Antibiotic bone cement accelerates diabetic foot wound healing: Elucidating the role of ROCK1 protein expression

Abstract

Clinical studies indicate antibiotic bone cement with propeller flaps improves diabetic foot wound repair and reduces amputation rates, but the molecular mechanisms, particularly key proteins' role remain largely unexplored. This study assessed the efficacy of antibiotic bone cement for treating diabetic foot wounds, focusing on molecular impact on ROCK1. Sixty patients were randomized into experimental (EXP, n = 40) and control (CON, n = 20) groups, treated with antibiotic bone cement and negative pressure. Wound healing rate, amputation rate, wound secretion culture and C‐reactive protein (CRP) changes, were monitored. Comprehensive molecular investigations were conducted and animal experiments were performed to further validate the findings. Statistical methods were employed to verify significant differences between the groups and treatment outcomes. The EXP group showed significant improvements in wound healing (${\chi }^2$ = 11.265, p = 0.004) and reduced amputation rates. Elevated levels of ROCK1, fibroblasts and VGF were observed in the trauma tissue post‐treatment in the experimental group compared to pre‐treatment and the control group (all p < 0.05). Improved trauma secretion culture and CRP were also noted in the EXP group (all p < 0.05). The study suggests that antibiotic bone cement enhances diabetic foot wound healing, possibly via upregulation of ROCK1. Further research is needed to elucidate the underlying molecular mechanisms and broader clinical implications.

1. INTRODUCTION

Diabetic foot ulcers present a major clinical challenge, with less than 70% healing ¹ and one‐third leading to lower limb amputation. ² The condition has reached epidemic proportions ³ and imposes significant clinical and quality‐of‐life burdens. Better prevention and treatment of the disease would reduce the rate of amputation by 80% ² Therefore, it is important to further improve the understanding of the mechanisms of diabetic foot wound formation and abnormal healing, and to find more effective treatment modalities.

Diabetic foot wounds present a challenging clinical landscape, primarily due to their complex underlying mechanisms. These wounds are often perpetuated by a triad of issues: chronic inflammation, poor blood flow and inadequate granulation tissue growth. ⁴ Addressing these wounds necessitates a multi‐faceted approach. The first line of treatment typically involves extensive surgical debridement, executed as promptly as possible. This intervention is particularly aimed at removing necrotic tissue—comprising degenerated muscle, tendon and bone—especially in cases where osteomyelitis has set in. Concurrently, blood glucose levels are rigorously managed to further facilitate the healing process. ⁴ However, the challenge does not end with debridement; the severity of the infection in diabetic foot wounds often precludes immediate flap or skin grafting. Instead, a subsequent phase focused on infection control and delayed wound repair is initiated. During this critical period, the clinical objective shifts towards preparing a favourable local environment. This involves promoting cell proliferation and fostering the rapid growth of fresh granulation tissue at the wound's base. Such preparatory steps are indispensable for creating a viable soft tissue bed, conducive for the subsequent application of flap or skin grafts. It is essential to highlight that the choice of temporary wound coverage can significantly impact clinical outcomes. ⁴

Within this context, the closed negative pressure technique stands out as the most widely adopted clinical strategy for post‐debridement wound management. This technique is generally considered safe and effective for treating diabetic foot wounds, with a relatively low risk of serious complications. ⁵ , ⁶ , ⁷ However, it is crucial to note that this method is not without limitations, particularly for cases involving chronic deep infections or higher‐grade diabetic foot wounds (Grades III–VI). These limitations ⁸ , ⁹ primarily encompass a weak local antimicrobial capacity; inadequate drainage frequently aggravates the infection and intensifies the necrosis of adjacent tissues. Additionally, there is an inability to maintain the viability of exposed bone, tendons and other tissues. Lastly, a common occurrence is that the granulation tissue often finds itself in a state of low viability.

Given these limitations, there is a pressing need for alternative methods that can address these challenges more effectively. One such promising development is the use of antibiotic bone cement. Inspired by the induced membrane technique (IMT), a method initially designed for bone defects, antibiotic bone cement, has recently been applied with satisfactory clinical results in treating various chronic refractory wounds, including diabetic foot ulcers. ¹⁰ , ¹¹ , ¹² Moreover, a clinical study found that repairing diabetic foot and ankle wounds with antibiotic bone cement combined with propeller flaps can provide better infection control with excellent results, which is beneficial to the repair of diabetic foot wounds. ¹³ This approach has shown promise in reducing the amputation rate and improving the healing rate. ¹⁴ However, the molecular mechanisms underlying the enhanced wound healing observed with antibiotic bone cement remain largely unexplored, particularly the role of key proteins in promoting tissue proliferation and angiogenesis.

Our research aims to bridge this gap be delving beyond the scope of existing studies, which are mostly confined to evaluating clinical outcomes. Specifically, we focus on understanding both the clinical efficacy and the molecular mechanisms by which antibiotic bone cement accelerates diabetic wound healing. This includes examining changes in the expression level of certain sensory effector proteins or genes in the wound and evaluating the enhancement in the proliferation and migration of fibroblasts and vascular endothelial cells. Moreover, a rigorous evaluation of clinical efficacy is undertaken to ascertain the impact on amputation rates, wound healing rates and infection control. Additionally, understanding the microenvironment dynamics in diabetic foot wounds post‐treatment with antibiotic bone cement is conducted. By doing so, we aspire to unveil the potential of antibiotic bone cement as a revolutionary therapeutic avenue for diabetic foot wounds, thereby contributing to the broader goal of reducing amputation rates and improving the quality of life for individuals afflicted with diabetic foot ulcers.

