Introduction

Traumatic osteomyelitis (TO) represents one of the most clinically challenging bone infections, with common etiologies including open fractures, hematogenous bacterial dissemination, and perioperative contamination during orthopedic procedures1.Infection constitutes the primary cause of nonunion in open fractures and the leading reason for failure in joint replacement surgeries2. Notably, approximately 40% of patients with bone infections develop recurrent or chronic infections3.Patients with TO typically require multiple surgical interventions, endure prolonged treatment courses, and face high recurrence rates with poor prognoses—all contributing to significant psychological distress and socioeconomic burdens4.

The treatment of traumatic osteomyelitis presents significant clinical challenges. While traditional approaches including repeated debridement and systemic antibiotic administration can provide partial infection control, they frequently fail to substantially shorten treatment duration or reduce recurrence rates5. Two pathophysiological barriers primarily account for these therapeutic difficulties: First, the compromised local microenvironment perpetuates infection. Trauma-induced tissue damage creates an ideal niche for bacterial colonization and dissemination6. The resulting ischemic conditions severely limit antibiotic penetration to therapeutic concentrations7, making protection of compromised tissues critically important8. This microenvironment not only fosters bacterial resistance but also complicates dose escalation due to systemic toxicity.Second, the virulence mechanisms of predominant pathogens prove particularly formidable. Staphylococcus aureus (S. aureus), responsible for approximately 75% of TO cases9, employs sophisticated evasion strategies including biofilm formation and host cell internalization10. Biofilms establish protective microenvironments within hours of infection, enhancing bacterial survival while promoting proliferation11,12. Through invasion of osteoblasts, fibroblasts, and endothelial cells, S. aureus evades immune clearance while inducing host cell necrosis—thereby generating new bacterial adhesion sites13. These mechanisms underscore why current research prioritizes enhancing local antibiotic delivery alongside surgical debridement14.

Local antibiotic delivery has emerged as a cornerstone therapeutic strategy, achieving high lesion-site concentrations while minimizing systemic exposure15. Since Buchholz and Engelbrecht’s seminal 1970 demonstration of antibiotic-loaded bone cement efficacy (reducing infection rates from 6–1.6%)16, this approach has gained widespread adoption. However, limitations including poor antibiotic elution, narrow spectrum coverage, and non-degradability17,18,19 may contribute to treatment failure and resistance development20. These shortcomings have driven innovation in biodegradable, sustained-release platforms21,22,23,24, among which poly(lactic-co-glycolic acid) (PLGA) excels due to its excellent biocompatibility, predictable degradation kinetics, and minimal immunogenicity25.

The bacteriological profile of TO demonstrates S. aureus predominance alongside environment-dependent pathogens26. In this era of escalating resistance, optimal antibiotic selection must balance broad-spectrum coverage with target specificity27. Moxifloxacin (MOX), a fourth-generation fluoroquinolone, exerts bactericidal effects through specific inhibition of bacterial DNA gyrase and topoisomerase IV. This mechanism effectively suppresses biofilm formation by Staphylococcus aureus and Pseudomonas aeruginosa while enabling penetration through established biofilm matrices28.Conversely, rifampicin (RIF)acts by binding the β-subunit of bacterial RNA polymerase, disrupting transcription. This agent demonstrates exceptional tissue penetration and intracellular bactericidal activity, including potent efficacy against biofilm-embedded microorganisms29,30.The combined application of MOX and RIF can significantly improve the antibacterial effect31.While numerous advanced drug delivery platforms have been developed for antibiotic encapsulation32,33,34, the majority of existing research has focused on single-antibiotic systems, with few reports addressing co-delivery of multiple antimicrobial agents.In our previous research, through technological innovation, we simultaneously encapsulated MOX and RIF in PLGA to prepare M/R-P microspheres. We evaluated and reported the characterization, drug sustained release, and in vitro antibacterial rate of the microspheres. Both quantitative and qualitative antibacterial assessments demonstrated that the dual drug-loaded microspheres exhibited superior antibacterial activity compared to single drug-loaded formulations35. This study aims to achieve the combined application of antibiotics by optimizing the PLGA vector technology, improve the therapeutic effect of local treatment for chronic bone infections, and provide a new idea for completely solving the problem of traumatic bone infections in the future.

Materials & methods

Materials

PLGA (50:50 lactide: glycolide, MW 10 kDa; Shandong Medical Instrument Co.). RIF and MOX were purchased from Sigma-Aldrich (USA). The Cell Counting Kit-8 (CCK-8) and alkaline phosphatase (ALP) assay kit were acquired from Beijing Solaibao Technology Co., Ltd. (China). All other reagents were obtained from Sigma-Aldrich and used without further purification.

