Bioinspired nanoparticles deliver DNase I to attenuate ulcerative colitis by degrading NETs and modulating PTGS2-mediated ER stress

Abstract

Ulcerative colitis (UC) is a chronic inflammatory bowel disease characterized clinically by recurrent abdominal pain and bloody diarrhea, for which there is currently a lack of curative therapeutic strategies. In recent years, neutrophil extracellular traps (NETs) have been identified as playing a key role in the pathogenesis of UC, emerging as a potential therapeutic target. Although deoxyribonuclease I (DNase I) can exert therapeutic effects by degrading the DNA backbone of NETs, its clinical application is limited by inherent drawbacks, including poor in vivo stability, low bioavailability, and inefficient penetration of biological barriers. Herein, this study constructed a biomimetic membrane-based nanodelivery system. Initially, zeolitic imidazolate framework-8 (ZIF-8) was used to encapsulate DNase I, forming ZD nanoparticles to enhance enzymatic stability. Subsequently, these nanoparticles were coated with bacterial outer membrane vesicles (BOMVs), yielding the final BVZD composite system, designed to improve targeting to inflammatory sites. The BVZD system exhibited suitable particle size, good biocompatibility, and a demonstrated ability to specifically accumulate in the inflamed intestinal regions of a UC model. Both in vitro and in vivo experiments confirmed that BVZD effectively degraded NETs, suppressed endoplasmic reticulum stress (ERS) by downregulating the expression of prostaglandin-endoperoxide synthase 2 (PTGS2), and promoted the repair of the intestinal epithelial barrier. These findings highlight the promising therapeutic potential of BVZD for the treatment of UC.

Graphical abstract

BVZD is a nanoparticle constructed by loading DNase I into ZIF-8 and modifying the surface with bacterial outer membrane vesicles (BOMVs). This nanoparticle can target neutrophils at inflammatory sites, degrade NETs, and further modulate the PTGS2-mediated endoplasmic reticulum stress pathway, thereby exerting a therapeutic effect against ulcerative colitis.

Highlights

• •BVZD enhances targeting ability through bacterial outer membrane vesicles.
• •BVZD degrades NETs, thereby alleviating endoplasmic reticulum stress.
• •BVZD degrades NETs to modulate PTGS2-mediated ERS, revealing a novel strategy for ulcerative colitis therapy.

1. Introduction

Ulcerative colitis (UC) is a type of inflammatory bowel disease characterized by non-specific inflammation, with a core pathological mechanism involving abnormal immune activation, leading to persistent inflammatory reactions and ulceration of the intestinal mucosa [1]. Clinically, UC patients often present with symptoms such as diarrhea, abdominal pain, and mucopurulent bloody stools, and may even develop complications including intestinal perforation, massive hemorrhage, and carcinogenesis [2]. In recent years, the global incidence of UC has continued to rise, making it a significant public health concern [3]. Current conventional therapeutics, such as 5-aminosalicylic acid preparations, glucocorticoids, and immunosuppressants, are often associated with considerable side effects. However, the diseased colon resected by surgical treatment is incapable of regeneration. Moreover, severe complications such as defecatory dysfunction, sexual dysfunction, and infertility are prone to occur postoperatively, which significantly impairs patients' quality of life. [4]. Given the limitations of existing treatment strategies, the development of low-toxicity anti-inflammatory agents, immunomodulators, and intestine-targeted delivery systems represents a critical direction for advancing UC therapy.

Recent studies have indicated that the pathogenesis of UC is closely associated with intestinal immune dysregulation, characterized by massive neutrophil infiltration into the inflamed mucosa and the formation of neutrophil extracellular traps (NETs) [5,6]. NETs have been shown to contribute not only to the development of systemic autoimmune diseases but also to exacerbating local inflammatory responses and promoting tissue damage through the release of histones, proteases, and other cytotoxic components [7]. The homeostasis of the endoplasmic reticulum (ER) is essential for maintaining cellular function [8]. Research has demonstrated that in UC, excessive NET formation and the release of their contents directly compromise the integrity of the ER in intestinal epithelial cells, thereby inducing or aggravating endoplasmic reticulum stress (ERS) [9]. This process can ultimately lead to apoptosis or programmed necrosis [10], exacerbating intestinal mucosal barrier dysfunction and inflammatory progression. Deoxyribonuclease I (DNase I), an endonuclease capable of digesting single- or double-stranded DNA, specifically cleaves the DNA backbone of NETs, playing a critical role in regulating NETs stability and clearance. In recent years, DNase I has demonstrated therapeutic potential in various disease models through NETs intervention. For example, Wang et al. [11] reported that DNase I-mediated NETs degradation alleviates chemotherapy-induced peripheral neuropathy, while Chen's group [12] utilized DNase I-functionalized nanoparticles to target and eliminate intravascular NETs, improving outcomes in acute ischemic stroke. Compared with conventional small-molecule drugs, DNase I exhibits higher specificity and bioactivity, showing considerable promise for therapeutic applications. However, its clinical translation is hampered by inherent limitations, including poor in vivo stability, low bioavailability, and inefficient penetration across biological barriers, owing to its large molecular size and complex structure.

Metal-organic frameworks (MOFs) are a class of three-dimensional porous materials formed by the self-assembly of metal ions or clusters and organic bridging ligands [13]. Characterized by an exceptionally high specific surface area and tunable porous structures, MOFs demonstrate a high loading capacity for various therapeutic agents, including small-molecule drugs, proteins, and nucleic acids. Furthermore, their ability to respond to specific microenvironmental stimuli enables controlled drug release, highlighting their significant potential in drug delivery applications [14]. However, the further biomedical translation of MOFs is hindered by limitations such as poor biomembrane penetration, short in vivo circulation half-life, and potential immunogenicity. In recent years, biomimetic nanosystems have become a research hotspot in drug delivery [15]. Bacterial outer membrane vesicles (BOMVs), derived from Gram-negative bacteria such as E. coli, are nano-scale, bilayer lipid membrane vesicles naturally released through outer membrane budding. Their surface is enriched with pathogen-associated molecular patterns, including outer membrane protein A (OmpA) and peptidoglycan, which confer upon BOMVs a natural tropism and accumulation capability towards inflammatory sites [16,17]. Surface modification of MOFs with BOMVs can effectively enhance their targeted delivery efficiency, physiological stability, and biocompatibility, facilitating the construction of a highly promising inflammation-targeted delivery system.

