Ultrasound-activated and P-selectin-targeted liposomes overcome biofilm barriers for surgical site infections therapy

Abstract

Pseudomonas aeruginosa-induced biofilm-associated surgical site infections (BSSIs) pose a dual therapeutic challenge: the dense extracellular polymeric substances (EPS) barriers hinder drug enrichment and penetration, whereas the immunosuppressive microenvironment within the biofilm impedes infection clearance, leading to persistent bacterial colonization and recurrence. This study developed an ultrasound-activated P-selectin-targeted liposome (SPCMPL) integrating natural sulfatide ligands (targeting P-selectin overexpressed in inflamed BSSI vasculature), sonosensitizer chlorin e6 (Ce6), meropenem prodrug, and perfluoropentane (PFP) to achieve efficient antibiotic delivery and BSSI treatment via breaking through the biofilm barriers and activating immunomodulation. SPCMPL employed ligand/receptor-mediated transcytosis for enrichment in BSSI lesions, where the PFP phase transition triggered by ultrasound disrupted the biofilm EPS structure. This process can both trigger the in-situ generation of reactive oxygen species (ROS) by Ce6 and loosen the EPS matrix. This degradation then facilitated meropenem release, allowing it to penetrate the biofilm more effectively and achieve antimicrobial concentrations throughout. Furthermore, the mass-produced ROS polarized macrophages to a pro-inflammatory M1 phenotype, thereby enhancing phagocytosis, remodeling the microenvironment, and inhibiting biofilm persistence. Ultrasound-triggered spatial control localized antibiotic release and immunomodulation to the infection site, optimizing local delivery while minimizing systemic toxicity and reducing the risk of systemic cytokine storms. The results demonstrated that the SPCMPL with ultrasound manipulation integrated biofilm disruption, targeted drug release, and immunomodulation to completely eradicate both planktonic and biofilm-embedded bacteria and effectively treat BSSI.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-025-03821-0.

Introduction

Biofilm-associated surgical site infection (BSSI) is one of the leading causes of postoperative infections, posing a heavy burden on healthcare systems and patients worldwide [1]. Pseudomonas aeruginosa (P. aeruginosa) is recognized for its robust ability to form biofilms by secreting copious amounts of extracellular polymeric substances (EPS) that encapsulate themselves in a three-dimensional, mushroom-shaped structure with unique spatial heterogeneity [2]. The high viscosity and density of biofilms not only slow down the diffusion and penetration of antimicrobial agents but also prevent the direct contact of immune cells with pathogens, thereby creating relatively immune-privileged zones [3, 4]. Concurrently, P. aeruginosa secretes virulence factors to suppress the host’s immune response, particularly by interfering with macrophage activation [5, 6]. Moreover, as a deep-seated infection within the body, the BSSI region is characterized by low vascular density and intact structure. Further, the endothelial cell junctions in this region tend to become even tighter under inflammatory stimulation [7]. These characteristics collectively limit the therapeutic efficacy of BSSI treatments. Conventional antibiotic therapies often only temporarily reduce bacterial load, with approximately 40% of patients still experiencing recurrent and persistent infections; they may even promote antibiotic resistance and biofilm formation [8]. Therefore, developing effective therapeutic strategies for completely eradicating P. aeruginosa BSSI has become imperative.

Lipid nanoparticles are garnering significant attention in the biomedical field due to their excellent biocompatibility and drug stability [9, 10]. Passive diffusion is not the main mechanism by which lipid nanoparticles cross the vascular barrier. In fact, up to 97% of nanoparticle transportation is accomplished through active processes in endothelial cells [11]. This implies that precise delivery to the site of infection can be achieved by functionalizing the surface of these nanoparticles with specific targeting ligands, thereby minimizing off-target effects. To date, numerous lipid nanoparticle formulations either have received clinical approval or are undergoing clinical trials [12]. However, although approximately 70% of intravenously administered nanoparticles reach the tumor tissue, only about 0.7% effectively accumulate and exert therapeutic effects [13, 14]. The cavitation effect induced by ultrasound (US) can generate substantial mechanical forces and transient high pressure, which are not limited by tissue depth [15]. This effect directly disrupts the physical structure of biofilms, weakens their defense mechanisms, and breaks the spatial heterogeneity within the biofilms, thereby promoting the uniform distribution and deep penetration of nanoparticle carriers within the biofilms [16, 17]. Based on this, our team previously constructed a liposome targeting the vascular cell adhesion molecule-1 (VCAM1) receptor on vascular endothelial cells. The drug could be precisely delivered to the site of infection through ligand/receptor-mediated endocytosis, reducing its distribution in nontarget tissues [18]. However, in this delivery system, VCAM1 receptor targeting was achieved by conjugating exogenous synthetic peptides to liposomes. This inevitably led to immunogenicity and accelerated blood clearance (ABC) phenomena after repeated administration, limiting its clinical application [19, 20].

We have continued to identify specific vascular endothelial cell targets with suitable endogenous ligands so as to integrate the high targeting specificity of local delivery with the repeatable dosing advantages of systemic delivery. P-selectin is a key adhesion molecule on the surface of vascular endothelial cells. It is capable of specifically recognizing and binding to glycosylated ligands on the cell surface or in the bloodstream, thereby mediating the recruitment and homing of immune cells to sites of inflammation [21, 22]. The stimulation of vascular endothelial cells with four major virulence factors from P. aeruginosa biofilms significantly increases the expression of P-selectin on the cell surface. Thus, P-selectin has emerged as an ideal target for drug delivery to the BSSI locus [23]. Among the various ligands for P-selectin, we have focused specifically on sulfatides. It is a class of acidic sphingolipids widely present in various human tissues, organs, and blood components. Sulfatides can specifically bind to P-selectin and are precisely delivered to infection sites via receptor/ligand-mediated endocytosis [24, 25]. As endogenous glycolipids, sulfatides exhibit high biocompatibility, effectively circumventing the immunogenicity and ABC phenomena associated with exogenous ligands. Moreover, it can be readily incorporated into liposomal structures through a straightforward preparation process [26]. These attributes make sulfatides an ideal material for constructing lipid nanoparticles.

This study aimed to develop a multifunctional liposome, designated as SPCMPL, which integrated the capabilities of biofilm eradication, pathogen killing, and immune activation. This liposome was primarily composed of sulfatides, chlorin e6 (Ce6), and a meropenem (MEM)-prodrug (MP), and encapsulated perfluoropentane (PFP) internally (Scheme 1A). SPCMPL could actively target inflamed vascular endothelial cells overexpressing P-selectin, traverse the vascular endothelial barrier, and achieve efficient accumulation of antimicrobial agents in the BSSI region via ligand/receptor-mediated endocytosis (Scheme 1B). Upon US stimulation, PFP underwent a liquid-gas phase transition, disrupting the EPS barrier and promoting the release and penetration of MP. Meanwhile, the reactive oxygen species (ROS) generated rejuvenated frustrated macrophages, restoring their phagocytic and bactericidal functions while triggering a sustained immune response (Scheme 1C). Collectively, these mechanisms enable effective elimination of BSSIs by thoroughly eradicating both planktonic and biofilm-embedded bacteria, preventing pathogen dissemination and biofilm reformation, and minimizing potential toxic side effects of the drugs.

Scheme 1.

Schematic representation of SPCMPL for efficient biofilm eradication in the deep BSSI model. (A) The fabrication of SPCMPL, alongside its size transformation and prodrug activation upon ultrasonic irradiation. (B) SPCMPL actively targeted the highly expressed P-selectin in vascular endothelial cells at BSSI site and utilized P-selectin-mediated transcytosis to enrich drug accumulation and reach the infection site. (C) Ultrasonic cavitation broke the EPS barrier to promote SPCMPL deep penetration, which worked with immune activation to destroy bacteria and clear biofilm-associated infections

Results and discussion

Expression of P-selectin in vivo and vitro BSSI model

P-selectin is an adhesion molecule highly upregulated on activated endothelial cells during inflammation, representing a promising target for improving the specificity and efficacy of in vivo anti-biofilm therapy [27]. The P-selectin-targeted drug delivery systems are promising, accurately quantifying the expression of P-selectin in BSSI environments and defining optimal in vitro modeling conditions are still highly challenging. To this end, BALB/c nude mice were used to establish the BSSI model using a validated two-step method previously reported by our research team [18]. First, a superficial wound was made, and then 10 µL of luciferase-labeled P. aeruginosa PAO1 suspension [PA-Luci; 10⁸ colony forming units (CFU)/mL] was applied to the wound. A visible white bacterial biofilm developed on the wound after 48 h, which was excised for later use. In the second step, the skin and peritoneum of another nude mouse were incised. The peritoneum was continuously sutured to simulate a surgical wound. A small piece of biofilm-bearing skin tissue was placed on this wound, followed by 10 µL of PA-Luci suspension. The skin was then sutured. After another 48 h, 10 µL of luciferase substrate was subcutaneously injected at the wound site. The bioluminescent changes were measured using the Caliper In Vivo Imaging System (IVIS) Lumina II in vivo imaging system. Mice with peak bioluminescence intensities of more than 3000 were considered successful BSSI models (Figure S1). Tissue samples (0.5 × 0.5 cm²) were randomly collected from biofilm-infected and healthy sites of three BSSI model mice for Western blot (WB) analysis. The statistical analysis revealed that the expression level of P-selectin at the site of infection was approximately three times that at the noninfected sites in the same mice, demonstrating the feasibility of this P-selectin-targeted drug delivery strategy (Figure S2).

