Controlled release of berberine modulates the wound microbiome to accelerate wound healing

Abstract

The wound microbiome has been shown to play a significant role in influencing the wound healing process. Coptis chinensis, a traditional Chinese medicine (TCM) known for its heat-clearing properties, contains berberine (BER) as major active ingredient, which exhibits notable antibacterial activity. In this study, we investigated the effect of BER on wound healing and wound microbiome through three distinct delivery strategies, including solution form, burst-release scaffolds (PCL/BER), and sustained-release scaffolds (PCL/PLGA/BER), compared with an untreated negative control (NC) group. Drug release studies confirmed that PCL/BER caused a pronounced burst release, while the incorporation of PLGA enabled sustained release of BER for up to 120 h. Further in vivo studies showed that the sustained BER release from the PCL/PLGA/BER resulted in the most effective improvement in wound healing. Microbiome analysis using 16S rRNA sequencing identified Staphylococcus xylosus (S. xylosus) as the key species influencing wound healing outcomes in response to BER delivery. S. xylosus overabundance in the NC group and its depletion in the BER solution and burst BER release groups impaired wound healing. In contrast, sustained BER delivery maintained an optimal S. xylosus abundance that promoted a favorable immune microenvironment by modulating CXCL10 and (IFN-α) expression. Our findings emphasize the importance of coordinating drug release kinetics with microbiome dynamics for optimal wound healing outcomes and provide valuable insights for developing future delivery systems for heat-clearing TCMs, with a focus on microbiome-modulation therapeutic strategies.

Graphical abstract

Burst-release and sustained-release scaffolds were prepared using electrospinning. The effects of different berberine (BER) delivery strategies, including solution form, burst-release scaffolds, and sustained-release scaffolds on wound healing and the wound microbiome were investigated in an acute wound model.

1. Introduction

Skin wounds are a globally recognized public health concern that significantly reduces quality of life and imposes a substantial economic burden [1]. One of the key factors contributing to impaired wound healing is an imbalance in the wound microbiome [2]. Although antibiotics are often used to treat infected wounds clinically [3], the complete clearance or reduced diversity of the wound microbiome is associated with delayed wound repair [[4], [5], [6], [7], [8], [9]]. For example, germ-free mice exhibited a slower wound healing rate compared to specific-pathogen-free mice, demonstrating that the absence of microorganisms may be detrimental to wound healing [8]. In contrast, increased microbial diversity is beneficial for wound healing [10], and certain key microbial species are essential for maintaining microbiome homeostasis [11]. These findings suggest that high-dose, broad-spectrum antibacterial treatment may be counterproductive. Instead, sustained local release of therapeutic agents that selectively eliminate the harmful bacteria while preserving the beneficial ones may offer a more targeted and effective approach for maintaining wound microbiome balance and improving wound healing [[12], [13], [14]].

Traditional Chinese Medicine (TCM), with its millennia-long history, represents a rich source of therapeutic agents. Coptis chinensis is a TCM known for its heat-clearing properties [15]. Berberine (BER), the primary active component in Coptis chinensis, has shown diverse pharmacological effects such as immunopotentiation, anti-inflammatory, anti-cancer and anti-ulcer properties [16]. In addition, it exhibits significant antibacterial activity and is widely used in the treatment of infections [17,18]. Interestingly, BER supplementation was shown to modify the gut microbiome [19], and this effect may be dose-dependent [20]. This raises the question of whether controlled release of BER could differentially modulate the wound microbiome and positively influence wound healing outcomes.

Electrospun scaffolds represent a promising class of wound dressing [21] as they enable the incorporation of poorly water-soluble drugs during fabrication and allow for sustained local release [22]. In addition, electrospun fibers closely mimic the natural structure of the skin extracellular matrix [23], which facilitates cell attachment, proliferation and migration, thereby supporting the wound healing process [24,25]. Polycaprolactone (PCL) and poly(lactic-co-glycolic acid) (PLGA) are the commonly used biomaterials for electrospinning due to their favorable biocompatibility and biodegradability, and approval by the U.S. Food and Drug Administration (FDA) for wound management [26,27].

In this study, we investigate the impact of different release mechanisms of BER on wound healing and the wound microbiome. By comparing multiple delivery strategies, including direct BER exposure in solutions, and the incorporation of BER in either burst-release or sustained-release electrospun scaffolds, we aim to elucidate how BER release kinetics influence microbial dynamics at the local wound site and subsequently, the healing process. To the best of our knowledge, this is the first study to systematically explore the effects of drug release mechanisms on the wound microbiome.

2. Materials and methods

2.1. Materials

PCL (Mw = 80,000 g/mol) and 1,1,1,3,3,3–hexafluoro–2–propanol (HFIP) were purchased from Sigma-Aldrich (St. Louis, MO, USA). PLGA (Mw = 100,000 g/mol, LA: GA = 50:50) was kindly provided by Jinan Daigang (Jinan, Shandong, China). Berberine was purchased from YuanYe (Shanghai, China).

2.2. Preparation of PCL/BER scaffolds

PCL and BER were separately dissolved in HFIP and stirred for 2 h to ensure complete dissolution. Then the solutions were mixed according to the ratio in Table 1 and continued stirring for 12 h to make the liquid evenly mixed to achieve the PCL/BER solution. Subsequently, the liquid was transferred to a syringe, and PCL/BER solution was electrospun at a rate of 0.5 ml/h, with a distance of 10 cm between the needle and the collector and a voltage of 18 kV. After electrospinning, the collected samples were dried at room temperature for 30 min and then stored at room temperature.

Table 1.

Sample concentrations used for the fabrication of PCL/BER scaffolds.

Names	Concentrations (%, w/v)
	PCL	BER
PCL	10	0
PCL/BER0.01	10	0.01
PCL/BER0.02	10	0.02
PCL/BER0.04	10	0.04
PCL/BER0.08	10	0.08

2.3. Preparation of PCL/PLGA/BER scaffolds

PCL and PLGA were dissolved in HFIP, as was BER, and all were stirred for 2 h until fully dissolved. Then the PCL/PLGA solution and BER solution were mixed according to the ratio in Table 2. The blended solution was stirred for 12 h to make the liquid evenly mixed to achieve a PCL/PLGA/BER solution. Then the PCL/PLGA/BER solution was transferred to a syringe, and electrospun at a rate of 0.5 ml/h, with a distance of 10 cm between the needle and the collector and a voltage of 18 kV. After electrospinning, the collected samples were dried at room temperature for 30 min and then stored at room temperature.