Preparation of microspheres

M-P (MOX-loaded) and M/R-P microspheres were prepared using our previously reported method35. Briefly, 240 mg PLGA and 30 mg RIF were dissolved in dichloromethane (DCM) as the oil phase (O). One milliliter of MOX solution (30 mg/mL) served as the internal water phase (W1), which was emulsified into the oil phase under continuous stirring to form a primary water-in-oil emulsion (W1/O). This emulsion was then added to 40 mL of 1% polyvinyl alcohol (PVA) solution and homogenized to form a double emulsion (W1/O/W2). The resulting emulsion was transferred into 800 mL of 0.5% PVA solution and stirred for 10 h to allow solvent evaporation.The microspheres were collected by centrifugation (12,000 rpm, 10 min), washed three times with deionized water, and lyophilized (Scientz-10 N, Ningbo Xinzhi Biotechnology Co., Ltd.) at −40 °C for 24 h. M-P microspheres were prepared similarly without rifampicin. All microspheres were stored in light-protected containers at 4 °C until use.The drug loading capacity of MOX-PLGA microspheres was 2.10% ± 0.12%, with an encapsulation efficiency of 17.35% ± 2.42%. For M/R-P microspheres, the loading capacities were 3.42% ± 0.74% (MOX) and 5.03% ± 0.85% (RIF), with encapsulation efficiencies of 33.25% ± 7.51% and 49.0% ± 11.25%, respectively. Prior to use, all microspheres were sterilized by γ-irradiation (25 kGy total dose, delivered at 5 kGy/h for 5 h) under nitrogen-purged vacuum packaging.

.

Cell culture

Human bone marrow-derived mesenchymal stem cells (hMSCs; Cyagen Biosciences, Shanghai, China) were cultured in α-minimum essential medium (α-MEM) supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 µg/mL streptomycin at 37 °C in a humidified 5% CO₂ atmosphere. The medium was refreshed every 48 h, and passage 3 cells were used for all experiments.

In vitro cytocompatibility evaluation

Cell viability assay (CCK-8)

hMSCs (1 × 10⁴ cells/well) were seeded in 96-well plates with 5 mg microspheres/well (n = 6 per group). Control wells contained cells only. After 3 and 7 days, cells were washed with PBS and incubated with 100 µL α-MEM containing 10 µL CCK-8 reagent for 30 min. Absorbance was measured at 450 nm using a microplate reader.

Alkaline phosphatase (ALP) activity

Following the same seeding protocol, after the cells adhered to the wall for 24 h, cells were cultured in osteogenic induction medium (Guangzhou Saiye Biotechnology Co., Ltd.) for 3 and 7 days. Cells were lysed with 0.1% Triton X-100, and ALP activity in the supernatant was measured using p-nitrophenyl phosphate as substrate (absorbance at 520 nm).

Scanning electron microscopy (SEM)

hMSCs (1 × 10⁴ cells/mL) were co-cultured with microspheres on cell culture slides for 3 and 7 days. Samples were fixed with 2.5% glutaraldehyde (12 h), dehydrated in an ethanol series, critical-point dried, sputter-coated with gold, and imaged using SEM (Hitachi SU8010, Japan).

Live/dead staining

Cells were stained with Calcein-AM (2 µM) and propidium iodide (4.5 µM) for 20 min at room temperature. Fluorescence images were acquired using confocal laser scanning microscopy (CLSM; Zeiss LSM 880, Germany).

Cytoskeletal staining

Cells were fixed with 4% paraformaldehyde, permeabilized with 0.4% Triton X-100, stained with FITC-phalloidin (1:200) and DAPI (1 µg/mL), and imaged by CLSM.

In vivo therapeutic efficacy assessment

Bacterial culture

Methicillin-resistant Staphylococcus aureus (MRSA; Purity Biotechnology, Wuhan, China) was cultured on LB agar at 37 °C for 24 h. Bacterial suspensions (1.2 × 10⁷ CFU/mL in PBS) were prepared for animal inoculation.

Initiate of animal model of TO

All animal experiments were approved by the Institutional Animal Protection and Use Committee of Ningxia Medical University (No. 2020-0001). Twenty-four male SD rats (4 weeks old, weighing 160–180 g, purchased from the Animal Experiment Center of Ningxia Medical University) were used to establish the animal model of osteomyelitis36.Rats stratified by weight using block randomization (GraphPad), Radiographic scoring performed by two independent blinded veterinarians, Buprenorphine (0.05 mg/kg) administered pre-op and q12h × 48 h, One control-group rat died at week 6 (infection) and was excluded.