In this study, we designed an inflammation-targeting nanoparticle (BVZD) for UC therapy (Scheme 1). BVZD employs a MOF as the core carrier, which is formed via the self-assembly of Zn²⁺ and 2-methylimidazole to create a porous structure for the encapsulation of DNase I. The resulting composite is further coated with BOMVs derived from msbB-deficient Escherichia coli. This design leverages the msbB-deficient BOMVs, which retain the inflammation-targeting capability mediated by OmpA while exhibiting reduced endotoxin toxicity, enabling efficient delivery to activated neutrophils. ZIF-8 not only provides a loading space for DNase I but also preserves its enzymatic activity, thereby ensuring efficient hydrolysis of the DNA backbone of NETs at the target site. Material characterization confirmed that BVZD possesses a suitable particle size, uniform morphology, and excellent biocompatibility. In vitro experiments demonstrated that BVZD effectively degrades NETs and downregulates PTGS2 expression, thereby alleviating ERS in NCM460 cells. In vivo studies revealed that BVZD specifically accumulates in inflamed colonic regions, releases DNase I to degrade NETs, and suppresses ERS by downregulating PTGS2, leading to a significant reduction in colonic inflammation and tissue damage. Collectively, these findings indicate that the BVZD holds considerable promise as a therapeutic strategy for UC.

Scheme 1.

Schematic diagram of BVZD preparation and its mechanism for treating ulcerative colitis.

2. Materials and methods

2.1. Materials

HL60 cells were purchased from Procell Life Science & Technology Co., Ltd. PLB-985 cells were purchased from Shanghai Yaji Biotechnology Co., Ltd. NCM460 cells were purchased from Wuhan Shang En Biotechnology Co., Ltd. 2-methylimidazole was purchased from Aladdin Biochemical Technology Co., Ltd. Zinc acetate dihydrate was purchased from Sigma Aldrich. Celecoxib (Cel) was purchased from Selleck Chemical LLC. The pcDNA3.1-PTGS2 plasmid were provide from Sangon Biotech. Lipofectamine™ 3000 was purchased from Thermo Fisher Scientific. Cell Counting Kit-8, Annexin V-FITC Apoptosis Detection Kit, Calcein/PI Cell Viability/Cytotoxicity Assay Kit and ER-Tracker Red were purchased from Beyotime Biotechnology. Anti-CitH3, anti-PAD4, anti-MPO, anti-GRP78, anti-ATF4, anti-CHOP, anti-ZO-1, anti-occludin and anti-Claudin-1 were purchased from Abcam Corporation. Anti-OmpA and anti-OmpC were purchased from Thermo Fisher Scientific.

2.2. Cell culture

HL60 cells were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS). To induce differentiation into neutrophil-like cells, HL60 cells were treated with 1.25% dimethyl sulfoxide (DMSO) for 6 days. The differentiated cells were then stimulated with 100 ng/mL phorbol 12-myristate 13-acetate (PMA) for 4 h to induce NET formation. PLB-985 cells were induced to differentiate by treatment with 600 ng/mL all-trans retinoic acid (ATRA) and 1.3% DMSO for 72 h. To trigger the formation of NETs, differentiated PLB-985 cells were stimulated with 30 ng/mL PMA for 4 h. Subsequently, different nanoparticles were introduced to assess their effects on NETs. NCM460 cells were cultured in DMEM/F12 (1:1) medium containing 10% FBS. To evaluate the impact of NET degradation on NCM460 cells, the supernatant from the aforementioned HL60 cell cultures was collected, mixed with an equal volume of serum-free medium, and applied to NCM460 cells for co-culture. All cell cultures were maintained at 37 °C in a humidified atmosphere containing 5% CO₂.

2.3. Preparation and characterization of BOMVs

BOMVs were isolated from msbB-deficient E. coli cultured to the logarithmic growth phase. Bacterial cells were removed by centrifugation at 5000 × g for 30 min. The supernatant was sequentially filtered through 0.45 μm filters to eliminate residual bacterial debris and impurities. The filtrate was then concentrated using a 100 kDa ultrafiltration device to enrich the detoxified BOMVs. The concentrated solution was subjected to ultracentrifugation at 130,000 × g for 2.5 h. The resulting pellet was resuspended in sterile PBS and filtered through a 0.22 μm membrane to remove residual contaminants, yielding purified BOMVs. To assess the biosafety of the obtained BOMVs, endotoxin content was measured using one-step chromogenic Limulus amebocyte lysate assay, and their impact on cell viability was evaluated via CCK 8 assay. Furthermore, Western blot (WB) analysis was performed to detect the expression of characteristic functional proteins, such as OmpA, to verify the structural integrity and functional activity of the BOMVs.

Endotoxin quantification: 20 μL of endotoxin-free water, endotoxin standard solution, or sample to be tested was pipetted into an endotoxin-free centrifuge tube. Subsequently, 10 μL of endotoxin detection reagent solution was added and incubated at 37 °C for 9 min in the dark. Following incubation, 20 μL of chromogenic substrate solution was added and incubated at 37 °C for 6 min. Finally, 40 μL of chromogenic stop solution was added and allowed to stand for 5 min. Absorbance was then measured at 405 nm using a microplate reader.

2.4. Synthesis and characterization of BVZD

Solution A was prepared by dissolving 656.8 mg of 2-methylimidazole in 50 mL of ddH₂O, followed by the addition of DNase I under stirring at room temperature for 10 min. Separately, Solution B was obtained by dissolving 74.35 mg of zinc acetate dihydrate in 25 mL of ddH2O. Solution B was then added dropwise to Solution A, and the reaction was allowed to proceed for an additional 10 min. The resulting mixture was centrifuged at 12,000 rpm for 20 min, and the precipitate was washed twice with ddH₂O to yield ZD nanoparticles. BVZD nanoparticles were prepared by co-extruding BOMVs and ZD at a 1:1 mass ratio through a 200 nm polycarbonate membrane for 15–20 cycles. The resulting nanoparticles were characterized for hydrodynamic diameter and zeta potential using dynamic light scattering, and their morphology was examined by transmission electron microscopy (TEM). Fourier transform infrared spectroscopy (FTIR) and X-ray diffraction (XRD) were employed to analyze the chemical composition and crystalline structure of the nanoparticles, respectively.