Table 2.

Sample concentrations used for the fabrication of PCL/PLGA/BER scaffolds.

Names	Concentrations (%, w/v)
	PCL	PLGA	BER
PCL/PLGA	7.5	7.5	0
PCL/PLGA/BER0.01	7.5	7.5	0.01
PCL/PLGA/BER0.02	7.5	7.5	0.02
PCL/PLGA/BER0.04	7.5	7.5	0.04
PCL/PLGA/BER0.08	7.5	7.5	0.08

2.4. Scanning electron microscopy

The surface morphologies of electrospun samples including PCL/BER and PCL/PLGA/BER scaffolds were examined by scanning electron microscopy (SEM) (Tescan Maia3 Gmu, Tescan Orsay Holding, CZ). The samples were coated with platinum at 40 nm prior to scanning by SEM at a voltage of 20 kV. Fiber diameter and pore size were quantified using Fiji ImageJ version 2.0 (National Institutes of Health, Bethesda, MD, USA). Twenty-five fibers and pores in one field of view were selected for analysis.

2.5. Porosity

The porosity of the electrospun scaffolds was measured using the liquid displacement method. Scaffolds were cut into 1 cm × 1 cm samples. The thickness of each sample was measured using a vernier caliper and was subsequently weighed. Scaffolds were then submerged in anhydrous ethanol. After 24 h, the scaffolds were taken out and gently blotted to remove excessive liquid on the surface and re-weighed. Porosity was calculated using the following formula.

$$\text{Porosity}(\%)=\frac{\frac{(\mathrm{W}1-\mathrm{W}2)}{\text{De}}}{\frac{\left(\mathrm{W}1-\mathrm{W}2\right)}{\text{De}}+\mathrm{d}\times \mathrm{S}}\times 100\%$$

Where, W1 refers to the wet weight of the scaffolds after immersion in anhydrous ethanol and W2 refers to the dry weight of the scaffolds before immersion in anhydrous ethanol. De is the density of anhydrous ethanol, d refer to the thickenss, and S refer to the area of the scaffolds.

2.6. FTIR-ATR

Fourier transform infrared spectroscopy - attenuated total reflectance (FTIR-ATR) was used to investigate the intermolecular interactions of BER, PCL/BER scaffolds and PCL/PLGA/BER scaffolds from 400 to 4000 cm⁻¹.

2.7. Mechanical properties

The tensile strength and Young’s modulus of PCL/BER and PCL/PLGA/BER scaffolds were measured using a tensile machine (TH-8203A, SUZHOU TOPHUNG MACHINERY CO., LTD). Each scaffold was cut into 1 × 2 cm samples and was stretched to break.

2.8. X-ray diffraction

The crystalline/amorphous states were scanned using X-ray diffraction (X-RD) (Rigaku SmartLab SE) to analyze and determine the presence of raw materials and components in PCL/BER and PCL/PLGA/BER scaffolds. Tested at a voltage of 40 kV, a current of 40 mA, a scanning range of 5 to 90 °, and a scanning speed of 10 °/min.

2.9. In vitro BER release

Scaffolds were placed into 10 ml centrifuge tubes containing 5 ml phosphate-buffered saline (PBS) (pH 7.4) and incubated in a shaker at 37 °C. At each time point, 1 ml PBS (pH 7.4) was removed from the centrifuge tube and substituted with 1 ml fresh PBS (pH 7.4). The release of BER was determined by comparing the U.V. (Unico UV-4800) absorbance at a maximum wavelength of 345 nm to the standard curve of BER in the same medium.

2.10. In vitro biocompatibility test and scratch assay

The PCL/PLGA/BER scaffolds were immersed in cell culture medium to prepare extracts. After incubation for 24, 48 and 72 h, respectively, the extracts were collected and filtered. For the cell viability assay, HaCaT cells were seeded in 96-well plates at a density of 5 × 10³ cells/well and subsequently treated with the extracts from different time points. Cell viability was then assessed using the CCK-8 kit. For the cell migration assay, HaCaTs were seeded in 6-well plates at a density of 5 × 10⁵ cells/well. a scratch wound was created, and the cells were treated with the PCL/PLGA/BER scaffold extract. Images of the scratch were captured at 0, 12, 24 and 48 h post-treatment, and the wound area was quantified.

2.11. PCL/PLGA/BER scaffolds and different release speeds of BER animal experiment

Male BALB/c mice, 12 weeks old, weighing 25–30 g each, were purchased from GemPharmatech (Nanjing, Jiangsu, China). All experimental procedures were executed according to the protocols approved by the Animal Ethics Committee, Nanjing University of Chinese Medicine (ethic no 202301A011&202404A018). A mouse model of an acute wound was established. BALB/c mice were under general anesthesia (3% isoflurane) during the surgery, where a small wound of 1 cm × 1 cm was created. After the model was successfully established, the animals were intraperitoneal injected with 1 ml warm saline and analgesic (carprofen 5 mg/kg). Each wound was treated with different kinds of scaffolds at variable time points (0, 3 and 7 d). All animals were then housed individually and monitored daily for any signs of distress or changes in physical appearance.

The PCL/PLGA/BER scaffolds animal experiment: all animals were randomly assigned to the following groups (n = 16, 4/group): NC (open wound), PCL/PLGA, PCL/PLGA/BER_0.02 and PCL/PLGA/BER_0.08, respectively. Wounds sizes of each group of mice were measured at variable time points (0, 3, 7 and 14 d). Mice were euthanized by cervical dislocation after 14 d treatment, and the wound tissues were harvested for histology and molecular analysis.