In vitro experiments were performed using human umbilical vein endothelial cells (HUVECs) to mimic the inflammatory microenvironment of BSSI. When HUVECs were stimulated with LPS at increasing concentrations, the production of nitric oxide (NO) by these cells exhibited a significant, dose-dependent increase. This upward trend in NO levels provided clear evidence of an increased inflammatory response, confirming the successful induction of an inflammatory state in the in vitro model (Figure S3). Concurrently, the expression level of P-selectin displayed a corresponding upsurge. Also, when HUVECs were stimulated with 50 µg/mL of LPS, the expression level of P-selectin protein was augmented approximately threefold (Fig. 1A and Figure S4). Considering the balance between P-selectin upregulation and potential cell toxicity at higher LPS concentrations, 50 µg/mL was deemed the optimal concentration and selected as the standard condition for establishing the in vitro BSSI inflammation model.

Fig. 1.

Characterizations and functional validation of SPCMPL for BSSI therapy. (A) Representative WB images (top) and quantitative analysis (bottom) of P-selectin overexpression in the IVEC model established by stimulating HUVECs with 50 µg/mL LPS for 12 h [mean ± standard deviation (SD), n = 3]. (B) ROS generation (red bars, left axis, mean ± SD) and MEM release (blue line graph, right axis, mean ± SD) profiles triggered by SPCMPL + US at different US intensities (0, 0.5, 1.0, 1.5, and 2.0 W/cm²). (C) Cell viability of HUVECs or IVECs (left axis) and hemolysis percentage (right axis) exposed to different concentrations of SPCMPL formulations (MP concentrations of 0.01, 0.1, 1, and 10 mg/mL). (D) Size distribution histograms and zeta potential analysis of SPCMPL using DLS, along with cryo-TEM images: (1) at 25 °C without US irradiation, scale bar = 100 nm, (2) at 37 °C with 30 s of US irradiation, scale bar = 5 μm, (3) at 37 °C with 10 min of US irradiation, scale bar = 100 nm. (E) CLSM images of HUVEC (LPS−) and IVEC (LPS+) conditions at 0 and 10 min post-SPCMPL incubation. Nuclei (DAPI, blue), ROS (FITC, green), and liposomes (Cy5, red) were visualized. Merged images show colocalization patterns. Scale bar = 100 μm. (F) Representative flow cytometry (FCM) histograms (left) and mean fluorescence intensities (MFI; right) of HUVECs and IVECs incubated with nontargeted liposomes (PCMPL, pink) and P-selectin-targeted liposomes (SPCMPL, blue). (G) Schematic representation of SPCMPL and PCMPL crossing the IVEC barrier (left) and MFI (right) of liposomes collected at different times in the lower chamber of the Transwell (1, 2, 3, and 4 h). Data are presented as mean ± SD (n ≥ 3), with ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, and ^****P < 0.0001 indicating statistical significance; ns denotes no significance

Preparation and characterization of SPCMPL

In this study, a novel liposome formulation, SPCMPL, was designed by incorporating the natural P-selectin ligand sulfatide, the photosensitizer Ce6, MP, and PFP, which exhibited liquid-gas phase transition properties, to enhance the efficiency and specificity of liposome-vascular binding while reducing immunogenicity. SPCMPL was prepared by a modified thin-film hydration method and subjected to extensive optimization. SPCMPL suspensions (1 mg/mL in PBS) were subjected to precise US parameter variations to systematically investigate the US-dependent release kinetics of ROS and prodrug conversion. The ROS generation was quantified using 2’,7’-Dichlorodihydrofluorescein diacetate (DCFH-DA) as a probe. SPCMPL was exposed to US at different intensity levels (0.5, 1.0, 1.5, and 2.0 W/cm²) using the ultrasonic processor (3 MHz frequency and 50% duty cycle) for 1 min, followed by incubation with DCFH-DA (10 µM final concentration) at 37 °C for 30 min. The level of ROS released by SPCMPL progressively increased with increasing US intensity, showing a significant rise at an intensity of 1.5 W/cm². For MEM release studies, high-performance liquid chromatography (HPLC) was used to characterize the drug release profile in response to sonication. SPCMPL (5 mg/mL) was sonicated at the same intensity gradient. The sample was collected immediately after sonication, centrifuged, and injected into the HPLC system with a C18 column, and the MEM was detected at 230 nm. The percentage release was calculated by comparing the peak area of the released MEM with the total MEM content in SPCMPL. The release of MEM also followed a similar trend, with 87% of MEM released at 1.5 W/cm² and a complete 100% released at 2.0 W/cm² (Fig. 1B). The US-dependent release kinetics of ROS and MEM endowed the system with precise spatiotemporal control, enabling synergistic antibacterial activity against both planktonic bacteria and biofilm-embedded pathogens. This dual-release mechanism leveraged ROS-mediated oxidative stress to disrupt bacterial membranes and biofilm matrix. MEM exerted its bactericidal effect through targeted inhibition of bacterial cell wall synthesis, providing a complementary strategy for the comprehensive eradication of P. aeruginosa-induced BSSIs.

The cytotoxicity and hemocompatibility of SPCMPL were rigorously evaluated to assess its clinical translatability. Using HUVECs and inflammatory vascular endothelial cell (IVECs) as models of the vascular endothelium, SPCMPL demonstrated minimal cytotoxicity as detected using the Cell Counting Kit-8 (CCK-8) assay (Fig. 1C), with cell viability remaining above 90% across all tested concentrations (0.01–10 mg/mL). This result indicated that SPCMPL could interact with the vascular system without damaging the endothelial lining, which was crucial for maintaining blood vessel integrity. The hemocompatibility was evaluated by measuring the hemolysis rate of red blood cells incubated with SPCMPL. The hemolysis rate was less than 2% for all concentrations, suggesting excellent blood compatibility and a low risk of adverse reactions such as hemolytic anemia. Collectively, these findings validated the safety profile of SPCMPL and provided a solid foundation for further preclinical and clinical studies, indicating the potential of SPCMPL to deliver therapeutic agents to target sites without causing severe systemic toxicity or blood-related complications.

Cryo-Transmission Electron Microscopy (Cryo-TEM) images and dynamic light scattering (DLS) analysis demonstrated that SPCMPL manifested as uniformly distributed spherical nanoparticles featuring a unilamellar lipid membrane architecture. These nanoparticles exhibited an average diameter of 184.76 ± 5.49 nm and a zeta potential of − 5.21 ± 0.39 mV, which indicated good colloidal stability (Fig. 1D). SPCMPL was exposed to US irradiation at an intensity of 1.5 W/cm², a frequency of 3 MHz, and a 50% duty cycle for 1 min to investigate the liquid-gas phase transition property. The particle size of SPCMPL increased dramatically to 2858.00 ± 545.26 nm immediately after US exposure, with the zeta potential shifting to − 7.18 ± 0.96 mV. Once the US cavitation ceased completely, the particle size of SPCMPL decreased to 139.67 ± 3.97 nm and the zeta potential reached − 7.39 ± 0.84 mV. In parallel, the control liposomes were prepared using a similar methodology. Each of these formulations lacked specific components: SCMPL did not contain PFP, PCMPL lacked sulfatides, and CMPL was devoid of both PFP and sulfatides. The particle sizes of the three liposomes were all in the range of 163–184 nm. Among these, PCMPL demonstrated a liquid-gas phase transition effect similar to that of SPCMPL, whereas the other two liposomes exhibited no significant changes in particle size under US stimulation (Table S1). This dynamic size alteration marked a significant departure from traditional static nanoparticles. The reversible nature of the phase transition, where the nanoparticles shrank to 139 nm following US exposure, played a crucial role in biofilm penetration. These nanoparticles underwent a transient expansion and subsequent contraction, thus first disrupting the biofilm matrix in the expansion phase and then penetrating its dense structure in the smaller, more mobile state. This unique property, induced by US, integrated targeted delivery with enhanced deep tissue penetration, thus overcoming the limitations of conventional drug carriers that often struggle to reach and act upon pathogens within biofilms [28].

The preparations were stored at 4 °C for 28 days to investigate the temporal stability of liposomes. The particle size of the liposomes was continuously monitored throughout the storage period using the DLS technique. The results showed that SPCMPL liposomes had excellent temporal stability and were suitable for long-term storage and practical applications. This stability was essential to ensure the integrity and efficacy of the liposomes during actual transportation and storage (Figure S5). The key factors governing the liquid-gas phase transition of liposomes were further investigated by monitoring the particle size changes using the DLS system after subjecting liposome suspensions to different storage temperatures or US intensities. Specifically, the liposome suspensions were stored at 4 °C, 25 °C, or 37 °C for 24 h or irradiated with US at 0.5, 1.0, 1.5, and 2.0 W/cm². As shown in Figure S5, the liposome particle size exhibited no significant changes at 25–37 °C but increased markedly upon US stimulation at 1.5 and 2.0 W/cm². This indicated that the mechanical perturbation generated by ultrasonic cavitation was the crucial factor inducing the liquid-gas phase transition of PFP in SPCMPL. The liposomes remained stable in a body-temperature environment in the absence of US irradiation. The stability of the SPCMPL liposomes under simulated physiological conditions was further investigated. Throughout the entire incubation period, both the particle size and the PDI of the liposomes remained essentially unchanged (Figure S6). This minimal change demonstrates the excellent colloidal stability of the SPCMPL formulation under physiological condition, confirming the robustness of the liposomal system.

