Biomimetic dual-crosslinked GelMA-SilMA hydrogel microneedle for delivering DSPE-PEG2K-TAT-modified quercetin liposomes to treat rheumatoid arthritis via the caspase-8/caspase-3/GSDME pathway Authors: Qingcong Zheng, Rongjie Lin, Rongsheng Chen, Weihong Xu Published: 2025 ABSTRACT Rheumatoid arthritis (RA) is a chronic inflammatory disease characterized by synovitis, whose pathogenesis remains unclear. Fibroblast-like synoviocytes (FLS) are central to the pathogenesis and progression of RA. This study aims to investigate the novel role of the pyroptosis mechanism mediated by the caspase-8/caspase-3/gasdermin-E (GSDME) pathway in RA-FLS, and further develop a transdermal drug delivery system (TDDS) to deliver quercetin (QUE) nanoparticles, which act on this pathway by targeting caspase-8. Specifically, we designed and constructed a 3D bioprinted bionic and architecture-mimetic double-crosslinked hydrogel dissolvable microneedle (MN) based on gelatin methacrylamide (GelMA) and silk fibroin methacryloyl (SilMA). This microneedle is used for transdermal delivery of quercetin liposomal nanoparticles modified with cell-penetrating peptide (CPP) (QUE-Lipo-TAT), so as to interrupt the vicious cycle between cell death and inflammation in RA-FLS. The results showed that caspase-8 is a potential therapeutic target for RA, and the novel caspase-8/caspase-3/GSDME pathway can mediate pyroptosis in RA-FLS. QUE-Lipo-TAT can not only block the caspase-8/caspase-3/GSDME axis by inhibiting caspase-8 via targeted action to prevent pyroptosis in RA-FLS, but also inhibit receptor-interacting protein kinase 3 (RIPK3) to prevent necroptosis in RA-FLS. Thereby, it suppresses inflammatory death, pro-inflammatory cytokine release, and disordered cell proliferation and migration of RA-FLS, achieving a “kill two birds with one stone” therapeutic effect. A novel hydrogel microneedle with bee stinger-mimetic and Oriental Pearl Tower-mimetic structures was prepared for the first time using GelMA-SilMA (GS) hydrogel. This microneedle possesses excellent biodegradability and mechanical properties, enabling it to easily penetrate the stratum corneum. Its favorable swelling performance and adhesiveness allow the double-sphere and single-ring microneedle structure to be firmly embedded in the skin. Additionally, the microneedle’s good degradability enables the sustained release of nanoparticles. QUE-Lipo-TAT delivered via this microneedle effectively alleviates hind paw swelling and improves joint conditions in collagen-induced arthritis (CIA) rats, reduces the levels of pro-inflammatory cytokines, inhibits synovial hyperplasia and bone and cartilage damage, and exhibits favorable biosafety. The novel target, new pathway, and new drug delivery method identified in this study provide a promising new approach for the precision treatment of RA.Graphical abstract Supplementary InformationThe online version contains supplementary material available at 10.1186/s12951-025-03831-y. Introduction Rheumatoid arthritis is a complex chronic inflammatory disease [1]. RA is characterized by persistent synovitis and progressive bone and cartilage damage, ultimately leading to widespread pain, chronic disability, and reduced life span [2, 3]. The prevalence of RA is approximately 1%, imposing a substantial global medical and economic burden [4]. In the pathogenesis of RA, fibroblast-like synoviocytes (FLS) in the synovial lining abnormally proliferate and transform into invasive RA-FLS. RA-FLS can participate in synovial pannus formation and bone and cartilage tissue destruction by secreting proinflammatory cytokines such as interleukin (IL)−1β and IL-18. These pathological behaviors show a significant positive correlation with the duration and severity of RA [5]. Currently, RA cannot be cured and is primarily managed with medication, including nonsteroidal anti-inflammatory drugs (NSAIDs), disease-modifying antirheumatic drugs (DMARDs), corticosteroids, and biologics. These drugs may cause adverse drug reactions of varying severity [2]. Additionally, oral administration results in insufficient bioavailability due to the first-pass effect, while injectable administration reduces patient compliance and carries a risk of inducing infection [6]. Therefore, we should delve into the novel pathogenic mechanisms of RA and develop corresponding drugs, while further improving