Different release speeds of BER animal experiment: all animals were randomly assigned to the following groups (n = 40, 5/group/time point including 3 and 14 d): NC (open wound), BER_0.08 solution, PCL/BER_0.08 scaffold and PCL/PLGA/BER_0.08 scaffold, respectively. Wound sizes were measured at variable time points (0, 3, 7, 10 and 14 d). The wound tissues were harvested for histology and molecular analysis after 3 and 14 d treatment.

2.12. Histopathological analysis

Wound tissues were embedded in paraffin and sectioned at a thickness of 5 µm. Each section was stained with hematoxylin and eosin (H&E) for histological analysis and Masson’s trichrome for collagen deposition.

2.13. 16S rRNA gene sequences

The wound area was wiped with a sterile cotton swab 20 times, followed by storage in a sterile and enzyme-free centrifuge tube at −80 °C. Wound microbiome samples collected from the mice were sent to Majorbio Bio-Pharm Technology (Shanghai, China) for Illumina 16S rRNA gene sequencing. Briefly, genomic DNA was extracted from wound samples, and microbial community 16S rRNA libraries were generated as previously described. After the PCR reaction, purified amplicons were pooled equimolarly and paired-end sequenced on an Illumina MiSeq PE300 platform (Illumina, San Diego, USA) according to the protocols. Raw FASTQ files were then demultiplexed using an in-house perl script and quality-filtered by fastp version 0.19.6 [28] and merged by FLASH version 1.2.11 [29]. The optimized sequences were clustered into operational taxonomic units (OTUs) using UPARSE [30] with a 97% sequence similarity level. To minimize the effects of sequencing depth on α and β diversity measure, the number of 16S rRNA gene sequences from each sample were rarefied to 39536. Based on the OTU abundance profile, α-diversity and β-diversity analyses were performed and analyzed using the online platform of Majorbio Cloud Platform (www.majorbio.com). The rarefaction curves of all samples rapidly increased with the increase of sequencing volume and eventually flattened, entering the plateau period. Within-sample α-diversity was assessed using Chao 1 and Shannon indices. The Chao 1 index is a metric of richness that estimates the total number of species in samples. The higher its value, the more abundant the α-diversity [31]. The Shannon index combines richness and evenness; the higher its value is, the more abundant the α-diversity [31]. Between-sample β-diversity was visualized by a principal coordinate analysis (PCoA) plot based on the binary Jaccard distance. Statistical significance of α-diversity was determined using the Kruskal-Wallis H test, and β-diversity using ANOSIM analysis. The relative abundance of key microbiomes on the species level was analyzed using the Kruskal-Wallis H test. The quality control metrics were as follows: raw sequencing reads were processed and cleaned through several steps. First, quality trimming was performed using a 10-bp sliding window, truncating reads when the average quality score within the window fell below 20; reads shorter than 50 bp after trimming or containing more than five ambiguous nucleotides (N) were discarded. Second, paired-end reads were merged into single sequences using their overlapping regions, requiring a minimum overlap of 10 bp. Third, merged sequences were filtered out if the mismatch ratio within the overlapping region exceeded 0.2. Finally, sequences were assigned to their respective samples based on exact barcode matches (0 mismatches allowed) and primer matches (up to two mismatches allowed), and their orientations were standardized.

2.14. BER co-incubation with S. xylosus in vitro

Viable Staphylococcus xylosus (S. xylosus) colonies were selected from the inclined culture medium, activated in Luria Bertani solid medium for 12 h, and subcultured in Luria Bertani liquid medium. After continuing to cultivate for 12 h, the obtained bacterial solution was diluted to a concentration with an optical density of 0.2 (OD_{630 nm}), and then diluted 10⁴ times with a gradient of liquid culture medium. The electrospinning bracket was cut in advance and irradiated with ultraviolet sterilization. We took a 4 ml shake flask and added 1 ml diluted bacterial solution. A total of 1 ml physiological saline was added to the blank control group, 1 ml BER physiological saline solution was added to the solution group, and 1 ml physiological saline and scaffold were added to the PCL/BER_0.08 and PCL/PLGA/BER_0.08 scaffold groups (excluding the blank control group, the final BER concentration for all groups was 80 µg/ml). The shaking tubes were incubated in a constant temperature shaker at 37 °C and 150 rpm. Take 100 µl microbiome solution at time points 0, 2, 4, 8, 10 and 12 h for analysis.

2.15. RNA isolation and quantitative real-time PCR

The wound tissues collected from mice after 3 d post-injury were used for RNA extraction using the Trizol (Invitrogen, Carlsbad, CA, USA) method. Total mRNA (1 µg) was reverse-transcribed to complementary DNA using the SensiFAST cDNA synthesis kit (Bioline, London, UK). Real-time polymerase chain reaction (PCR) analysis was performed using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad, Hercules, CA, USA). The efficiency of DNA amplification was evaluated using the mean cycle threshold (Ct) method. The ΔCt value was calculated from Ct values of different interest genes by subtracting the Ct value of the housekeeping gene, β-actin. The resulting relative mRNA expression was shown as fold change (2−ΔΔCt) relative to the expression in baselines. The primer sequence is mentioned in Table 3.

Table 3.

Primer sequence (Sangon Biotech, Shanghai, China) is set as follows:

Gene	Primer	Sequence (5′−3′)
Actb	Actb-F	CCACCATGTACCCAGGCATT
	Actb-R	AGGGTGTAAAACGCAGCTCA
Cxcl10	Cxcl10-F	CCAAGTGCTGCCGTCATTTTC
	Cxcl10-R	GGCTCGCAGGGATGATTTCAA
Ifna2	Ifna2-F	TACTCAGCAGACCTTGAACCT
	Ifna2-R	CAGTCTTGGCAGCAAGTTGAC

2.16. Enzyme-linked immunosorbent assay (ELISA)

The wound tissues were put into RIPA Lysis Buffer (Biosharp Life Sciences, Beijing, China) and then centrifuged to collect supernatants. Supernatants were used for measuring chemokine (C-X-C motif) ligand 10 (‌CXCL10) and interferon-α‌ (IFN-α) protein levels by Mouse CXCL10 ELISA Kit (MultiSciences Biotech Co., Ltd) and Mouse IFN-α ELISA Kit (ELK Biotechnology CO., LTD) according to the manufacturer’s instructions.

