Genome editing of PAR2 through targeted delivery of CRISPR-Cas9 system for alleviating acute lung inflammation via ERK/NLRP3/IL-1β and NO/iNOS signalling

Abstract

Excessive and uncontrollable inflammatory responses in alveoli can dramatically exacerbate pulmonary disease progressions through vigorous cytokine releases, immune cell infiltration and protease-driven tissue damages. It is an urgent need to explore potential drug strategies for mitigating lung inflammation. Protease-activated receptor 2 (PAR2) as a vital molecular target principally participates in various inflammatory diseases via intracellular signal transduction. However, it has been rarely reported about the role of PAR2 in lung inflammation. This study applied CRISPR-Cas9 system encoding Cas9 and sgRNA (pCas9-PAR2) for PAR2 knockout and fabricated an anionic human serum albumin-based nanoparticles to deliver pCas9-PAR2 with superior inflammation-targeting efficiency and stability (TAP/pCas9-PAR2). TAP/pCas9-PAR2 robustly facilitated pCas9-PAR2 to enter and transfect inflammatory cells, eliciting precise gene editing of PAR2 in vitro and in vivo. Importantly, PAR2 deficiency by TAP/pCas9-PAR2 effectively and safely promoted macrophage polarization, suppressed pro-inflammatory cytokine releases and alleviated acute lung inflammation, uncovering a novel value of PAR2. It also revealed that PAR2-mediated pulmonary inflammation prevented by TAP/pCas9-PAR2 was mainly dependent on ERK-mediated NLRP3/IL-1β and NO/iNOS signalling. Therefore, this work indicated PAR2 as a novel target for lung inflammation and provided a potential nanodrug strategy for PAR2 deficiency in treating inflammatory diseases.

Graphical abstract

TAP/pCas9-PAR2 alleviated acute lung inflammation through ERK/NLRP3/IL-1β and NO/iNOS signalling via genome editing of PAR2.

1. Introduction

Inflammation as a naïve defence response of the body can eradicate pathogens to the host organism^1,2. However, sustained and excessive inflammatory responses provoke a range of adverse effects, including tissue damage, edema and cytokine storm. Acute lung inflammation (ALI) is critically initiated by extensive inflammatory responses in the lung as a core respiratory organ, largely impairing respiratory system³. Emerging evidence has revealed that ALI is usually associated with potent proinflammatory cytokine release, protease and immune cell accumulation, and edema in alveoli, exhibiting the mortality rate of 35%–45%^3,4. Importantly, coronavirus disease 2019 (COVID-19) as a novel type of pneumonia possesses elevated pulmonary inflammation with aggravation of cytokine storm^5,6. The blockade of lung inflammatory responses and massive proinflammatory cytokines may dramatically mitigate the severe exacerbation of COVID-19. Indeed, it is an emergent need for probing drug targets and strategies in treating pulmonary inflammatory diseases.

Protease-activated receptor 2 (PAR2) belonging to G protein-coupled receptors (GPCRs) actively participates in various inflammatory diseases, such as paw edema, irritable bowel syndrome^7,8. PAR2 can be cleaved and activated by certain proteases to modulate downstream inflammatory signal transduction, while ALI can evoke abundant secretions of proteases^4,9, suggesting the possible role of PAR2 in lung inflammation. Additionally, PAR2 activated by proteases or synthetic ligands can powerfully initiate plentiful proinflammatory cytokine releases in flamed tissues¹⁰ which are also involved in lung inflammatory diseases. However, there is rare investigations on the relationship between PAR2 and pulmonary inflammation, despite targeting PAR2 as a potential and extraordinary therapeutical strategy for other inflammatory diseases.

To probe the participation of PAR2 in pulmonary inflammatory diseases, CRISPR-Cas9 technology can be utilized to evoke specific gene editing for accurate PAR2 inhibition. As a novel and powerful genome-editing tool, the plasmid-based CRISPR-Cas9 all-in-one system that encodes Cas9 and sgRNA targeting specific gene has been widely used for gene knockout in disease treatment¹¹. Due to the obstacles of large size, negative charge and nuclease degradation, plasmid itself cannot trigger effective and safe gene editing, while various approaches including nanoparticles have been applied for CRISPR-Cas9 plasmid delivery¹². Nanoparticle delivery system for CRISPR-Cas9 mainly focuses on cationic lipid nanoparticles, such as liposome, cationic polymers and lipid-based nanoparticles, which can condense plasmid and robustly improve gene editing efficiency of CRISPR-Cas9 for preventing disease progression13, 14, 15. However, such cationic nanoparticles usually exhibit high systemic toxicity that largely restricted effectiveness and safety of CRISPR-Cas9¹⁶. Human serum albumin (HSA) as natural biomaterials with excellent biocompatibility can be modified to encapsulate gene drugs for enhanced cell uptake and transfection efficacy in vitro and in vivo with low toxicity, suggesting that it is a potential alternative candidate for CRISPR-Cas9 delivery^17,18.

In this work, we developed an anionic albumin-based nanoparticle to encapsulate CRISPR-Cas9 plasmid encoding Cas9 and PAR2 sgRNA (pCas9-PAR2). Moreover, it has reported that elevated levels of collagen accumulated in inflamed sites and a collagen-binding peptide (CBP) can precisely target to inflammatory tissues via high collagen affinity¹⁹. Therefore, CBP was conjugated to the surface of HSA nanoparticles to enhance inflammation-targeting efficiency, consequently constructing inflammation-targeted albumin-PEI nanoparticles that encapsulated pCas9-PAR2 (TAP/pCas9-PAR2). Although it has been considered that cationic nanoparticles but not anionic materials can preferably improve cell uptake and transfection of genes via interacting with negative charge of cell membrane²⁰, the anionic TAP/pCas9-PAR2 successfully delivered CRISPR-Cas9 into cells with high cell transfection efficiency, and subsequently provoke gene editing of PAR2 in inflammatory cells. PAR2 deficiency initiated M1 to M2 macrophage polarization and prevented PAR2-specific or unspecific paw edema. Furthermore, TAP/pCas9-PAR2 safely and effectively targeted to inflamed lungs and evoked PAR2 knockout in vivo to eliminate PAR2-induced ALI, revealing the new role of PAR2 in lung inflammation. More importantly, we uncovered its underlying mechanism of this nanodrug, in which PAR2 deficiency by TAP/pCas9-PAR2 alleviated PAR2-mediated lung inflammation via ERK/NLRP3/IL-1β and NO/iNOS signalling.

2. Materials and methods

2.1. Materials and cell culture

Human serum albumin (HSA) was purchased from Sigma (Cas No. 70024-90-7, St. Louis, MO, USA), while poly-(ethylene imine) (PEI 10K), glutaraldehyde, λ-carrageenan were obtained from Adamas (Cas No. 9002-98-6, 111-30-8, 9064-57-7, Shanghai, China). 2-Furoyl-LIGRL-NH₂ (2f-LIGRL-NH₂, 2f, 98%) and collagen-binding peptide (CBP, LRELHLNNNC, 98%) were custom-made by Apeptides (Shanghai, China). The primary antibodies against total/phospho-ERK were purchased from Cell Signaling Technology (Cat No. 4695S, 4370S, Boston, MA, USA). The primary antibodies against PAR2, NLRP3, iNOS and GAPDH were obtained from Abcam (Cat No. ab180953, ab263899, ab178945, ab181602, Cambridge, MA, USA). The secondary antibodies HRP-conjugated anti-rabbit IgG was obtained from Invitrogen (Cat No. 31460, Carlsbad, CA, USA). The eSpCas9-2A-Puro (PX459) V2.0 plasmid encoding Cas9 and human or murine PAR2 sgRNA, and the eSpCas9-2A-GFP (PX458) plasmid with Cas9, GFP and PAR2 sgRNA were constructed by GenScript (Nanjing, China).