2.5. Coomassie brilliant blue staining assay

The protein concentration of the samples was quantified using the BCA assay. Proteins were then separated by SDS-PAGE. Following electrophoresis, the gel was stained with Coomassie Brilliant Blue staining solution for 30 min. The gel was then destained by covering it with ddH₂O for 10 min, after which the ddH₂O was replaced. This short washing step was repeated twice. Subsequently, the ddH₂O was replaced every hour for a total of three washes. Finally, the gel was left to destain in fresh ddH₂O on a shaker overnight.

2.6. Over-expression of PTGS2

NCM460 cells at approximately 70% confluence were transfected with the pcDNA3.1-PTGS2 plasmid using Lipofectamine™ 3000. The medium was replaced after 24 h, and the cells were cultured for an additional 24 h. The overexpression efficiency of PTGS2 was subsequently validated by qPCR and WB analysis.

2.7. Western bolt

Total protein was extracted from cells and tissues using RIPA lysis buffer, and its concentration was determined by the bicinchoninic acid (BCA) assay. Protein samples were separated by molecular weight via electrophoresis on polyacrylamide gels and subsequently transferred onto polyvinylidene fluoride (PVDF) membranes. The membranes were blocked with 5% skim milk for 2 h at room temperature and then incubated overnight at 4 °C with primary antibodies against specific targets, including CitH3, PAD4, MPO, NE, GRP78, ATF4, CHOP, OmpA, OmpC, and PTGS2. Following incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies at room temperature, protein bands were visualized using a chemiluminescence imaging system and quantified with ImageJ software.

2.8. Enzyme-linked immunosorbent assay (ELISA)

To evaluate the degradation of NETs and the expression levels of inflammatory factors, the concentrations of CitH3, MPO, PAD4, NE, TNF-α, IL-1β, and IL-6 in cell culture supernatants and serum samples were measured using ELISA. The procedure was performed as follows: standards and samples were added to the pre-coated 96-well plates, followed by the sequential addition of a biotinylated detection antibody working solution, HRP-conjugated working solution, and a substrate solution. After each incubation step, the plates were thoroughly washed to remove unbound components. Following the addition of the stop solution, the absorbance of each well was immediately measured at a wavelength of 450 nm using a microplate reader. The concentration of each analyte was calculated based on the corresponding standard curve.

2.9. CCK 8 assay

HL60 cells were seeded in a 96-well plate at a density of 5 × 10³ cells per well and incubated with different nanoparticles. Subsequently, 10% CCK-8 solution was added to each well, followed by further incubation for 1.5 h. The absorbance of each well was measured at a wavelength of 450 nm using a microplate reader. Cell viability was calculated according to the following formula:

$$\text{Cell}\text{viability}(\%)=\left({\mathrm{A}}_{\text{Sample}}-{\mathrm{A}}_{\text{Blank}}\right)/\left({\mathrm{A}}_{\text{Control}}-{\mathrm{A}}_{\text{Blank}}\right)\times 100\%$$ (1)

2.10. Cell cytotoxicity assay

After treatment with different nanoparticles, HL60 cells were washed with PBS and incubated with a Calcein-AM/PI working solution at 37 °C for 30 min in the dark. Following incubation, images were acquired using a fluorescence microscope. Viable cells exhibit green fluorescence due to the conversion of Calcein-AM to calcein by intracellular esterases, whereas dead cells with compromised membranes are labeled with PI, emitting red fluorescence.

2.11. Cellular uptake

To evaluate cellular uptake of the nanoparticles, neutrophil-like cells were incubated with DiO-labeled ZD and BVZD nanoparticles, respectively. Following incubation, the cells were fixed in 4% paraformaldehyde and stained with DAPI. Fluorescence microscopy was subsequently performed to compare the intensity of DiO fluorescence within cells across groups and to analyze nanoparticle internalization.

2.12. Flow cytometry

To assess the effect of nanoparticles on cell apoptosis, cells subjected to different treatments were collected and washed twice with PBS to remove residual culture medium. Subsequently, according to the instructions of the Annexin V-FITC Apoptosis Detection Kit, the cell pellet was resuspended in a working solution containing FITC-labeled Annexin V and PI, followed by incubation at room temperature for 10-15 min in the dark. After incubation, an appropriate volume of binding buffer was added to resuspend the cells, which were then immediately analyzed by flow cytometry.

2.13. RT-qPCR

Total RNA was extracted from cells using TRIzol reagent and subsequently reverse-transcribed into cDNA. RT-qPCR amplification was performed using a premixed SYBR Green fluorescence detection system, with the reaction program set according to the manufacturer's instructions. The mRNA expression level of GAPDH in the same sample was used as an internal reference to normalize target gene expression. The specific primer sequences employed in this study are listed in Table 1.

Table 1.

List of primers.

Gene	Orientation	Sequence (5′- 3′)
GRP78	Forward	CCCAGATGAAGCTGTAGCGT
	Reverse	TGGTCATGACACCTCCCACA
ATF4	Forward	TGTGGATGGGTTGGTCAGTC
	Reverse	CCCAACAGGGCATCCAAGTC
CHOP	Forward	GGAGCTGGAAGCCTGGTATG
	Reverse	AAGCCAGAGAAGCAGGGTCA
GAPDH	Forward	TGGAGTCCACTGGCGTCTT
	Reverse	TGCAGGAGGCATTGCTGAT

2.14. Immunofluorescence analysis

Cell and tissue samples collected after treatment with different nanoparticles were fixed in 4% paraformaldehyde. After washing with PBS, the samples were permeabilized with 0.1% Triton X-100. Subsequently, nonspecific antigen sites were blocked by incubating with 10% goat serum at room temperature for 30-60 min. Following removal of the blocking solution, the samples were incubated with corresponding primary antibodies at 4 °C overnight. After PBS washes, fluorescently labeled secondary antibodies were applied and incubated at room temperature for 30-60 min in the dark. Finally, nuclei were stained with DAPI. Images were acquired and observed using a fluorescence microscope.

2.15. In vivo experimentation

To evaluate the in vivo targeting capability of the nanoparticles, the UC model was established by administering 4% dextran sulfate sodium (DSS) in drinking water to C57BL/6 mice continuously for 7 days. The mice were then randomly divided into two groups (n = 3): the ZD group and the BVZD group. Each group received a tail vein injection of the corresponding fluorescently labeled nanoparticles. In vivo fluorescence imaging was performed at 6, 12, and 24 h post-injection. After 24 h, the mice were euthanized, and the colon tissues were excised for fluorescence imaging to compare the accumulation of the two nanoparticle formulations in the inflamed colon.