2.17. Statistical analysis

Statistical analyses were performed using the GraphPad Prism software version 9.0 with data shown as means ± SD (GraphPad, ISI Software Inc., San Diego, CA, USA). Student’s t-test analysis was utilized to determine significant differences between the two groups. One-way analysis of variance (ANOVA) was utilized to determine significant differences between multiple groups. P < 0.05 was considered as statistical significance. (^⁎P < 0.05, ^⁎⁎P < 0.01, ^⁎⁎⁎P < 0.001, ^⁎⁎⁎⁎P < 0.0001).

3. Result and discussion

3.1. Fabrication and characterization of PCL/BER scaffolds

To construct scaffolds with varying BER release rates, we first used PCL as the sole polymer to fabricate electrospun scaffolds loaded with different BER concentrations, including PCL, PCL/BER_0.01, PCL/BER_0.02, PCL/BER_0.04 and PCL/BER_0.08 (corresponding to BER concentrations of 0.00%, 0.01%, 0.02%, 0.04% and 0.08% respectively). SEM showed that PCL and all PCL/BER scaffolds exhibited uniform, bead-free fibers with smooth surfaces and dispersed pores. No visible drug aggregates or deposition were observed on the surfaces, suggesting that BER was well distributed within the PCL fibers (Fig. 1A). The fiber diameters were measured as: PCL (141.3 ± 37.0 nm), PCL/BER_0.01 (109.3 ± 35.0 nm), PCL/BER_0.02 (191.1 ± 48.7 nm), PCL/BER_0.04 (105.0 ± 14.2 nm), PCL/BER_0.08 (179.5 ± 40.5 nm); Pore sizes as: PCL (527.9 ± 37.0 nm), PCL/BER_0.01 (567.8 ± 150.6 nm), PCL/BER_0.02 (443.4 ± 162.3 nm), PCL/BER_0.04 (256.4 ± 124.6 nm), PCL/BER_0.08 (550.1 ± 175.8 nm) (Fig. 1B–1C). These scaffolds also exhibited similar porosities around 60% (Fig. 2A). Together, these results indicate that incorporation of BER at various concentrations had minimal effect on the scaffold fabrication and morphology, and produced favorable fiber diamter, pore size, porosity to support cell proliferation, migration and adhesion [32], with potential to modulate the wound microbiome [33] to promote wound healing. Furthermore, the appropriate porosity and pore size are considered optimal for wound dressing functions such as exudation absorption and gas exchange [34] to help maintain a moist wound environment for optimal wound healing [35].

Fig. 1.

Characterization of different PCL/BER scaffolds. (A) SEM images. Scale bar: 2 µm; (B) Fiber width (n = 25); (C) Pore size (n = 25).

Fig. 2.

Characterization of different PCL/BER scaffolds. (A) Porosity (n = 3); (B) In vitro drug release profile (n = 3); (C) Mechanical properties (n = 3); (D) FTIR spectra; (E) XRD patterns.

Next, we evaluated BER release from the PCL/BER scaffolds in vitro in PBS and found all PCL/BER scaffolds experienced burst release with over 80% of BER released within 2 h and >90% released by 12 h (Fig. 2B). This indicates PCL/BER scaffolds are unable to sustain BER delivery over time for the wound healing process and therefore are used as burst-release scaffolds.

We further characterized other physicochemical properties, including mechanical strength, FTIR and XRD using PCL, PCL/BER_0.02 and PCL/BER_0.08 scaffolds. PCL (2.3 ± 0.8 MPa), PCL/BER_0.02 (2.3 ± 0.2 MPa) and PCL/BER_0.08 (2.5 ± 0.3 MPa) scaffolds exhibited comparable tensile strengths (Fig. 2C), indicating that the BER incorporation did not compromise mechanical strength and the scaffolds retained adequate strength and flexibility requirements as skin wound dressings [36]. The FTIR spectrum of PCL scaffold exhibited a characteristic bimodal peak between 2790 cm^-1 and 2950 cm^-1, corresponding to the C—H stretching vibrations of methylene groups. A prominent peak at 1725 cm^-1 was attributed to the carbonyl (C = O) stretching vibration. The spectral band of 700–1600 cm^-1 displayed multiple absorption bands associated with the skeletal vibrations of the polymer chains, including bending, swinging, stretching of methylene groups, as well as contributions from isophthalic esters and trans-isomerization of ester groups. The FTIR spectrum of pure BER exhibited a broad absorption band 3000–3600 cm^-1, indicative of O—H or N—H stretching, and distinct peaks at 1600 cm^-1, 1505 cm^-1 and 1229 cm^-1, corresponding to C = C stretching in vinyl and aromatic groups (Fig. 2D). After BER was incorporated into the PCL scaffold, no new peaks were observed in the spectrum, which possibly due to the low BER concentration. The crystal structures analyzed by XRD showed that the BER powder exhibited characteristic sharp diffraction peaks around 9°, indicating its crystalline nature. PCL showed distinct peaks at 21.3° and 24.1°, consistent with its semi-crystalline structure [37]. In contrast, the PCL/BER_0.02 and PCL/BER_0.08 scaffolds did not display the characteristic BER peaks (Fig. 2E), which is probably because of the low BER concentration.

