 + Tregs in CD4 + lymphocytes could be significantly increased in the cocultures of spleen T cells with the biofilmtreated MDSCs (Fig 8). These results therefore suggested that biofilm-treated MDSCs could suppress immunity through multiple different ways. Taken together, our findings support that S. aureus biofilm mediate both MDSC expansion and activation.
In conclusion, since recalcitrant PJIs are still a difficult challenge for orthopaedic surgeons, a detailed understanding of the relationship between MDSCs, M2-macrophages and biofilms may further expand the possibility of immunotherapies in treating PJIs in the future.