The human non-small cell lung cancer A549 cell line and mouse leukemia cells of monocyte-macrophage RAW264.7 cell line were obtained from Sichuan University. Cells were cultured in DMEM medium (Cat No. 10-013-CVR, Corning, New York, NY, USA), supplemented with 10% fetal bovine serum (Gibco, Carlsbad, CA, USA) and 100 U/mL penicillin/streptomycin at 37 °C in a humidified incubator containing 5% CO₂.

2.2. Preparation of nanoparticles

HSA nanoparticles encapsuled CRISPR-Cas9 encoding PAR2 gRNA and Cas9 (pCas9-PAR2) were constructed by a modified desolvation technique as described previously^17,18,21. Briefly, HSA and PEI were dissolved in ddH₂O (pH 8.0), respectively. HSA, PEI and pCas-PAR2 were vortexed for 20 s and then incubated for 20 min. Subsequently, 2 mL ethanol was added dropwise to the mixture under stirring at 500 rpm (Cat No. 0005019825, Ika, Germany) at room temperature for 30 min. The mixture was then crosslinked with 100 μL 2.5% glutaraldehyde under stirring for 30 min. Finally, collagen-binding peptide (CBP) was added and stirred at 220 rpm at 4 °C overnight for forming CBP-grafted HSA-PEI-pCas9-PAR2 nanoparticles (TAP/pCas9-PAR2). For preparing albumin-PEI nanoparticles without CBP that encapsulated pCas9-PAR2 (AP/pCas9-PAR2), HSA nanoparticles were crosslinked with 2.5% glutaraldehyde under stirring at 220 rpm at 4 °C overnight. All nanoparticles were washed and redispersed with ddH₂O through repeated ultracentrifugation for separating unencapsulated plasmid.

2.3. Physicochemical characterization of nanoparticles

The size, PDI and Zeta potential of nanoparticles were measured by the dynamic light scattering (DLS) technique (ZEN3600, Malvern, UK). For morphology, nanoparticles were observed by transmission electron microscope (TEM, JEM-2100F, Tokyo, Japan). To investigate their stability, the size and PDI of nanoparticles were measured over 30 days. The fluorescent dye SYBR Green was used to measure plasmid concentration. Free plasmids with various concentrations were used to establish a standard curve, and the unencapsulated pCas9-PAR2 was collected after ultracentrifugation. All plasmids were incubated with SYBR Green for 10 min, while the unencapsulated plasmid concentration was evaluated by detecting the fluorescence of SYBR Green-conjugated plasmids. The weight of encapsulated plasmid was subtraction of unencapsulated plasmid from total plasmid. The encapsulation efficiency and drug loading capacity were calculated as in Eqs. (1) and (2):

$$\mathrm{E}\mathrm{n}\mathrm{c}\mathrm{a}\mathrm{p}\mathrm{s}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y}(\%)=\frac{\mathrm{W}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t}\mathrm{o}\mathrm{f}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{a}\mathrm{p}\mathrm{s}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{p}\mathrm{l}\mathrm{a}\mathrm{s}\mathrm{m}\mathrm{i}\mathrm{d}}{\mathrm{W}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t}\mathrm{o}\mathrm{f}\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{p}\mathrm{l}\mathrm{a}\mathrm{s}\mathrm{m}\mathrm{i}\mathrm{d}}\times 100$$ (1)

$$\mathrm{D}\mathrm{r}\mathrm{u}\mathrm{g}\mathrm{l}\mathrm{o}\mathrm{a}\mathrm{d}\mathrm{i}\mathrm{n}\mathrm{g}\mathrm{c}\mathrm{a}\mathrm{p}\mathrm{a}\mathrm{c}\mathrm{i}\mathrm{t}\mathrm{y}(\%)=\frac{\mathrm{W}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t}\mathrm{o}\mathrm{f}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{a}\mathrm{p}\mathrm{s}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{p}\mathrm{l}\mathrm{a}\mathrm{s}\mathrm{m}\mathrm{i}\mathrm{d}}{\mathrm{W}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t}\mathrm{o}\mathrm{f}\mathrm{N}\mathrm{P}\mathrm{s}}\times 100$$ (2)

2.4. In vitro cellular uptake

Cy5-DNA as a fluorescent DNA substitution for pCas9-PAR2 to evaluate cell uptake efficiency of nanoparticles. Cells were seeded at 24 well plates overnight and were treated with Cy5-DNA and different nanoparticles (equal to 1 μg Cy5-DNA) for 1 h. Cells were then washed three times with PBS, and cellular uptake efficiency was determined by flow cytometry (AccuriTM C6 flow cytometer, Becton Dickinson, NJ, USA).

2.5. In vitro cellular transfection

GFP-tagged pCas9-PAR2 (GFP-pCas9-PAR2) was used for detection of cell transfection. Cells were seeded in 24 well plates overnight and treated with GFP-pCas9-PAR2 and nanoparticles (equal to 1 μg GFP-pCas9-PAR2) for 4 h. Cells were then washed with PBS and medium was replaced with fresh medium for further 44 h. After that, cells were washed with PBS three times and determined by flow cytometry (Becton Dickinson). Alternatively, after transfection, cells were stained with DAPI and observed by confocal microscopy (FV-OSR, Olympus, Tokyo, Japan).

2.6. In vitro cytotoxicity

Cells were seeded at a 96 well plate overnight. Cells were then treated with nanoparticles for 4 h and medium was replaced with fresh medium for further 44 h. Alternatively, cells were treated with nanoparticles for 48 h experiment. After that, cells were incubated with CCK8 (Cat No. K101826433EF5E, APExBIO, Houston, TX, USA) and measured at 450 nm.

2.7. DNase protection assay

pCas9-PAR2 and nanoparticles were incubated with DNase (Cat No. 1078S, TaKaRa, Kyoto, Japan) for 30 min at 37 °C, and then added with 1M EDTA to stop the reaction. Cells were incubated with DNase-treated and untreated pCas9-PAR2 and nanoparticles for 4 h. Cells were then thoroughly washed three times with PBS and incubated with a fresh medium for another 44 h. Fluorescence images were observed by fluorescence microscopy.

2.8. T7 endonuclease I (T7EI) assay and sanger sequencing analysis

Cells were seeded at 24 well plates overnight and treated with different nanoparticles for 48 h. Genomic DNA was extracted and amplified according to the manufacturer's instructions. PCR products diluted in reaction buffer were annealed in a thermocycler using the following conditions (95 °C for 5 min, −2 °C/s to 85 °C and −0.1 °C/s to 25 °C, and then cooling down to 4 °C). After that, T7EI (Cat No. 017E229GA, Vazyme, Nanjing, China) was added and incubated at 37 °C for 15 min. The cleavage reaction was stopped by 0.25 mol/L EDTA. The fragmented DNA were subjected to electrophoresis and was analyzed by ImageJ. The cleavage percentage was calculated by Eq. (3):

Cleavage (%) = (1‒(1‒Fraction cleaved)^1/2) × 100 (3)

The primers are listed in Supporting Information Table S1. PCR products were also subcloned by a TA/Blunt-Zero cloning kit (Cat No. 017E2210IA, Vazyme), and were randomly picked for sanger sequencing by Tsingke Biotechnology Co., Ltd.

2.9. Real-time quantitative PCR (qRT-PCR)

Cells were seeded at 24 well plates overnight and were incubated with different nanoparticles containing 2 μg pCas9-PAR2 for 30 h. 25 μmol/L 2f-LIGRL-NH₂ and 1 μg/mL LPS was then added for further 18 h. RNA was isolated using TRIzol™ Reagent (Cat No. 15596026, Invitrogen) according to manufacturer’s instructions. One microgram of isolated RNA was reverse transcribed using HiScript III 1st Strand cDNA Synthesis Kit (Cat No. 7F581J1, Vazyme) for cDNA synthesis. qRT-PCR analysis was performed using ChamQ Universal SYBR qPCR Master Mix (Cat No. 027E1280LA, Vazyme) with suitable primers. The primers used are listed in Table S1.