2. MATERIALS AND METHODS

This section unfolds a two‐pronged approach: Clinical Evaluation and Molecular Mechanism Investigation. Initially, the Clinical Evaluation encompasses participant selection and clinical procedures, focusing on the treatment and observational analysis of diabetic foot ulcers (DFUs). Following this, the Molecular Mechanism Investigation is introduced, which delves into protein workup, bioinformatics analysis, histologic and immunohistochemical analyses and animal experiments to explore the molecular underpinnings and potential therapeutic targets for DFUs. Lastly, the Section 2.7 outlines the methodologies employed for data scrutiny and validation, ensuring the robustness and reliability of the findings.

2.1. Participant selection and description

A total of 60 patients were enrolled from January 2021 to December 2022 for this study. All patients were hospitalized in the Affiliated Hospital of Zunyi Medical University, Zunyi, China. Ethical approval for this study was granted by the Ethics Committee of the Affiliated Hospital of Zunyi Medical University. Informed consent was obtained from each participant prior to their involvement in the study.

For the experimental group (labelled as EXP), 40 patients were randomly assigned using simple randomization techniques. Demographically, the group consisted of 21 males and 19 females with a mean age of 57.1 ± 8.4 years. Wagner classification was applied, including 18 patients with Grade III and 22 patients with Grade IV diabetic foot ulcers. The control group (labelled as CON) consisted of 20 patients, 12 of whom were male and remaining were female. The mean age was 56.5 ± 7.7 years. According to the Wagner classification, 13 patients were at Grade III and seven patients were at Grade IV. Statistical analyses confirmed that there were no significant differences in general demographics or Wagner classifications between the experimental and control groups (p > 0.05).

The participants eligible for this study are individuals aged between 18 and 70 years, regardless of gender, who have been diagnosed with Type 2 diabetes according to the criteria set by the WHO Diabetes Expert Committee in 2005. Additionally, they should have a confirmed diagnosis of a diabetic foot ulcer, classified as either Wagner Grade III or IV, with no serious complications or other significant diseases such as severe liver or kidney insufficiency or sepsis. It is imperative that there are no absolute contraindications to surgery and that the participants voluntarily join the study, confirmed through informed consent.

On the other hand, individuals will be excluded from the study if they have difficult to control hyperglycaemia (e.g., HbA1c >12.0%), diabetic ketoacidosis, non‐ketotic hyperosmolar coma or concurrent malignancy. Pregnant or lactating individuals, as well as those with diseases or psychiatric disorders that could interfere with treatment evaluation or successful study completion, are also excluded. The exclusion extends to individuals with any serious uncontrolled disease or acute systemic infections, significant organ insufficiencies including liver, kidney, heart, lung or brain diseases, and those who have participated in other clinical trials within 3 months prior to this study.

2.2. Clinical procedures and observations

Upon fulfilling the inclusion criteria, patients received a standardized supportive care regimen, which included glycaemic management, limb perfusion enhancement, electrolyte imbalance correction and treatment for hypoproteinaemia following admission and requisite preoperative preparations. Debridement procedures were conducted on day 2 ± 1 post‐admission for the experimental group to systematically excise necrotic tissues, encompassing both tendinous and osseous necrosis. For cases incorporating antibiotic‐loaded bone cement, the cement was applied to the debrided area as an adjunctive measure, whereas in the control group, post‐debridement wound irrigation was performed. Subsequent wound management involved either routine dressing changes or the utilization of Vacuum‐Assisted Closure® (VAC®) system (KCI, USA) for Continuous negative pressure wound therapy (NPWT).

In both the experimental and control arms, standardized pre‐debridement protocols were adhered to: collection of wound exudates for microbial culture, blood analyses focusing on infection markers such as C‐reactive protein (CRP), and photographic documentation of wound status. Following saline irrigation of the wound, collected samples were immediately cryopreserved in liquid nitrogen. By postoperative day 7 ± 2, wounds were reassessed to ascertain the need for additional debridement procedures. In the experimental group, newly exposed wound areas were managed with antibiotic‐loaded bone cement, either retained or replaced as clinically indicated. As the wounds reached the 21 ± 2‐day milestone, additional exudate cultures were performed, wound dimensions were documented and photographic records were updated. Biopsy specimens from various tissue layers were cryopreserved in liquid nitrogen. At this stage, formation of healthy granulation tissue generally signalled that the wound was conducive for closure via skin graft, suture or flap graft techniques. Another round of exudate culture was executed, followed by granulation tissue sampling for cryopreservation.

Venous blood was drawn from each patient's antecubital fossa following an 8‐h fasting period, prior to 8:00 AM, for assessment of key parameters including CRP, fasting plasma glucose (FPG), and bacterial cultures. Clinical outcomes were evaluated at the 3‐month mark using the Wagner's classification index, and efficacy rates were computed based on the formula: Total Efficacy Rate = (Number of Cured + Efficacious + Effective cases)/Total cases × 100%.