Enhanced binding and transport capabilities of SPCMPL in IVECs

A series of experiments were conducted to evaluate the binding and transport capabilities of SPCMPL to investigate the targeting ability of SPCMPL to the infected site and its capacity to traverse the vascular barrier. IVECs were established by stimulating HUVECs with 50 µg/mL LPS for 12 h. First, immunofluorescence staining with FITC was performed to compare basal P-selectin expression between HUVECs and IVECs, which served as a prerequisite for evaluating SPCMPL binding specificity. As illustrated in Fig. 1E, P-selectin in normal HUVECs exhibited only a few sparse green punctate fluorescence signals within the cytoplasm. In contrast, P-selectin displayed a bright green ring-shaped fluorescence at the cell membrane in IVECs. This difference in the localization and intensity of P-selectin staining was crucial because it reflected the upregulation and membrane presentation of P-selectin in inflamed endothelial cells, which are characteristic of the microenvironment at the infection site [29].

After co-incubation with Cy5-labeled SPCMPL for 10 min, confocal laser scanning microscopy (CLSM) revealed that IVECs took up a significantly greater amount of the drug while HUVECs showed sparse cytoplasmic signals. This phenomenon was further confirmed using FCM, which revealed that IVECs internalized more than twice the amount of SPCMPL compared with HUVECs (4.29 ± 0.09 vs. 2.11 ± 0.20 [a.u., 10⁴], P < 0.001). In contrast, the uptake by IVECs was dramatically lower for PCMPL without ligand modification than for SPCMPL (4.29 ± 0.09 vs. 2.09 ± 0.05 [a.u., 10⁴], P < 0.01). It did not differ significantly from that by HUVECs (P > 0.05, Fig. 1F). These results clearly demonstrated a key role of the sulfatide modification on SPCMPL in the targeted binding and internalization of liposomes. The sulfatide ligand specifically recognized and bound to the P-selectin on the surface of IVECs, facilitating the entry of SPCMPL into these cells through receptor-mediated endocytosis.

Transwell assays were conducted to investigate the transendothelial transport efficiency of sulfatide-modified SPCMPL in IVECs. HUVECs were cultured in Transwell chambers until they established a strong endothelial barrier, indicated by a trans-epithelial electrical resistance (TEER) exceeding 400 Ω/cm². Following this, they were stimulated with or without LPS (Figure S7). Cy5-labeled SPCMPL or PCMPL was added to the upper chambers, and the fluorescence intensity in the lower chambers was measured after 1–4 h to quantify transvascular transport. As shown in Fig. 1G, SPCMPL liposomes with ligand modification were more efficiently and rapidly transported through the vascular barrier by IVECs. Specifically, the transport rate of SPCMPL across the IVEC monolayer reached 2.63 ± 0.33 a.u./s, which was 4.1-fold higher than that of PCMPL (1.28 ± 0.33 a.u./s, P < 0.01). At the 4-h time point, the cumulative fluorescence intensity in the basolateral compartment for SPCMPL was 4.54 ± 0.34 [a.u., 10⁴], significantly exceeding the intensity of 2.85 ± 0.32 [a.u., 10⁴] observed for PCMPL (P < 0.001). This is of great significance for the delivery of drugs to the infection site. In vivo, the vascular endothelium acts as a barrier that prevents the free passage of drugs. However, the upregulation of adhesion molecules such as P-selectin provides an opportunity for targeted drug delivery in inflamed tissues such as those at the infection site. The efficient transport of SPCMPL across the endothelial barrier through receptor-mediated endocytosis implies that it can carry therapeutic agents directly to the infected area, thus increasing the local drug concentration and enhancing the therapeutic effect.

Collectively, these results strongly demonstrated that SPCMPL modified with sulfatide was efficiently bound, internalized, and transported by IVECs, highlighting the role of ligand/receptor-mediated endocytic transport. This targeted and active transport mechanism is a major advantage of SPCMPL over traditional drug delivery systems. Traditional systems often rely on passive diffusion or nonspecific accumulation, leading to low drug concentrations at the site of infection and potential side effects due to off-target drug distribution [30, 31]. In contrast, the ligand/receptor-mediated approach of SPCMPLs ensures that the drug is delivered precisely to the area where it is needed, thus maximizing its efficacy while minimizing adverse effects. This mechanism also provides a theoretical basis for the further development and optimization of SPCMPL as a therapeutic agent for BSSIs.

Biofilms disruption and antibacterial efficacy

A dual-barrier model simulating inflamed blood vessels and dense biofilms (Fig. 2A) was employed to evaluate the in vitro penetration and antibacterial efficacy. This model recapitulated the physiological complexity of infected tissues, where bacteria were shielded by both the endothelial barrier and a protective extracellular matrix. The 3D-reconstructed CLSM images provided visual evidence of the superior biofilm penetration ability of the SPCMPL + US (Figure S8–S9). It was clearly observed that SPCMPL + US group achieved significantly greater penetration depth into the GFP-labeled PAO1 biofilm than the SPCMPL group without US and other control groups. This enhancement could be attributed to US-induced cavitation, which generated mechanical forces to disrupt the biofilm structure and propel liposomes into deeper sites.

Fig. 2.

Evaluation of biofilm disruption and bacterial killing efficacy under different treatments. (A) Schematic of IVEC barrier establishment, biofilm formation, and different treatment processes. (B) Line graph showing the OD₆₀₀ over a 12-h period, reflecting the growth kinetics of biofilms under different treatments. The OD value was an indicator of biofilm biomass, where a lower OD suggested effective inhibition of biofilm growth. (C) Representative CLSM images (top and side views) and crystal violet–stained images of biofilms after 12 h of different treatments. (D) Representative colony count images and live-dead-stained CLSM images within biofilms after 12 h of different treatments (top and side views, dead bacteria in green and live bacteria in red). (E) Four bar graphs showing quantitative data on biofilm thickness, relative ONPG content in the supernatant, biofilm biomass, and antimicrobial rate for different treatment groups. (F) Representative SEM images of biofilms after different treatments. Scale bar = 5 μm. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, and ^****P < 0.0001; ns means no significance

Following group-wise administration and US irradiation, the absorbance at optical density (OD₆₀₀) was monitored hourly over a 12-h period to characterize the dynamic changes in biofilm and planktonic bacterial density. The SPCMPL + US group exhibited a consistent decline in OD₆₀₀ values, dropping to 0.58 ± 0.01 by the 6-h mark, which was a 40% reduction from baseline and significantly lower than that in the MEM + US group (1.03 ± 0.01, P < 0.05), SCMPL + US group (0.79 ± 0.02, P < 0.01) and PCMPL + US group (0.93 ± 0.04, P < 0.05). This trend persisted for 12 h, stabilizing at 0.35 ± 0.06, indicating sustained biofilm disruption and bacterial killing. In contrast, the PBS control group showed a 1.5-fold increase in OD₆₀₀ (from 1.13 ± 0.07 to 1.67 ± 0.06), thus reflecting unimpeded biofilm growth. However, the OD₆₀₀ value in the SPCMPL group (without US) increased to 1.30 ± 0.05, thereby highlighting the vital role of US in triggering drug release. The value in the MEM + US group plateaued at 0.88 ± 0.06 after 8 h, suggesting incomplete biofilm eradication (Fig. 2B).

We evaluated the therapeutic effect of SPCMPL combined with US after 12 h of US irradiation from the viewpoints of biofilm disruption and bacterial killing. After removing planktonic bacteria, the biofilm was subjected to crystal violet staining and fluorescent matrix staining (Fig. 2C). Crystal violet staining showed the most pronounced biofilm disruption in the SPCMPL + US group, with absorbance at 595 nm 28.21% lower than that in the SCMPL + US group (P < 0.001) and 21.51% lower than that in the MEM + US group (P < 0.01), providing quantitative evidence of a significant reduction in the density of EPS. CLSM further corroborated this finding, demonstrating that the biofilm matrix in the SPCMPL + US group was markedly sparser and thinner, with an 82.02% reduction in thickness compared with that for untreated controls (SCMPL + US: 44.45% reduction, SPCMPL: 16.03% reduction). This indicated substantial degradation of the mature biofilm structure and a notable decrease in biomass accumulation. This structural disruption was attributed to the US-triggered phase transition of PFP, which generated microbubbles that mechanically disrupted the EPS network, as evidenced by fragmented matrix architecture and reduced biofilm compactness.

The live/dead staining results of SPCMPL + US treatment showed effective elimination of both planktonic and biofilm-embedded bacteria. In contrast, other treatment groups exhibited limited antibacterial action, primarily affecting bacteria on the biofilm surface without achieving significant penetration into its interior. The CFU counting experiment accurately assessed the remaining viable bacterial counts after different treatments. The CFU count was significantly lower in the SPCMPL + US group than in the other groups, indicating the strong antibacterial ability of SPCMPL + US to effectively reduce the number of viable bacteria in the biofilm (Fig. 2D and E). The biochemical analysis of culture supernatants from different treatment groups provided additional mechanistic insights. Specifically, 1.65 ± 0.34 mM potassium ions and 1.04 ± 0.14 U/mL β-galactosidase were released in the SPCMPL + US group, which were higher than those in the MEM + US group (0.94 ± 0.03 mM, P < 0.05; 0.66 ± 0.04 U/mL, P < 0.01; Figure S10). These elevated levels strongly indicated that the SPCMPL + US strategy significantly disrupted bacterial cell walls, leading to the leakage of intracellular contents and loss of cellular integrity.

