Background

Staphylococcus aureus (S. aureus) is the main causative pathogen of osteomyelitis in long bones and vertebrae [Full size image

CXCL9/10 can be expressed and secreted by a variety of cells, including myeloid cells and T cells [24, 25]. As T-cell cytokines in S. aureus-infected bone exhibit a late activation pattern typically observed between 7 and 14 days post-infection [26, 27], we did not assess the expression of CXCL9/10 in T cells by day 3 post-infection in our study. To evaluate cellular sources of CXCL9/10 in bone marrow myeloid cells during S. aureus infection, we isolated monocytes, bone marrow-derived macrophages (BMDMs) and neutrophils, which are main components of S. aureus abscess [28], from the femoral bone marrow in 8-week-old mice and 10-month-old mice, and challenged them with various MOI of S. aureus. The purity of isolated monocytes, neutrophils, and BMDMs from bone marrow were determined by flow cytometry with surface markers before further experiments. There was an average purity of 88.7% for CD11b+Ly6C+ monocytes, 88.37% for CD11b+Ly6G+ neutrophils and 94.93% for CD11b+F4/80+ macrophages, respectively (Fig. S4a-S4f), confirming the reliability of our cell isolations. In response to S. aureus challenge, monocytes represented the major cell type that had strikingly increased mRNA expression of both CXCL9 and CXCL10 in response to S. aureus challenge, while expression of CXCL9/10 was only partially increased or unchanged in macrophages and neutrophils upon S. aureus challenge (Fig. 3i and j). Notably, S. aureus failed to activate the expression of CXCL9/10 in all three types of cells from 10-month-old mice (Fig. 3i and j), consistent with the reduced protein levels of CXCL9/10 in the S. aureus-infected femurs from 10-month-old mice. In sum, our data suggest that monocytes might be the major cellular sources of CXCL9/10 upon S. aureus infection in the bone marrow, and that the local CXCL9/10 production response might be reduced in monocytes from middle-aged mice after S. aureus infection.

Monocyte-derived CXCL9/10 is critical for the bactericidal function of neutrophils and macrophages from bone marrow

Recent studies have indicated critical roles of monocytes in immune escape, immune tolerance and wound repair during S. aureus infection [29,30,31]. Therefore, before evaluation of the function of CXCL9/10, we evaluated whether the effect of monocytes from young mice on innate immune responses might be different from that from middle-aged mice upon S. aureus infection. We isolated bone marrow monocytes from 8-week-old and 10-month-old mice, and challenged them with S. aureus or vehicle for 12 h, before we collected and filtered the conditioned medium (CM) for bacterial killing assay in neutrophils or macrophages. Flow cytometry analysis showed that predominant neutrophils isolated from Percoll were CMFDA+PI viable cells and 30 min of S. aureus challenge had no much effect on the percentage of CMFDA+PI neutrophils (Fig. S5a and S5b). In addition, there was a notable increase in DHE-labelled ROS levels compared with non-infected neutrophils (Fig. S5c and S5d), indicating reactiveness of the isolated neutrophils to S. aureus infection. We found that neutrophils treated with CM of S. aureus-challenged monocytes from 8-week-old mice (CM-8 W Mono-S. aureus) had significantly fewer extracellular and intracellular bacterial colonies compared with those treated with CM of vehicle-treated monocytes from 8-week-old mice (CM-8 W Mono-Vehicle) (Fig. 4a-c). Noticeably, neutrophils treated with CM of S. aureus-challenged monocytes from 10-month-old mice (CM-10 M Mono-S. aureus) had similar extracellular and intracellular bacterial loadings with those treated with CM of vehicle-treated monocytes from 10-month-old mice (CM-10 M Mono-Vehicle), and had a strikingly higher bacterial loading compared with those treated with CM-8 W Mono-S. aureus (Fig. 4a-c). Similar to the observations in neutrophils, a significant decreased bacterial loading was observed in macrophages treated with CM-8 W Mono-S. aureus compared with those treated with CM-8 W Mono-Vehicle. Meanwhile, CM-10 M Mono-S. aureus-treated bone marrow-derived macrophages (BMDMs) had a similar bacterial loading with CM-10 M Mono-Vehicle-treated ones, but had a strikingly higher bacterial expansion compared with CM-8 W Mono-S. aureus-treated ones (Fig. 4d and e). These results suggest that S. aureus-challenged monocytes from young mice may improve the bactericidal function of neutrophils and macrophages, but those from middle-aged mice may not.