drug delivery to achieve optimal therapeutic outcomes. The pathogenesis of RA remains unclear, and current views suggest that dysregulated programmed cell death (PCD) plays a significant role in rheumatic diseases [7]. Pyroptosis is a pro-inflammatory form of PCD mediated by the gasdermin protein family, characterized morphologically by membrane blebbing, cellular swelling, and the efflux of intracellular contents [8]. Pyroptosis in RA joint cells leads to plasma membrane rupture and the release of proinflammatory cytokines, serving as a key pathological mechanism driving local inflammatory responses [9]. Clinical studies have reported that high levels of caspases, IL-1β, IL-18, and lactate dehydrogenase (LDH) can be detected in the serum and synovial fluid of patients with RA [10]. Genome-wide association studies report that the caspase-8 encoding gene locus (rs6715284) is a RA risk loci [11]. Notably, caspase-8 acts as a molecular switch governing multiple forms of cell death, including pyroptosis and necroptosis [12]. Caspase-8 plays a crucial role in the canonical, non-canonical, and alternative pathways of pyroptosis [13]. The absence of caspase-8 leads to the activation of receptor-interacting protein kinase 3 (RIPK3) and mixed lineage kinase domain-like (MLKL), which mediate necroptosis. Therefore, necroptosis is a form of PCD negatively regulated by caspase-8 [14]. In summary, caspase-8 positively regulates pyroptosis, negatively regulates necroptosis, acts as an upstream positive regulator of IL-1β and IL-18, and serves as a potential therapeutic target for RA, and we are highly interested in investigating the role of the caspase-8/caspase-3/GSDME pathway in RA. Quercetin is a natural flavonoid compound with biological activities including anti-rheumatic, anti-inflammatory, analgesic, and hepatoprotective effects [15]. QUE demonstrates superior efficacy to methotrexate (MTX) in protecting rat joints from inflammatory damage [16]. QUE effectively alleviates arthritis by downregulating IL-1β levels and further suppresses inflammation when used in combination with MTX [17]. MTX causes elevated levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST), while the combination of QUE and MTX effectively reduces hepatotoxicity, demonstrating that QUE exhibits anti-inflammatory and hepatoprotective effects [18].In terms of molecular mechanisms, QUE inhibits the binding of ASC and caspase-8 to suppress inflammasome activation, thereby inhibiting pyroptosis in hepatic macrophages [19]. QUE can downregulate caspase-8 and caspase-3 levels, which in turn restores testicular function [20]. QUE inhibits RIPK3/MLKL-mediated necroptosis in oligodendrocytes following spinal cord injury to promote repair [21]. However, there are no relevant studies on the treatment of RA by QUE through its effects on pyroptosis and necroptosis pathways. Although QUE demonstrates promising anti-rheumatic potential and safety, its single-dose oral bioavailability is less than 3.6%, with poor solubility and bioavailability QUE can bind with sugar molecules to form derivatives such as gluconic acid, resulting in loss of biological activity. It readily precipitates in the gastrointestinal tract, further reducing bioavailability and significantly limiting its clinical application [22, 23]. As a non-invasive drug delivery mode, transdermal drug delivery systems (TDDS) enable effective absorption of drugs through the skin’s microcirculation. Applying QUE via TDDS strategies to RA holds greater potential for safer and more efficient treatment [24, 25]. The second-generation TDDS employs nano-carriers and drug modification strategies to significantly enhance drug penetration efficiency, while the third-generation TDDS utilizes microneedle strategies to achieve effective drug delivery [26]. This study aims to design and implement nano-carrier packaging and nanoparticle surface modification for QUE, while improving the microneedle system to achieve effective transdermal drug delivery. Liposomes effectively encapsulate QUE, not only enhancing the stability of the complex but also improving drug delivery and controlled release, significantly increasing solubility and bioavailability [27]. CPP can efficiently mediate interactions between nanoparticles and cell membranes, serving as effective passive permeation enhancers for transdermal drug delivery [28, 