2.10. Western blotting

Cells were seeded at 12 well plates overnight, and were then incubated with different nanoparticles containing 2 μg pCas9-PAR2 for 30 h. 25 μmol/L 2f-LIGRL-NH₂ and 1 μg/mL LPS was then added for 18 h. Cells were lysed with RIPA buffer containing PMSF. Protein samples were separated using SDS-PAGE and transferred to the NC membrane. After blocked, membrane was incubated overnight with primary antibodies and HRP-conjugated goat anti-rabbit IgG secondary antibody was added for 2 h. Bands were visualized using a BeyoECL Moon (Cat No. P0018AS, Beyotime, Shanghai, China), and the software ImageJ was applied for quantitative analysis.

2.11. PAR2 gene and protein expression detection

Cells were seeded at 24 well plates overnight and were then treated with different nanoparticles for 48 h. For gene expression, total RNA of cells was isolated and then PAR2 gene expression was detected by RT-PCR. For cell surface expression of PAR2, cells were incubated with the primary antibody against PAR2 in PBS for 1 h and were then incubated with Cy3-conjugated affinipure goat anti-rabbit IgG followed by fluorescence detection using flow cytometry analysis (Becton Dickinson). For PAR2 protein expression, lung tissues were lysed and detected by Western blotting technique after drug treatments.

2.12. PAR2 agonist-induced and λ-carrageenan-induced paw edema

All experimental procedures were executed according to the protocols approved by Southwest Jiaotong University Animal Care and Use Committee. Before administration, the thickness and width of left hind paw of each male mouse (4 weeks) was measured as an internal control. Mice were intravenously injected with pCas9-PAR2, AP/pCas9-PAR2 or TAP/pCas9-PAR2 containing 10 μg pCas9-PAR2. After 24 h, the left hind paw pad was injected with 2-furoyl-LIGRL-NH₂ (2f, 350 μg/paw) for PAR2-specific paw edema, λ-carrageenan for non-specific paw edema or saline control. Paw measurements were measured within different time points post injection. The paw was cut off and stained with hematoxylin and eosin (H&E) for histopathological analysis. The paw swelling degree was calculated as thickness × width normalized to maximal swelling (100%).

2.13. Establishment of acute lung inflammation (ALI) model

For acute lung inflammation model, mice (C57BL/6, male, 6 weeks) were randomly divided into six groups. Mice were anesthetized and nasally administrated with saline, LPS or LPS+2f (LPS: 50 μg, 2f: 250 μg). After 2 h, mice were intravenously injected with saline, pCas9-PAR2, AP/pCas9-PAR2 or TAP/pCas9-PAR2 containing 10 μg pCas9-PAR2. After 48 h, mice underwent bronchus-alveolus lavage with saline, and the supernatant of bronchoalveolar lavage fluid (BALF) was collected and stored at −80 °C. The precipitated cells were resuspended with PBS, and the total cell numbers of BLAF were counted on a hemocytometer and observed with Wright‒Giemsa staining. The protein concentrations in the BALF were determined by BCA kit. Alternatively, lungs were fixed with 4% formaldehyde for histopathological analysis, or were collected for detecting protein levels by Western blotting.

2.14. ELISA

BALF was collected from ALI model and the concentrations of inflammatory factors (IL-6, TNF-α and IL-1β) in BALF were quantified by ELISA Kit (Cat No. 1210602, 1217202, 1210122, DAKEWE, Shenzhen, China) according to the manufacturer's instructions.

2.15. Pharmacokinetic studies

Healthy mice were intravenously injected with Cy5-DNA and Cy5-DNA-loaded nanoparticles. The blood from orbit was collected at 0, 1, 5, 10, 30 min, 1, 2, 3, 6, and 24 h after administration. Free Cy5-DNA with different concentrations were used to prepare a Cy5 standard curve (Supporting Information Table S2). The blood was diluted, and fluorescence was detected (Ex = 640 nm, Em = 670 nm). Pharmacokinetics parameters were calculated to fit a two-compartment model by DAS software.

2.16. In vivo biodistribution and targeting efficacy in inflamed lungs

Acute lung inflammation model was constructed as described above. Mice were intravenously injected with 10 μg Cy5-DNA and Cy5-DNA-loaded nanoparticles. Mice were sacrificed and organs (heart, liver, spleen, lung, and kidney) were photographed with the imaging system at 4 and 24 h. Moreover, the Cy5 fluorescence was detected in the lung at different time points (0.5, 4, 8, 24, 48, and 72 h) after ALI mice injected with TAP nanoparticles by the IVIS Spectrum imaging system (PerkinElmer, Waltham, MA, USA). Additionally, lungs were fixed with 4% formaldehyde in a dark place. The tissues were labelled with collagen-I primary antibody and a FITC-labelled secondary antibody. Finally, DAPI was applied to label nucleus and lung tissues were imaged with a fluorescence microscope.

2.17. In vivo safety measurement

Healthy mice were intravenously injected with different nanoparticles for 24 h or 2 weeks and the blood was collected for detecting CRE, ALT and AST concentrations according to the kit instructions (Cat No. C011-2-1, C010-2-1, C009-2-1, Jiancheng, Nanjing, China). The livers were collected and stained with H&E for histopathological analysis.

2.18. Statistical analysis

All statistical analyses were performed using GraphPad Prism 8. One-way ANOVA followed by Dunnett's multiple comparisons test was used for comparison of more than two groups. Unpaired two-tailed Students' t-test was used for comparison of two groups, Data are presented as mean ± standard error of mean (SEM) (n = 3). A P value <0.05 is considered as significant difference (∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001).

3. Results

3.1. The construction and characterization of TAP/pCas9-PAR2

The CRISPR-Cas9 plasmid alone fails to induce PAR2 deficiency due to its large size, negative charge and instability in vitro or in vivo. For improving the efficiency of gene editing, the natural and anionic human serum albumin (HSA) with excellent biocompatibility was utilized to encapsulate CRISPR-Cas9 plasmid encoding PAR2 sgRNA and Cas9 (pCas9-PAR2) in the assistance of PEI10K (Fig. 1A). In addition, the conjugation of collagen-binding peptide (CBP) endowed HSA nanoparticles with inflammatory site-targeting capacity (TAP/pCas9-PAR2), since abundant collages were exposed in the inflamed tissues¹⁹. Importantly, we discovered that pH was of most importance for HSA nanoparticle formula. HSA solution at pH 8 constructed nano-sized TAP/pCas9-PAR2 particles at different N/P with average diameters of ∼100 nm, whereas HSA solution at pH 4 or 6 formed large size particles (Fig. 1C and D, Supporting Information Table S3). It may because HSA solution at pH 8, which was much higher than the isoelectric point of HSA, led to the repulsion of anionic HSA molecules, consequently facilitating the formation of smaller nanoparticles^22,23. As shown in Fig. 1B, PEI/pCas9-PAR2 and AP/pCas9-PAR2, TAP/pCas9-PAR2 nanoparticle exhibited strong Tyndall effects, suggesting its extraordinary colloidal property. The morphology of TAP/pCas9-PAR2 also showed nearly spherical and uniform distribution by transmission electron microscopy (TEM) (Fig. 1E). Moreover, DLS analysis indicated that the average diameter of TAP/pCas9-PAR2 was 105.59 ± 3.91 nm with a polydispersity index (PDI) of 0.209 ± 0.013, indicating relatively high uniformity, while the surface charge was negative with Zeta potential of −20 ± 0.78 mV (Fig. 1F). The SYBR Green staining tool was applied to detect unencapsulated plasmid after nanoparticle preparation, revealing that the drug loading capacity of TAP/pCas9-PAR2 reached 0.195 ± 0.00003% and the encapsulation efficiency was 99.97 ± 0.014% (Fig. 1F). Additionally, the agarose gel electrophoresis result showed that TAP/pCas9-PAR2 did not exhibit visual plasmid band compared to pCas9-PAR2 alone, and there was nearly no plasmid band (unencapsulated pCas9-PAR2) detected after ultracentrifugation of TAP/pCas9-PAR2 (Supporting Information Fig. S1), which was because plasmid concentration was too low to be detected. This result visually indicated the excellent encapsulation capacity of TAP/pCas9-PAR2.