3.2. Fabrication and characterization of PCL/PLGA/BER scaffolds

Previous studies have shown that incorporating PLGA into electrospun scaffolds can enhance the sustained release profile [38], likely due to the amphiphilic property of PLGA [39]. We propose that the incorporation of PLGA into the hydrophobic scaffold PCL allows the encapsulation of originally surface-bound BER, leading to a sustained release from the PCL/PLGA/BER scaffold. We optimized electrospinning using 7.5% PCL and 7.5% PLGA with BER incorporated at concentrations of 0.01%, 0.02%, 0.04% and 0.08% (w/v), resulting in the fabrication of PCL/PLGA, PCL/PLGA/BER_0.01, PCL/PLGA/BER_0.02, PCL/PLGA/BER_0.04 and PCL/PLGA/BER_0.08 electrospun scaffolds, respectively. SEM analysis showed that the fiber morphology of PCL/PLGA/BER was comparable to that of PCL/PLGA scaffold. The fiber surfaces of PCL/PLGA/BER scaffolds appeared smoother than those of PCL/BER scaffolds (Figs. 3A and 1A). Fiber diameters ranged from 100 to 500 nm, and pore sizes ranged from 500 to 2000 nm with similar distributions (Fig. 3B and 3C). Compared to the PCL/BER scaffolds (Fig. 1B and 1C), the PCL/PLGA/BER scaffolds exhibited slightly increased fiber diameters and pore sizes. Because the addition of PLGA the viscosity of the spinning solution increased, which consistent with previous studies [40,41]. The porosity was measured as 65.3% ± 1.0% for PCL/PLGA, 64.9% ± 1.3% for PCL/PLGA/BER_0.01, 66.9% ± 1.1% for PCL/PLGA/BER_0.02, 65.7% ± 0.8% for PCL/PLGA/BER_0.04, and 67.2% ± 1.64% for PCL/PLGA/BER_0.08 (Fig. 4A). Compared to the PCL/BER scaffold, incorporating PLGA increased permeability, which can better promote gas exchange and help maintain a moist wound environment. This potentially further improves the wound microbiome balance, preserves the integrity of skin barriers, and thereby enhances resistance to pathogen invasion [42].

Fig. 3.

Characterization of different PCL/PLGA/BER scaffolds. (A) SEM images. Scale bar: 10 µm; (B) Fiber width (n = 25); (C) Pore size (n = 25).

Fig. 4.

Characterization of different PCL/PLGA/BER scaffolds. (A) Porosity (n = 3); (B) In vitro drug release profile (n = 3); (C) Mechanical properties (n = 3); (D) FTIR spectra. Arrows: main characteristic peaks; (E) XRD patterns. Arrows: main characteristic peaks.

We subsequently conducted in vitro release studies and showed the PCL/PLGA/BER scaffolds achieved sustained release of BER for 120 h (Fig. 4B). Within the first 12 h (the highest release rate), cumulative BER release from PCL/PLGA scaffolds increased with loading concentration, which were 24.2% ± 1.0%, 34.6% ± 3.1%, 44.9% ± 2.5% and 57.5% ± 1.3% for 0.01%, 0.02%, 0.04% and 0.08% (w/v), respectively. This is a significant improvement compared to PCL/BER scaffolds that exhibited a rapid 80% BER release within 2 h and nearly complete release by 12 h. The sustained release of BER from the PCL/PLGA/BER scaffolds suggests their potential to reduce the frequency of dressing changes in clinical settings, thereby promoting healing while minimizing disturbance to the wound site. We selected PBS at 37 °C as the release medium for the in vitro studies because it is an internationally recognized and standardized model for preliminary screening. It offers a stable, controllable and reproducible environment for evaluating scaffold material characteristics. However, we acknowledge that PBS alone does not fully replicate the complexity of the in vivo wound microenvironment. Wounds are rich in proteases, including matrix metalloproteinases (MMPs), which are involved in tissue remodeling and may also directly degrade the polymer scaffold used in our study. This could potentially lead to a faster drug release rate compared to that observed in PBS. Additionally, pH is a critical factor, as variations in local pH can influence both the release behavior of BER and the degradation rate of polymer.

Mechanical properties, FTIR and XRD were also conducted for the PCL/PLGA, PCL/PLGA/BER_0.02, and PCL/PLGA/BER_0.08 scaffolds. The tensile strengths of the PCL/PLGA, PCL/PLGA/BER_0.02 and PCL/PLGA/BER_0.08 scaffolds were recorded at 4.5 ± 0.3 MPa, 4.4 ± 0.3 MPa and 4.3 ± 1.1 MPa, respectively, indicating that the incorporation of PLGA enhanced the mechanical strength compared to PCL/BER scaffolds, but there were no significant differences among PCL/PLGA/BER scaffolds (Fig. 4C). The FTIR spectra of PCL scaffold and BER were consistent with those shown in Fig. 2D The FTIR spectrum of PLGA displayed a strong absorption peak between 1750–1740 cm^-1 corresponding to the stretching vibration of the carbonyl group (C = O). Characteristic C—O stretching bands were observed in the range of 1300–1150 cm^-1. A weak absorption band between 3000–2900 cm^-1 was attributed to the stretching vibrations of CH, CH₂ and CH₃ groups. Additional absorption peaks between 1500–1000 cm^-1 were assigned to bending vibrations of CH₂ and CH₃, as well as rocking vibrations of CH and CH₂ groups. The FTIR spectrum of PCL/PLGA scaffold presented a superimposition of spectras of PCL and PLGA, with no new absorption peaks detected (Fig. 4D), indicating that the two polymers were physically blended during electrospinning without chemical interaction or cross-linking. Similarly, the FTIR spectra of PCL/PLGA/BER_0.02 and PCL/PLGA/BER_0.08 scaffolds revealed no new absorption peaks following BER incorporation, which probably because of the low BER concentration. The XRD patterns of the PCL/PLGA/BER scaffold was presented in Fig. 4E. PCL is a semi-crystalline polymer and PLGA is an amorphous polymer. After co-electrospun of PCL and PLGA, the crystallinity of the scaffolds decreases. Compared to the patterns shown in Fig. 2E, a reduction in crystallinity was observed, as evidenced by the decreased intensity of the diffraction peaks relative to the PCL scaffold. This reduction is attributed to the incorporation of the amorphous polymer PLGA into the semi-crystalline PCL matrix during electrospinning. Additionally, no characteristic diffraction peaks of BER were detected in any of the PCL/PLGA/BER scaffolds, regardless of drug concentrations, which probably because of the low BER concentration.