29]. Trans-activator of transcription (TAT) is a cationic CPP derived from HIV-1 that effectively mediates intracellular delivery of nanoparticles, significantly enhancing cellular uptake efficiency [30]. The TAT peptide alone lacks the hydrophobic structures necessary for binding to the liposomes membrane and cannot achieve stable anchoring; 1,2-Distearoyl-sn-glycero-3-phosphoethanolamine (DSPE) can embed into the liposomes membrane via hydrophobic interactions, enabling stable modification of TAT peptides on the carrier surface; hydrophilic polyethylene glycol (2000) (PEG2K) confers “stealth” properties to liposomes, evading immune clearance; The DSPE-PEG2K-TAT conjugate combines stability, prolonged circulation, and active penetration capabilities, enabling superior nanoparticle delivery [31]. Hydrogels derived from natural materials exhibit excellent substance exchange capability, biocompatibility, and biodegradability, making them ideal materials for delivering nanoparticles and developing microneedle structures [32]. Gelatin modified with methacrylamide groups forms gelatin methacrylamide (GelMA), which retains the inherent properties of the natural gelatin material and exhibits excellent biocompatibility. GelMA exhibits ideal thermosensitive properties, existing as a solid below 25 °C and as a gel above 37 °C. Additionally, GelMA exhibits low cytotoxicity, which helps maintain cell viability. However, GelMA degrades too rapidly and exhibits poor mechanical properties, rendering it unsuitable for fabricating microneedle with adequate mechanical performance [33, 34]. Given its substantial hydrogen bonding, and the high protein crystallinity provided by β-sheet crystals, silk fibroin exhibits superior in vivo stability and mechanical strength compared to other natural materials. Modified methacrylated silk fibroin (SilMA) can significantly enhance the mechanical properties and structural stability of hydrogels, effectively regulate drug release behavior, and reduce the burst release effect, making it an excellent microneedle support material. However, SilMA suffers from concentration limitations and low viscosity issues [35, 36]. The composite hydrogel formed by GelMA and SilMA can realize synergistic optimization in terms of moldability, mechanical properties, degradation rate, and swelling behavior, laying the material foundation for preparing hydrogel microneedle [37]. Microneedle can effectively enhance drug delivery efficiency and achieve sustained-release effects, improve patient treatment compliance, significantly reduce gastrointestinal reactions, and circumvent hepatic first-pass metabolism effects [38]. N should combine excellent convenience, degradability, and biocompatibility to effectively promote drug delivery and enhance the safety and efficiency of the absorption process [39]. Hydrogel microneedle feature a unique crosslinked structure that enables them to absorb interstitial fluid and expand upon insertion into the skin. Combined with their inherent adhesive properties, this allows MN to embed firmly within the tissue. The swelling hydrogel microneedle form a transport pathway with the capillary circulation, functioning as a rate-controlling membrane that effectively enhances drug permeation efficiency and bioavailability. Subsequently, the hydrogel microneedle can form an in situ hydrogel within the body, ensuring sustained release of nanoparticles without damaging the skin barrier or causing skin-related side effects. Additionally, hydrogel microneedle offer the advantages of painless treatment, easy sterilization, and effortless removal of the patch backing layer from the skin surface [40–42]. Traditional microneedle exhibit inadequate mechanical properties and poor skin fixation, particularly in frequently moving joint areas where they are prone to dislodgement or extrusion. Inspired by bee stingers, the micro-needle structure easily interlocks with tissue, preventing premature detachment that leads to insufficient drug delivery [43]. This study intends to design microneedle with a double-sphere and single-ring structure, which mimics both the bee stinger and the Oriental Pearl Tower. The spherical connection design of these microneedle can rely on their bionic properties to achieve skin anchoring and drug delivery effects. By virtue of the morphological advantages of the architectural-mimetic