For microbial load quantification and species differentiation, swab samples were obtained from the chronic ulcerative lesions of the diabetic foot patients. These samples were vigorously agitated in 5 mL of Luria‐Bertani (LB) medium to liberate bacterial entities. Subsequent dilutions were performed at varying concentrations (undiluted, 1:10, 1:50, 1:100) to ensure thorough bacterial quantification. Aliquots of 100 μL were plated on non‐selective LB agar plates and incubated overnight at 37°C. Colonies were enumerated the following day, and bacterial isolates were harvested from 500 mL LB medium after 6–8 h of incubation. Strains were preserved with 30% glycerol and stored at −20°C for future analyses. Select isolates were submitted to the China Type Culture Collection Centre for definitive identification via 16S rDNA sequencing, thereby providing a robust analytical framework for our investigation.

2.3. Protein workup

2.3.1. Protein extraction and concentration determination

The initial phase of protein processing entailed thawing the samples on ice at 4°C, followed by the addition of lysis buffer, composed of 8 M urea, in a 1:10 weight to volume ratio to facilitate protein extraction. The samples were then subjected to ultrasonic extraction on ice, utilizing a power setting of 25%, post grinding at 60 Hz for 90 s. Subsequent to this, the samples underwent centrifugation at 14 000g for 20 min at 4°C to separate the proteins. Post centrifugation, the supernatants were collected for ensuing analysis.

Post extraction, the protein concentrations were quantified via Bradford protein assay, conforming to the manufacturer's instructions. A total of 15 μg of protein from each sample were mixed with 5× loading buffer respectively. The proteins were then separated by SDS‐PAGE gel electrophoresis and visualized by Coomassie brilliant blue staining, as per standard procedures. ¹² , ¹⁵ This separation and visualization enabled further examination and identification of the proteins present in the samples in subsequent sections.

2.3.2. Liquid chromatography‐mass spectrometry analysis for protein identification

Continuing from the previous sections, the protein workup process advanced to the liquid chromatography‐mass spectrometry (LC–MS) analysis for protein identification. The dried peptides were first dissolved in 0.1% formic acid (FA) and centrifuged at 14 000g for 10 min. Peptide concentrations were determined from the A280 measurement, with aliquots diluted in 0.1% FA, and 0.5 μg loaded onto a column for mass spectrometric analysis.

The LC–MS analysis was carried out using a Q Exactive HF‐X mass spectrometer (Thermo Fisher Scientific, USA) linked to a Nano‐electrospray Flex ion source (Thermo Fisher Scientific, USA) and an Easy‐nLC 1200 (Thermo Fisher Scientific, USA). Peptides were separated on a homemade reverse phase C18 column (1.9 μm, 75 μm × 250 mm) with a flow rate of 300 nL/min. The mobile phases comprised of 0.1% FA (phase A) and 0.1% FA in 80% acetonitrile (phase B). The linear gradient for peptide separation was as follows: 0–2 min, 5%–8% B; 2–102 min, 8%–23% B; 102–110 min, 23%–40% B; 110–112 min, 40%–100% B; 112–120 min, 100% B.

Peptide ionization was facilitated by a 2.0 kV spray voltage, with the capillary temperature maintained at 300°C. MS1 scans ranged from 300 to 1800 m/z with an automatic gain control (AGC) target value of 3e6 and maximum ion injection times of 50 ms. This was followed by high‐energy collisional dissociation‐based fragmentation at a normalized collision energy of 27 eV. A resolution of 60 000 was employed for MS1 scans. Up to 20 of the most abundant precursor ions were dynamically selected for fragmentation with an isolation window of 1.5 m/z. The MS2 data had a resolution of 15 000 (AGC target value 1e5, maximum ion injection times 50 ms). The dynamic exclusion time was fixed at 30 s, and only precursors with charge states of 2 were selected for fragmentation. ¹¹ , ¹⁶

2.3.3. Protein identification and quantification

Following the LC–MS analysis, the RAW data files were processed using Proteome Discoverer software v2.4.1.15 (Thermo Fisher Scientific, Massachusetts, USA) and cross‐referenced with the RefSeq Human protein database (24 078 sequences, release 2017_03) via the SEQUEST algorithm. The search parameters were configured as follows: Fixed modifications: Carbamidomethylation of cysteine (+57.02146 Da); Variable modifications: Oxidation of methionine (+15.99492 Da); Precursor ion mass tolerance: 20 ppm; Fragment ion mass tolerance: ± 0.05 Da; Enzyme: Trypsin, allowing up to two missed cleavages.

A reversed sequence decoy strategy was employed to regulate peptide false discovery. Identification validation was carried out using the Percolator software, stipulating a false discovery rate (FDR) of 0.01 for both proteins and peptide‐spectrum matches (PSMs). ² Protein quantification incorporated a background correction followed by a median normalization for adjusting the raw protein quantification data to account for background noise or other systematic biases. Only results exhibiting a minimum of one unique peptide were selected for further analysis.

In this study, the samples were categorized into four distinct groups to facilitate a comparative analysis of protein expression pre and post‐treatment in both experimental and control groups. The groups are defined as follows:

Group C1: Represents the protein samples collected from the experimental group before the treatment.

Group C2: Represents the protein samples collected from the experimental group after the treatment.

Group A1: Represents the protein samples collected from the control group before the treatment.

Group A2: Represents the protein samples collected from the control group after the treatment.

The intention behind this grouping is to enable a systematic analysis of the differential protein expression and to understand the impact of the treatment on protein expression levels in both experimental and control groups. The comparative analysis between these groups (C1 vs. C2, A1 vs. A2, C1 vs. A1 and C2 vs. A2) will provide insights into the molecular mechanisms underlying the observed clinical outcomes and the potential therapeutic efficacy of the treatment administered.