To evaluate the therapeutic efficacy of BVZD in UC, mice were randomly assigned to 7 groups (n = 6): control, DSS, BOMVs, ZIF-8, DNase I, ZD, and BVZD groups. Except for the control group, UC was induced in all other groups by administering 4% DSS in drinking water. The respective nanoparticles were administered via tail vein injection every two days (DNase I = 250 U) [18], while the control group received an equivalent volume of normal saline. The treatment continued for two weeks. Upon completion of the experiment, euthanasia was performed, and colon tissues were collected for length measurement and photographic documentation. Subsequently, the colon tissues were fixed in 4% paraformaldehyde for subsequent histological and pathological analyses, including HE staining, immunofluorescence, and immunohistochemistry. Histopathological damage was evaluated and scored based on the extent of epithelial injury and inflammatory cell infiltration. The epithelial damage score was assigned as: 0 (none), 1 (minimal goblet cell loss), 2 (extensive goblet cell loss), 3 (minimal crypt loss with extensive goblet cell loss), 4 (extensive crypt loss). The infiltration score was defined as: 0 (none), 1 (infiltrate around crypt bases), 2 (infiltrate reaching muscularis mucosa), 3 (extensive infiltrate in muscularis mucosa with edema), 4 (infiltration extending into submucosa) [18,19]. Histopathological scoring was conducted by pathologists blinded to the experimental groups. The animal experiments were approved by the Animal Ethics Committee of Guangzhou Seyotin Biotechnology Co., Ltd. (SYT2025053)

2.16. Statistical analysis

All experimental data were analyzed and plotted using GraphPad Prism 9.5. Data are presented as the mean ± SD. Differences between groups were assessed by one-way analysis of variance (ANOVA). Statistical significance was defined as *p < 0.05, **p < 0.01, and ***p < 0.001. All experiments were performed with at least three independent replicates.

3. Results and discussion

3.1. ERS mediated by NETs in UC

Studies have demonstrated that neutrophils are activated during inflammatory responses and recruited to lesion sites, where they release NETs to eliminate pathogens [20,21]. However, under pathological conditions, excessive NETs formation or impaired clearance can exacerbate tissue injury and compromise mucosal barrier integrity [22]. As an endonuclease that specifically degrades the DNA backbone within NETs, DNase I represents a potential therapeutic strategy for intervening in this process. To elucidate the role of NETs in UC and evaluate the therapeutic potential of DNase I, we established a murine UC model using DSS and systematically assessed the interventional effects. Mice exposed to DSS exhibited a progressive decrease in body weight, accompanied by a significant elevation in the disease activity index (DAI) score. Following DNase I treatment, the mice showed a gradual recovery of body weight, with the DAI score displaying a marked downward trend (Fig. 1A and B). Macroscopic examination revealed a significant shortening of colon length in DSS group compared to control group, whereas DNase I treatment restored colon length to near-normal levels (Fig. 1D and E). HE staining further showed that DSS group had disrupted colonic mucosal structure and severe damage to epithelial integrity, while DNase I treatment significantly reduced the pathological damage and promoted epithelial structure repair (Fig. 1C–F). To confirm NETs involvement, we examined the expression of key NETs markers. Immunofluorescence staining showed a substantial increase in CitH3 expression of DSS group, indicating abundant NETs formation (Fig. 1G and H). This effect was reversed by DNase I treatment, as evidenced by significantly reduced CitH3 expression. WB analysis further demonstrated that DSS induction markedly upregulated the expression of CitH3, PAD4, and MPO, all of which were effectively suppressed by DNase I (Fig. 1J). These findings collectively indicate that UC pathogenesis involves aberrant NETs accumulation, which can be effectively degraded by DNase I. Further immunofluorescence analysis revealed a significant increase in the expression of the ERS marker GRP78 in DSS group, which was inhibited following DNase I treatment (Fig. 1G–I). Correspondingly, WB results showed elevated protein levels of ATF4 and CHOP in DSS group, both of which were downregulated in DNase I group (Fig. 1J). These results suggest that NETs accumulation may contribute to the activation of ERS in UC, and that clearance of NETs by DNase I can alleviate ERS.

Fig. 1.

Evaluation of DNase I in vivo. (A) The weight changes of mice(n = 6); (B) Disease activity index in control, DSS, and DNase I groups (n = 6); (C) The pathological score results of the control, DSS, and DNase I groups (n = 6); (D) Digital images of cecal-colon tissues in control, DSS, and DNase I groups; (E) The length of cecal-colon tissues in control, DSS, and DNase I groups (n = 6); (F) Representative images of HE staining in control, DSS, and DNase I groups; (G) Representative images of CitH3 immunohistochemistry and GRP78 immunofluorescence in control, DSS, and DNase I groups; (H) Quantitative analysis of CitH3 expression (n = 3); (I) Quantitative analysis of GRP78 expression (n = 3); (J) Western blot bands of protein expression in different groups and their quantitative analysis (n = 3). Data are presented as the mean ± SD. *p < 0.05, **p < 0.01, and ***p < 0.001, ns, no significance.

3.2. Characterization of BVZD

According to the workflow illustrated in Fig. 2A, ZIF-8 was first constructed via the self-assembly reaction between Zn²⁺ and 2-methylimidazole, with DNase I encapsulated within their interior to form ZIF-8@DNase I (ZD) nanoparticles. Subsequently, BOMVs were coated onto the surface of ZD through a 200 nm polycarbonate membrane, resulting in the final fabrication of BOMVs@ZIF-8@DNase I (BVZD) nanoparticles. To evaluate the physicochemical properties of BVZD, TEM was first employed to observe its morphology. The results showed that BOMVs exhibited a spherical vesicle-like structure, ZD displayed a regular rhombic dodecahedral morphology, and BVZD presented a typical core-shell architecture with a distinct membrane coating clearly visible on the surface. DLS measurements showed that the hydrodynamic diameters of BOMVs and ZD were approximately 100 nm, while that of BVZD was about 120 nm. Their corresponding zeta potentials were −10.39 mV, 15.74 mV, and −2.35 mV, respectively (Fig. 2B–D). Then, BVZD exhibited minimal changes in hydrodynamic diameter across different physiological buffers, indicating its excellent colloidal and serum stability (Fig. S1A). The FTIR spectrum of ZIF-8 exhibited characteristic peaks at 1582.90 cm⁻¹ (C=N) and 416.36 cm⁻¹ (Zn–N), confirming the successful formation of the framework. After synthesizing ZD, a shoulder peak appeared at 1660.44 cm⁻¹, attributed to the overlap of the amide I band (C=O) of the enzyme with the C=N vibration of ZIF-8 (Fig. S1B). Furthermore, XRD analysis revealed sharp diffraction peaks at 7.16°, 10.23°, and 12.15°, indicating the successful synthesis of highly crystalline ZIF-8. The XRD pattern of ZD retained the characteristic diffraction peaks of ZIF-8 without significant peak shifts (Fig. S1C). Then, the encapsulation efficiency and drug loading efficiency of DNase I in ZIF-8 were 79.68% and 8.02%, respectively.