Fig. 4
figure 4

Conditioned medium of S. aureus-challenged monocytes from young mice improves the bactericidal function of neutrophils and macrophages. (a) Representative images of bacterial colonies on agar plates and quantification of extracellular bacterial burden in supernatants (b) and intracellular bacterial burden (c) of neutrophils. After neutrophils were pre-stimulated with conditioned medium of monocytes culture treated with PBS (CM Mono vehicle) or S. aureus (CM Mono S. aureus) for 4 h, S. aureus (MOI = 10) was co-cultured with pre-stimulated neutrophils for 30 min. CM-8 W Mono and CM-10 M Mono represent CM of monocytes isolated from the bone marrow of 8-week-old and 10-month-old mice, respectively. (d) Representative images of bacterial colonies on agar plates and (e) quantification of intracellular bacterial burden of BMDMs. After BMDMs were pre-stimulated with CM Mono vehicle or CM Mono S. aureus for 4 h, S. aureus (MOI = 10) was co-cultured with pre-stimulated BMDMs for 1 h. After removal of the extracellular S. aureus, the BMDMS were continued to be cultured with conditioned medium for 12 h. n = 3/ group. One-way ANOVA with Bonferroni post hoc test, *p < 0.05, **p < 0.01, ***p < 0.001

The chemotaxis function of CXCL9/10 is known as its action on its cognate receptor C-X-C motif chemokine receptor 3 (CXCR3) in immune cells such as macrophages, neutrophils and other immune cells [32]. Moreover, recent literature has demonstrated the protective role of CXCR3 blockade against virus infection [33, 34]. These prompted us to determine whether the presence of CXCL9/10 upon S. aureus infection might be functionally relevant to the bactericidal function of innate immune cells. We pretreated neutrophils in vitro with recombinant mouse CXCL9, CXCL10, and a combination of them to measure the bacterial killing capacity of neutrophils. Interestingly, we found that the number of extracellular bacterial colonies was significantly decreased in the wells treated by CXCL9 or CXCL10 (Fig. 5a and b). Meanwhile, we also observed a significant decline in intracellular bacterial colonies in the neutrophils treated by CXCL9 or CXCL10 (Fig. 5a and c). Whereas, the combination of CXCL9 and CXCL10 treatment didn’t further reduce either extracellular or intracellular bacterial load of neutrophils compared with treatment with either CXCL9 or CXCL10 (Fig. 5a-c). In addition, there was a significant decrease in the number of bacterial colonies in the BMDMs treated by either CXCL9 or CXCL10, and a combination of CXCL9 and CXCL10 further reduced bacterial load in BMDMs (Fig. 5d and e).

Fig. 5
figure 5

Monocyte-derived CXCL9/10 evoked by S. aureus enhance the bactericidal function of neutrophils and macrophages. (a) Representative images of the remaining extracellular and intracellular S. aureus colonies inoculated in TSA plates, and quantitative analysis of extracellular (b) and intracellular (c) bacterial colonies of neutrophils culture. After neutrophils were pre-stimulated with 100 ng/ml of recombinant CXCL9 (r-CXCL9), recombinant CXCL10 (r-CXCL10), or a combination of them for 4 h, cells were challenged with S. aureus at MOI of 10 for 30 min. n = 3/ group. One-way ANOVA with Bonferroni post hoc test, **p < 0.01, ***p < 0.001. (d) Representative images of intracellular S. aureus of BMDMs inoculated in TSA and (e) quantification of bacteria colonies of BMDMs pre-treated with r-CXCL9, r-CXCL10, and a combination of them. n = 3/ group. One-way ANOVA with Bonferroni post hoc test, ***p < 0.001. (f) Representative images of the remaining extracellular and intracellular S. aureus colonies inoculated in TSA plates from neutrophils culture, and quantitative analysis of extracellular (g) and intracellular (h) bacterial colonies. After neutrophils were pre-stimulated for 4 h with CM-8 W Mono S. aureus that had been treated with recombinant siRNA for CXCL9 (si-CXCL9), CXCL10 (si-CXCL10), a combination of si-CXCL9 and si-CXCL10, or negative control (si-NC), cells were then challenged with S. aureus at MOI of 10 for 30 min. CM-8 W Mono-S. aureus represents the CM of monocytes isolated from the bone marrow of 8-week-old and challenged by S. aureus. n = 3/ group. One-way ANOVA with Bonferroni post hoc test, *p < 0.05, **p < 0.01. (i) Representative images of intracellular S. aureus colonies of BMDMs and (j) quantification of bacterial colonies of BMDMs pretreated with CM-8 W Mono-S. aureus with the knockdown of CXCL9 and CXCL10. n = 3/ group. One-way ANOVA with Bonferroni post hoc test, *p < 0.05