2.4. Bioinformatics analysis

The bioinformatics analysis commenced with a principal component analysis (PCA) to illustrate the variance across distinct biological conditions. Subsequently, an enrichment analysis was carried out to discern the augmentation of specific biological functions or pathways. For this, we utilized Fisher's exact test on differentially expressed proteins against the gene ontology (GO) ³ and Kyoto Encyclopedia of Genes and Genomes (KEGG) ¹ databases. Protein–protein interaction analyses were conducted using STRING software. ¹⁷ The visualization of these interaction networks was achieved with the igraph package. ⁴

To verify our findings, a thorough literature search was undertaken across several databases, including PubMed, Elsevier's ScienceDirect and LUBsearch. Our search strategy incorporated terms specifically pertinent to diabetic foot and chronic wound healing: ‘wound’, ‘bone cement’, ‘biomarkers’, ‘induced membrane technique’, ‘diabetic foot’ and ‘healing’. Only articles adjudged of high quality and relevance were selected for review. From these articles, we extracted data on relevant repair factors or proteins. The extracted data was then processed using a Python script to screen and align it with the differentially expressed proteins identified in our proteomic transcription dataset.

2.5. Histological and immunohistochemical analyses and validation

Wound tissues procured from patients were fixed in 4% paraformaldehyde for a duration of 24 h at 4°C, followed by embedding in paraffin. Sections of 4‐μm thickness were prepared and stained using haematoxylin and eosin, alongside Masson's trichrome staining, adhering to the manufacturer's instructions. Each stained section was meticulously examined under an optical microscope (Olympus BX53; Olympus Corp., Shinjuku, Japan) at 400× magnification, focusing on three distinct regions: right, middle and left.

For quantitative assessments: Collagen density was assessed by measuring the areas stained by Masson's trichrome and identifying collagen fibres, using the ImageJ software. Subsequently, a ratio representing collagen density was calculated. Inflammatory cell density was derived from the areas stained by haematoxylin and eosin, with the ratio being computed using ImageJ.

Specimens derived from both experimental and control cohorts were retrieved from liquid nitrogen storage and sectioned in a frozen state. Following this, these sections were fixed in 4% PFA and thoroughly rinsed. To minimize non‐specific antibody binding, sections were then blocked using 10% FCS/PBS. Thereafter, they were incubated with primary antibodies: anti‐ROCK1 (1:1200, ab134181; Abcam, Cambridge, UK) and anti‐CD31 (1:500, GB113151; Servicebio, China) overnight at 4°C. Note that ROCK1 represents Rho‐associated coiled‐coil containing protein kinase 1. Post rinsing, the sections were exposed to secondary antibodies for a duration of 1 h at room temperature.

For visualization purposes, images of the treated sections were captured at magnifications of 100× and 400×, focusing on three distinct regions: right, middle and left of each specimen. The mean values extracted from these regions were computed for further analysis.

2.6. Animal experiments

2.6.1. Diabetic mouse wound healing intervention with ROCK1 and inhibitor

Fifteen 6‐week‐old male diabetic mice (db/db) with a. average glucose level of 479 ± 17 mg/dL were obtained from Huafukang (Beijing, China). They were methodically separated into three distinct groups: ROCK1 protein (RO), GSK429286A inhibitor (RO + GS) and a control (CON), each comprising five mice. The experimental protocols received institutional approval from our animal care committee (Approval No. 2017–0028).

For the intervention, each mouse was anaesthetised, and their dorsal fur was carefully removed. Subsequently, two excisional wounds with a diameter of 10‐mm were created using a sterile biopsy tool from Miltex. The wounds of the RO, RO + GS and CON groups were treated with ROCK1, GSK429286A and normal saline, respectively, each in equivalent volumes. The dosage and timing for the administration of ROCK1 and GSK429286A were informed by a preceding study, ¹⁸ which demonstrated the benefits of injecting 20 mg/kg of either substance into mice every 0–3 days, without eliciting cytotoxicity symptoms. Aligning with this, our study encompassed three subcutaneous injections of 20 mg/kg for each substance per wound, administered immediately post‐wounding and again 3 days thereafter.

Following the treatments, wounds were protected using Vaseline dressings from 3 M. Observations of the wound's healing progression were recorded on days 0, 3, 5, 7 and 14. Each wound's dimensions were traced onto transparent acetate films, which were then digitally captured with an MF633Cdw scanner. These scanned wound images were analysed using ImageJ 1.36b software, ensuring precise measurements expressed in square millimetres.

2.6.2. Immunohistochemical staining

Immunohistochemical staining was performed on 4‐μm paraffin sections of day‐14 wounds. The skin sections were fixed in 4% paraformaldehyde and blocked with 5% goat serum for 1 h at room temperature, followed by incubation with anti‐cMet (1:100, ab51067; Abcam, Cambridge, United Kingdom), antismooth muscle actin (SMA) (1:2000, 14 395‐1‐AP;Proteintech, Rosemont, IL), anti‐CD31 (1:300, MCA341R; Bio‐Rad, Hercules, CA) overnight at 4°C. After washing, the sections were incubated with secondary antibodies for 1 h at room temperature. Images were captured at 100× and 400× magnifications from three different regions (right, middle and left) of each specimen using a Leica DM2500 microscope (Leica Microsystems, Wetzlar, Germany), and the mean values were calculated.