Finally, high-resolution images of biofilms after various treatments were captured using scanning electron microscope (SEM) to conduct a detailed microscopic examination of biofilm structure and bacterial morphology. For sample preparation, the biofilms were first fixed with 2.5% glutaraldehyde and then dehydrated through an ethanol gradient, followed by sputter coating with a gold-palladium layer. They were finally imaged using Hitachi SU8010. SEM images demonstrated that the EPS matrices in both SPCMPL + US and PCMPL + US groups had undergone severe disintegration, appearing as sparse, thread-like fragments rather than the thick, intact membranous structures observed in the control groups. Compared with the dense, continuous EPS networks embedding bacteria in the MEM + US group, SPCMPL + US-treated biofilms exhibited nearly complete EPS degradation, leaving only isolated, free-floating bacteria (Fig. 2F). High-magnification SEM further revealed that these bacteria displayed distinct signs of cell wall damage, including membrane blebbing, surface roughening, and discrete lysis sites, which are morphological hallmarks of cytoplasmic leakage and cellular disintegration (Figure S11). These microscopic observations provide direct, visual evidence of the dual-mode efficacy of the treatment, underscoring its capacity to both disrupt the protective EPS matrix and induce lethal damage to bacterial cells.

Collectively, the strategic integration of crystal violet staining, CLSM, live/dead viability assays, bacterial colony count, biochemical marker analyses, and SEM constituted a robust, multimodal analytical framework. This convergence of qualitative and quantitative data from diverse experimental modalities unequivocally established the dual-action mechanism of SPCMPL + US. This innovative treatment strategy has overcome the key resistance mechanisms of biofilm-associated infections by synergistically disrupting the protective biofilm architecture through US-triggered PFP phase transition and eradicating embedded bacteria via targeted delivery of ROS and meropenem. The multimodal evidence base not only validates the treatment efficacy but also positions SPCMPL + US as a promising therapeutic approach for combating recalcitrant biofilm-related diseases.

Promotion of macrophage activation

Macrophage-associated innate immunity serves not only as the primary defense mechanism against pathogen invasion but also as a central driving force for adaptive immune responses [32]. However, P. aeruginosa-related biofilm infections often induce “frustrated phagocytosis” and immune suppression in macrophages. This dysfunction highlights the urgent need for targeted strategies to re-activate macrophage activity [6]. Extensive exploration has demonstrated a crucial role of ROS, as key immunomodulatory signaling molecules, in activating innate immune responses by regulating macrophage polarization and phagocytic function, especially under the heterogeneous immune conditions within biofilms [33]. In this study, RAW 264.7 murine macrophages were subjected to various treatments, including SPCMPL, PCMPL, SCMPL, CMPL, MEM, and PBS controls, followed by US treatment (1.5 W/cm², 3 MHz, 50% duty cycle, 1 min) for the US-treated groups. WB analysis confirmed that the US-activated Ce6-containing liposomal formulation significantly upregulated key macrophage polarization-related signaling molecules, including the phosphorylation of NF-κB p65 (p-NF-κB p65/total NF-κB p65) and STAT1 (p-STAT1/total STAT1), as well as the protein level of HIF-1α (Fig. 3A and B and S12). These findings demonstrated that ROS production upon US irradiation functioned as a critical upstream mediator, simultaneously activating the NF-κB, HIF-1α, and STAT1 pathways, thereby synergistically promoting and sustaining M1 macrophage polarization [34–36]. Consistent with these molecular findings, FCM analysis further demonstrated that Ce6-containing liposomal formulations significantly enhanced the expression of M1-type surface markers CD80 and CD86 on macrophages under US irradiation (SPCMPL + US: 63.5% ± 14.93%; SPCMPL: 10.72% ± 2.11%, P < 0.001; MEM + US: 16.19% ± 11.51%, P < 0.001). This enhancement indicated a phenotypic shift toward a more pro-inflammatory and immunologically active state, which was crucial for effective immune responses. In contrast, neither US irradiation alone nor SPCMPL without US stimulation produced a discernible shift in macrophage phenotype (Fig. 3C).

Fig. 3.

Macrophage phenotype polarization, migration, and phagocytosis assays through activation of key signaling pathways. (A) Representative WB images of HIF-1α, p-STAT1, Total STAT1, p-NF-κB p65, Total NF-κB p65, and β-actin in different treatment groups. (B) Quantitative analysis of relative expression levels for p-NF-κB p65/NF-κB p65, HIF-1α, and p-STAT1/STAT1. ^*P < 0.05, ^**P < 0.01, and ^***P < 0.001. (C) FCM analysis of macrophage M1 phenotype markers (CD80/CD86). The percentage of double-positive (CD80⁺CD86⁺) cells (gated regions) was indicated in each plot, representing the proportion of macrophages polarized to the pro-inflammatory M1 phenotype. (D) Crystalline violet-stained bright-field microscope images of macrophage migration in Transwell after different treatments. Purple-stained cells represented migrated macrophages, with deeper staining indicating stronger migration ability. Scale bar = 100 μm. (E) Representative FCM histograms quantifying phagocytosis by macrophages after different treatments by measuring the percentage of green fluorescent proteins (GFP)-positive (indicating internalized GFP-P. aeruginosa) cells. (F) Bar graph summarizing the percentage of CD80⁺CD86⁺ macrophages in all treatment groups. (G) Bar graph showing macrophage migration rates by treatment group. (H) Bar graph quantifying macrophage phagocytosis (percentage of GFP⁺ macrophages) by treatment group. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, and ^****P < 0.0001

Transwell migration assays further validated the functional impact of the phenotypic change. Macrophages treated with SPCMPL + US demonstrated a 2.37-fold increase in migration toward a P. aeruginosa-conditioned media compared with control groups, reflecting enhanced chemotactic responsiveness. FCM-based phagocytosis assays, using fluorescently labeled P. aeruginosa, quantified a 2.25-fold increase in phagocytic activity in the SPCMPL + US group compared with control groups (Fig. 3D and H). These results collectively indicated a crucial role of US-triggered ROS release from SPCMPL liposomes in reactivating “frustrated” macrophages [37]. This treatment strategy overcame immune suppression by restoring macrophage polarization and function, thereby enabling sustained phagocytosis and bactericidal activity against biofilm-embedded pathogens. In addition, the cytotoxicity assessments on RAW 264.7 cells were performed using the CCK-8 assay. The cell viability in RAW 264.7 cells exceeded 85% in all treatment groups, indicating the favorable biocompatibility of liposomes (Figure S13).

BSSI-targeting and enhanced accumulation of SPCMPL

We employed a rigorous in vivo methodology using BALB/c nude mice bearing BSSIs to systematically evaluate the targeting efficiency of sulfatide-modified liposomes and determine the optimal timing for US irradiation to enhance SPCMPL accumulation (Fig. 4A). Four Cy5-labeled liposome formulations (SPCMPL, PCMPL, SCMPL, and CMPL) were intravenously injected into separate groups (n = 3 per group), with fluorescence imaging conducted 1, 2, 4, 6, and 8 h after injection using an IVIS Lumina system (excitation/emission: 649/670 nm). Quantitative region-of-interest analysis was performed using Living Image software, normalizing signal intensity to background levels and correcting for autofluorescence. The kinetic analysis revealed that SPCMPL and SCMPL groups exhibited significantly higher fluorescence intensity at BSSI sites compared with PCMPL and CMPL groups across all time points (e.g., 6 h: SPCMPL 7.73 ± 1.04 vs. PCMPL 5.79 ± 0.47 [(p/s)/(cm²/sr), 10⁹], P < 0.05), confirming the vital role of sulfatide modification in enhancing endothelial cell targeting via P-selectin binding (Fig. 4B). The signal kinetics followed a biphasic pattern, peaking at 6 h postinjection before declining, likely reflecting the balance between active targeting, tissue penetration, and systemic clearance (Fig. 4C).

Fig. 4.