Fig. 2.

Characterization of BVZD nanoparticles. (A) Schematic diagram of the preparation of BVZD; (B) Representative TEM images of BOMVs, ZD, and BVZD; (C) Particle size of BOMVs, ZD, and BVZD; (D) Zeta potential of BOMVs, ZD, and BVZD (n = 3); (E) WB analysis of the protein expression of OmpA and OmpC; (F) Quantification of endotoxin levels in wild-type and msbB knockout strains (n = 3); (G) CCK8 analysis of BOMVs on HL60 and NCM460 (n = 3). Data are presented as the mean ± SD. *p < 0.05, **p < 0.01, and ***p < 0.001, ns, no significance.

WB analysis further confirmed that BVZD expressed OmpA and OmpC, the characteristic outer membrane proteins of BOMVs (Fig. 2E). Concurrently, Coomassie brilliant blue staining revealed clear protein bands in both BOMVs and BVZD, indicating effective retention and purification of the protein components during the construction of the delivery system (Fig. S1D). These findings collectively indicate that BVZD possesses favorable structural uniformity and suitable nanoscale dimensions, with BOMVs successfully modified onto the surface of ZD particles. To verify the biosafety of BOMVs derived from msbB-deficient Escherichia coli, one-step chromogenic Limulus amebocyte lysate assay was performed to detect their endotoxin content (Fig. 2F). The results showed that the endotoxin content of msbB-deficient strains was significantly lower compared to that of wild-type strains. CCK 8 further demonstrated that BOMVs exerted no significant toxic effects on cells (Fig. 2G). Subsequently, agarose gel electrophoresis analysis confirmed that DNase I released from the nanocarrier retained strong DNA hydrolytic activity (Fig. S1E). These results collectively support the feasibility and biosafety of utilizing msbB-deficient BOMVs as a modification material. Furthermore, the drug release assay results demonstrated that the cumulative release of DNase I from BVZD at pH 5.4 (∼60%) was significantly higher than that at pH 7.4 (∼20%) (Fig. S1F). This finding confirms that BVZD possesses an acidic environment-responsive drug release property, enabling the controlled release of DNase I in vivo and thereby providing a guarantee for the exertion of its biological functions.

3.3. Biocompatibility of BVZD

The biosafety of BVZD was systematically evaluated through in vitro cytotoxicity assays, in vivo hematological parameters, and histopathological examination of major organs. CCK 8 assay revealed no significant reduction in cell viability following BVZD treatment, with values remaining above 90% compared to control group, indicating low cytotoxicity (Fig. 3A). The apoptosis assay did not find any induction of apoptosis by BVZD (Fig. 3B). Additionally, Calcein-AM/PI double staining further confirmed that the majority of cells exhibited green fluorescence (live cells), while red fluorescence (dead cells) was scarcely detected, suggesting minimal damage induced by BVZD (Fig. 3C). In vivo, hematological analysis showed no statistically significant differences between BVZD and control group, with all parameters within normal physiological ranges (Fig. 3D). In addition, blood biochemical analysis revealed that mice in the BVZD group exhibited normal liver and kidney function (Fig. S2). Histopathological evaluation of the heart, liver, spleen, lungs, and kidneys via HE staining revealed well-preserved tissue architecture without noticeable pathological alterations (Fig. 3E). Taken together, these results demonstrate that BVZD possesses favorable biocompatibility.

Fig. 3.

Biosafety analysis. (A) CCK8 analysis in each group (n = 6); (B) Apoptosis analysis and qualification in each group (n = 3); (C) Live/dead staining and qualification of HL60 cells (n = 3); (D) Blood routine testing in each group (n = 3); (E) HE staining analysis in each group. Data are presented as the mean ± SD. *p < 0.05, **p < 0.01, and ***p < 0.001, ns, no significance.

3.4. Targeting ability of BVZD

Studies have shown that BOMVs, as natural nanovesicles derived from Gram-negative bacteria, are enriched with characteristic components such as OmpA protein and adhesins on their surface [23]. These components can bind to specific receptors on the neutrophil membrane, mediating precise recognition between BOMVs and neutrophils [24]. As key effector cells in the innate immune response, neutrophils are rapidly activated during the early stages of inflammation and are recruited in large numbers to inflammatory or infected sites via chemotaxis [25]. Leveraging their surface ligands, BOMVs can efficiently bind to neutrophils that have migrated to inflammatory regions, thereby achieving targeted delivery to sites of inflammation [26].

To evaluate the targeting capability of BOMVs, fluorescently labeled ZD and BVZD nanoparticles were incubated with neutrophil-like cells, and their cellular uptake was observed by fluorescence microscopy. The results indicated a significantly higher uptake of BVZD compared to ZD nanoparticles. Furthermore, the addition of a TLR2 inhibitor significantly reduced the cellular uptake of BVZD, indicating that TLR2-mediated recognition of OmpA plays a critical role in the internalization process of BVZD (Fig. 4A and B). Then, ZD and BVZD were administered via tail vein injection to a mouse model of UC induced by DSS. In vivo imaging revealed that at 6 h post-injection, the fluorescence intensity in the colon region of BVZD group reached its peak and was significantly stronger than that in ZD group. By 24 h, a clear fluorescent signal remained detectable in BVZD group, whereas the signal in ZD group had markedly weakened (Fig. 4C and D). Ex vivo imaging of isolated organs further confirmed significantly greater accumulation of BVZD in inflamed colon tissue than in ZD (Fig. 4E and F). In addition, results from colon tissue sections revealed that DiO-labeled BVZD significantly accumulated in the inflamed regions of the colon and exhibited a high degree of co-localization with the neutrophil marker Ly6G in these areas (Fig. S3). These consistent findings demonstrate that surface modification with BOMVs substantially enhances targeted accumulation and prolongs nanoparticle retention at intestinal inflammatory sites.