The animals were anesthetized with isoflurane (RWD, China), and the proximal tibial metaphysis was exposed through an anteromedial incision of the right knee. A bone marrow cavity was created using a bone drill, into which a Kirschner wire (approximately 5.0 × 1.0 mm) was implanted. Subsequently, 0.2 mL of MRSA suspension (1.2 × 10⁷ CFU/mL) was injected into the cavity. The defect was sealed with bone wax, irrigated three times with normal saline, and the wound was sutured. No antibiotics were administered during or after the procedure.

Postoperatively, three rats were randomly selected for radiographic confirmation of Kirschner wire placement. Then all the rats were randomly divided into 3 groups(n = 8 per group). At week 4, three rats from each group underwent imaging examination followed by euthanasia for pathological specimen collection. The remaining animals received surgical intervention consisting of Kirschner wire removal and thorough debridement to excise all necrotic tissue. The treatment groups received either M/R-P or M-P microsphere implantation, while the control group received debridement only. At week 8, three rats per group were sacrificed after imaging for final pathological evaluation.Clarify that 8-week data = 4 weeks post-treatment.

At experimental endpoints, rats were humanely euthanized via intraperitoneal injection of pentobarbital sodium (150 mg/kg) following AVMA guidelines. Death was confirmed by absence of corneal reflex and cessation of heartbeat.

Radiographic analysis

X-ray and CT scans were performed at 0, 4, and 8 weeks. osteomyelitic changes were scored using Norden’s criteria37.

Bacterial load quantification

Infected bone tissues were homogenized, serially diluted, plated on LB agar, and incubated at 37 °C for 24 h. Colony-forming units (CFU) per gram of bone were calculated.

Histopathological examination

Tissues were fixed in 4% paraformaldehyde, decalcified in 10% EDTA (pH 7.4) for 6 weeks, paraffin-embedded, sectioned (5 μm), and stained with: H&E: Inflammatory cell infiltration and bone destruction; Masson’s trichrome: Collagen deposition and bone remodeling.Images were acquired using a Nikon Eclipse Ti microscope (Japan).

Statistical analysis

Data were analyzed by two-way ANOVA using GraphPad Prism 9.0 and presented as mean ± SD (n ≥ 3). Significance was set at *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001; “ns” denotes not significant. Intergroup differences at the same time point are indicated by *, while temporal differences within groups are marked by #.

Result

In vitro cytocompatibility of microspheres

The CCK-8 assay demonstrated progressive cell proliferation across all groups over time (P < 0.0001). No significant differences in cell viability were observed among groups at either 3-day or 7-day time points (P > 0.05, Fig. 1A), indicating that microsphere addition of microspheres did not affect hMSC proliferation.

Fig. 1
figure 1

After 3 and 7 days of co-culture of cells and microspheres, cell activity was detected by CCK-8 kit (Fig. 1A) and cell ALP value was detected by ALP kit (Fig. 1b).

Full size image

Alkaline phosphatase (ALP) activity, a key marker of early osteogenic differentiation, showed temporal increases in all groups under osteoinductive conditions (P < 0.0001). Similar to proliferation results, no intergroup differences in ALP activity were detected at either time point (P > 0.05, Fig. 1B), confirming that microspheres did not impair osteogenic potential.

The SEM image (Fig. 2) observed that hMSCs could attach to microspheres for proliferation. 3D images of live and dead cells after staining (green fluorescence indicates live cells and red fluorescence indicates dead cells), as shown in Fig. 3, rarely can the red fluorescence symbolizing dead cells be observed.

Fig. 2
figure 2

SEM photos of M/R-P microspheres and M-P microspheres and cells co-cultured for 3 and 7 days, respectively.

Full size image
Fig. 3
figure 3

3D images of living and dead cells stained under CLSM after co-culture of M/R-P microspheres and M-P microspheres with cells for 3 and 7 days, respectively. The white arrows indicate the positions of the microspheres.

Full size image

The three groups of samples at two time periods were subjected to Cytoskeletal staining using TRITC-Phalloidin AF555 staining solution, and the cell nuclei were labeled with DAPI staining solution. The morphology of the cytoskeleton was observed. The 3D image (Fig. 4) shows that the hMSCs skeleton around the microspheres is intact.