2.6.3. Quantitative real‐time PCR

To elucidate the potential role of ROCK1 in enhancing wound healing through the transformation of fibroblasts into myofibroblasts, we evaluated the mRNA expression levels of α‐SMA, COL‐1 and COL‐3, given their pivotal roles in wound healing processes. Total RNA was harvested from full‐thickness tissue samples of mice using the TRIzol reagent (Servicebio, China), adhering to the manufacturer's prescribed protocol. The extracted RNA was then transcribed into cDNA using the M‐MLV Reverse Transcriptase (RNase H) kit (Servicebio, China). Quantitative real‐time PCR (qRT‐PCR) was executed as presented in a prior study by Sheng et al., 2019. ¹⁹ The specific primer sequences employed for this analysis are presented in Table 1.

TABLE 1.

Primer sequences.

Name	Sequence (5′–3′)	Length
M‐GAPDH‐S	CCTCGTCCCGTAGACAAAATG	133
M‐GAPDH‐A	TGAGGTCAATGAAGGGGTCGT
M‐a‐SMA(2)‐S	GTACCACCATGTACCCAGGC	152
M‐ơ‐SMA(2)‐A	GAAGGTAGACAGCGAAGCCA
M‐COLLAGEN I‐S	TCCTGGCAAAGACGGACTCA	159
M‐COLLAGEN I‐A	GGCAGGAAGCTGAAGTCATAACC

2.7. Statistical analysis

Statistical analysis was performed using SPSS 20.0 software. The Shapiro–Wilk test was used to test the normality of the measures. Data that conformed to a normal distribution were expressed as $\overline{x}$ ± s, while non‐normally distributed data were presented as median (Q1, Q3). For comparisons between groups: the two independent samples t‐test was employed for normally distributed data, whereas the Wilcoxon test was utilized for non‐normally distributed data. Differences were considered statistically significant at p < 0.05. Two independent samples t‐test and fold change method were used to identify differentially expressed proteins. The screening criteria for differentially expressed proteins between different sample groups: fold change <−2.0 or fold change >2.0 and p < 0.05.

3. RESULTS

3.1. Baseline comparison: Antibiotic cement versus conventional treatment

In evaluating the clinical outcomes, a comparison was made between the baseline conditions of patients treated with antibiotic bone cement (i.e., EXP group) and those subjected to conventional treatment (i.e., CON group). During the analysis period, 60 patients fulfilling the pre‐determined inclusion criteria were assessed. The findings indicated no substantial differences between the antibiotic bone cement group and the conventional treatment group with respect to several crucial variables, including age, duration of diabetes history, presence of diabetic foot and certain laboratory tests such as fasting glucose, CRP and albumin levels. Moreover, no significant discrepancies were observed between the two groups concerning complications like peripheral neuropathy and lower limb arterial ischaemic disease. A detailed depiction of these findings is provided in Table 2.

TABLE 2.

Comparison of preoperative conditions among patients.

	EXP	CON	t/Z × 2	p‐Value
Clinical feature
Age, years ($\overline{x}$ ± s)	57.1 ± 8.4	56.5 ± 7.7	0.268	0.790
Duration of diabetes mellitus, months ($\overline{x}$ ± s)	122.0 ± 70.0	87.6 ± 64.3	1.841	0.071
Duration of ulcer, months	1.0 (0.5 ~ 3.0)	1.0 (0.5 ~ 2.0)	0.008	0.993
The results of the laboratory tests
FBG (fasting blood‐glucose) ($\overline{x}$ ± s, mmol/L)	15.9 ± 3.7	15.8 ± 3.9	0.122	0.903
CRP ($\overline{x}$ ± s, μmol/L)	111.2 ± 57.7	75.0 ± 77.4	1.465	0.162
Albumin ($\overline{x}$ ± s, g/L)	25.9 ± 3.4	25.5 ± 4.8	0.313	0.755
Comorbidity
Peripheral neuropathy [n (%)]	33 (82.5)	16 (80.0)	0.000	1.000
Diabetic nephropathy [n (%)]	15 (37.5)	6 (30.0)	0.330	0.566
Peripheral arterial disease [n (%)]	28 (70.0)	14 (70.0)	0.000	1.000

Note: EXP group is experimental group, while CON group is control group.

Preoperative bacterial cultures were performed in both groups, with the results shown in Table 3. The distribution of bacteria types infecting the trabeculae was comparable between the groups and bore no statistical significance. Staphylococcus aureus was the primary bacterium implicated in the trauma infection, followed by Enterococcus faecalis.

TABLE 3.

Composition of preoperative infectious organisms in both groups (n [%]).