double-sphere structure, the microneedle enhance adhesion and interlocking with skin tissue, thereby facilitating nanoparticle drug delivery and improving drug delivery efficiency [44]. The two-photon polymerization process enables material polymerization via laser technology to form three-dimensional structures, which can be applied to construct microneedle arrays [45]. In conclusion, the purpose of this study is to conduct an in-depth investigation into caspase-8/caspase-3/GSDME as a novel pathway in RA. It aims to explore the impact of QUE-Lipo-TAT on RA by regulating pyroptosis and necroptosis of RA-FLS. The study will employ 3D bioprinting to fabricate microneedle with bionic and bio-inspired architectural structures. It will further investigate the therapeutic efficacy of dual-crosslinked hydrogel microneedle loaded with QUE-Lipo-TAT in CIA rats. This work intends to provide novel insights and robust experimental evidence for the basic research, clinical diagnosis, and treatment of RA (Fig. 1). Fig. 1Experimental schematic diagram. (A) the preparation process of the bionic and architecture-mimetic hydrogel microneedle patch; (B) the application of GelMA-SilMA double-crosslinked hydrogel microneedle for loading QUE-Lipo-TAT nanoparticles; (C) the use of GS hydrogel microneedle loaded with QUE-Lipo-TAT in the treatment of arthritis in CIA rats; and (D) the molecular mechanism by which QUE-Lipo-TAT treats RA by inhibiting pyroptosis mediated by the caspase-8/caspase-3/GSDME pathway and necroptosis mediated by RIPK3 in RA-FLS Experimental schematic diagram. (A) the preparation process of the bionic and architecture-mimetic hydrogel microneedle patch; (B) the application of GelMA-SilMA double-crosslinked hydrogel microneedle for loading QUE-Lipo-TAT nanoparticles; (C) the use of GS hydrogel microneedle loaded with QUE-Lipo-TAT in the treatment of arthritis in CIA rats; and (D) the molecular mechanism by which QUE-Lipo-TAT treats RA by inhibiting pyroptosis mediated by the caspase-8/caspase-3/GSDME pathway and necroptosis mediated by RIPK3 in RA-FLS Materials and methods Preparation of QUE-Lipo-TAT DSPE-PEG2K-TAT was synthesized according to the procedure illustrated in Figure S1. 100 mg DSPE-PEG2K-MAL was dissolved in 3 mL dimethylformamide, followed by the addition of 1.1 eq. of TAT peptide. The reaction mixture was stirred at room temperature for 12 h after complete dissolution. The mixture was transferred to a dialysis bag for dialysis in pure water for 24 h, and DSPE-PEG2K-TAT was obtained after lyophilization. QUE-Lipo-TAT was prepared using the thin-film hydration method. (1) Rotary evaporation. Accurately weigh 50 mg of soybean lecithin, 10 mg of cholesterol, 20 mg of DSPE-PEG2K-TAT, and 5 mg of QUE, and dissolve them together in 5 mL of chloroform. The mixture is then subjected to reduced-pressure evaporation in a sample vial under rotary evaporation conditions at 37 °C to form a thin film. (2) Hydration and sonication. Add 4 mL of pure water to the system for hydration. The mixture is then treated with sonication (power: 200 W, frequency: 20 kHz, interval: 2 s, duration: 20 s) and a liposomes extruder. After that, it is dialyzed using a nano-dialysis device for 3 h to remove residual drug, and pure water is added to adjust the volume to 5 mL, resulting in the QUE-Lipo-TAT nanoparticles. For the preparation of QUE-Lipo nanoparticles, DSPE-PEG2K was replaced with DSPE-PEG2K-TAT, and the same preparation method was adopted. A lyoprotectant was added to the synthesized nanoparticles, which were then stored at −20 °C for cryopreservation. Characterization and physicochemical properties of QUE-Lipo-TAT Determination of particle size, polydispersity index (PDI), and zeta potential Particle size distribution, PDI, and zeta potential of QUE-Lipo and QUE-Lipo-TAT were measured using NanoBrook 90Plus PALS. To prevent phospholipid degradation and ensure the uniform dispersion of liposome vesicles during the measurement, all tests were conducted within 12 h after the completion of liposome preparation, and the liposomes were always stored in a light-protected environment at 4 °C. Entrapment efficiency of QUE-Lipo-TAT The drug loading percentage (DL) and entrapment efficiency (EE) of the nanoparticles were determined using an ultraviolet (UV) spectrophotometric method. The relationship between the concentration and absorbance value of the QUE standard sample was determined, and a standard curve between absorbance and concentration was established. 