Profiling of inflammatory and anti-inflammatory cytokines in osteomyelitic bone tissue

Introduction

Osteomyelitis (OM) is a clinically and surgically demanding condition characterized by infection of bone tissue, often accompanied by impaired healing and long-term morbidity. Internal fixation materials, such as plates, screws, or intramedullary nails, can exacerbate infection by promoting biofilm formation, which protects pathogens from host immunity and antibiotics [1], 2]. OM may result from open fractures, surgical procedures, chronic wounds, or can be caused by hematogenous spread, and typically requires radical debridement, prolonged antibiotic therapy, and delayed reconstruction [3]. Treatment is resource-intensive and associated with high healthcare costs [4].

Despite its relevance, the molecular microenvironment within infected human bone tissue remains insufficiently characterized. Most current knowledge is derived from in vitro cell models or from preclinical animal studies [5]. However, these models fail to fully replicate the complexity of human OM, particularly the interplay between immune response and bone regeneration. Marriott et al. showed that osteoblasts from infected human bone express IL-6 [6], while Klosterhalfen et al. documented local and systemic cytokine release in OM patients [7]. Yet, cytokine profiling in native human OM tissue, especially in relation to the underlying pathogens, is lacking. In this study, we focused on three surgically relevant subtypes of OM: sternal, vascular, and posttraumatic OM. Each of these entities presents unique challenges. Vascular OM arises predominantly in patients with chronic wounds caused by persisting pressure (decreasing local blood flow as in pressure ulcers), diabetes, or peripheral arterial disease and is typically accompanied by compromised perfusion leading to insufficient blood and oxygen supply to the underlying bone tissue [8]. Posttraumatic OM is frequently associated with open fractures or hardware infections and often leads to segmental bone loss requiring complex reconstruction [9]. Sternal OM represents a hybrid entity, typically developing after cardiothoracic surgery. It is frequently associated with internal fixation materials (e.g., sternal wires), reduced vascular supply due to internal mammary artery harvest, thin soft tissue coverage in the sternal midline region, and sometimes even the risk of mediastinal involvement in terms of mediastinitis [10]. These three subtypes were selected to reflect a representative spectrum of etiologies, tissue perfusion status, and reconstructive needs in clinical practice.

Five cytokines involved in inflammation (interleukin-6 (IL-6), interleukin-8 (IL-8), monocyte chemoattractant protein-1 (MCP-1)) and tissue remodeling (osteopontin, SPARC (secreted protein acidic and rich in cysteine, also known as osteonectin)) were measured and analyzed across subgroups.

IL-6 and IL-8 are key proinflammatory cytokines in OM [11], 12]. MCP-1 plays a critical role in monocyte/macrophage chemotaxis and may reflect a more chronic inflammatory environment [13]. Osteopontin and SPARC, by contrast, are matricellular proteins involved in bone matrix remodeling, osteoblast function, and tissue regeneration [14], 15].

Materials and methods

Results

Across all OM cases (n=21), Staphylococcus epidermidis was the most frequently detected organism. In sternal OM, it was identified both as a monoinfection and in mixed infections with Pseudomonas aeruginosa and Corynebacterium tuberculostearicum. Vascular OM cases showed a broader pathogen spectrum, including S. epidermidis, Corynebacterium striatum, Enterobacter cloacae, and Proteus mirabilis. One vascular case showed no microbiological pathogen detection, but OM was confirmed by histopathological examination. Posttraumatic OM included monoinfections with Staphylococcus aureus, S. epidermidis, E. cloacae, and Klebsiella aerogenes. Additionally, three polymicrobial infections were observed (S. aureus and P. mirabilis; S. epidermidis and C. striatum; Corynebacterium propinquum and S. aureus).

Inflammatory response (IL-6, IL-8, MCP-1)

IL-6 concentrations were significantly elevated in patients with OM compared to the controls (Figure 1A). Subgroup analysis revealed significantly higher IL-6 levels in all three OM subtypes vs. controls, but no significant differences occurred between the OM subgroups (Figure 2A).

Figure 1:

Comparison of cytokine concentration between total osteomyelitis and healthy control bone tissue: (A) IL-6, (B) IL-8, (C) MCP-1, (D) osteopontin, and (E) SPARC. Boxplots showing the concentrations of IL-6, IL-8, MCP-1, osteopontin, and SPARC measured by ELISA in osteomyelitic (OM) bone samples compared to healthy controls. Statistical differences were assessed using the Mann–Whitney U test (*=p<0.05, **=p<0.01, ***=p<0.001). Boxplots show cytokine concentrations (pg/mL) in osteomyelitic (OM) bone tissue (n=21) and uninfected controls (n=23). OM includes all subtypes combined (sternal, vascular, and posttraumatic).