Figure 1.

The physicochemical property of TAP/pCas9-PAR2 nanoparticle. (A) The preparation of TAP/pCas9-PAR2; (B) The Tyndall effect of different nanoparticles; (C, D) The comparison of size and PDI of TAP/pCas9-PAR2 under different N/P ratio and pH; (E) The TEM image of TAP/pCas9-PAR2. Scale bar = 1 μm (low-magnified); Scale bar = 200 nm (high-magnified); (F) The Size, polydispersity index (PDI), Zeta, drug loading rate and encapsulation efficiency of TAP/pCas9-PAR2; (G) TAP/pCas9-PAR2 protected cell transfection of pCas9-PAR2 from DNase I degration in RAW264.7. Scale bar = 50 μm; (H, I) The stability of TAP/pCas9-PAR2 during one month by detecting changes in the size (H) and PDI (I) of nanoparticles.

More importantly, it has shown that TAP/pCas9-PAR2 induced high transfection efficiency of GFP-pCas9-PAR2 with or without DNase I, but pCas9-PAR2 alone failed to induce such effects (Fig. 1G), suggesting the excellent protection of pCas9-PAR2 by TAP/pCas9-PAR2 under enzyme conditions. TAP/pCas9-PAR2 has been also stable over 30 days since the particle size and polydispersity of nanoparticles were not significantly changed at room temperature for 30 days (Fig. 1H, I, and Supporting Information Table S4). These results suggest that TAP/pCas9-PAR2 possessed outstanding physicochemical properties and preferably protected plasmid from nuclease degration, supporting its potency for gene drug delivery.

3.2. TAP/pCas9-PAR2 effectively delivered pCas9-PAR2 to evoke gene deficiency in vitro

Before gene editing by CRISPR-Cas9 in nucleus, the Cas9-PAR2 plasmid should enter and effectively transfect cells. Near-infrared Cy5-labeled DNA was used instead of pCas9-PAR2 to quantitatively evaluate the cell uptake of nanoparticles (Fig. 2B). The results showed TAP/pCas9-PAR2 group markedly promoted uptake efficiency of pCas9-PAR2 but pCas9-PAR2 itself failed to enter cells. Importantly, TAP/pCas9-PAR2 with N/P = 16:1 exhibited the most excellent efficiency of cell uptake among different N/P ratios. Nanoparticles with CBP (TAP/pCas9-PAR2) showed higher uptake efficiency than AP/pCas9-PAR2 in RAW264.7 and A549 cells, suggesting that this collagen-binding peptide significantly assisted TAP/pCas9-PAR2 to deliver plasmid into inflammatory cells.

Figure 2.

The cell uptake, transfection and gene editing of TAP/pCas9-PAR2. (A) Representative confocal images of cell transfection of pCas9-PAR2 and TAP/pCas9-PAR2. DAPI: blue, nucleus. GFP: green, GFP-pCas9-PAR2. Scale bar = 10 μm; (B) Cell uptake efficiency of nanoparticles with different N/P ratios in RAW264.7 and A549 cells; (C) Transfection efficiency of NPs with different N/P ratios in RAW264.7 and A549 cells; (D) The CRISPR/Cas9 plasmid encoding different guide RNA sequences for targeting mouse PAR2 (mF2RL1) or human PAR2 (hF2RL1); (E) Gene editing of PAR2 locus after cells treated with pCas9-PAR2, TAP/pCas9-PAR2 or lipofectamine 3000 were detected by T7EI assay. The original PCR product size is 555 bp; (F) For estimating gene editing by CRISPR/Cas9, RT-PCR technique was applied for detecting F2rl1 (Par2) gene expression in RAW264.7 cells; (G) The cell surface expression of PAR2 was detected among different groups by FACS. (H) The cell viability of cells treated with different nanoparticles were detected for evaluating their toxicity. Cells were treated with nanoparticles at different N/P ratios for 4 h and replaced with complete medium for further 44 h or were directly incubated with nanoparticles for 48 h. There was no statistical significance among all groups. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 vs. LPS+2f. ^###P < 0.001 vs. pCas9-PAR2. ˆˆP < 0.01, ˆˆˆP < 0.001 vs. AP/pCas9-PAR2.

To further investigate the transfection efficiency of TAP/pCas9-PAR2, we prepared pCas9-PAR2 plasmid encoding green fluorescent protein (GFP) (GFP-pCas9-PAR2). The expression of GFP allowed the convenient readout of cell transfection efficiency. Under confocal fluorescence microscopy, pCas9-PAR2 itself could not evoke any transfections by cells whereas GFP-tagged TAP/pCas9-PAR2 exhibited strong green fluorescence, indicating that TAP/pCas9-PAR2 intensively promoted cell transfections of pCas9-PAR2 (Fig. 2A). Furthermore, consistent with cell uptake results, TAP/pCas9-PAR2 at N/P = 16:1 triggered higher transfection efficiency than nanoparticles at other N/P ratios (8:1, 32:1), suggesting 16:1 was the optimized N/P ratio (Fig. 2C). Importantly, the CBP also dramatically enhanced cell transfections of nanoparticles, since TAP/pCas9-PAR2 preferably improved ∼30-fold transfection efficiency of pCas9-PAR2 compared to plasmid alone and notably triggered ∼1.5-fold higher transfection efficiency than AP/pCas9-PAR2. Although PEI polymer induced cell transfection of pCas9-PAR2, the transfection efficiency of TAP/pCas9-PAR2 was remarkedly higher (∼5-fold) than PEI/pCas9-PAR2, confirming TAP/pCas9-PAR2 as a superior transfection reagent for CRISPR-Cas9.

The CRISPR-Cas9 plasmid encoding guide RNA targeting mouse PAR2 (mF2RL1) and human PAR2 (hF2RL1) were designed and constructed, respectively (Fig. 2D). T7E1 assay was applied to evaluate if the genome editing of CRISPR-Cas9 at targeted PAR2 genome locus through recognizing and cleaving the mismatched DNA by T7EI endonuclease. Similar to lipofectamine 3000 group, two cleavage bands were clearly observed in the PAR2 locus of TAP/pCas9-PAR2 group with the cleavage efficiency of ∼25.5% whereas pCas9-PAR2 alone showed the original band without any cleavage bands (Fig. 2E). Additionally, Sanger sequencing results confirmed that indels were presented at PAR2 loci with different mutation sites around protospacer adjacent motifs (PAMs) (Supporting Information Fig. S2), indicating the precise genome editing of TAP/pCas9-PAR2. Additionally, compared to naïve control, PEI/pCas9-PAR2 and AP/pCas9-PAR2, RAW264.7 cells treated with TAP/pCas9-PAR2 encoding mF2RL1 sgRNA triggered markable reductions in F2rl1 (Par2) gene expression and PAR2 surface expression (Fig. 2F and G), while TAP/pCas9-PAR2 encoding hF2RL1 gRNA also evoked strong PAR2 gene editing in mRNA and surface expression levels in A549 cells (Supporting Information Fig. S3), suggesting successful gene editing of PAR2. Alternatively, cells were treated with TAP/pCas9-PAR2 at different N/P ratios for 4 h and then replaced with fresh medium for further 44 h, which did not evoke obvious cell toxicity, while TAP/pCas9-PAR2 also elicited excellent cell viability after 48 h incubation, supporting the superior safety of TAP nanoparticles (Fig. 2H, Supporting Information Fig. S4). Taken together, TAP/pCas9-PAR2 at N/P ratio = 16:1 could efficiently deliver CRISPR-Cas9 plasmid into cells and successfully evoke PAR2 gene editing in vitro.