The water vapor transmission rate of PCL/PLGA/BER scaffolds was measured, indicating that they possess good gas permeability (Fig. S1). The biocompatibility of PCL/BER_0.08 and PCL/PLGA/BER_0.08 scaffolds was further evaluated using hemolysis assays (Fig. S2). The result showed that hemolysis of PCL/BER_0.08 and PCL/PLGA/BER_0.08 are 0.64% ± 0.19% and 0.45% ± 0.36%, which were significantly lower than Triton (100.00% ± 1.07%), confirming their hemocompatibility and suitability for use as wound dressings.

3.3. PCL/PLGA/BER scaffolds improved wound healing

To investigate the effect of different BER sustained release rates on wound repair, we evaluated wound healing efficacies of PCL/PLGA, PCL/PLGA/BER_0.02 and PCL/PLGA/BER_0.08 scaffolds using an acute wound model in BALB/c mice (Fig. 5A). After 3 d treatment, the PCL/PLGA/BER_0.08 group exhibited an accelerated closure rate (22.2% ± 3.8%), significantly higher than the negative control (NC) group (−6.6% ± 6.9%). After 7 d treatment, wound closure in the PCL/PLGA/BER_0.08 group (48.4% ± 3.2%) remained significantly higher than in the NC group (27.7% ± 1.8%). Although the PCL/PLGA and PCL/PLGA/BER_0.02 groups showed a trend for improved closure, the difference were not statistically significant. After 14 d treatment, nearly complete wound closure was observed in all groups (Fig. 5B–5D). H&E staining was subsequently performed to demonstrate the quality of wound healing, which showed that the PCL/PLGA/BER_0.08 scaffold group exhibited the epithelial structure is most closely attuned to the wound bed (Fig. 5E). However, no significant difference was observed in re-epithelialization among the groups after 14 d treatment (Fig. 5G). Masson’s trichrome staining revealed markedly more collagen deposition in the PCL/PLGA/BER_0.08 group compared to other groups (Fig. 5F), evidenced by a significantly higher collagen density quantified relative to the NC group (Fig. 5H). These results suggest that the PCL/PLGA/BER_0.08 scaffold improved wound healing and was therefore selected for further investigation. The cytotoxicity assessment and cell migration assay of PCL/PLGA/BER_0.08 scaffold were evaluated in vitro using HaCaT cells. The results demonstrate that the scaffold exhibit good biocompatibility and promote cell migration (Fig. S3A–S3C).

Fig. 5.

PCL/PLGA/BER scaffolds-treated wound healing in vivo. (A) Flowchart of wound model; (B) Images of the wounds at variable time points (0, 3, 7 and 14 d). Scale bar: 1 cm; (C) Traces of wound area after 14 d treatment; (D) Relative wound closure rate (n = 4); (E) H&E staining of wound tissues after 14 d treatment. Scale bars: 1000 µm (left), 400 µm (right); (F) Masson’s trichrome staining of wound bed after 14 d treatment. Scale bars: 1000 µm (left), 100 µm (right); (G) Re-epithelialization of wounds after 14 d treatment (n = 4); (H) Collagen deposition of wounds after 14 d treatment (n = 4). ^⁎P < 0.05, ^⁎⁎P < 0.01.

3.4. Sustained BER release enhances wound healing compared to burst release

To investigate the different effects of burst release vs. sustained release on wound repair, we compared the wound healing efficacy of 0.08% (w/v) BER solution (BER_0.08, drug-only), PCL/BER_0.08 (burst release) and PCL/PLGA/BER_0.08 (sustained release) using a BLAB/c acute wound model (Fig. 6A). We observed that wounds treated with both PCL/BER and PCL/PLGA/BER scaffolds exhibited cleaner surfaces with reduced exudate compared to NC and BER groups, highlighting the desired capability of the electrospun scaffolds to absorb wound fluids (Fig. 6B–6C). Notably, the PCL/PLGA/BER_0.08 scaffold group showed significantly enhanced wound closure compared to the NC group at variable time points (0, 3, 7 and 10 d). After 7 and 10 d treatment, the PCL/PLGA/BER_0.08 scaffold group also showed significantly improved wound healing compared to the PCL/BER_0.08 scaffold group (Fig. 6D). Furthermore, after 10 d treatment, the wound closure rate in the PCL/PLGA/BER_0.08 group was significantly higher than that of BER_0.08group. These results indicated that the PCL/PLGA/BER_0.08 scaffold provided superior therapeutic outcomes compared to both the BER_0.08 solution and the PCL/BER_0.08 scaffolds. Indeed, H&E staining revealed that PCL/PLGA/BER_0.08 treated group’s epithelial structure is most closely attuned to the wound bed, no significant difference was observed in re-epithelialization among the groups after 14 d treatment (Fig. 6E and 6G). Masson’s trichrome staining confirmed significantly higher collagen density in the PCL/PLGA/BER_0.08 group compared to the NC group (Fig. 6F and 6H). These findings suggest that the sustained BER release from PCL/PLGA/BER_0.08 outperforms burst release or solution approaches, suggesting its considerable potential in enhancing wound healing and facilitating tissue remodeling. One possible explanation is that sustained release of BER from the PCL/PLGA/BER_0.08 scaffold avoided excessive local drug concentration and minimized the risk of over-sterilization, which can disrupt the wound microbiome and impair the healing process [8]. In contrast, BER delivered in high doses, such as those in the BER solution and PCL/BER_0.08 scaffold, may have adversely affected the microbial balance at the wound site, potentially hindering optimal repair.

Fig. 6.

Different release speeds of BER-treated wound healing in vivo. (A) Flowchart of wound model; (B) Images of the wounds at variable time points (0, 3, 7, 10 and 14 d). Scale bar: 1 cm; (C) Traces of wound area after 14 d treatment; (D) Relative wound closure rate (n = 5); (E) H&E staining of wound tissues after 14 d treatment. Scale bars: 1000 µm (left), 400 µm (right); (F) Masson’s trichrome staining of wound bed after 14 d treatment. Scale bars: 1000 µm (left), 100 µm (right); (G) Re-epithelialization of wounds after 14 d treatment (n = 5); (H) Collagen deposition of wounds after 14 d treatment (n = 5). ^⁎P < 0.05, ^⁎⁎P < 0.01, ^⁎⁎⁎P < 0.001.