Figure 2:

Comparison of cytokine concentrations between osteomyelitis subgroups and healthy controls: (A) Il-6, (B) IL-8, (C) MCP-1, (D) osteopontin, and (E) SPARC. Boxplots illustrating cytokine concentrations (IL-6, IL-8, MCP-1, osteopontin, and SPARC) across different osteomyelitis subtypes (sternal, vascular, posttraumatic) compared to healthy controls. Statistical differences were determined by Kruskal–Wallis tests with FDR correction (∗=p<0.05, ∗∗=p<0.01, ∗∗∗=p<0.001). Cytokine concentrations of IL-6, IL-8, MCP-1, osteopontin, and SPARC in three OM subtypes: sternal OM (n=8), vascular OM (n=6), posttraumatic OM (n=7), and uninfected controls (n=23).

IL-8 levels were also significantly increased in osteomyelitic bone tissue compared to the control (Figure 1B). The Kruskal–Wallis H-test confirmed significant differences across all groups (p<0.001). Post-hoc analysis showed that all three OM subgroups differed significantly from the control, while no significant differences were found between the OM subgroups (Figure 2B).

MCP-1 was significantly elevated in OM compared to controls (Figure 1C). The Kruskal–Wallis test indicated overall group differences for MCP-1 levels (p<0.001). Post-hoc testing showed that all OM subgroups had significantly higher MCP-1 levels compared to controls, with a significant difference also observed between sternal and vascular OM (Figure 2C).

Pathogen-related response

To further explore the immunological response patterns associated with different bone infections, cytokine concentrations were compared across all confirmed pathogen combinations, including monomicrobial and polymicrobial infections. A heatmap was generated using min–max normalization per cytokine to highlight relative concentration patterns within each cytokine-pathogen combination group (Figure 3).

Figure 3:

Heatmap showing the relative levels of IL-6, IL-8, MCP-1, osteopontin, and SPARC measured in osteomyelitic bone samples. Samples are grouped according to the identified pathogen species, including both monomicrobial and polymicrobial infections. Color intensity reflects the relative cytokine concentration (pg/mL). Staphylococcus epidermidis monoinfection (n=6) and control samples (n=23) are the only groups with more than one case; all other pathogen combinations represent individual cases (n=1).

Notably, S. epidermidis and polymicrobial infections involving common skin bacteria (e.g., S. epidermidis, C. striatum) exhibited elevated IL-6 and MCP-1 levels. In contrast, pathogen combinations containing gram-negative species (e.g., E. cloacae, P. mirabilis, Enterobacter coli) showed high relative IL-8 and osteopontin levels.

The control group displayed uniformly low levels across all cytokines, supporting the specificity of these findings. These patterns might indicate the presence of pathogen-specific cytokine profiles, particularly distinguishing gram-positive from gram-negative infections and mono- from polymicrobial etiologies.

Cytokine correlation

Spearman correlation analysis revealed strong positive correlation between IL-6, IL-8, and MCP-1 (all p<0.001), suggesting a coordinated proinflammatory response. Osteopontin correlated significantly with all three inflammatory markers, while SPARC showed a strong correlation only with osteopontin (p<0.001), indicating a distinct regenerative pattern (Figure 4).

Figure 4:

Heatmap illustrating pairwise Spearman correlation coefficients (ρ) between cytokine concentrations across all samples (n=44; OM: n=21, control: n=23). Significance levels after false discovery rate (FDR) correction are indicated as follows: ∗=p<0.05, ∗∗=p<0.01, ∗∗∗=p<0.001. Color intensity reflects the strength of the correlation.

Discussion

This study provides novel insights into the infectious and regenerative microenvironment of human osteomyelitic bone tissue and is the first cytokine-based profiling of human OM.

The observed elevations of osteopontin and SPARC in infected samples, despite ongoing inflammation, support the hypothesis that osteomyelitic bone retains regenerative capacity. This finding challenges the prevailing assumption that chronic infection necessarily precludes tissue regeneration and may influence timing and strategy in reconstructive surgery [16].

The simultaneous upregulation of inflammatory markers (IL-6, IL-8, and MCP-1) across all OM subtypes confirms a robust and coordinated innate immune response. These cytokines are well-established mediators of neutrophil and monocyte/macrophage recruitment and activation by infection [11], [12], [13]. The strong correlations between IL-6, IL-8, and MCP-1 observed in our data support this interaction and validate our methodology as sensitive to biologically relevant immune pathways.