In vivo biodistribution and targeting assessment of different liposomes in the BSSI mouse model. (A) Schematic illustration of the two-step procedure for constructing the BSSI mouse model. (B) In vivo fluorescence images of mice 1, 2, 4, 6, and 8 h after injection of different liposomes. Luminescence indicates areas of high bacterial aggregation (left column), and epifluorescence indicates areas of high liposome aggregation (right column). (C) Quantitative analysis of in vivo MFI dynamics of liposomal formulations at the BSSI site over time. (D) In vivo fluorescence images showing the biodistribution of different liposomal formulations 24 h after intravenous injection (top row). Ex vivo fluorescence images of major organs (BSSI site/heart/liver/spleen/lung/kidney) harvested from treated mice (bottom row). Bar graph quantifying the MFI of each organ and BSSI tissue (right). ^**P < 0.01

Based on this finding, we subjected the BSSI sites to US irradiation (intensity 2 W/cm², frequency 3 MHz, duty cycle 50%, and duration 10 min) 6 h after injection. This parameter combination was selected to align with safety limits established for diagnostic US applications, which did not induce any significant thermal damage or histological abnormalities. Also, we included a group that received only SPCMPL without US for further evaluating the in vivo drug distribution of these liposomes. The in vivo and ex vivo fluorescence imaging revealed the highest drug accumulation at BSSI sites in the SPCMPL + US group 18 h after injection. The quantitative analysis showed a 7.78-fold greater accumulation in the SPCMPL + US group compared with the PCMPL + US group (6.46 ± 3.20 vs. 0.83 ± 0.16 [(p/s)/(cm²/sr), 10⁹], P < 0.001), directly demonstrating the superior targeting efficiency of sulfatide modification. This enhancement stemmed from the specific binding of sulfanilamide to P-selectin, which is upregulated by inflammatory endothelial cells, as confirmed by CLSM and FCM (Fig. 1E and G). The drug accumulation at the BSSI sites was 4.82-fold (P < 0.001) and 8.25-fold (P < 0.001) higher in the SPCMPL + US group compared with the SCMPL + US group (devoid of PFP) and the SPCMPL group without US irradiation, respectively. This stark difference highlighted the dual mechanistic advantage of US-triggered PFP phase transition: (1) the liquid-to-gas conversion generating microbubble expansion forces that disrupted the biofilm EPS matrix, as observed in SEM images and CLSM (Fig. 2F and S8); and (2) the cavitation effect reducing liposome hydrodynamic diameter from 184.76 ± 5.49 nm to 139.67 ± 3.97 nm (Fig. 1A), enabling deeper penetration through biofilm pores.

Apart from the expected hepatic and renal clearance pathways, fluorescence intensity accumulation of SPCMPL in BSSI infection site was 9.40–42.74 times greater than those of in other major organs, proving the specific targeting to inflammatory tissues (Fig. 4D). Meanwhile, no significant difference was found in the distribution of liposomes in the heart, liver, spleen, lungs, and kidneys compared with the other liposomes. This finding contrasted with previous studies using untargeted liposomes, which often exhibited off-target accumulation, thus highlighting the safety advantages of our targeting strategy. This spatial confinement of drug delivery was attributed to the synergistic effects of sulfatide-mediated targeting and US-localized activation. They jointly overcame the two major bottlenecks, vascular endothelial barrier and biofilm EPS barrier, in BSSI treatment. The results collectively demonstrated that the combination of sulfatide modification and US-triggered PFP phase transition enabled precision drug delivery to infected sites, thereby maximizing therapeutic efficacy while minimizing systemic toxicity.

Promotion of macrophage enrichment and M1 polarization in BSSI lesions

The ideal therapeutic strategy for P. aeruginosa biofilm infections requires effective disruption of the protective biofilm barrier and simultaneously activation of the immune system to counteract the large number of pathogens released [38]. This is because the physical barrier of EPS and the immunosuppression and immune depletion caused by chronic inflammation led to immune isolation. Following three consecutive treatments (24-h intervals) in BSSI-bearing mice model, FCM of infected tissue homogenates was performed using a panel of fluorochrome-conjugated antibodies (F4/80, CD80, and CD206) and analyzed with FlowJo software. The quantitative analysis revealed that, regardless of the severity of infection, the macrophages in the PBS control group failed to achieve effective enrichment or infiltration into the BSSI lesions, with the total infiltration (F4/80⁺ cells) remaining at only 2.07% ± 0.16% of viable cells. In contrast, the SPCMPL + US group exhibited a significant 1.91-fold increase in total macrophage infiltration at the BSSI sites (3.95% ± 0.69%, P < 0.01). This indicated that this strategy effectively breached the biofilm barrier and markedly enhanced macrophage recruitment to the BSSI site. In comparison, the proportion of M1 was 3.05% ± 1.27% in the PCMPL + US group and 3.22% ± 0.18% in the SCMPL + US group, further underscoring that the combination of targeted modification and US cavitation significantly promoted drug enrichment and penetration at the BSSI site (Fig. 5A). Of greater significance, the SPCMPL + US group exhibited a markedly higher proportion of pro-inflammatory M1-phenotype macrophages (CD80⁺CD206⁺: 52.0% ± 9.22% of F4/80⁺ cells) compared with the PBS control group. This was dominated by immunosuppressive M2-phenotype macrophages (CD80⁺CD206⁺: 31.1% ± 7.48% of F4/80⁺ cells, P < 0.01; Fig. 5B), indicating a robust transition from an immunoinhibitory to an antibacterial microenvironment. The immunofluorescence staining further corroborated this finding, revealing dense clusters of iNOS-expressing M1 macrophages (Alexa Fluor 555, red) infiltrating the disrupted biofilm matrix in SPCMPL + US-treated tissues. However, PBS controls displayed abundant CD206⁺ M2 macrophages (Alexa Fluor 488, green) embedded within intact EPS networks (Fig. 5C and S14).

Fig. 5.

Macrophage phenotype profiling in BSSI tissues after treatment. (A) Representative FCM contour plots illustrating the proportion of F4/80⁺ macrophages in the total live cell population in BSSI tissues across various treatment groups. Adjacent to the contour plots, a bar graph quantifies the percentage of F4/80⁺ cells (mean ± SD, n = 3). (B) Representative FCM contour plots showing the proportion of M1-like (CD80⁺, CD206⁺) macrophages in the total F4/80⁺ cell population in BSSI tissues across different treatment groups. A bar graph adjacent to the contour plots quantifies the percentage of CD80⁺ and CD206⁺ cells (mean ± SD, n = 3). (C) Representative immunofluorescence images of BSSI tissues stained for M1-like (iNOS, red) and M2-like (CD206, green) macrophage markers, with DAPI (blue) counterstaining for nuclei. Scale bar = 100 μm

The enhanced macrophage recruitment and M1 polarization elicited by the SPCMPL + US strategy could be ascribed to two synergistic mechanisms. Initially, US-induced PFP phase transition physically disrupted the biofilm EPS matrix and heterogeneous immunosuppressive microenvironment. This disruption exposed embedded bacterial antigens and pathogen-associated molecular patterns, which activated macrophage pattern recognition receptors to initiate pro-inflammatory responses [39, 40]. In parallel, the second mechanism involved US-triggered ROS release from SPCMPL. Specifically, the energy transfer from excited-state Ce6 converted ground-state oxygen into singlet oxygen, inducing M1 polarization through multiple metabolic pathways, as previously documented [41, 42]. US-induced cavitation further amplified this effect by increasing oxygen supply and Ce6 excitation efficiency.

Collectively, these mechanisms synergized to form a self-reinforcing cycle, endowing the SPCMPL + US strategy with unique spatiotemporal precision that set it apart from conventional systemic immunostimulants. This precision allowed for the targeted reactivation of macrophages within biofilm-infected niches, effectively restoring their bactericidal capabilities and reinstating a robust immune response against BSSI lesions. The strategy disrupted the cycle of “immune paralysis” by circumventing the EPS-mediated immunosuppression that typified chronic biofilm infections, significantly enhancing the recognition and eradication of bacteria. As a result, it efficiently controlled the release of pathogens during biofilm disruption, thereby preventing their systemic spread and subsequent recolonization, and offering a comprehensive solution to combat recalcitrant biofilm-associated infections.

Antibiofilm and antibacterial efficacy in BSSI mice

This study strictly adhered to the technical roadmap depicted in Fig. 6A, systematically evaluating the anti-biofilm and antibacterial efficacy of the SPCMPL + US combinatorial strategy in a P. aeruginosa-induced BSSI model. Mice with peak in vivo fluorescence intensity > 3000 counts were selected and randomly assigned to eight treatment groups (n = 5): PBS + US control, MEM + US, three control liposome (CMPL, SCMPL, and PCMPL) + US, SPCMPL + US, only SPCMPL, and ciprofloxacin (CIP). The only-SPCMPL group was included to isolate the effect of US, thereby deconstructing the critical contribution of US in this combined strategy. Additionally, a CIP treatment group, representing another alternative first-line antibiotic, was included to enable comparative assessment of SPCMPL + US against a clinically relevant antibiotic regimen. All groups received intravenous administration of respective formulations on days 1–3. Six hours postinjection, US irradiation (2 W/cm², 3 MHz, 50% duty cycle, 10 min) was administered to treatment groups, except the only-SPCMPL group. The longitudinal monitoring of BSSI lesion fluorescence and animal body weight was performed on days 1, 3, 5, and 7 using an in vivo imaging system (Fig. 6B and C). This experimental design enabled a multidimensional assessment of the efficacy of targeted liposomes, the US-triggered effect, and the antibiotic synergistic action through parallel comparisons between the control and experimental groups. It also provided further translational insights and facilitated the mechanistic dissection of how acoustic energy modulated biofilm disruption and immunomodulatory responses by directly comparing the SPCMPL + US strategy with various traditional antibiotic therapies.

Fig. 6.