3.3. PAR2 deficiency by TAP/pCas9-PAR2 mitigated PAR2-induced paw edema and non-specific paw edema in vivo

Apart from in vitro PAR2 knockout by TAP/CRISPR-Cas9, we further evaluated its capacity of gene editing in vivo. The pharmacokinetics of TAP/pCas9-PAR2 was firstly investigated by detection of the average plasma concentration of Cy5-DNA after intravenous injection of nanoparticles into healthy mice (Fig. 3A). These pharmacokinetic parameters were also summarized in Supporting Information using two-compartment model analysis. Results showed that the distribution half-life (t_1/2ɑ) of TAP/pCas9-PAR2 was 3.5-fold higher than pCas9-PAR2, while TAP/pCas9-PAR2 robustly prolonged the elimination half-life (t_1/2β) of pCas9-PAR2 in vivo by ∼30 h. Furthermore, AUC_0‒t of TAP/pCas9-PAR2 treatment exhibited ∼3-fold higher than pCas9-PAR2, and TAP/pCas9-PAR2 with enhanced peak plasma concentration and persistent time showed ∼100-fold increase of AUC_0‒∞ for pCas9-PAR2. Therefore, TAP/pCas9-PAR2 possessed superior pharmacokinetic properties and significantly improved the bioavailability of pCas9-PAR2 for effective gene editing.

Figure 3.

TAP/pCas9-PAR2 initiated in vivo gene editing to alleviate PAR2-specific and non-specific paw edema. (A) Pharmacokinetic study of TAP/pCas9-PAR2 in healthy mice; (B) Schematic diagram of paw edema experiments; (C, D) The paw edema images and further H&E staining analysis of nanoparticle treatment groups in 2f-LIGRL-NH₂-induced specific short-term paw edema at 0.5 h (C) and λ-carrageenan-stimulated paw edema model at 4 h (D). Red arrow: edema, blue arrow: lymphocytes, green arrow: dead collagen fibers, black arrow: neutrophils. Scale bar = 100 μm (100 × ), scale bar = 10 μm (400 × ); (E) Time courses of nanoparticles-treated PAR2-specific short-term paw edema; (F) The comparison of different treatments for PAR2-specific paw edema models at 0.5 and 1 h; (G) Time courses of nanoparticles-treated non-specific paw edema; (H) The comparison of different treatments in non-specific paw edema models at 4 h, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 vs. 2f or λ-carrageenan, ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 vs. pCas9-PAR2, ˆˆP < 0.01 vs. AP/pCas9-PAR2.

Alternatively, we constructed two mice paw edema models as traditional animal inflammatory models which were highly related to PAR2^24,25. The classic PAR2 agonist, 2-furoyl-LIGRL-NH₂ (2f-LIGRL-NH₂, 2f), stimulated PAR2-specific paw edema as a short-term model, while λ-carrageenan that is a soluble complex extracted from carrageenaceae can evoke non-specific paw edema modeling (Fig. 3B). The swelling degree were monitored and measured using electronic vernier calipers at fixed intervals over 8 h, discovering that it peaked at 0.5 h for PAR2-specific paw edema and at 4 h for non-specific paw edema, and rapidly decreased after peak time (Fig. 3E and G). To precisely investigate the anti-inflammatory effects, nanoparticles were administrated intravenously 24 h before stimulation of paw edema, which could guarantee enough time for PAR2 gene editing. Results showed that TAP/pCas9-PAR2-treated mice remarkedly alleviated PAR2-specific and non-specific foot swelling, which were more potent than AP/pCas9-PAR2 group (Fig. 3C‒H). It is worth mentioning that pCas9-PAR2 exhibited a stronger paw edema response than stimulators probably due to plasmid-induced inflammation, suggesting that TAP nanoparticle could prevent immunogenicity of pCas9-PAR2. Meanwhile, H&E staining results showed that no obvious edema or inflammatory infiltration was observed in TAP/pCas9-PAR2 and control groups (Fig. 4C and D), but pCas9-PAR2 alone had the most severe inflammatory infiltration, which also reflected the advantage of using human serum albumin as nanoparticles with low toxicity. These results indicated that TAP/pCas9-PAR2 possessed the extraordinary capacity of in vivo PAR2 gene editing and potential therapeutical effects for PAR2-mediated inflammatory diseases.

Figure 4.

TAP/pCas9-PAR2 modulated macrophage polarization, proinflammatory cytokine expressions through ERK/NLRP3/Caspase-1/IL-1β and NO/iNOS signalling. (A) A schematic diagram for PAR2-mediated macrophage polarization; (B) The changes in mRNA levels of Cd206, Il6, and Tnfα after RAW264.7 cells were treated with different nanoparticles detected by RT-PCR; (C–E) The changes in protein expressions of ERK, phosphorylated ERK and NLRP3 after RAW264.7 cells were treated with different nanoparticles detected by Western blotting; (F) The comparison of mRNA levels of Il1β and Caspase1 after RAW264.7 cells were treated with different treatments; (G) The protein expression of iNOS in RAW264.7 cells after different treatments; (H) The changes in NO productions after RAW264.7 cells with different treatments; (I–K) The changes in protein expressions of ERK, phosphorylated ERK and NLRP3 after A549 cells were treated with different nanoparticles. (L, M) The changes in mRNA levels of IL6, TNFα, IL1β and CASPASE1 after A549 cells were treated with different nanoparticles; (N). The schematic diagram of TAP/pCas9-PAR2-mediated downstream inflammatory signalling pathways. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 vs. LPS+2f. ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 vs. pCas9-PAR2. ˆP < 0.05, ˆˆˆP < 0.001 vs. AP/pCas9-PAR2.

3.4. PAR2 gene editing by TAP/pCas9-PAR2 evoked anti-inflammatory activity via alleviating ERK/NLRP3/caspase-1/IL-1β and NO/iNOS signalling

In attempt to probing underlying mechanism of TAP/pCas9-PAR2-mediated anti-inflammatory capacity, we further investigated intracellular signalling pathways and responses. It has been widely reported that macrophages exhibited different types, including M0 macrophages, M1 pro-inflammatory macrophages and M2 anti-inflammatory macrophages²⁶. During inflammatory progress, different macrophages initiate transformation and polarization that could be a crucial indicator for inflammation. Importantly, TNF-α and IL-6 cytokine are M1 macrophages-related cytokines whereas CD206 represents M2-like macrophages (Fig. 4A). As shown in Fig. 4A and B, LPS and 2f-LIGRL-NH₂ (2f) significantly attenuated mRNA expressions of Cd206 and enhanced Tnfα and Il6 expressions, suggesting that LPS plus PAR2 activation promoted M2 to M1 macrophage polarization. Moreover, TAP/pCas9-PAR2, but not pCas9-PAR2, PEI/pCas9-PAR2 or AP/pCas9-PAR2, evoked PAR2 gene editing to robustly mitigate PAR2-induced enhancement of Tnfα and Il6 expressions (∼3.5-fold) as well as reduction of Cd206 expressions (∼2.5-fold) in RAW264.7 cells. It indicated that PAR2 promoted transformation of M2 to M1 proinflammatory macrophages and PAR2 deficiency induced by TAP/pCas9-PAR2 remarkably reversed macrophage polarization, exerting superior anti-inflammatory responses.