Species of bacteria	EXP	CON
Staphylococcus aureus	16 (40.0)	9 (42)
Efaecalis	6 (15)	4 (19)
Proteus	3 (7.5)	2 (9)
Pseudomonas aeruginosa	3 (7.5)	1 (4.8)
Kleber pneumoniae	3 (7.5)	1 (4.8)
Baumanii	2 (5.0)	1 (4.8)
Streptococcus agalactiae	2 (5.0)	1 (4.8)
Streptococcus haemolyticus	2 (5.0)	1 (4.8)
Staphylococcus epidermidis	1 (2.5)	1 (4.8)
Corynebacterium striatum	2 (5.0)	0 (4.8)
Total	40 (100.0)	21 (100.0)

3.2. Efficacy of antibiotic bone cement

All patients enlisted in this study were identified with Wagner classification levels 3–4, signifying a pronounced risk for above‐ankle amputation. Upon the execution of rigorous debridement and the subsequent application of vancomycin antibiotic bone cement in the EXP group, noticeable healing was manifested distally at the ankle joint, thus facilitating limb preservation. This is vividly illustrated through the progression of wound healing in two distinct cases as depicted in Figures 1 and 2. A comparative analysis revealed that this tailored intervention markedly accelerated the healing trajectory in contrast to the conventional treatment administered to the CON group. This is further substantiated by the data encapsulated in Table 4, which shows a significant enhancement in healing rates (87.5% in EXP vs. 50.0% in CON, p = 0.004) and a consequential reduction in amputation rates (5.0% in EXP vs. 30.0% in CON). However, the short‐term mortality rates between the two cohorts did not exhibit a significant variance, with a 5.0% mortality rate observed in both groups.

FIGURE 1.

Progression of diabetic foot wound healing post debridement and antibiotic bone cement application (Case 1). (A) Pre‐treatment diabetic foot condition; (B) immediate post‐debridement phase, showcasing extensive necrotic tissue extending to muscle and bone; (C) 1 week post vancomycin antibiotic bone cement application, with absence of inflammatory signs such as periwound erythema; (D) 2 weeks post second debridement and vancomycin antibiotic bone cement application, displaying a clean wound and induction membrane formation; (E) immediate post‐debridement phase with a free flap graft for wound repair; (F) 6‐month post‐trauma repair, exhibiting a fully healed wound devoid of rupture and pus discharge.

FIGURE 2.

Progression of diabetic foot wound healing post debridement and antibiotic bone cement application (Case 2). (A) Pre‐treatment diabetic foot condition; (B) immediate post‐debridement wound treatment with vancomycin antibiotic bone cement; (C) 14 days post vancomycin antibiotic bone cement application, showing a clean wound with visible induction membrane formation; (D) 2 weeks post wound repair with a free skin graft.

TABLE 4.

Comparison of therapeutic effects between two treatment modalities [n (%)].

Title 1	Healing	Improvement	Amputation	Death	χ 2	p
EXP	35 (87.5)	1 (2.5)	2 (5.0)	2 (5.0)	11.265	0.004
CON	10 (50.0)	3 (15.0)	6 (30.0)	1 (5.0)

3.3. Bioinformatics analysis to identify differential proteins in DFW

3.3.1. Quantitative proteomic analysis

In this study, mass spectrometry analysis was employed to identify a total of 5489 proteins, among which 4018 were deemed quantifiable. The term ‘quantifiable proteins’ refers to those proteins for which quantitative information was available in at least one comparison group. To mitigate the effects of inter‐sample variations attributable to spiked volume and instrument manipulation, the protein quantification values were normalized utilizing the median method. Subsequently, the K‐nearest neighbour (KNN) algorithm was applied to impute missing values, ensuring a more robust dataset for analysis.

The distribution of protein expression across samples was visualized using a box plot, facilitating the assessment of data dispersion before and after the background correction process. As depicted in Figure 3A, the consistency in the level of protein quantification values across all corrected samples is evident, with the median line remaining at a uniform level, thereby indicating a satisfactory correction outcome.

FIGURE 3.

Distribution, statistical analysis and group comparison of quantified proteins. C2 and C1 represent the protein groups post‐treatment and pre‐treatment, respectively, in the experimental group, while A2 and A1 denote the protein groups post‐treatment and pre‐treatment, respectively, in the control group. (A) Box plot illustrating the distribution of protein quantification values post‐correction, showcasing the consistency in data dispersion across samples. (B) Statistical representation of the count of differential proteins. The X‐axis denotes the count of differential proteins, while the Y‐axis represents group comparison conditions. The red and blue bars signify the number of upregulated and downregulated differential proteins respectively. (C). Venn diagram depicting the distribution of differential proteins among groups. Each circle in the plot symbolizes a differential protein between groups; overlapping regions denote differential proteins shared among multiple groups, whereas non‐overlapping areas highlight group‐specific differential proteins.

The protein differential analysis plays a pivotal role in the identification of novel biomarkers, thereby enhancing the precision of biomarker detection. This methodology holds significant implications for elucidating molecular mechanisms, understanding pathological conditions and identifying potential drug targets.

Initially, the fold change is calculated by taking the ratio of the mean values of all biological measurements for each protein between two comparative sample sets. To assess the statistical significance of observed variations, a t‐test is applied to the quantitative values of each protein within these sample sets. For comparisons involving more than two groups, a one‐way analysis of variance (ANOVA) is utilized, and the resulting p‐value is computed as an index of statistical significance. Subsequently, this p‐value is adjusted using the Benjamini–Hochberg false discovery rate (BH FDR) correction.

The ratio of the mean values of all biological measurements for each protein in the two intercomparison sample groups was used as the fold change of the difference. Figure 3B presents the statistical analysis of differential protein expression across various groups. In the comparison of C2 and C1, a significant portion of proteins, precisely 64.3%, were upregulated, while 35.7% were downregulated. Similarly, in the comparison of C2 and A2, 61.6% of proteins were upregulated, with 38.4% showing downregulation. In both the C1 versus A1 and A2 versus A1 comparisons, there was no significant variation in the percentage of proteins upregulated as opposed to those downregulated.