3.5. BER release mechanism alters the wound microbiome composition and structure during wound healing

To determine whether BER release mechanism differentially affects the wound microbiome during wound healing, we performed 16S rRNA sequencing analysis on collected wound samples. When evaluating microbiome diversity, α-diversity is a key metric that reflects both microbial richness and evenness, which can be presented with the Shannon index for overall diversity and the Chao 1 index for microbial richness. Analysis across all four treatment groups at three different time points revealed no significant differences in either Chao1 or Shannon indices (Fig. 7A–7C), indicating that BER treatment, regardless of release rate, had no effect on the α-diversity of the wound microbiome.

Fig. 7.

Analysis of the microbiome in wounds. (A) α-diversity index and β-diversity after 3 d treatment. β-diversity (ANOSIM: R = 0.279, P = 0.004); (B) α-diversity index and β-diversity after 7 d treatment. β-diversity (ANOSIM: R = 0.235, P = 0.013); (C) α-diversity index and β-diversity after 10 d treatment. β-diversity (ANOSIM: R = 0.335, P = 0.011). n = 4–5.

β-diversity reflects the differences in microbial community composition among groups. To assess this difference, we performed principal coordinate analysis (PCoA) based on β-diversity metrics after 0, 3, 7 and 10 d treatment. After 3 d treatment, the variance explained by PC₁ and PC₂ was 41.62% and 15.68%, respectively, with no statistically significant differences observed among the four groups. However, after 7 d treatment, PC₁ and PC₂ accounted for 52.04% and 20.72% of the variance, respectively. The NC group exhibited significant differences from the other three groups in the PC₁ direction and also differed significantly from the PCL/PLGA/BER_0.08 group in the PC₂ direction. After 10 d treatment, PC₁ and PC₂ accounted for 51.54% and 21.93% of the variance, respectively. At this time point, the NC group showed significant separation from the BER_0.08 group along the PC₁ axis, and from both the BER_0.08 and PCL/BER_0.08 groups along the PC₂ direction. These findings suggest that treatments with BER_0.08 solution, PCL/BER_0.08, PCL/PLGA/BER_0.08 scaffolds dynamically and significantly altered the composition and structure of the wound microbiome compared to the NC group (Fig. 7A–7C).

To further investigate the impact of BER_0.08, PCL/BER_0.08 and PCL/PLGA/BER_0.08 treatments on the wound microbiome, we conducted a systematic analysis of wound microbial composition at variable time points (3, 7 and 10 d) at both the phylum and genus levels, which are commonly used taxonomic ranks in microbiome research. At the phylum level, Proteobacteria, Firmicutes and Actinobacteriota were identified as the predominant phyla across all time points (Fig. 8A–8C). At the genus level, several genera exhibited relative abundances greater than 1%, while those below this threshold were grouped under “Others.” Among the dominant genera, Corynebacterium, Acinetobacter and Staphylococcus, were consistently identified as the most abundant across at variable time points (3, 7 and 10 d) (Fig. 8D–8F). These genera were previously reported as key components of the normal human microbiome [43], and key microbial species within the genera can play a crucial role in maintaining the microbiome homeostasis [11].

Fig. 8.

Wound microbiome analysis. (A–C) Percentage of community composition on the phylum level after 3, 7 and 10 d treatment (n = 4–5); (D–F) Percentage of community composition on the genus level after 3, 7 and 10 d treatment (n = 4–5).

We therefore further explored the changes in genera composition following treatment with BER_0.08, PCL/BER_0.08 and PCL/PLGA/BER_0.08 groups by performing a differential analysis of microbial communities across 3, 7 and 10 d samples, which identified S. xylosus as a key species, which is a commensal microorganism that resides on the skin surface. Upon skin injury, it colonizes the wound area and can influence the wound healing process. The relative abundance of S. xylosus followed a consistent pattern at all time points among the treatment groups: NC > PCL/PLGA/BER_0.08 > BER_0.08 ≈ PCL/BER_0.08. Specifically, the average proportion of S. xylosus in the BER_0.08 and PCL/BER_0.08 groups was below 1%, while it was approximately 22%−30% in the NC group and 2%−7% in the PCL/PLGA/BER_0.08 group (Fig. 9A–9C). These findings suggest that the different BER release mechanisms significantly influenced the relative abundance of S. xylosus, which may in turn affect the overall balance of the wound microbiome. A previous study showed that community-level rested critically upon certain species being present [11]. The S. xylosus is the key species for wound healing in our study.

Fig. 9.

(A–C) Relative abundance of key microbiomes on the species level (n = 4–5) after 3 d, 7 d and 10 d treatment; (D) In vitro co-incubation of scaffolds with S. xylosus. ^⁎P < 0.05, ^⁎⁎⁎P < 0.001, ^⁎⁎⁎⁎P < 0.0001.

To determine whether the observed changes in S. xylosus abundance were not attributed to the PLGA component in scaffolds, we analyzed the wound genera composition in the PCL/PLGA scaffold without BER. The abundance of S. xylosus in this group was comparable to that in the NC group, with no statistically significant difference observed (Fig. S4). The results indicate that the PCL and PLGA alone did not alter the abundance of S. xylosus in the wound environment similar to a previous report [33], and suggest that BER is the key component that modulated S. xylosus levels. To verify this, we further conducted in vitro co-incubation of S. xylosus with BER_0.08 solution, PCL/BER_0.08 scaffolds and PCL/PLGA/BER_0.08 scaffolds. Results showed that S. xylosus proliferation remains unaffected in the absence of BER whereas both the BER_0.08 solution and the PCL/BER_0.08 scaffolds effectively inhibited the growth of S. xylosus. This can be attributed to direct BER exposure and burst BER release, respectively (Fig. 2B). Meanwhile, the PCL/PLGA/BER_0.08 scaffold demonstrated a moderate inhibitory effect, evidenced by the OD_630nm absorbance value that fell between those of the NC and BER_0.08 solution and PCL/BER_0.08 groups. This is likely due to the sustained release of BER from the PCL/PLGA/BER_0.08 scaffold, which enables the gradual antibacterial actions against S. xylosus. Collectively, these findings support the hypothesis that sustained release of BERs from electrospun scaffolds regulates the abundance of S. xylosus in the wounds, thereby promoting a more balanced wound microbiota to improve healing (Fig. 9D). We have also summarized our findings (Fig. 10).