Fig. 4.

Targeting ability of BVZD nanoparticles. (A-B) Representative image and qualification of cell uptake (n = 3); (C-D) In vivo fluorescence imaging and qualification of ZD and BVZD at 6, 12, and 24 h (n = 3); (E-F) In vitro fluorescence images and qualification of colons at 24 h (n = 3). Data are presented as the mean ± SD. *p < 0.05, **p < 0.01, and ***p < 0.001, ns, no significance.

3.5. Degradation of NETs by BVZD in HL60 cells

During the UC, compromised intestinal mucosal barrier function leads to dysregulation of the local immune microenvironment. This results in substantial neutrophil recruitment to sites of intestinal inflammation, mediated by chemotactic signals. Infiltrated neutrophils become activated upon exposure to inflammatory stimuli, leading to the release of NETs, which are composed of DNA, histones, NE, MPO, PAD4, and other components. Excessive NET deposition exacerbates mucosal inflammation and disrupts normal tissue repair processes [27]. Du et al. developed genetically engineered cellular nanovesicles as a targeted DNase I delivery system to clear NETs in acute lung injury [28]. Similarly, Zhu et al. constructed C-176 loaded Ce DNase nanoparticles that synergistically suppress the cGAS–STING pathway for the treatment of ischemic stroke [29]. These studies provide important evidence supporting the therapeutic potential of DNase I delivery in disease intervention.

To evaluate the ability of BVZD to degrade NETs at the cellular level, HL-60 cells were first differentiated into neutrophil-like cells using DMSO, followed by stimulation with PMA to establish an in vitro NETs model (Fig. 5A). During NET formation, neutrophil membrane integrity is compromised, and intracellular organelles such as mitochondria and lysosomes undergo structural disintegration, ultimately leading to cell death [30]. CCK 8 assay results showed that cell viability was significantly reduced in PMA group, whereas BVZD treatment partially restored viability, suggesting that PMA successfully induced NET formation and associated cytotoxicity, and that BVZD may alleviate this toxic effect by degrading NETs (Fig. S4A). To further validate the successful establishment of the NETs model and assess the interventional effects, immunofluorescence co-staining for CitH3 and DNA was performed on neutrophil-like cells. The results demonstrated that NET formation was significantly induced in the PMA group, whereas the ZIF-8 treatment group showed no significant difference compared to the PMA group. In contrast, the BVZD group exhibited the most pronounced inhibitory effect on NET formation. (Fig. 5B). ELISA further demonstrated that PMA stimulation significantly increased the levels of MPO, PAD4, and NE, and these increases were effectively reversed by both ZD and BVZD treatments (Fig. 5C). Consistent with these findings, WB analysis yielded corroborative results. In summary, an in vitro NETs model was successfully established (Fig. 5D). In addition, in PLB-985 cells, WB analysis demonstrated that the protein expression levels of CitH3, MPO, NE, and PAD4 were significantly upregulated in PMA group, while BVZD treatment markedly downregulated the expression of these proteins (Fig. S6A). The results confirm that both ZD and BVZD effectively promote NET degradation, with BVZD showing the highest efficacy. This enhanced effect is likely attributable to the BOMV coating, which promotes cellular internalization of the nanoparticles, thereby improving their intracellular bioavailability.

Fig. 5.

Degradation of NETs by BVZD in HL60 cells. (A) Schematic diagram of HL-60 cells differentiating into neutrophil-like cells; (B) Representative fluorescence images of CitH3/DNA; (C) The expression of MPO, NE, and PAD4 (n = 3); (D) Representative protein bands and qualification of CitH3, MPO, NE, and PAD4 (n = 3). Data are presented as the mean ± SD. *p < 0.05, **p < 0.01, and ***p < 0.001, ns, no significance.

3.6. BVZD alleviates ERS in NCM460 cells by degrading NETs

The aberrant accumulation of NETs triggers a cascade of pathological responses in host cells, disrupting cellular homeostasis and disturbing the balance of the tissue microenvironment [31]. For instance, the abundant bioactive components released from NETs can directly compromise cellular barriers, induce endoplasmic reticulum dysfunction, and subsequently activate stress responses, ultimately leading to programmed cell death in target cells [32]. Furthermore, the release of endogenous damage-associated molecular patterns (DAMPs) from lysed cells recruits additional inflammatory cells, amplifying the inflammatory cascade and exacerbating local tissue damage and functional impairment [33]. Bai et al. developed inflammation-responsive cell-membrane-camouflaged nanoparticles that modulate ERS and oxidative stress to treat liver fibrosis [34]. This finding suggests that biomimetic membrane-engineered nano-delivery systems hold promise for inflammatory diseases such as UC, as they may enable synergistic therapy by concurrently targeting NETs and their downstream ERS pathways.

To investigate the impact of NETs on colonic epithelial cells, culture supernatants from HL60 cells subjected to different treatments were collected and co-cultured with NCM460 cells for 24 h (Fig. 6A). CCK 8 assay results demonstrated that the viability of NCM460 cells incubated with NETs-containing supernatant was significantly reduced, whereas treatment with ZD or BVZD led to a notable recovery of cell viability (Fig. S4B). Furthermore, flow cytometric analysis of apoptosis revealed that NETs could induce apoptosis in NCM460 cells, while both ZD and BVZD treatments reduced the apoptosis rate, with the BVZD group exhibiting the most pronounced inhibitory effect (Fig. 6B). Based on previous findings of NETs accumulation associated with ERS in UC tissues, this study further investigates the effect of NETs on the endoplasmic reticulum in colonic epithelial cells at the cellular level. By ER-Tracker staining, it was observed that the fluorescence intensity of the endoplasmic reticulum was significantly enhanced in the NETs-treated group, indicating endoplasmic reticulum dilation (Fig. S5). This phenomenon was reversed after ZD and BVZD treatment. Meanwhile, the experimental results in PLB-985 cells were also consistent with those in HL60 cells (Fig. S6B). GRP78, a key molecular marker of ERS, is significantly upregulated under stress conditions. Immunofluorescence results showed that supernatants collected after PMA stimulation of neutrophil-like cells induced increased GRP78 expression in NCM460 cells, whereas treatment with ZD and BVZD downregulated GRP78 expression, with BVZD exhibiting the most pronounced effect (Fig. 6C). Further analysis by WB and qPCR revealed that the expression of ERS-related molecules GRP78, ATF4, and CHOP was upregulated at both protein and mRNA levels upon NETs induction, whereas ZD and BVZD treatment effectively inhibited this trend (Fig. 6D and E). Concurrently, WB analysis demonstrated that NETs induced phosphorylation of IRE1α and PERK, thereby activating the ERS. In contrast, treatment with BVZD significantly suppressed the phosphorylation levels of these proteins (Fig. 6D). These results indicate that NETs can induce ERS in colonic epithelial cells, leading to cell damage and death; meanwhile, BVZD significantly alleviates endoplasmic reticulum stress by effectively degrading NETs, thereby exerting a cytoprotective effect.