Figure 3B illustrates the differential protein expression in various comparative scenarios. Specifically, 241 differential proteins were identified in the C2‐vs.‐C1 group, which compares samples before and after the application of antibiotic bone cement. Meanwhile, 232 differential proteins were observed in the A2‐vs.‐A1 group, contrasting the effects of antibiotic bone cement treatment with conventional methods. These disparities in protein expression were graphically represented using a Venn diagram, as shown in Figure 3C. The diagram delineates both the unique and shared proteins that exhibited differential expression across the groups.

A salient observation emerged from the analysis: 63 proteins exhibited differential expression when comparing pre‐ and post‐treatment samples in the group treated with antibiotic bone cement relative to the conventionally treated group. Notably, these proteins were predominantly localized within the induced membrane.

Upon integrating our findings with the existing literature, we scrutinized 63 specific differential proteins for association with wound repair. Among them, ROCK1 exhibited a notably higher expression degree in the induced membranes associated with antibiotic bone cement application for diabetic foot wounds, as opposed to those associated with the conventional treatment (Figure 4A). A significantly higher degree of ROCK1 expression was also observed during the post‐treatment phases compared to the pre‐treatment phases within the experimental group, as depicted in Figure 4B.

FIGURE 4.

Histogram of differential protein ROCK1 expression. (A) Significantly higher ROCK1 expression in the experimental group after treatment compared to the pre‐treatment group. (B) Significantly higher ROCK1 expression in the experimental group after treatment compared to the control group (p < 0.05).

3.3.2. Functional classification and interaction network of differential proteins

The differential proteins identified in each comparison group were analysed for enrichment at the GO classification and protein structural domain levels. This analysis aimed to discern if these proteins were significantly enriched in specific functional types. The bubble plots represent the enrichment analysis, where the vertical axis denotes the functional classification or pathway, and the horizontal axis displays the Log2‐transformed value of the change in the proportion of differential proteins in that functional type compared to the proportion of identified proteins. The colour of the circles indicates the enrichment significance p‐value, and the size of the circles represents the number of differential proteins in the functional class or pathway. The top 20 most significantly enriched categories are presented in the bubble plots. Notably, the differential protein ROCK1, implicated in cell adhesion signalling pathways, showcased significant involvement, especially in the induced membranes associated with antibiotic bone cement application to diabetic foot wounds (Figure 5).

FIGURE 5.

Analysis results of differential protein GO‐BP (A) before and after treatment in the experimental group and that (B) after treatment in the experimental and control groups.

The interaction relationships among the identified differential proteins were explored using the String DB protein interaction database. Network diagrams were constructed using igraph software to visualize the interaction dynamics. The network graph constructed based on pathway enrichment results elucidates the interaction relationship between the identified pathways and the differential proteins, as illustrated in Figure 6. The network graph showcases the central role of ROCK1 in mediating signalling pathways critical to the cellular responses observed in the experimental group. The interaction dynamics revealed through this analysis provide insights into the molecular mechanisms underpinning the observed clinical outcomes, thereby enriching our understanding of the therapeutic potential of antibiotic bone cement in diabetic foot wound treatment.

FIGURE 6.

The linkage of the differential protein ROCK1 to other pathways through the Q13464 signalling pathway (A) before and after treatment in the experimental group and (B) after treatment between experimental group and control group.

3.3.3. Validation of differential proteins through histopathological analysis

Post antibiotic bone cement treatment, an induction membrane was observed to form on the wound surface, as depicted in Figure 7A. A comparative analysis using MASSON staining of wound tissue pre and post‐treatment in the experimental group revealed a significant and organized arrangement of fibrous tissue within the induction membrane, indicating a progression in the wound healing process.

FIGURE 7.

Histopathological analysis of wound healing. (A) HE staining showcasing the induced membrane formation in the experimental group post‐treatment (in the red circle); (B) Mason staining effect diagram depicting the organized arrangement of fibrous tissues post‐treatment in the experimental group; (C–F) expression analysis of CD31 and Rock1 pre and post‐treatment in both the negative pressure group and cement group, revealing a significant increase in ROCK1 and CD31 expression post antibiotic bone cement treatment in the experimental group (*p < 0.05). Data represented as mean ± SEM of three independent experiments; n = 3.

To further validate the role of ROCK1 in promoting wound healing, immunohistochemical assays for ROCK1 and CD31, crucial factors for wound healing, were performed. A comparative analysis of pre and post‐treatment samples from both the experimental and control groups demonstrated a significantly higher expression of ROCK1 and CD31 post‐treatment in the experimental group.

3.4. Impact of ROCK1 on diabetic mice wound healing

The pivotal role of Rho‐associated coiled‐coil containing protein kinase 1 (ROCK1) in wound healing was explored in diabetic mice. Enhanced wound healing was observed in the ROCK1 protein group as compared to both the control and ROCK1 inhibitor groups, underscoring the therapeutic potential of ROCK1 (Figure 8A,B).

FIGURE 8.