Additionally, we discovered that PAR2-mediated inflammation was highly related to ERK and NLRP3 inflammasome signalling. Although ERK signalling was one of the most common pathways modulated by PAR2 activation, the relationship between PAR2 and NLRP3 has been rarely reported. LPS and 2f-LIGRL-NH₂ significantly initiated ERK phosphorylation and NLRP3 expression (∼2-fold increase) in both RAW264.7 (Fig. 4C‒E) and A549 cells (Fig. 4I‒K). Importantly, PAR2-evoked ERK phosphorylation was time-dependent, which peaked at 30 min in RAW264.7 cells and at 10 min in A549 cells (Supporting Information Fig. S5). TAP/pCas9-PAR2 could not only dramatically deliver CRISPR-Cas9 to edit PAR2 gene for alleviating PAR2-induced phosphorylated ERK and NLRP3 expressions, but also triggered much more potent inhibitions than PEI/pCas9-PAR2 and AP/pCas9-PAR2. Moreover, Caspase-1 and IL-1β as NLRP3 inflammasome downstream proteins were detected for supporting NLRP3 signalling pathway^27,28. Unsurprisingly, as shown in Fig. 4F and L, TAP/pCas9-PAR2 attenuated PAR2-induced mRNA expression levels of Caspase1 and Il1β compared to simulators and pCas9-PAR2 group in RAW264.7 and A549 cells, confirming the active participation of NLRP3/Caspase-1/IL-1β signalling. TAP/pCas9-PAR2, but not pCas9-PAR2, PEI/Cas9-PAR2 and AP/Cas9-PAR2, also markedly prevented PAR2-mediated proinflammatory cytokine expressions, TNFα and IL6, in A549 cells, further indicating its strong anti-inflammatory activity (Fig. 3M). Furthermore, PAR2 activation dramatically facilitated ∼20-fold enhancement of iNOS protein expressions in RAW264.7 (Fig. 4G) and ∼2-fold increase in A549 cells (Supporting Information Fig. S6), while LPS and 2f-LIGRL-NH₂ promoted ∼14-fold increase in nitric oxide (NO) production (Fig. 4H). Compared to pCas9-PAR2, PEI/pCas9-PAR2 and AP/pCas9-PAR2, TAP/pCas9-PAR2 markedly mitigated PAR2-induced NO/iNOS pathways, confirming that TAP/pCas9-PAR2 with CBP could inhibit PAR2-mediated NO/iNOS signalling. Importantly, consistent with previous results, pCas9-PAR2 stimulated inflammation-related cytokine release, protein expression and NO productions, suggesting plasmid itself might stimulate inflammatory responses. TAP nanoparticles could prevent pCas9-PAR2 from such spontaneous inflammation and trigger anti-inflammatory activity.

Taken together, TAP/pCas9-PAR2 could enter and transfect cells to initiate PAR2 gene editing by CRISPR-Cas9, and then alleviated PAR2 and LPS-induced ERK/NLRP3/Caspase-1/IL-1β signalling and NO/iNOS pathways, consequently mitigating PAR2-induced inflammatory responses (Fig. 4N).

3.5. PAR2 deficiency by TAP/pCas9-PAR2 alleviated PAR2-mediated acute lung inflammation in vivo

We also probed the novel value of PAR2 and anti-inflammatory effects of TAP/pCas9-PAR2 in another mice inflammatory model. Acute lung inflammation (ALI) is one of the most common LPS-induced inflammation related to cytokine storm and lung inflammation. Our study firstly investigated the relationship between PAR2 and ALI through LPS plus 2f-LIGRL-NH₂-induced lung inflammation (Fig. 5A). The addition of 2f-LIGRL-NH₂ into LPS could robustly enhance LPS-induced lung inflammation, uncovering that PAR2 activation worsened ALI progression (Fig. 5). More importantly, ALI could be initially induced at 2 h while the severity of this disease probably peaked at 24 or 48 h, and then decreased. Meanwhile, CRISPR-Cas9 plasmid required at least 24 h for transfection and gene editing of PAR2. Therefore, we used LPS+2f to stimulate initial ALI at 2 h and then added nanoparticles for investigating PAR2 gene editing in treating ALI progression. Before evaluating therapeutic efficacy, the in vivo biodistribution of nanoparticles was investigated for verifying inflamed lung-targeting efficiency of TAP/pCas9-PAR2 (Supporting Information Fig. S7A‒S7C). At 4 h after intravenous injection of Cy5-labeled nanoparticles, it was observed that TAP/pCas9-PAR2 groups with addition of CBP that was the collagen-targeting peptide possessed higher fluorescence intensity in the lung tissue than AP/pCas9-PAR2, whereas there was no fluorescence in lung section from mice treated with pCas9-PAR2. Interestingly, TAP/pCas9-PAR2 group still maintained a high fluorescence intensity in lung tissue, but the fluorescence signal of AP/pCas9-PAR2 was gradually decreased after 24 h. These results could be explained by the high affinity of collagen-binding peptide CBP for abundant collagens at inflammation sites. When acute lung inflammation occurs in mice, the intercellular substance secretes collagen to repair, and consequently excessive collagen aggregation promotes pulmonary fibrosis¹⁹. As expected, TAP/pCas9-PAR2 group exhibited strong Cy5 fluorescence intensity (red) which was preferably co-located with the green fluorescence of collagens in inflamed lungs whereas pCas9-PAR2 failed to trigger such effects (Fig. 5B), further supporting the superior targeting efficiency of TAP/pCas9-PAR2. Moreover, AP/pCas9-PAR2 without CBP largely induced the half-reduction of mean fluorescence intensity or fluorescence area percentages compared to TAP/pCas9-PAR2 (Fig. 5B, Fig. S8), confirming that CBP assisted abundant TAP nanoparticles to deeply arrive at the damaged pulmonary sites via binding to pathological collagens. It also supported that TAP nanoparticles could trigger the abundant accumulations of pCas9-PAR2 and Cas9 positive cells for efficient gene editing in vivo. More importantly, we detected Cy5 fluorescence in the lung at different time points (0.5, 4, 8, 24, 48, and 72 h) after ALI mice injected with TAP nanoparticles for revealing the dynamic manner of plasmid delivery in lungs (Fig. 5C, Fig. S7D). Results uncovered that TAP nanoparticles could deliver pCas9-PAR2 to the lung at 0.5 h, while the fluorescence was largely decreased at 48 h and totally could not be detected at 72 h. It suggested that TAP/pCas9-PAR2 could rapidly arrive at the inflamed lung for initiating gene editing of CRISPR-Cas9, and its persistent time of 24∼48 h could guarantee sufficient PAR2 deficiency since it was enough for plasmid transfection. Therefore, TAP/pCas9-PAR2 with CBP targeting the increased collagen in lungs exhibited higher retention and targeting capacity in the inflamed lung, which is conducive to the accurate PAR2 gene editing and low side effects in vivo.

Figure 5.

TAP/pCas9-PAR2 evoked PAR2 deficiency to mitigate acute lung inflammation. (A) A schematical diagram of LPS+2f-induced acute lung inflammation model and nanoparticles administration; (B) The detection of Cy5-labelled nanoparticles co-located with collagen (green) in inflamed lungs of ALI mice injected with different treatments by an immunofluorescence assay. Cy5-DNA as a fluorescent substitution for pCas9-PAR2. Scale bar = 50 μm; (C) The Cy5 fluorescence was detected in the lung at different time points (0.5, 4, 8, 24, 48, and 72 h) after ALI mice injected with TAP nanoparticles; (D) The comparison of lung tissue images of LPS+2f-induced ALI mice treated with different nanoparticles; (E) The H&E staining of inflamed lung tissues. The scale bar = 100 μm (100 × ); (F, G) Gisam staining of cells and cell number count in the BALF; (H) Total protein content in the BALF; (I) The changes in body weight of ALI mice with different treatment groups. There was no statistical significance between AP/pCas9-PAR2 and TAP/pCas9-PAR2 group. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 vs. LPS+2f. ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 vs. pCas9-PAR2.