Fig. 6.

BVZD alleviates ERS in NCM460 cells by degrading NETs. (A) Schematic illustration of the interaction between degraded NETs and NCM460 cells; (B) Flow cytometry analysis of apoptosis in different groups and their quantitative results (n = 3); (C) Representative fluorescence images and quantitative analysis of GRP78 (n = 3); (D) Representative protein bands and qualification of GRP78, ATF4, CHOP, p-IRE1α, IRE1α, p-PERK, and PERK(n = 3); (E) qPCR analysis of GRP78, ATF4, and CHOP mRNA levels (n = 3). Data are presented as the mean ± SD. *p < 0.05, **p < 0.01, and ***p < 0.001, ns, no significance.

3.7. BVZD regulates PTGS2-mediated ERS in NCM460 cells through degradation of NETs

To elucidate the molecular mechanism by which BVZD suppresses ERS through degrading NETs, RNA sequencing transcriptome analysis was performed on control and NETs-treated NCM460 cells. Heatmap analysis revealed a significant upregulation of PTGS2 gene expression in the NETs-treated group (Fig. 7A, S7). PTGS2, a key rate-limiting enzyme in the arachidonic acid metabolism pathway, exhibits marked stress-inducible characteristics [35]. And its product PGE₂ plays a crucial role in sustaining chronic inflammation and modulating cellular stress responses [36,37]. Although other differentially expressed genes (e.g., MMP1, TP63) may contribute to disease progression at different levels, the central position of PTGS2 in integrating inflammatory signals and cellular stress responses establishes it as the focal mechanistic target for in-depth investigation in this study. KEGG pathway enrichment analysis indicated that the differentially expressed genes (DEGs) were primarily associated with inflammatory response and cellular metabolism-related pathways (Fig. 7B). Recent studies have shown that PTGS2 can be rapidly activated under various stimuli, such as inflammation, oxidative stress, and cellular damage, and is involved in regulating inflammatory responses, cell survival, and tissue repair processes [38,39]. WB and immunofluorescence results further confirmed that NETs induced elevated PTGS2 protein expression in NCM460 cells, whereas treatment with different nanoparticles reduced expression, with the BVZD group showing the most pronounced inhibitory effect (Fig. 7C and D). Previous studies have demonstrated a close association between CitH3 and the regulation of PTGS2 expression [40,41]. Meanwhile, DNA fragments within NETs can activate downstream inflammatory pathways through the TLR9 receptor, thereby driving PTGS2 expression [42]. Additionally, MPO and NE in NETs may exacerbate ROS to amplify inflammatory signals, consequently cooperating to induce PTGS2 [41,43,44]. Based on the above evidence, we hypothesize that histones and DNA are the key molecular components in NETs that contribute to the upregulation of PTGS2 expression in NCM460 cells.

Fig. 7.

Mechanism of BVZD degradation of NETs in regulating ERS. (A) Heat map of DEGs; (B) KEGG pathway; (C) WB analysis of PTGS2 protein levels (n = 3); (D) Immunofluorescence analysis and quantification of PTGS2 (n = 3). Data are presented as the mean ± SD. *p < 0.05, **p < 0.01, and ***p < 0.001, ns, no significance.

Both WB and qPCR results demonstrated that the inhibition of PTGS2 could suppress ERS (Fig. S8A and B). To elucidate the role of PTGS2 in NETs-induced ERS, this study employed PTGS2 overexpression (OE-PTGS2) to examine its impact on ERS-related markers. Both qPCR and WB results confirmed the successful establishment of the OE-PTGS2 model (Fig. S9A and B). Then, CCK 8 assay results indicated that PTGS2 overexpression in NCM460 cells partially reversed the BVZD-mediated reduction in cytotoxicity (Fig. S9C). Flow cytometric analysis further revealed a significant increase in the apoptosis rate in BVZD + OE-PTGS2 group compared to BVZD + OE-NC group (Fig. 8A). ER-Tracker staining demonstrated enhanced fluorescence intensity upon PTGS2 overexpression, suggesting aggravated ERS (Fig. 8B). Subsequent WB and qPCR analyses showed that protein and mRNA levels of GRP78, ATF4, and CHOP were upregulated in BVZD + OE-PTGS2 group relative to BVZD + OE-NC group (Fig. 8C and D). These findings collectively demonstrate that PTGS2 overexpression attenuates the inhibitory effect of BVZD on ERS, indicating that BVZD likely modulates NETs-induced ERS through regulation of PTGS2 expression. It has been shown that ROS generated via PTGS2 metabolism can directly impair endoplasmic reticulum folding capacity [45]. Concurrently, as the enzymatic products of PTGS2, PGE₂ can disrupt intracellular calcium homeostasis via EP receptors [46]. The combined effect of ROS and calcium dysregulation collectively activates the PERK/eIF2α and IRE1α/XBP1s unfolded protein response pathways, ultimately leading to the ERS [47,48].

Fig. 8.

BVZD attenuates PTGS2-mediated ERS by degrading NETs. (A) Flow cytometry analysis of apoptosis and quantitative results (n = 3); (B) Representative fluorescence images and quantitative analysis of ER-Tracker (n = 3); (C) WB analysis of GRP78, ATF4, and CHOP protein levels (n = 3); (D) qPCR analysis of GRP78, ATF4, and CHOP mRNA levels (n = 3). Data are presented as the mean ± SD. *p < 0.05, **p < 0.01, and ***p < 0.001, ns, no significance.