We next investigated whether the elevated expression of CXCL9/10 in monocytes in response to S. aureus infection might be responsible for the elevated bactericidal activity of neutrophils and macrophages. We isolated primary monocytes from 8-week-old mice bone marrow, and treated them with siRNA fragments targeting CXCL9 or CXCL10 or control ones for 48 h before S. aureus infection. The knock-down efficiency of siRNA fragments was evaluated by qPCR, and si-CXCL9-3 and si-CXCL10-2 fragments were chosen for further experiments (Fig. S6). CM-8 W Mono-S. aureus with knockdown of CXCL9, CXCL10, or both of them was collected after 12 h of infection before treatment with neutrophils or macrophages. We found that CM-8 W Mono-S. aureus with knockdown of either CXCL9 or CXCL10 significantly increased the numbers of both extracellular and intracellular bacterial colonies of S. aureus-infected neutrophils (Fig. 5f-h). However, knockdown of both CXCL9 and CXCL10 didn’t further increase the bacterial load in neutrophils compared with knockdown of either CXCL9 or CXCL10 alone (Fig. 5f-h). In addition, the number of bacterial colonies in BMDMs was also considerably increased by the treatment of CM-8 W Mono-S. aureus with knockdown of CXCL9 or CXCL10 (Fig. 5i and j). Interestingly, knockdown of both CXCL9 and CXCL10 further increased the bacterial load in BMDMs compared with knockdown of either CXCL9 alone (Fig. 5i and j). The above data indicate that CXCL9/10 production from monocytes activated by S. aureus infection may enhance the bactericidal function of neutrophils and macrophages.

CXCR3 signaling in neutrophils and macrophages promotes their bactericidal function

To confirm the role of monocytes-originated CXCL9/10 in the activities of neutrophils and macrophages, we then determined whether blocking CXCR3 signaling could lead to an alteration in their bactericidal function. Blocking CXCR3 with AMG487 increased both the extracellular and intracellular bacterial burdens of neutrophils (Fig. 6a-c). A similar phenotype with an increased S. aureus burden was observed in AMG487-treated BMDMs (Fig. 6d and e). A recent study indicated a bactericide/bacteriostatic function of Teleost CXCL10 against fish pathogens [35]. Thus, to exclude the possible direct effect of CXCL9/10 on suppressing S. aureus, we explored whether inhibition of CXCR3 might block the effect of CXCL9/10 on the bactericidal function of neutrophils and macrophages. As expected, AMG487 treatment significantly ablated the positive effect of CXCL9 or CXCL10 on extracellular and intracellular bacterial killing of neutrophils (Fig. 6f-k). In addition, either CXCL9- or CXCL10-driven antimicrobial activity of macrophages against S. aureus was also blocked by AMG487 (Fig. 6l-n). The above data demonstrated that S. aureus-challenged monocytes from young mice may enhance bactericidal activity of neutrophils and macrophages through the CXCL9/10-CXCR3 axis.