More importantly, it was observed that the LPS, LPS+2f and pCas9-PAR2 treatment group elicited pulmonary edema, and surfaces of lung tissues were congested and covered with blood clots (Fig. 5D). However, the lung of mice treated with TAP/pCas9-PAR2 was normal and smooth without congestion. Furthermore, the H&E analysis validated that lung tissues of LPS, LPS+2f, pCas9-PAR2 and AP/pCas9-PAR2 group were dense due to alveolar injury and inflammatory cell infiltration, whereas TAP/pCas9-PAR2 exhibited loose and regular lung tissues (Fig. 5E). More importantly, the in vivo anti-inflammatory effects of TAP/pCas9-PAR2 on LPS+2f-induced ALI was compared with a commercial carrier, Lipofectamine 3000, as a positive control (Supporting Information Fig. S9). Results showed that TAP/pCas9-PAR2 exhibited smoother lungs without any edema or congestion than the lipofectamine 3000 group while H&E results uncovered more normal and looser lung tissues in TAP/pCas9-PAR2 group, although both groups triggered potent anti-inflammatory responses. When the lungs of mice are injured, immune cells (macrophages, neutrophils, and lymphocytes) are mobilized to accumulate in the lungs to engulf bacteria and suppress inflammatory response, but they also induce an inflammatory infiltration. As shown in Fig. 5F‒G, LPS and 2f-LIGRL-NH₂ robustly increased cell number in BALF, while AP/pCas9-PAR2 and TAP/pCas9-PAR2 treatment largely abolished their effects compared to LPS+2f group. Moreover, due to a large amount of protein accumulation in inflamed lungs, the protein secretion content in mouse bronchoalveolar lavage fluid (BALF) can characterize the pathological status of ALI mice. As shown in Fig. 5H, the protein content of mice treated with LPS+2f was increased by 54% compared to LPS-induced ALI, indicating that PAR2 activation could enhance protein secretion conducive to acute lung inflammation. AP/pCas9-PAR2 and TAP/pCas9-PAR2 group were remarkably decreased by 32% and 40% in protein production in BALF compared to LPS+2f group. Alternatively, although mice with nanoparticle treatments showed slight weight loss, TAP/pCas9-PAR2 only exhibited the least weight loss (∼5%) among all groups and there was ∼20% weight loss in LPS+2f group (∗P < 0.05), suggesting that TAP/pCas9-PAR2 significantly improved body weight of ALI mice (Fig. 5I). We also detected the lung wet/dry weight ratio which was evaluated to indicate the pulmonary edema. However, 2f + LPS increased wet/dry weight ratio from ∼4.2 to 4.5 without statistical significance, and TAP/pCas9-PAR2 treatment had no obvious effect (Supporting Information Fig. S10). It was probably because the experiment time of our ALI model (48 h) was longer than ALI model at 24 h, initiating the partial transformation of pulmonary edema into injury which led to less observation in the lung edema. There was no significant difference of lung edema (wet/dry ratio) among all groups. Therefore, it firstly revealed that PAR2 activation contributed to ALI and TAP/pCas9-PAR2 possessed PAR2 deficiency to trigger anti-inflammatory effects in lung inflammatory diseases.

3.6. TAP/pCas9-PAR2 attenuated acute lung inflammation via NLRP3 inflammasome/IL-1β and iNOS signalling

For ALI, the massive cytokine releases are directly related to pneumonia and may worsen into the cytokine storm. The proinflammatory cytokine productions in BALF were detected to further investigate inflammatory status (Fig. 6A‒C). Compared with control, LPS and 2f-LIGRL-NH₂ group robustly enhanced TNF-α, IL-6, and IL-1β secretion by 20-, 8- and 5-fold, respectively, which were also more potent than LPS alone, indicating PAR2-mediated serious inflammatory infiltration in lungs. Meanwhile, TNF-α, IL-6 and IL-1β cytokine releases in TAP/pCas9-PAR2 group were significantly down-regulated by 66%, 85% and 65%, respectively, compared to LPS+2f group, preventing the deterioration of inflammatory response. Alternatively, key protein expressions in lung tissues were evaluated to verify molecular mechanism of the inhibitory effect of TAP/pCas9-PAR2 in inflammation in vivo. Importantly, lung tissues from TAP/pCas9-PAR2 group exhibited two cleavage bands with the cleavage efficiency of ∼22.0%, which was similar to lipofectamine 3000 group, suggesting the efficient in vivo gene editing of PAR2 (Supporting Information Fig. S11). As shown in Fig. 6D, E, and G, LPS and LPS+2f group significantly enhanced the protein levels of NLRP3 and iNOS, whereas TAP/pCas9-PAR2 markedly decreased PAR2 expression via CRISPR-Cas9 editing and subsequently alleviated NLRP3 and iNOS signalling. These results further supported PAR2-mediated inflammatory responses via NLRP3 inflammasome and iNOS signalling in vivo.

Figure 6.

TAP/pCas9-PAR2 mitigated cytokine releases and lung inflammation via NLRP3 inflammasome and iNOS signalling with low short-time systemic toxicity. (A–C) IL-6, TNF-α and IL-1β productions in the BALF from ALI mice with different nanoparticle administrations were examined by ELISA; (D, E) The protein expressions of NLRP3 and iNOS in lung tissues from ALI mice with different administrations; (F) The H&E staining of liver tissues from ALI mice with different administrations. The scale bar = 100 μm (100 × ); (G) PAR2 expressions in lung tissues from ALI mice with different administrations after 24 h; (H–J) The level of CRE, AST and ALT in serum from ALI mice with different administrations after 2 weeks. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 vs. LPS+2f. ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 vs. pCas9-PAR2. ns: no significant difference.

Finally, we investigated in vivo safety and biocompatibility of TAP/pCas9-PAR2 nanoparticles in healthy mice. According to the in vivo biodistribution results of nanoparticles (Supporting Information Fig. S7A‒S7C), apart from inflamed lung tissues, some TAP/pCas9-PAR2 were accumulated in liver and kidney tissues. We detected the possible toxicity of nanoparticles in liver and kidney. As shown in Fig. 6F, the H&E staining showed that there was no significant toxic damage in liver tissues after nanoparticles were intravenously injected into mice for 24 h. More importantly, the systematic toxicity of nanoparticles was also evaluated after two-week administration. As shown in Supporting Information Fig. S12, there was no obvious toxicity in heart, liver, spleen, lung and kidney tissues of TAP/pCas9-PAR2 group. TAP/pCas9-PAR2 did not exhibit significant differences from control and pCas9-PAR2 group in AST and ALT as serum liver function indicators and CRE as a renal function index (Fig. 6H‒J), further suggesting TAP/pCas9-PAR2 did not induce severe two-week toxicity. These results indicated that this HSA-based nanoparticle exhibited low in vivo short-term toxicity, suggesting its potential therapeutical application in treatment of inflammatory diseases.

4. Discussion

Due to the outbreak of COVID-19 pandemic, exploring forceful therapeutical strategies for pulmonary inflammation becomes increasingly emergent and concerned⁶. GPCRs account for ∼40% of molecular targets for approved clinical drugs, since they including PAR2 are widely distributed through human body and vigorously involve in a variety of physiological responses and diseases²⁹. Although the principal role of PAR2 in diseases is the robust modulation of inflammatory responses⁹, PAR2 has rarely been probed in lung inflammatory progression. Interestingly, it has been reported that proteases, such as TMPRSS2, cathepsin, elastase, and protease-induced tissue damages are highly associated with the pathogenesis of lung inflammatory diseases^4,30, while PAR2 is a membrane receptor primarily activated by proteases to evoke signal transductions⁸, suggesting possible relationship of PAR2 and pulmonary diseases. This work focused on revealing the novel role of PAR2 and underlying molecular mechanism in lung inflammatory diseases.