In vivo therapeutic efficacy evaluation of liposomes in BSSI-bearing mice model. (A) Schematic timeline of the in vivo experiment. (B) Changes in mouse body weight over the experimental period. (C) Serial in vivo bioluminescence images of BSSI-infected mice on days 1, 3, 5, and 7 posttreatments. The color scale represented bioluminescence radiance [(p/s)/(cm²/sr)], where higher radiance indicated a greater bacterial load. The bottom row showed representative CFU plates from tissue homogenates in each group, visually confirming the bacterial burden. (D) Bar and line graphs quantifying bacterial load in BSSI mice. Left: Maximum (Max FI) and mean fluorescence intensities (Mean FI) of bioluminescence signals in 7 days (mean ± SD, n = 5 per group). Right: Antibacterial rates (%) calculated as the percentage reduction in bacterial load on day 7 compared with day 1 in the PBS control group. (E) Histopathological analysis of BSSI tissues. HE staining visualized general tissue morphology, inflammation, and cellular infiltration. Masson’s trichrome staining highlighted collagen deposition (blue) to assess tissue fibrosis. Giemsa staining aided in the identification of bacteria and inflammatory cells. The scale bar (bottom right) = 100 μm, and images are representative of five independent samples per group. (F) Bar graphs quantifying key histopathological parameters in BSSI tissues: Neutrophil proportion (%), collagen proportion (%), and bacterial proportion (%). (G) Quantitative analysis of bacterial CFU from tissue homogenates in each group. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, and ^****P < 0.0001

Continuous tracking using the in vivo fluorescence imaging system revealed significant declines in both the peak and mean fluorescence intensities of BSSI lesions in the SPCMPL + US group, with the maximum intensity decreasing by 70.7% and the average intensity by 57.7% on day 5 (Fig. 6C and D). These findings robustly confirmed the exceptional capability of the strategy to disrupt biofilm architecture and exert dual antimicrobial effects against both planktonic and biofilm-embedded bacteria. In contrast, the control liposome + US groups (CMPL, SCMPL, and PCMPL) and clinical antibiotic monotherapy groups (MEM and CIP) exhibited a more gradual decline in mean fluorescence intensity, with average reductions of only 26.4%–30.9% and peak intensity reductions of 19.0%–32.1% by day 5. By day 7, BSSI lesion fluorescence in the SPCMPL + US group further decayed to near the instrument detection threshold, achieving anti-biofilm and antibacterial rates of 90.6% and 92.6%, respectively. Concomitantly, the steady recovery of mouse body weight starting from day 3 indicated the restoration of host metabolic function. On the contrary, other treatment groups exhibited a plateau in fluorescence intensity from day 5 (fluctuation < 10%), with some even showing rebounds. This suggested that, although the bacterial burdens were reduced in these groups, intact biofilm structures enabled P. aeruginosa to persistently colonize the BSSI site, thus sustaining and exacerbating the infection. The US-untreated group displayed paradoxical fluorescence intensification. When combined with in vitro liposome release assays, this finding highlighted the critical dependency of the nanotherapeutic system on US-triggered smart responsiveness.

Anatomical observations revealed that the SPCMPL + US treatment group achieved primary intention healing, marked by robust tissue repair, absence of intra-abdominal adhesions, and no adverse reactions [43]. In contrast, other treatment groups showed varying degrees of congestion, serous exudation, and abnormal intra-abdominal adhesions (Figure S15). The histopathological analysis using HE staining showed that the neutrophil infiltration density in the SPCMPL + US group was 55.0% lower than that in the model group. In contrast, other groups presented numerous neutrophil apoptotic bodies, suggesting that the infection had entered a chronic phase. Masson staining indicated a 2.8-fold increase in collagen fiber deposition in the SPCMPL + US group, with fibers arranged in line with myofibrils, thus suggesting an optimized wound remodeling process. Giemsa staining combined with tissue homogenate colony counting revealed that the bacterial density at the infection site in the SPCMPL + US group was less than 10³ CFU/g, thus meeting the clinical healing criteria (Fig. 6E and F). Taken together, these multidimensional data demonstrated that the SPCMPL + US strategy achieved dual enhancement of anti-biofilm and antibacterial efficacy, enabling efficient control of refractory BSSI and promotion of surgical site wound healing. From a translational perspective, the safety profile of the strategy was validated through a comprehensive toxicological assessment: major organs showed no histological damage, and the biochemical indicators of liver and kidney function remained within the normal range, with no apparent toxic reactions detected (Figure S16–S17). While these results proved a favorable safety profile, the long-term biosafety evaluation might be further performed on chronic toxicity, biodistribution kinetics, and immunogenicity to ensure the clinical translation of SPCMPL.

The differential efficacy observed between SPCMPL + US and control groups emphasized the vital role of synergistic mechanisms in fighting biofilm infections. Compared to standard-dose monotherapy with antibiotics (MEM and CIP), SPCMPL + US achieved superior therapeutic outcomes at equivalent dosages, demonstrating 92.6% antibacterial efficacy on day 7. This was attributed to the combined effects of targeted drug delivery, US-triggered drug release, biofilm matrix degradation, and immune activation. The paradoxical phenomenon of increased fluorescence in the non-US-treated group highlights a major limitation of conventional nanocarriers: passive drug release may exacerbate inflammation without disrupting the biofilm structure, leading to relapse after transient bacterial inhibition. The immunomodulatory effects of SPCMPL + US are particularly notable, given the central role of macrophage dysfunction in chronic infections. The strategy not only enhances phagocytosis but also creates a pro-inflammatory milieu hostile to biofilm persistence by polarizing macrophages to the M1 phenotype [44]. The spatial control afforded by US-triggering is critical because systemic M1 polarization can induce cytokine storms, which is a risk mitigated by confining immunomodulation to the infection site [45]. From a clinical translational perspective, the achievement of primary intention healing in SPCMPL + US-treated animals highlight the potential of SPCMPL + US for managing biofilm-related surgical site complications, which are a leading cause of postoperative morbidity. Future studies should prioritize optimizing US parameters for clinical translation, evaluating long-term efficacy in large-animal models, and developing combination therapies to address multi-species biofilm infections.

Conclusions

In summary, this study developed a kind of P-selectin endogenous sulfatide ligand-modified transformable liposome, SPCMPL, for BSSI treatment. The liposome enabled active targeting and accumulation at deep infection sites through ligand/receptor-mediated endocytosis. Under US irradiation, SPCMPL encapsulating PFP underwent a potent liquid-gas phase transition, disrupting P. aeruginosa biofilms barriers to promote deep penetration and uniform distribution of antibiotics within the biofilm matrix. Concurrently, ROS generation activated the immune system, promoting macrophage recruitment and activation for sustained phagocytic and bactericidal effects. The SPCMPL with US assistance demonstrated significant antibacterial efficacy in both in vitro and in vivo models, thus effectively eradicating bacterial biofilms through enhanced phagocytosis while inhibiting pathogen proliferation and biofilm reformation. US-activated and P-selectin-targeted liposomes are practical and efficient carriers for active BSSI targeting drug delivery.

Materials and methods

Materials

All chemical reagents, unless specified otherwise, were purchased from Sigma-Aldrich Inc. or Aladdin Reagent Inc. 1,2-Dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), 3β-(N-(N, N-dimethylaminoethane)carbamoyl)cholesterol hydrochloride (DC-Chol), and 1,2-Distearoyl-sn-Glycero-3-Phosphoethanolamine-N-[Methoxy (Polyethylene Glycol)−2000] (DSPE-PEG2000) were obtained from Avanti Polar Lipids, Inc. (Alabaster, USA). Ce6-modified DSPE-PEG2000 (DSPE-PEG-Ce6) and Cyanine5 (Cy5)-labeled DSPE-PEG2000 (DSPE-PEG-Cy5) were acquired from Xi’an Ruixi Biological Technology Co., Ltd. (Xi’an, China). Luria-Bertani (LB) broth, MEM, and CIP were obtained from Sigma-Aldrich Inc. (St. Louis, USA). Dulbecco’s modified eagle medium (DMEM), fetal bovine serum (FBS), PBS and 0.25% trypsin solution were procured from Gibco (Grand Island, USA). Human lipopolysaccharide (LPS) recombinant protein and FilmTracer SYPRO Ruby were purchased from Thermo Fisher Scientific Inc. (Waltham, USA). The bicinchoninic acid (BCA) protein assay kit, dimethyloctylamine (DMAO), Propidium Iodide (PI), 4’,6-Diamidino-2-Phenylindole (DAPI), DCFH-DA ROS Assay Kit, and NO Assay Kit were obtained from Beyotime Biotechnology (Nanjing, China). Anti-P-selectin, anti-β-actin antibodies and FITC-conjugated secondary antibody were purchased from Abcam (Cambridge, UK). F4/80-BV421, CD80-PE, CD86-BV785, CD206-APC, iNOS-Alexa Fluor 555, and CD206-FITC antibodies were purchased from BD Biosciences (Franklin, USA).

Establishment of BSSI model and targeting receptor validation

A BSSI model was established employing male BALB/c nude mice (aged 6–8 weeks) (Figure S1). A two-step process was followed for the model establishment. First, a superficial wound was created, and 10 µL of PA-Luci suspension (10⁸ CFU/mL) was applied to the wound. A visible white bacterial biofilm developed on the wound after 48 h, which was then excised and reserved for subsequent use. Second, the skin and peritoneum of another nude mouse were incised, and the peritoneum was sutured continuously to simulate a surgical wound. A small piece of biofilm-bearing skin tissue was placed on this wound, followed by the addition of 10 µL of PA-Luci suspension. The skin was then sutured. Subsequently, 10 µL of luciferase substrate was injected subcutaneously at the wound site after another 48 h. The bioluminescent changes were measured using the in vivo imaging system of the Caliper IVIS Lumina II (PerkinElmer, USA). Mice with peak bioluminescence intensities exceeding 3000 counts were considered successful BSSI models. We obtained 0.5 × 0.5 cm² tissue samples at random from both biofilm-infected and healthy sites of three BSSI model mice. The samples were homogenized using a tissue mincer for 2 min, followed by the addition of 500 µL of RIPA lysis buffer and incubation on ice for 30 min. They were then centrifuged at low temperature. The supernatant was collected for WB analysis to investigate the expression differences of P-selectin protein between the BSSI and normal sites. The objective of this experiment was to assess the feasibility of P-selectin as a target for drug delivery to BSSI sites.