Effect of ROCK1 and its inhibitors on wound healing factors (CD31, α‐SMA, Collagen I, Collagen III) in diabetic mice. (A) Representative images of full‐thickness skin defects in rats of NC group, ROC group, and CCG group immediately, and on days 0, 3, 5, 7, 10 and 14 days postoperatively. (B) Wound closure rate (%). (C) Representative images of (50×, scale bar 200 μm) of Immunohistochemistry sections of the NC group, ROC group and CCG group 14 days postoperatively. (D) Quantitative analysis of CD31 and Rock1 expression. (E) qRT‐PCR analysis detecting α‐SMA, Collagen I, Collagen III gene expression in diabetic tissues of mice. Comparisons were performed using t‐test. *p < 0.05; **p < 0.01. Error bars represent SD (n = 5).

At the 14‐day mark, immunohistochemical analysis revealed a marked elevation in the expression of ROCK1 protein and the angiogenic marker CD31 in the wounds of mice in the ROCK1 protein group, as opposed to the other groups (Figure 8C,D). Further investigation into the fibroblast‐to‐myofibroblast conversion process, a critical aspect of wound repair, was conducted. PCR analysis demonstrated a significant upregulation of α‐SMA, a myofibroblast marker, along with vascular repair factors COL‐1 and COL‐3 in the ROCK1 protein group (Figure 8E).

In summary, our findings advocate for the instrumental role of ROCK1 in accelerating wound healing and granulation tissue formation in diabetic mice. This acceleration is potentially mediated through the promotion of neovascularization and fibroblast to myofibroblast differentiation, shedding light on a promising therapeutic avenue for diabetic wound management.

4. DISCUSSION

This study explored the therapeutic potential of antibiotic bone cement in promoting the healing of diabetic foot wounds, revealing a notable reduction in amputation rates and an enhanced healing process in the experimental group when compared to the negative pressure control group. The uniformity in the clinical treatment of patients across both groups ensured a fair comparison, shedding light on the promising role of antibiotic bone cement in managing diabetic foot wounds.

Diabetic foot wounds present a clinical challenge due to their association with prolonged infection, complex treatment requirements, and extended healing duration, often complicated further by concurrent lower limb vascular disease. ²⁰ Historically, the Masquelet technique has been employed to address bone defects, utilizing bone cement to induce membrane structure, thereby promoting healing. ¹⁶ , ²¹ , ²² Our adaptation of this technique demonstrated that antibiotic bone cement could offer a localized, sustained antibiotic release, mitigating infection risks and systemic antibiotic side effects while creating a conducive microenvironment for wound healing. The results are consistent with others. ¹⁶ , ²³ , ²⁴ , ²⁵ , ²⁶

A crucial finding in our study was the increased expression of ROCK1 in the wounds treated with antibiotic bone cement. ROCK1, known for its role in cell adhesion, migration, proliferation and angiogenesis, ²⁷ was markedly upregulated, suggesting a potential mechanism underlying the enhanced wound healing observed. ¹⁴ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ The detection of elevated CD31 levels, a known marker for neovascularization, further bolstered the argument for improved angiogenesis contributing to wound healing.

Our proteomic transcriptional analysis identified 63 differentially expressed proteins within the induced membranes. The particular role of ROCK1 and its association with the Q13464 signalling pathway emerged as a significant finding, opening avenues for further exploration of the molecular mechanisms driving wound healing in diabetic foot ulcers. The association between ROCK1 and the Q13464 signalling pathway is particularly intriguing. The Q13464 signalling pathway is known for its involvement in various cellular processes crucial for wound healing. The interaction between ROCK1 and this signalling pathway may potentially modulate cellular behaviours such as migration, proliferation and angiogenesis, which are essential for wound repair and regeneration. The upregulation of ROCK1 could possibly enhance the activation of the Q13464 signalling pathway, thereby promoting a favourable wound healing environment.

The study, however, is not devoid of limitations. The primary focus on induced membranes post antibiotic bone cement application in a specific microenvironment of diabetic foot wounds necessitates a broader understanding of these membranes, especially the cellular populations involved. Additionally, a more detailed analysis of other potential molecular pathways involved in wound healing could yield a comprehensive understanding, enabling the development of novel therapeutic strategies.

In conclusion, this study significantly unveils the therapeutic advantage of antibiotic bone cement in improving diabetic foot wound healing, with a notable linkage to ROCK1‐mediated mechanisms. The noticeable reduction in amputation rates and accelerated healing in the experimental group, compared to the control group, highlight the innovation and clinical relevance of this intervention. A key discovery was the elevated expression of ROCK1, a pivotal player in cell adhesion, migration, proliferation and angiogenesis, hinting at a molecular mechanism underpinning the enhanced healing observed. The intertwining of ROCK1 with the Q13464 signalling pathway, known for its essential role in crucial cellular processes for wound healing, emerged as a significant finding, inviting further in‐depth molecular explorations. This revelation not only enhances our molecular understanding but also moves us a step closer to creating novel, effective therapeutic strategies for managing diabetic foot wounds. The promise shown herein necessitates extended investigations to carefully unravel the precise molecular dialogues and validate the clinical utility of antibiotic bone cement across a wider spectrum of patient populations, thereby potentially introducing a new era of innovative therapeutic interventions in diabetic foot wound management.

FUNDING INFORMATION

This research was funded by Collaborative Innovation Center of Chinese Ministry of Education (2020‐39) and National Natural Science Foundation of China (82360445).

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Yang C, Wang D. Antibiotic bone cement accelerates diabetic foot wound healing: Elucidating the role of ROCK1 protein expression. Int Wound J. 2024;21(4):e14590. doi: 10.1111/iwj.14590.

DATA AVAILABILITY STATEMENT

Data are unavailable due to privacy and ethical restrictions.