Unlike traditional PAR2 inhibitors in previous studies, we firstly constructed an all-in-one CRISPR-Cas9-PAR2 sgRNA plasmid (pCas9-PAR2) for investigating PAR2 blockade in diseases. CRISPR-Cas9 as a powerful and superior genome-editing technique can precisely evoke gene editing¹¹, while gene therapy exhibits abundant advantages over chemotherapeutic agents, such as extraordinary specificity and targeting^31,32, suggesting that PAR2 deficiency by CRISPR-Cas9 could be a novel candidate for PAR2-based therapeutics. However, a key restriction for gene therapy is the lack of a delivery system for drug safety and efficacy. While cationic nanoparticles have been widely applied for delivering CRISPR-Cas9 system, high toxicity still limited their applications^14,33,34. Apart from cationic nanoparticles, it has been proposed that the cell membrane is relatively dynamic and negatively-charged nanoparticles can be adhere to the positive-charged position of cells³⁵, indicating anionic nanoparticles also elevate cell uptake and transfection efficacy of gene drugs. Our study firstly utilized HSA-based nanoparticles to condense and encapsulate pCas9-PAR2 (TAP/pCas9-PAR2) with excellent particle size and stability, whereas pCas9-PAR2 itself exhibited large size and was readily degraded by DNase (Fig. 1). Moreover, TAP/pCas9-PAR2 successfully facilitated pCas9-PAR2 to enter and transfect inflammatory cells, consequently provoking PAR2 knockout by CRISPR-Cas9 in vitro and in vivo (Figure 2, Figure 6G, Fig. S11). Alternatively, mouse paw edema model as standard PAR2-mediated inflammatory model was used for verifying PAR2 deficiency by TAP/pCas9-PAR2 in vivo. TAP/pCas9-PAR2, but not pCas9-PAR2, robustly eliminated 2f-LIGRL-NH₂- or λ-carrageenan-induced paw edema, confirming that TAP/pCas9-PAR2 precisely evoked gene editing of PAR2 to prevent inflammatory responses in vivo. Therefore, we successfully developed anionic nanoparticles to deliver pCas9-PAR2 to initiate PAR2 deficiency by CRISPR-Cas9.

More importantly, this study revealed that PAR2 agonists could exacerbate LPS-induced acute lung inflammation, suggesting the participation of PAR2 activation (Fig. 5). To enhance drug targeting in inflamed sites, the collagen-binding peptide (CBP) was also grafted to HSA-based nanoparticles since massive collagen accumulated in inflamed tissues¹⁹. TAP/pCas9-PAR2 with CBP exhibited the highest inflamed lung-targeting efficiency compared to pCas9-PAR2 and AP/pCas9-PAR2 (Fig. 5B, C, and Figs. S7 and S8). Surprisingly, HSA-based nanoparticles without CBP (AP/pCas9-PAR2) also showed relatively high fluorescence of pCas9-PAR2 in lung, probably because it has been reported that supramolecular arrangement of proteins, such as HSA, in a nanoparticle structure can promote lung accumulation in ALI mouse, preferably triggering anti-inflammatory responses⁴. However, it is still not clear about the mechanism of inflamed lung-targeting capacity and anti-inflammatory effects of the HSA nanoparticle itself, which requires further explorations. Additionally, TAP/pCas9-PAR2 largely prolonged the half-time and enhanced peak plasma concentration of pCas9-PAR2, endowing its excellent bioavailability for potent gene editing of PAR2 (Fig. 3A, Table S5). With elevated pharmacokinetic properties, TAP/pCas9-PAR2 robustly promoted PAR2 knockout by CRISPR-Cas9 to alleviate lung inflammation of ALI mouse, playing a protective role in inflamed lungs (Fig. 5). By contrast, it has been reported that PAR2-mediated cAMP accumulation activated by thrombin acutely prevented TLR4-dependent Ca²⁺ signalling and inflammation in alveolar macrophages³⁶. It is probably because there are the Yin‒Yang roles of PAR2, pro-inflammatory and anti-inflammatory effects that are context-dependent (cell type, ligands and the environment⁹. Moreover, PAR2 deficiency by TAP/pCas9-PAR2 powerfully mitigated various proinflammatory cytokine secretions, including IL-6, TNF-α, IL-1β, in lungs with low short-term toxicity, revealing that it is conducive to preventing the deterioration of cytokines storm and lung inflammation (Fig. 6, Fig. S12).

Alternatively, we further explored the underlying mechanism of TAP/pCas9-PAR2-inhibited lung inflammation (Fig. 4). PAR2 deficiency by TAP/pCas9-PAR2 initiated transformation of M1 to M2 anti-inflammatory macrophages and suppressed abundant cytokine expressions in inflammatory cells (Figs. 4A and B, L, and 6A‒C). Importantly, we uncovered the active involvement of NLRP3 inflammasome signalling, since TAP/pCas9-PAR2 remarkably prevented PAR2-induced NLRP3 inflammasome, and subsequently alleviated its downstream Caspase-1 and IL-1β secretion in both inflammatory cells and ALI mouse (Figs. 4 and 6C–D). Although NLRP3 inflammasome as a cytoplasmic community signaling complex drastically participates in the pathogenesis of inflammatory diseases³⁷ and its activation can be modulated through some GPCRs, such as GRP40/120, MT1/2, EP4 and TGR5³⁸, there is a lack of studies on PAR2 and NLRP3 signal cascade. Indeed, our findings provided a novel direction for elucidating PAR2-mediated inflammation. It has been reported that ERK/2 signnalling was a classical PAR2-mediated pathway³⁹, and also participated in NLRP3 inflammasome activation⁴⁰. Consistently, TAP/pCas9-PAR2 dramatically suppressed PAR2-induced classical ERK1/2 signalling pathway (Fig. 4C, D, I, and J), suggesting the participation of ERK-mediated NLRP3/Caspase-1/IL-1β signalling. Additionally, PAR2 knockout by TAP/pCas9-PAR2 largely mitigated another inflammatory signalling, NO/iNOS pathway, through inhibiting PAR2-induced iNOS and NO productions (Fig. 4G, H, and Fig. S6). There are also emerging evidence that NLRP3 inflammasome and iNOS signalling cooperatively modulated various inflammatory responses⁴¹, supporting our results of ERK/NLRP3 inflammasome and NO/iNOS pathways involved in PAR2-mediated inflamm-ation.

5. Conclusions

This study constructed a novel anionic nanoparticle to encapsulate pCas9-PAR2, exhibiting superior stability and pharmacokinetic properties. TAP/pCas9-PAR2 remarkably elevated cell uptake and transfection efficiency of pCas9-PAR2, consequently evoking PAR2 deficiency in vivo and in vivo. Interestingly, we revealed a novel proinflammatory role of PAR2 in ALI, and TAP/pCas9-PAR2 initiated precise PAR2 knockout to suppress PAR2-induced ALI via ERK/NLRP3/Caspase-1/IL-1β signalling and NO/iNOS pathways (Fig. 7). Therefore, we firstly developed CRISPR-Cas9-based nanoparticles for gene editing of PAR2 as a gene therapy strategy with extraordinary anti-inflammatory effects in ALI, providing a novel value of PAR2 deficiency and underlying signalling mechanism for treating inflammatory diseases.

Figure 7.

Schematical diagram of TAP/pCas9-PAR2 alleviating ALI and related signaling mechanism.

Author contributions

Yuhong Jiang designed the research. Yuhong Jiang, Xin Zhuo, Yue Wu and Xiujun Fu carried out the experiments and performed data analysis. Jianbin Li, Yuxin Xiang, Xiaoyu Liang participated part of the experiments. Xin Zhuo drafted the manuscript and Yuhong Jiang edited the manuscript. Yuhong Jiang and Canquan Mao revised the manuscript. All of the authors have read and approved the final manuscript.

Conflicts of interest

The authors have no conflicts of interest to declare.

Appendix A. Supplementary data

The following is the Supplementary data to this article.