Fig. 6
figure 6

CXCR3 signaling in neutrophils and macrophages promotes their bactericidal function. (a) Representative images of the remaining extracellular and intracellular S. aureus colonies of neutrophils inoculated in TSA, and quantitative analysis of extracellular (b) and intracellular (c) bacterial colonies. Neutrophils were pre-stimulated with AMG487 (500 µM), an inhibitor of CXCR3, for 1 h, and then cultured in CM-8 W Mono-S. aureus for 4 h. Next, cells were challenged with S. aureus (MOI = 10) for 30 min, and bactericidal function were evaluated. n = 3/ group. Student’s t test, *p < 0.05, **p < 0.01. (d) Representative images of the remaining intracellular S. aureus colonies of BMDMs inoculated in TSA, and (e) quantification of bacterial colonies. BMDMs were pre-stimulated with 500 µM AMG487 for 1 h, and then further stimulated with CM-8 W Mono-S. aureus for 4 h. Next, cells were challenged with S. aureus (MOI = 10) for 1 h. After removal of the extracellular S. aureus, culture of the BMDMS was continued with CM-8 W Mono-S. aureus and AMG487 for 12 h and bacterial loading were evaluated. n = 3/ group. Student’s t test, *p < 0.05. (f and i) Representative images of the remaining extracellular and intracellular S. aureus colonies of neutrophils inoculated in TSA, and quantitative analysis of extracellular (g and j) and intracellular (h and k) bacterial colonies. Neutrophils were pre-stimulated with 500 µM AMG487 for 1 h, and then further treated with 100 ng/ml of r-CXCL9 or r-CXCL10 for 4 h. Next, S. aureus (MOI = 10) was co-cultured with pre-stimulated neutrophils for 30 min. Finally, the extracellular and intracellular bacterial loading were evaluated. n = 3/ group. Student’s t test, *p < 0.05, ***p < 0.001. (l) Representative images of remaining intracellular bacteria of BMDMs inoculated on TSA and (m and n) quantification of bacterial colonies in BMDMs. BMDMs were pre-stimulated with 500 µM AMG487 for 1 h, and then further stimulated with 100 ng/ml of r-CXCL9 or r-CXCL10 for 4 h. Next, cells were challenged with S. aureus (MOI = 10) for 1 h. After removal of the extracellular S. aureus, culture of the BMDMS was continued with AMG487 and with the presence or absence of r-CXCL9 or r-CXCL10 for 12 h. n = 3/ group. Student’s t test, *p < 0.05, **p < 0.01

Reduced production of CXCL9/10 in monocytes upon S. aureus challenge contributes to compromised antimicrobial function in middle-aged mice

To test the protective role of CXCL9/10 in an early stage of S. aureus osteomyelitis in mice, we administrated recombinant CXCL9 or CXCL10 to 10-month-old mice with S. aureus osteomyelitis. We found that either CXCL9- or CXCL10-treated mice had a strikingly smaller bacterial burden in bone marrow compared with vehicle-treated ones, as revealed by immunofluorescence staining of S. aureus in infected femurs (Fig. 7a, b, e and f). Additionally, vehicle-treated mice showed extensive cell necrosis and bone marrow lesions around the S. aureus-infected site, while either CXCL9 or CXCL10 treatment significantly improved cellular morphology and tissue structure in S. aureus-infected bone marrow (Fig. 7c, d, g and h).

Fig. 7
figure 7

CXCL9/CXCL10-CXCR3 signaling mediates age-related lesions in the acute phase of S. aureus osteomyelitis in mice. (a) Representative images of immunofluorescence staining for S. aureus and (b) quantification of S. aureus-positive stained area per area of field of view (FOV). (c) Representative images of H&E staining and (d) quantification of necrotic area per area of FOV. 10-month-old mice were injected with 1 µl r-CXCL9 (100 ng/µl) into the bone marrow where the implant was placed and S. aureus was injected, and right femurs were collected by day 3 after surgery for further analysis. n = 3/group. Scale bar = 50 μm. (e) Representative images of immunofluorescence staining for S. aureus and (f) quantification of S. aureus-positive stained area per FOV area. (g) Representative images of H&E staining and (h) quantification of necrotic area per area of FOV. 10-month-old mice were injected with 1 µl r-CXCL10 (100 ng/µl) into the bone marrow where the implant was placed and S. aureus was injected, and right femurs were collected by day 3 after surgery for further analysis. n = 3/group. Scale bar = 50 μm. (i) Representative images of immunofluorescence staining for S. aureus and (j) quantification of S. aureus-positive stained area per FOV area. (k) Representative images of H&E staining and (l) quantification of necrotic area per area of FOV. After being implanted and infected with S. aureus or treated with vehicle in right femurs, 8-week-old mice were injected subcutaneously with AMG487 (5 mg/kg, twice a day), right femurs were collected for further analysis by day 3 after surgery. n = 3/group, Student’s t test. *p < 0.05, **p < 0.01, ***p < 0.001. Red fluorescence indicates S. aureus positive staining, and blue DAPI stained nucleus. The red star indicates the cellular lytic changes, the green arrow nuclear fragmentation

To determine whether enhanced production of CXCL9/10 in monocytes from young mice might account for increased bactericidal function in neutrophils and macrophages, we applied AMG487 to 8-week-old mice with S. aureus osteomyelitis. Immunofluorescence staining for S. aureus showed significant expansion of bacteria in the bone marrow in AMG487-treated mice compared with vehicle-treated ones (Fig. 7i and j). In line with the increased bacterial loading in bone marrow, cells necrosis and bone marrow lesions were considerably increased after AMG487 treatment in 8-week-old mice (Fig. 7k and l).