Fig. 4
figure 4

3D images of cytoskeletal protein staining after co-culture of M/R-P microspheres and M-P microspheres with cells for 3 and 7 days, respectively. The white arrows indicate the positions of the microspheres.

Full size image

In vivo therapeutic efficacy

Imaging evaluation

Serial radiographic assessments at 0, 4, and 8 weeks (Fig. 5) demonstrated: Week 0: Successful Kirschner wire placement in tibial marrow cavity. Week 4: Characteristic osteomyelitic changes in all groups, including: reduced bone density, peri-implant abscess formation, trabecular thinning/erosion, necrotic bone formation, periosteal reaction.Week 8:Control group: Progressive bone destruction with irregular reactive bone formation. M-P group: Mild remission. M/R-P group: The osteomyelitic symptoms demonstrated marked improvement, with near-complete resolution of both osseous abscesses and lytic bone lesions.One control animal succumbed to severe infection during the study (Table 1).Clarify that 8-week data = 4 weeks post-treatment.

Fig. 5
figure 5

X-ray and CT images of each group at week 0, week 4 and week 8.

Full size image
Table 1 Schedule.
Full size table

Using Norden’s radiographic scoring system37(Fig. 6A; Table 2):Week 4: All groups scored ≥ 4.5 (P ≥ 0.05), confirming model establishment.Week 8:M/R-P: Scores ≤ 0.5 (P < 0.0001 vs. week 4), indicating near-complete resolution.M-P: Scores ≤ 3 (P < 0.0001 vs. week 4), showing partial improvement.Control: Scores ≥ 6 (P = 0.0095 vs. week 4), demonstrating progression.Intergroup comparisons at week 8 showed significant differences (P < 0.0001).

Fig. 6
figure 6

(A) Imaging scores. (B) Statistical analysis of colony culture counts in the control group and the two microsphere groups.

Full size image
Table 2 Oentgenographic criteria.
Full size table

Bacterial quantification

Bacterial loads (CFU/g bone tissue)(Fig. 6B): Week 4:M/R-P: 833 ± 96 × 10³.M-P: 873 ± 45 × 10³.Control: 789 ± 29 × 10³.(P > 0.05 among groups).Week 8:M/R-P: 8 ± 3 × 10³ (P = 0.0085 vs. week 4).M-P: 353 ± 72 × 10³ (P = 0.1268 vs. week 4).Control: 3143 ± 727 × 10³ (P < 0.0001 vs. week 4).Significant differences at week 8:M/R-P vs. control: P < 0.0001.M-P vs. control: P < 0.0001.Notably, M/R-P vs. M-P: P = 0.2545.

Histopathological assessment

H&E staining (Fig. 7):Week 4: all groups showed: extensive inflammatory infiltration, interstitial hemorrhage.abscess formation.Week 8:M/R-P: Reduced inflammation dramatically, Visible osteoblast rimming, Partial marrow reconstitution.M-P: Reduced inflammation mildly, Organized lymphoplasmacytic infiltrates.Control: Severe neutrophilic infiltration, Confluent abscess formation, Prominent osteolysis with empty lacunae, Marrow space obliteration by inflammatory exudate.

Fig. 7
figure 7

Pathological images of HE staining of rats in each group.

Full size image

Masson’s trichrome staining (Fig. 8):Week 4: All groups exhibited: Disrupted collagen architecture (blue stain), Minimal osteoid formation, Fragmented trabeculae.Week 8:M/R-P: Robust trabecular bone formation, Active osteoid seams, Bridging trabeculae formation.M-P: Organized collagen deposition.Control: Few collagen deposition, Advanced bone loss.

Fig. 8
figure 8

Masson staining pathological images of rats in each group.

Full size image

Discussion

Antibiotics play a pivotal role in osteomyelitis management; however, systemic intravenous administration often yields suboptimal therapeutic outcomes in TO due to multiple barriers including bacterial biofilm formation and ischemic conditions in infected regions10,38. Localized sustained-release antibiotic delivery systems have emerged as a promising strategy to overcome these challenges, offering enhanced therapeutic efficacy while minimizing systemic toxicity39. While antibiotic-loaded poly(methyl methacrylate) (PMMA) remains the current clinical gold standard for osteomyelitis treatment, it presents several inherent limitations: (1) the release of antibiotics is limited and uneven, (2) incompatibility with a variety of antibiotics, (3) necessity for secondary removal surgery and potential reinfection risk. These drawbacks have driven the development of biodegradable alternatives capable of localized antibiotic delivery while promoting osteogenesis40. In our previous study, we successfully co-encapsulated RIF and MOX in PLGA microspheres (M/R-P), demonstrating favorable drug loading efficiency, sustained release kinetics, and potent antibacterial activity in vitro35.The current study further validates the cytocompatibility and in vivo therapeutic efficacy of these microspheres.