Our findings expand upon previous work by Crisologo et al., who reported elevated serum levels of IL-6, IL-8, and MCP-1 in patients with diabetic foot osteomyelitis. While their study focused on systemic inflammatory responses [17], our analysis of cytokines directly within infected bone tissue provides complementary and more localized insight into the osteomyelitic microenvironment.

Importantly, no statistically significant differences in regenerative markers (osteopontin, SPARC) were found between the different OM subgroups (sternal, vascular, and posttraumatic), meaning that the regenerative response is preserved across OM types, regardless of anatomical site or underlying etiology. This supports the notion that all OM subtypes may possess reconstructive potential when adequately debrided and revascularized, aligning with clinical observations that successful outcomes are achievable across diverse OM presentations [18], 19].

Nonetheless, descriptive differences in inflammatory cytokine concentration suggest subtype-specific immune dynamics. For instance, higher MCP-1 levels in sternal OM may reflect a monocyte-dominated response, possibly influenced by proximity to prosthetic material [13]. This finding could also be attributed to S. epidermidis, which was identified in all sternal OM specimens.

Although not statistically tested, we found trends toward pathogen-specific cytokine profiles. Gram-negative and polymicrobial infections tended to present elevated IL-8 and osteopontin levels, reflecting a pronounced neutrophil-driven inflammatory reaction. Gram-positive and low-virulence pathogens such as skin commensals appeared to be associated with a higher MCP-1 concentration indicating a mononuclear-dominated immune response. These patterns are consistent with the distinct immune profiles induced by different bacterial classes: neutrophil-driven responses in gram-negative infections vs. chronic, macrophage-dominated inflammation in gram-positive or low-virulence organisms [20], [21], [22], [23], [24]. This suggests that cytokine profiling, by identifying specific patterns of inflammatory and regenerative markers in infected bone tissue, could provide insights into the severity and phase of infection. This might help determine the optimal timing for surgical intervention, guide the choice and duration of antibiotic therapy based on the dominant immune response, and predict the potential for successful tissue regeneration during reconstruction.

From a regenerative standpoint, the consistent elevation of osteopontin, and to a lesser extent SPARC, across all OM subtypes implies that matrix remodeling and osteoblast activity were not entirely suppressed in infected tissue. Osteopontin’s significant correlations with IL-6, IL-8, and MCP-1 suggested that it might have served as a mediator linking inflammation to repair. In contrast, SPARC showed a strong correlation only with osteopontin, but not with inflammatory markers, indicating a potentially more autonomous regenerative role. This differentiation could reflect separate phases or parallel pathways of bone remodeling, with SPARC possibly marking later-stage extracellular matrix stabilization [25].

Some limitations must be acknowledged. First, the sample size is relatively small, particularly within the subgroup analyses, limiting statistical power to detect differences between OM subtypes. Second, confounding factors such as patient comorbidities, duration of infection, and prior antibiotic therapy could not be fully controlled or standardized across the cohort, which may have influenced local cytokine concentrations. Third, this study is cross-sectional, capturing a single time point per patient. Longitudinal sampling would be necessary to assess cytokine dynamics over the course of infection, debridement, and healing and lastly there is a lack of matrix effect validation (e.g., cytokine spiking). Last, a limitation of this study is the difficulty in distinguishing true infection from contamination, particularly in cases involving skin commensals such as S. epidermidis or Corynebacterium spp. However, all cases in this study were histopathologically confirmed, supporting the interpretation that the detected organisms represent true pathogens. Still, the possibility of colonization or mixed flora, especially in polymicrobial infections, cannot be entirely excluded.

Despite these limitations, the present findings offer proof of concept for cytokine profiling as a clinically relevant approach to assess local immune and regenerative states in osteomyelitic bone. Future studies should integrate clinical outcomes to help validate specific markers as predictors for reconstructive success or risk of reinfection. The strong correlations between IL-6, IL-8, and MCP-1 confirm a coordinated inflammatory response, consistent with their known roles in neutrophil and monocyte recruitment [26]. The link between osteopontin and all inflammatory markers further supports its dual function in immune modulation and matrix turnover [27]. In contrast, SPARC’s independent correlation profile underscores its potential as a specific marker of regenerative matrix activity independent from the tested inflammatory markers [28].

Conclusions

Osteomyelitic bone tissue exhibits both inflammatory activation and signs of regenerative potential. Elevated osteopontin and SPARC concentrations may serve as indicators of ongoing remodeling and tissue repair even in a poorly vascularized environment. Cytokine profiling in infected bone may support individualized surgical strategies and improve therapeutic outcomes in reconstructive procedures.