Preparation of SPCMPL

SPCMPL, a liposome containing sulfatide, Ce6, and MP, and encapsulating PFP, was synthesized using the following procedure. An 18 mg lipid mixture, comprising DPPC (3.0 mg, 4.08 µmol), Dc-Chol (3.0 mg, 5.56 µmol), sulfatides (3.0 mg, 4.64 µmol), DSPE-PEG-Ce6 (6.0 mg, 1.74 µmol), and MP (3.0 mg, 3.46 µmol), was placed in a 10-mL flask and dissolved in 3 mL of chloroform. The mixture was evaporated for 30 min using a rotary evaporator, followed by 1 h of vacuum drying at room temperature to form a lipid film. The lipid film was cooled to 4 °C, and then 100 µL of PFP was added and allowed to soak into the lipid film. The film was then rehydrated with 5 mL of PBS containing 5% glycerol and mechanically agitated for 30 min at 4 °C. Subsequently, the mixture was agitated for 1 h at 30 °C in a heated water bath to remove any residual PFP. The control liposomes SCMPL, PCMPL, and CMPL were fabricated using an analogous procedure, with each formulation omitting specific components: SCMPL did not contain PFP, PCMPL lacked sulfatides, and CMPL was devoid of both PFP and sulfatides. The mass of the omitted components was adjusted by incorporating an equivalent amount of DSPE-PEG. Concurrently, Cy5-labeled liposomal nanodroplets were fabricated by substituting an equivalent mass of DSPE-PEG with Cy5-conjugated DSPE-PEG (0.30 mg, 0.09 µmol), representing approximately 1.7% of the total lipid mass within the formulation, following the previously detailed methodology.

Characterization of liposomes

The Zetasizer Advance Range particle size analyzer, based on DLS technology, was used to measure the size, potential, and dispersion characteristics of the liposomes, which helps in assessing the uniformity and stability of nanoparticle distribution. The same tests were conducted on the other three types of liposomes. Cryo-transmission electron microscope (cryo-TEM) was used to precisely observe the morphology and distribution characteristics of the lipid nanoparticles. Subsequently, we further investigated the changes in size in SPCMPL under various storage durations, temperature stimuli, and ultrasonic intensity conditions. We also analyzed the stability variations of the liposomes under these influencing factors and explored the key elements for liquid-gas phase transitions. The liposomes were suspended in a simulated medium supplemented with 10% FBS and incubated with gentle shaking at 37 °C. The particle size and stability of the liposomes were monitored at predetermined time intervals throughout the incubation period.

Release of ROS and MEM

The DCFH-DA ROS Assay Kit was employed as an ROS indicator probe. A 100-µL volume of SPCMPL (10 mM) was subjected to US irradiation at intensities of 0.5, 1.0, 1.5, and 2.0 W/cm² for 10 min. Subsequently, a 100-µL aliquot of a 10 nM solution of the DCFH-DA probe was added, and the samples were incubated at 37 °C for 30 min. ROS oxidized the nonfluorescent DCFH-DA to generate fluorescent 2’,7’-Dichlorofluorescein (DCF). The amount of ROS released by the liposomes was determined by measuring the fluorescence intensity at 525 nm using a multimode microplate reader. The reaction was allowed to proceed for an additional hour. Then, a 100-µL aliquot of the solution was subjected to HPLC analysis to quantify the formation of MEM. The HPLC system included a 1525 binary pump, a 2475 multi-λ-fluorescence detector, a 2998 photodiode array detector, and a SunFire C18 column (4.6 × 250 mm², 5 μm). The mobile phase consisted of methanol and water in a ratio of 20:80 (v/v), with a flow rate of 1.0 mL/min at 35 °C. The detection wavelength was set at 296 nm.

Development of IVEC model in vitro

LPS was used as a tool to mimic bacterial infections and inflammatory responses, thereby stimulating HUVECs to construct an in vitro IVEC model. Initially, HUVEC cells were cultured to the logarithmic growth phase and seeded at a density of 20,000 cells per well in a six-well plate. They were then stimulated with LPS at concentrations of 0, 0.1, 1, 10, 50, and 100 µg/mL for 12 h. After removing the supernatant and washing the cells three times with PBS, a NO fluorescent probe was added to detect the inflammation levels of HUVECs. Subsequently, total proteins were extracted from each well using RIPA lysate buffer for WB analysis to measure the expression level of P-selectin. Relative quantification was performed using β-actin as an internal reference to explore the relationship between different levels of inflammation and P-selectin expression.

Evaluation of biosafety

HUVECs and IVECs served as cellular models to ascertain the biocompatibility of liposomes. SPCMPL solutions were introduced at distinct concentration gradients of 0.01–10 mg/mL (MEM-equivalent dose: from 0.04 µM to 41.70 µM)), allowing for 48 h of substantial cell-drug interaction to mimic the prolonged effects of the material within a biological environment. Following the incubation period, the CCK-8 assay was employed to assess cellular viability. The absorbance of each group was measured at 450 nm using a spectrophotometer, and the cell survival rate was calculated to evaluate the biocompatibility of SPCMPL with HUVECs and IVECs. Hemocompatibility was predominantly assessed using a hemolysis test. A 5% erythrocyte suspension was combined with SPCMPL solutions at varying concentrations and incubated at 37 °C on a constant-temperature shaker for 4 h. The mixture was then centrifuged at 1000g for 5 min to precipitate the erythrocytes, and the supernatant was carefully removed. The degree of hemolysis was preliminarily determined by observing the changes in the color of the supernatant. Quantitative analysis was performed by measuring the absorbance at 540 nm using a spectrophotometer.

Binding capacity of ligand and receptor

A total of 20,000 HUVECs were added to confocal plates. After the cells adhered to the wall, 50 µg of LPS was added and incubated for 12 h to activate and transform these cells into IVECs. The cells were treated with cold methanol for 15 min at 4 °C to achieve fixation, followed by incubation with 5% bovine serum albumin for 30 min to block nonspecific binding. Subsequently, the P-selectin primary antibody was introduced, and the samples were incubated overnight to allow for the antibody to bind. On the following day, the samples were incubated with a FITC-conjugated secondary antibody for 1 h at room temperature. The nuclei were stained with DAPI for 5 min, and the difference in FITC fluorescence between HUVECs and IVECs was observed by CLSM. Then, Cy5-labeled SPCMPL was added and the cells were rinsed with PBS after 10 min to observe the binding of Cy5 fluorescence to HUVECs and IVECs. Similarly, Cy5-SPCMPL and Cy5-PCMPL were added to the media containing HUVECs and IVECs and incubated at 37 °C for 2 h. After rinsing the cells three times with PBS, FCM analysis was performed to record the Cy5 fluorescence intensity in different groups.

Ligand/receptor-mediated transport

A total of 50,000 HUVECs were added to each Transwell chamber (24 units with 3.0-µm pore polycarbonate membrane and 6.5-mm inserts, tissue culture-treated). The culture medium was changed every other day until the resistance difference between the upper and lower chambers exceeded 400 Ω/cm², confirming the formation of a tight HUVEC barrier. Subsequently, the chambers were replaced with a serum-free medium containing 50 µg/mL LPS and stimulated for 12 h to establish an IVEC barrier model. Then, Cy5-labeled SPCMPL and PCMPL were added to the upper chambers and incubated at 37 °C for 1–4 h. The Cy5 fluorescence intensity in the lower chamber was measured using a multifunctional microplate reader.

Establishment of antibacterial models and antibacterial curves in vitro

The in vitro antibacterial model comprised three parts. On the one hand, we established barrier chambers for HUVECs and IVECs in the same way as described earlier. On the other hand, 1 mL of P. aeruginosa PAO1 suspension was added to each well of a 24-well plate, and the cells were cultured for 72 h until a dense biofilm developed at the air–liquid interface. Finally, the chambers containing HUVECs and IVECs were placed into the well plate containing biofilms, and different drugs were added to the chambers. The chambers were removed after 4-h incubation, and half of the biofilm was subjected to US irradiation (1.5 W/cm², 3 MHz, 50% duty cycle, 1 min). The OD values of the biofilms were recorded using a microplate reader over a period of 0–12 h to plot the antibacterial curves. Further in vitro antibacterial efficacy evaluation experiments were conducted at the 12-h treatment endpoint following administration.

Evaluation of biofilm penetration ability

A direct assessment of liposomal biofilm penetration was conducted using the Transwell model. The upper chamber was loaded with Cy5-labeled liposomes, and the lower chamber was pre-seeded with a GFP-tagged PAO1 biofilm. After a 6-hour incubation, the entire biofilm structure was visualized via Z-stack scanning with a CLSM, using specific filter sets for GFP (503/530 nm) and Cy5 (635/670 nm) to distinguish the spatial distribution of the liposomes within the biofilm matrix.