Fig. 10.

Schematic figure summarizing the findings from Fig. 7, Fig. 8, Fig. 9.

3.6. Sustained BER release improves wound healing via S. xylosus-modulated CXCL10 and IFN-α expression

Previous studies have suggested that S. xylosus may promote wound healing through stimulating CXCL10 and type Ι IFN secretion [5], but their overexpression is associated with chronic inflammation [44,45]. We suspected that sustained BER release modulated CXCL10 and type I IFN secretions to levels beneficial to wound healing. To investigate this hypothesis, we assessed the gene and protein expression of CXCL10 and IFN-α in wound tissues after 3 d treatment. At the mRNA level, the expression of Cxcl10 closely reflected the relative abundance of S. xylosus in each treatment group. The NC group exhibited the highest Cxcl10 expression, which was significantly higher than that in the BER_0.08, PCL/BER_0.08 and PCL/PLGA/BER_0.08 groups. The PCL/PLGA/BER_0.08 group showed significantly higher Cxcl10 expression compared to the BER_0.08 and PCL/BER_0.08 groups. A similar trend was observed for Ifna2 expression but no significant difference was observed (Fig. 11A–11B). At the protein level, the expression of CXCL10 was consistent with mRNA expression patterns. The NC group again showed significantly higher CXCL10 levels than the other groups, confirming the regulatory role of S. xylosus in modulating CXCL10 expression. ‌IFN-α expression at the protein level followed the same pattern as CXCL10, with the NC group exhibiting significantly higher levels than the BER_0.08 and PCL/BER_0.08 groups. The PCL/PLGA/BER_0.08 group showed significantly higher IFN-α levels compared to the PCL/BER_0.08 group and a trend toward higher levels than the BER_0.08 group (Fig. 11C–11D). Interestingly, despite the elevated expression of CXCL10 and IFN-α, the NC group demonstrated a slower wound healing rate, which may be attributed to excessive inflammation due to the overexpression of these factors [44].

Fig. 11.

CXCL10 and IFN-α expression in wound tissue after 3 d treatment. (A) Expression of Cxcl10 on the gene level (n = 3); (B) Expression of Ifna2 on the gene level (n = 3); (C) Expression of CXCL10 on the protein level (n = 3); (D) Expression of IFN-α on the protein level (n = 3). ^⁎P < 0.05, ^⁎⁎P < 0.01, ^⁎⁎⁎P < 0.001, ^⁎⁎⁎⁎P < 0.0001.

Taken together, these findings suggest that S. xylosus is a key regulator of wound microbiome and immune signaling. The rapid accumulation of BER in the BER_0.08 and PCL/BER_0.08 groups likely suppressed S. xylosus to levels insufficient to elicit beneficial inflammation responses. In contrast, the PCL/PLGA/BER_0.08 scaffold achieved sustained BER release, preserving an appropriate abundance of S. xylosus, which in turn modulated the expression of CXCL10 and IFN-α at levels conducive to wound healing.

We observed that the PCL/PLGA/BER scaffold significantly accelerated wound closure and improved histological architecture, that may be partly attributed to the broad anti-inflammatory properties of BER [46]. As a multitarget compound, BER can directly inhibit key inflammatory signaling pathways, such as NF-κB [47], thereby mitigating tissue damage and promoting a restorative microenvironment. However, our data indicate that the pronounced healing effects cannot be explained solely by BER’s anti-inflammatory action; modulation of the microbiota appears to play a critical role. Microbiome analysis revealed that both free BER and PCL/BER scaffold treatments reduced the abundance of S. xylosus, whereas the PCL/PLGA/BER group showing the best healing outcome due to higher abundance of S. xylosus. This finding suggests that BER’s effects are multifaceted. We therefore hypothesize that while BER provides anti-inflammatory benefits, its sustained release from PCL/PLGA/BER scaffolds also modulates S. xylosus in the wound microbiome, which may serve as a key mechanism driving healing. The persistent presence of an optimal abundance of S. xylosus could support a pro-healing microenvironment, potentially through competitive exclusion of pathogens and continuous regulation of local immune responses.

4. Conclusions

In this study, we employed electrospinning technology to fabricate two types of BER-loaded scaffolds: a burst-release scaffold (PCL/BER) and a sustained-release scaffold (PCL/PLGA/BER), to investigate the impact of BER release kinetics on wound healing. In vivo results demonstrated that the PCL/PLGA/BER scaffold significantly enhanced wound healing and increased collagen deposition in the wound bed compared to other groups. Microbiome analysis revealed that both excessive and insufficient levels of S. xylosus were unfavorable for wound repair. Among all tested scaffolds, only the PCL/PLGA/BER scaffold-maintained S. xylosus at beneficial levels, suggesting that sustained BER release transforms its antibacterial activity into a microbiome-modulating effect. These findings highlight the importance of tuning drug release profiles when designing antibacterial formulations, particularly for heat-clearing TCMs, to support microbiome balance and enhance wound healing. The findings of this study offer valuable insights into the development of advanced delivery systems tailored to harness the therapeutic potential of TCM herbal components.

CRediT authorship contribution statement

Zeyu Xu: Writing – review & editing, Writing – original draft, Validation, Methodology, Investigation, Formal analysis, Data curation. Lixiang Zhang: Methodology, Investigation, Data curation. Juntong Guo: Investigation. Qing Xia: Investigation. Zhengping Ge: Investigation. Ziyu Wang: Visualization. Ruoyu Mu: Supervision. Jie Dong: Supervision. Zhiguo Qin: Supervision. Jun Chen: Writing – review & editing, Funding acquisition, Conceptualization. Yiwei Wang: Writing – review & editing, Writing – original draft, Supervision, Funding acquisition, Conceptualization.

Conflicts of interest

The authors declare that there is no conflicts of interest.