PLGA has gained prominence in biomedical applications due to its excellent biocompatibility and tunable degradation properties. As its degradation products (lactic and glycolic acids) are endogenous metabolites, PLGA offers superior biosafety profiles41,42. However, excessive accumulation of acidic degradation byproducts may locally decrease tissue pH, potentially triggering inflammatory responses and impairing osteogenesis—thereby hindering bone infection resolution9,43. Our in vitro cytocompatibility assessments with hMSCs revealed no significant adverse effects of PLGA microspheres or their degradation products on cell proliferation, adhesion, or osteogenic differentiation (as measured by ALP activity). These findings align with most published reports, though it should be noted that existing literature suggests potential cell viability inhibition due to excessive acidification from rapid PLGA degradation44. To avoid this problem, we chose high-molecular-weight PLGA for the preparation of microspheres, as the degradation rate of PLGA is related to its own molecular weight. The higher the molecular weight, the slower the degradation45.

Severe open fractures and implant-related procedures represent major risk factors for traumatic bone infections46, with S. aureus being the predominant pathogen2,13,45. In this study, we successfully established a clinically significant TO model by surgically implanting Kirschler needles and introducing MRSA into the tibia of rats47. In clinical practice, the acute phase of TO (prior to sequestrum formation) typically responds well to surgical debridement combined with systemic intravenous antibiotics39, making this stage less therapeutically challenging. Consistent with previous studies, our intervention targeted the chronic phase of bone infection characterized by radiographic evidence of bone destruction, sequestrum formation, and sclerotic changes—conditions where systemic antibiotic penetration becomes severely limited36,48. The in vivo results demonstrated superior infection control in the M/R-P microsphere group compared to controls. The limited therapeutic efficacy observed in the M-P group may be attributed to the relatively low encapsulation efficiency of MOX, resulting in suboptimal local drug concentrations. Notably, during microsphere preparation, we found that co-encapsulation of hydrophobic RIF with hydrophilic MOX significantly enhanced drug loading capacity compared to MOX encapsulation alone. However, whether such hydrophilic-hydrophobic drug co-encapsulation universally improves encapsulation efficiency remains unexplored in current literature, with no existing reports addressing this specific phenomenon.

The M/R-P treatment group exhibited the most pronounced symptomatic improvement, likely attributable to comprehensive bacterial clearance that established a pro-regenerative microenvironment for endogenous bone repair. While RIF has been reported to promote M2 macrophage polarization—potentially contributing to both anti-inflammatory effects and osteogenesis49—the exact immunomodulatory mechanisms in TO warrant further investigation to fully harness its therapeutic potential in bone repair50.

Although positive experimental results were achieved, this experiment has many limitations: In the staining experiment of live and dead cells, we only set up the M/R-P group, the M-P group, and the control group with only cells, and did not set up a negative control group. In order to observe the growth state of cells and microspheres after co-culture, all our fluorescence images were 3D images. Due to technical limitations, fluorescence quantitative analysis could not be carried out.In vivo PLGA degradation and local pH changes were not quantified; future studies should measure inflammatory cytokines (e.g., IL-6, TNF-α).Furthermore, in animal experiments, the number of animals we used was relatively small, quantitative radiographic analysis was not systematically performed, while bone regeneration was inferred from histology but requires µCT confirmation. Moreover, traumatic bone infections are usually combined with osteonecrosis and bone defects. Although the use of M/R-P microspheres alone can effectively control the infection, it does not promote osteogenesis. In our future research, adding bone-promoting components such as magnesium ions to microspheres, or combining them with scaffolds such as hydroxyapatite/chitosan, collagen/BMP-2, etc., to have both anti-infection and bone regeneration capabilities will be our research focus.

Conclusion

Based on the previous studies, this research further conducted studies on the cytocompatibility and in vivo therapeutic effects of microspheres, and completed a series of studies on the preparation, properties, in vitro antibacterial and cytocompatibility, and in vivo treatment of traumatic bone infections in animals of M/R-P microspheres. Although the experimental results have many limitations, this study proved the feasibility of the concept of using PLGA as a carrier encapsulate two antibiotics and achieve the combined application of antibiotics in local treatment of TO. It provides a new idea for thoroughly solving the problem of TO in the future.