Staining and quantification of biofilm matrix

The medium was aspirated, and 500 µL of 4% paraformaldehyde was added to each well to fix the cells. This fixation process was allowed to proceed for 15 min. The wells were air-dried after removing paraformaldehyde. Subsequently, 500 µL of a 1% crystal violet solution was added for staining for 15 min. The crystal violet solution was then aspirated, and the wells were photographed under a microscope. The wells were washed three times with PBS until they were colorless. Next, 200 µL of 95% ethanol was added to dissolve crystal violet, and the OD at 595 nm was measured using a microplate reader. Meanwhile, a 200-µL volume of FilmTracer SYPRO Ruby Biofilm Matrix Stain was used to stain the biofilms for 30 min at room temperature, protected from light. Subsequently, the biofilms were rinsed three times with PBS. A CLSM equipped with excitation/emission maxima settings of 450/610 nm was used for visualization. The overall biofilm was examined from various angles, and its thickness was measured from the side profile.

Live/Dead staining and colony counting

Bacterial live/dead staining solution was prepared by mixing 5 µL of DMAO stain, 5 µL of PI stain, and 90 µL of buffer. Then, 10 µL of staining solution was added to 300 µL of each treated bacterial suspension and incubated in the dark for 30 min. The live and dead bacteria in the biofilm were observed by CLSM at wavelengths 503/530 nm and 535/617 nm. Subsequently, the biofilm was processed using an ultrasonic homogenizer for 30 s. Further, 100 µL of the biofilm was diluted 10⁶, 10⁷, and 10⁸ times and then uniformly spread on LB agar plates. After air-drying, the plates were inverted and incubated at 37 °C for 18 h. A camera was used to photograph and record the number of CFUs in each group.

Detection of the leakage of bacterial intracellular substances

The bacterial suspensions were centrifuged at 12,000 rpm for 5 min after various treatments to achieve complete sedimentation of P. aeruginosa PAO1. Then, 100 µL of the supernatant was mixed with a sodium tetraphenylborate solution at a concentration of 1 mg/mL for 10 min. The sodium tetraphenylborate reacted with potassium ions to form a white precipitate, which was quantified by measuring the absorbance at 560 nm. Similarly, the colorless o-nitrophenyl-β-D-galactopyranoside (ONPG) substrate was inoculated with β-galactosidase at 37 °C for 30 min, resulting in the formation of a yellow product detectable by colorimetric assay. The OD was measured at 420 nm.

Scanning electron microscopy

P. aeruginosa PAO1 was inoculated onto silicon wafers and cultured at 37 °C for 72 h to allow biofilm formation. The wafers were carefully removed after 6 h of treatment in different subgroups, washed with PBS to remove planktonic bacteria, and fixed with 2.5% glutaraldehyde solution overnight. Then, they were dehydrated through a graded series of ethanol solutions, desiccated using critical point drying, sputter-coated with gold, and observed for biofilm and bacterial morphology using a cold field emission SEM (Hitachi SU8010).

Immunomodulatory experiments in vitro

RAW 264.7 murine macrophages were seeded in appropriate culture plates and allowed to adhere overnight and then subjected to various treatments. Selected treatment groups were further exposed to US irradiation using a parameter of 1.5 W/cm², 3 MHz, 50% duty cycle for a duration of 1 min. Following treatments, cells were incubated for a specified period before subsequent analysis. To investigate the activation of signaling pathways, protein extracts were harvested from treated cells using RIPA lysis buffer containing protease and phosphatase inhibitors. Equal amounts of protein were separated by SDS-PAGE and transferred onto PVDF membranes. The membranes were blocked and then incubated overnight at 4 °C with primary antibodies against phospho-NF-κB p65, total NF-κB p65, HIF-1α, phospho-STAT1, and total STAT1, followed by incubation with appropriate HRP-conjugated secondary antibodies. The expression of M1-type surface markers (CD80 and CD86) on macrophages was assessed by FCM. For each group, 1 µL of anti-mouse CD80-PE and 1 µL of anti-mouse CD206-APC were added for staining at room temperature in the dark for 30 min. The expression of CD80 and CD206 on RAW264.7 cells was then measured using FCM. Posttreatment, macrophages were placed in the upper chamber of a Transwell with a diameter of 6.5 mm and a pore size of 3.0 μm and co-cultured with the lower chamber containing the P. aeruginosa biofilm for 1 h. Then, the upper chamber was gently removed, washed with PBS, fixed with 4% paraformaldehyde (PFA) for 30 min, and stained with crystal violet for 2 h. The cells remaining inside the upper chamber were wiped away, and macrophage migration was photographed under a microscope and quantitatively analyzed using ImageJ software. The macrophages under different treatments were co-cultured with GFP-labeled P. aeruginosa PAO1 for 2 h to further investigate the altered phagocytic capacity. The culture medium containing 200 µg/mL MEM was then used to replace the suspension for further incubation at 37 °C for 1 h to eliminate extracellular bacteria. After washing and collection, the fluorescence intensity of macrophages in each group was measured using FCM. The cell viability of RAW 264.7 cells exposed to the four liposomal formulations or MEM medium, with or without US, was assessed by the CCK-8 assay.

Targeting ability and biodistribution in vivo

A total of 12 BSSI model mice were randomly divided into 4 groups and administered Cy5-labeled SPCMPL, PCMPL, SCMPL, and CMPL, respectively. They were imaged 1, 2, 4, 6, and 8 h after drug administration using the apparent fluorescence imaging capability of the IVIS spectral imaging system at 640/670 nm. The targeting ability of sulfide-modified liposomes was evaluated by comparing the enrichment time and intensity of the four liposomes in the BSSI region. Moreover, the time point at which liposome enrichment at the BSSI site was most pronounced was used to determine the optimal time for sonication stimulation. The Mettler Sonicator-740 therapeutic US device (Mettler Electronics Corp., CA, USA) was used to irradiate the BSSI region at the time point of greatest liposome enrichment. The specific parameters were set to a sound intensity of 2 W/cm², a frequency of 3 MHz, a duty cycle of 50%, and a duration of 5 min. A control group was set up that received only SPCMPL injections and no US irradiation. The mice were imaged again after 18 h using the IVIS spectral imaging system to monitor the overall distribution of liposomes in the body. Subsequently, they were euthanized and the heart, liver, spleen, lungs, kidneys, and BSSI regions were excised for in vitro fluorescence imaging and weighed to document the distribution of the drug among the groups.

Immunomodulatory experiments in vivo

On the fourth day of the experiment, following three consecutive days of drug administration, BSSI tissues were harvested from the mice to assess macrophage accumulation in vivo. After washing, the tissues were minced into small pieces and incubated with 5 mL of a dermal digestion solution (containing 1 mg/mL collagenase IV, 50 µg/mL DNase, and 10% FBS culture medium) at 37 °C on a shaker at 180 rpm for 90 min. The digested tissue was then passed through a 70-µm mesh to collect the single-cell suspension. Subsequently, red blood cell lysis buffer was used to treat the cells for 5 min at room temperature, followed by resuspension in 1 mL of PBS. After blocking Fc receptors with CD16/32 antibodies, 1 µL each of anti-mouse F4/80-BV421 and anti-mouse CD80-PE and anti-mouse CD206-APC staining solutions were added to each tube and incubated at room temperature in the dark for 15 min. Then, 100 µL of DAPI viability stain was added and further incubated for another 15 min under similar conditions. The cells were analyzed using the FCM after final washing with PBS.

Antibacterial assay in vivo

The BSSI model was established using BALB/c nude mice. Mice with maximum fluorescence intensity exceeding 3000 counts were selected and randomly assigned to the following treatment groups (n = 5): a PBS control group, an MEM group, three control liposome groups (CMPL, SCMPL, and PCMPL), and an SPCMPL group. US irradiation was administered at a setting of 2 W/cm², 3 MHz, and 50% duty cycle for 10 min after 6 h of intravenous injection. Additionally, another SPCMPL group without US irradiation was incorporated to explore the impact of US stimulation within this drug delivery strategy. Also, a CIP + US group, representing another clinical first-line agent, was included to assess efficacy and clinical potential. The mice were treated consecutively on days 1, 2, and 3, with fluorescence area and intensity changes, as well as body weight, recorded on days 1, 3, 5, and 7. On day 7, the mice were euthanized and the BSSIs were dissected after collecting the orbital blood. BSSI tissues were divided into two parts: one part was homogenized for bacterial colony counting, and the other was fixed in 4% PFA for histological analysis using hematoxylin and eosin (HE), Masson, and Giemsa staining. Furthermore, the heart, liver, spleen, lungs, and kidneys were also harvested for subsequent HE staining to assess potential organ damage from the treatment.

Statistical analysis

Quantitative image analysis was conducted using Image-J software (National Institutes of Health, Bethesda, USA). Data were presented as mean ± SD and statistically analyzed using GraphPad Prism 10.0 (GraphPad Software Inc., San Diego, USA). An unpaired two-tailed Student’s t-test was employed for comparisons between two independent groups. For comparisons among three or more independent groups, a one-way analysis of variance (one-way ANOVA) was performed, followed by post-hoc Tukey’s honestly significant difference test for pairwise comparisons. Statistical significance was defined as P < 0.05 for all tests.

Author contributions

Xiao Liu and Zixuan Huang designed and performed most of the experiments and contributed to writing-original draft. Chenlu Hu, Yanan Zhao, Huili Pan, Yinxin Wu, Xia Fang, and Jifan Chen performed some of the experiments, contributed to data curation, formal analysis and validation. Yajing Liu, Guowei Wang and Pintong Huang designed the study, provided the funding, participated in the supervision and coordination of the study, contributed to the review and revision of the manuscript. All authors read and approved the final manuscript.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.