3.8. In vivo treatment efficacy against UC of BVZD

C57BL/6 mice were randomly divided into 7 groups (n = 6): control, DSS, BOMVs, ZIF-8, DNase I, ZD, and BVZD. To evaluate the therapeutic effect of BVZD on UC, the experimental procedure was conducted as outlined in Fig. 9A. Briefly, the UC model was established by administering 4% DSS in drinking water for 7 consecutive days, followed by tail vein injection of the respective nanoparticles. During modeling, mice in DSS group exhibited a progressive decrease in body weight and an elevation in the DAI score. In contrast, mice treated with DNase I, ZD and BVZD exhibited a gradual recovery in body weight and a progressive decrease in DAI scores, indicating alleviation of colitis symptoms (Fig. 9B and C). Macroscopic analysis of colon morphology revealed that DSS group had significantly shorter colon lengths compared to control group, while all treatment groups showed varying degrees of recovery, with BVZD group displaying colon lengths nearly comparable to those of control group (Fig. 9D and E). HE staining further demonstrated severe mucosal architecture disruption and impaired epithelial integrity in DSS group, whereas nanoparticle treatments ameliorated these pathological changes, with BVZD group exhibiting the most pronounced improvement (Fig. 9F–S10). ELISA results indicated that BVZD treatment significantly reduced the expression levels of pro-inflammatory cytokines TNF-α, IL-1β, and IL-6 in colon tissues, suggesting a potent anti-inflammatory effect (Fig. S11A). Moreover, immunofluorescence staining showed marked downregulation of the tight junction proteins ZO-1, Occludin, and Claudin-1 in DSS group, indicating impaired intestinal epithelial barrier function (Fig. 9G–S11B). In contrast, BVZD treatment upregulated the expression of these proteins, demonstrating its role in restoring intestinal epithelial barrier integrity.

Fig. 9.

Evaluation of BVZD in vivo. (A) Schematic diagram of the construction and treatment of UC mice; (B) The weight changes of mice in control, DSS, BOMVs, ZIF-8, DNase I, ZD, and BVZD groups (n = 6); (C) Disease activity index in control, DSS, BOMVs, ZIF-8, DNase I, ZD, and BVZD groups (n = 6); (D) The length of cecal-colon tissues in control, DSS, BOMVs, ZIF-8, DNase I, ZD, and BVZD groups (n = 6); (E) Digital images of cecal-colon tissues in different groups; (F) Representative images of HE staining in different groups; (G) Representative images of ZO-1, Occludin, and Claudin-1. Data are presented as the mean ± SD. *p < 0.05, **p < 0.01, and ***p < 0.001, ns, no significance.

To gain deeper insight into the inflammatory microenvironment of UC, we first examined the expression of the neutrophil marker Ly6G in colon tissues by immunofluorescence. The results showed a significant increase in Ly6G signal in DSS group, indicating substantial neutrophil infiltration at the inflammatory site (Fig. 10A and B). In contrast, BVZD treatment markedly suppressed Ly6G expression, suggesting attenuation of neutrophil recruitment to the intestinal mucosa. Further analysis by immunohistochemistry, ELISA, and WB revealed that the expression of CitH3, PAD4, and MPO was significantly upregulated in the colon tissues of DSS group, reflecting abundant NET formation (Fig. 10C–D, S11C-D). Following treatment, the expression of these NETs markers decreased to varying degrees, with BVZD group exhibiting the most pronounced inhibitory effect, confirming that BVZD effectively promotes NETs degradation. Moreover, immunofluorescence demonstrated upregulation of the ERS marker GRP78 upon DSS induction, which was reversed by BVZD treatment (Fig. 10A and B). WB analysis further indicated that BVZD significantly downregulated the expression of ATF4 and CHOP, demonstrating its efficacy in alleviating ERS (Fig. 10D). Additionally, immunofluorescence analysis showed that PTGS2 expression was significantly upregulated in DSS group and markedly downregulated after BVZD intervention (Fig. 10A and B). In summary, BVZD exerts its therapeutic effects in the UC model by degrading NETs and subsequently modulating the PTGS2-mediated ERS pathway.

Fig. 10.

BVZD degrades NETs and alleviates ER stress in vivo. (A) Representative immunofluorescence staining of Ly6G, GRP78, and PTGS2; (B) The expression of CitH3, MPO, and PAD4 (n = 3); (C) ELISA assay of CitH3, MPO, and PAD4 (n = 3); (D) WB analysis of CitH3, MPO, PAD4, ATF4, and CHOP protein levels (n = 3). Data are presented as the mean ± SD. *p < 0.05, **p < 0.01, and ***p < 0.001, ns, no significance.

As a macromolecular protease, orally administered DNase I is highly susceptible to rapid degradation and inactivation by gastric acid and various intestinal hydrolases. Moreover, the ZIF-8 carrier is prone to structural collapse in the strongly acidic gastric environment, leading to premature leakage and inactivation of the encapsulated enzyme [49,50]. Therefore, intravenous injection via the tail vein was selected in this study to bypass these complex oral barriers, thereby enabling a precise assessment of BVZD's targeting capability, in vivo enzymatic activity retention, and anti-inflammatory potential. Nonetheless, given that oral formulations are the mainstream for clinical UC therapy, BVZD holds considerable promise for adaptation into an oral dosage form. In the future, coating the nanoparticles with pH-responsive biopolysaccharides, alginate, or montmorillonite to construct oral microspheres could help overcome the two major barrier, ultimately facilitating the translation of BVZD from intravenous injection to a clinically convenient oral administration route [50,51].

4. Conclusions

In this study, we successfully developed a BVZD nanoparticle system capable of targeted delivery to activated neutrophils. This system effectively treats UC by degrading NETs and suppressing ERS. The three-dimensional porous structure of ZIF-8 provides a stable loading environment for DNase I, significantly enhancing the enzymatic stability. Modification with BOMVs further enhanced the nanoparticles’ inflammatory targeting ability and biocompatibility. Furthermore, BVZD effectively degraded NETs, alleviated local intestinal inflammation, and promoted the recovery of intestinal epithelial barrier function. Further mechanistic investigations revealed that BVZD regulates PTGS2 expression through NET clearance, thereby inhibiting ERS and providing a therapeutic effect against UC. These findings demonstrate that BVZD nanoparticles hold significant potential for the treatment of UC.

CRediT authorship contribution statement

Zhipeng Li: Data curation, Investigation, Validation, Visualization, Writing – original draft. Yaoyao Wu: Data curation, Investigation, Validation, Visualization. Jialin Wu: Validation, Visualization. Cheng Zhang: Methodology. Xinsheng Cheng: Data curation. Hongzhen Zhang: Supervision. Shikai Wang: Funding acquisition, Resources, Supervision, Writing – original draft, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the supplementary data to this article.

Data availability

Data will be made available on request.