0.1 mL of the sample solution was pipetted into 2.4 mL of methanol. The absorbance value of QUE was measured at its maximum absorption wavelength, and the QUE concentration in the diluted sample was obtained by substituting the absorbance into the standard curve. This concentration was then used to calculate the QUE concentration in the original sample solution. The mass of QUE in the sample was determined by multiplying the QUE concentration by the volume of the sample solution, which was subsequently substituted into the following formulas to calculate DL% and EE%.\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\text{D}\text{L}\left(\text{\%}\right)=\raisebox{1ex}{${\text{M}}_{\text{Q}}$}\!\left/\:\!\raisebox{-1ex}{${\text{M}}_{\text{T}}$}\right.\times\:100\text{\%}$$\end{document}\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\text{E}\text{E}\left(\text{\%}\right)=\raisebox{1ex}{${\text{M}}_{\text{Q}}$}\!\left/\:\!\raisebox{-1ex}{${\text{M}}_{0}$}\right.\times\:100\text{\%}$$\end{document} MQ represents the mass of QUE in the sample, MT denotes the total mass of the drug and the carrier, and M0 is the initial mass of quercetin in the system. Surface morphology The microstructure characteristics of the nanoparticles were observed using transmission electron microscopy (TEM). After the nanoparticles were sonicated in a water bath for 60 s, they were diluted and thoroughly mixed. A drop of the solution was then pipetted onto a carbon film-coated copper grid. After natural drying, the grid was stained with a 1% phosphotungstic acid solution for 2 min, followed by observation of the nanoparticles’ surface morphology. Fourier transform infrared (FTIR) spectroscopy analysis FTIR-8400 S was used to evaluate the interaction between QUE and the carrier. The fully dried sample was mixed with potassium bromide at a mass ratio of 1:100, followed by further grinding and tableting. Scanning and analysis were then performed at room temperature over a wavenumber range of 500–4000 cm⁻¹. Differential scanning calorimetry (DSC) analysis The samples were analyzed using HITACHI DSC200. An accurate mass of 5 ± 0.5 mg of each sample was weighed into an aluminum pan, which was then hermetically sealed and placed in the sample chamber of the DSC instrument. Under a nitrogen atmosphere, the temperature scanning range was set to 30–200 °C. The samples were held at a constant temperature of 30 °C for 4 min, followed by heating to 200 °C at a heating rate of 10 °C/min and maintaining this temperature for another 4 min. Experimental data were recorded and processed. Preparation of GelMA-SilMA composite hydrogel Synthesis of GelMA: Dissolve 10 g of gelatin in 100 mL of PBS and stir until completely dissolved. At 50 °C, add 10 mL of methacrylic anhydride dropwise at a rate of 0.5 mL/min to the solution while stirring thoroughly. Dilute with Dulbecco’s phosphate-buffered saline (DPBS) and filter to remove impurities. Dialyze the solution in distilled water for one week to obtain GelMA. SilMA was purchased from EFL company (EFL-SilMA-001). Preparation of double-crosslinked GS hydrogel: Weigh 2 g each of freeze-dried GelMA and SilMA separately and add to 20 mL of PBS. Stir and dissolve at 60 °C in the dark for 1 h. After complete dissolution, centrifuge at 37 °C (3000 rpm, 5 min) to obtain the GS hydrogel. Filter the reaction solution, then dialysis it. Freeze-dry the hydrogel and store it at −20 °C. Characterization of hydrogels Microstructure of hydrogels Scanning electron microscopy (SEM) was used to observe the internal structure and characteristics of the GS hydrogel. The GS hydrogel was treated with liquid nitrogen, followed by re-dehydration and fragmentation to obtain its cross-section. Place the sample on the sample stage. After spraying gold for 60 s, acquire the scanning image at a current of 10 mA. Swelling behavior of hydrogels Evaluate the swelling behavior of three hydrogels (G: S = 1:0, G: S = 1:1, and G: S = 0:1). The three hydrogel solutions were subjected to photopolymerization, freeze-drying, and weighing (W0), respectively. Immerse the sample at specified time intervals in PBS maintained at human body temperature (37 °C). After blotting the surface dry with filter paper, weigh the sample (W1). Conduct three replicate tests. The swelling rate formula is as follows:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\text{D}\text{e}\text{g}\text{r}\text{a}\text{d}\text{a}\text{t}\text{i}\text{o}\text{n}\:\text{r}\text{a}\text{t}\text{i}\text{o}\left(\text{\%}\right)=({\text{W}}_{1}-{\text{W}}_{0})/{\text{W}}_{0}\:\times\:100\text{\%}$$\end{document} Mechanical properties testing Weigh 400 mg of GS lyophilized pellets and add them in small portions to 2 mL of 0.25% LAP photoinitiator solution. Dissolve by shaking at 60 °C in the dark for 1 h, then centrifuge to remove bubbles, forming the GS hydrogel. After mixing, dispense 1 mL of GS hydrogel into the dumbbell-shaped standard sample mold. Photopolymerize by exposing to ultraviolet light at an intensity of 25 mW/cm² for 2 min. Select Stretch mode using the P/TG tensile probe of the Universal TA Texture Analyzer to perform a constant-speed mechanical tensile test. The P/TG probe moves at a speed of 0.1 mm/s with a trigger force of 0.049 N. Maintain the probe parallel to the film, gradually increase the displacement, and stop when the sample breaks during tensile testing. Plot the load-displacement curve and stress-strain curve according to Eqs. (1) and (2). Calculate Young’s modulus (E) using Eq. (3).1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\text{S}\text{t}\text{r}\text{e}\text{s}\text{s}=\text{F}/(\text{W}\times\:\text{T})$$\end{document}2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\text{S}\text{t}\text{r}\text{a}\text{i}\text{n}={\Delta\:}\text{L}/\text{L}$$\end{document}3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\text{E}=(\text{F}\times\:\text{L})/(\text{A}\times\:{\Delta\:}\text{L})$$\end{document} F represents the tensile force applied during the tensile test; W, L, and T denote the initial width, length, and thickness of the nanoparticle-loaded hydrogel sample, respectively; A is the cross-sectional area of the sample; and ΔL is the length deformation. In vitro drug release assay QUE-Lipo-TAT nanoparticles and GS hydrogels loaded with QUE-Lipo-TAT were immersed in pH 7.4 PBS solution, respectively. After shaking at 37 °C, samples were collected at predetermined time points (1, 2, 3, 6, 12, 24, 48, 72, 96, 120, 144 h). An equal volume of medium was added to the original solution. The absorbance of each supernatant sample was measured using a UV spectrophotometer. The cumulative drug release rate was calculated to establish a release curve. Modeling and fabrication of biomimetic and architecture-inspired hydrogel microneedle Microneedle structure design The design of the microneedle in this study draws inspiration from both bee stingers and the Oriental Pearl Tower. Combining the advantages of both, we employed 123D Design software to create a biomimetic, architecture-inspired microneedle featuring a double-sphere-and-ring configuration modeled after the Oriental Pearl Tower—a first-of-its-kind design. The individual microneedle structure consists of a base, a semicircular ring, a large sphere, a middle section, a small sphere, and an upper cone. The lower circle has a radius of 175 μm, the upper circle has a radius of 100 μm, and the height is 200 μm. The center of a semicircle with a radius of 25 μm is positioned at the top of the cylindrical surface, 100 μm vertically above the base. This semicircle is rotated to form a semicircular ring. The large sphere has a radius of 150 μm, with its center positioned 100 μm vertically above the center of the base circle. The lower radius of the middle section is 100 μm, the upper radius is 50 μm, and the height is 200 μm. The center of the lower circle in the middle section is located 100 μm vertically above the center of the large sphere. The radius of the small sphere is 75 μm. The center of the small sphere is positioned 50 μm vertically above the center of the upper central circle. The radius of the base circle of the upper cone is 50 μm. The needle tip diameter is ≤ 10 μm, with a height of 250 μm. The center of the base circle is located 50 μm vertically above the center of the sphere. The overall structure height is 950 μm, with a bottom circle diameter of 350 μm and a needle tip spacing of 850 μm. The microneedle array consists of 10 × 10 needles, with a groove depth of 2 mm. Each needle tip measures 10 μm in diameter, and the needle distribution area spans 8 × 8 mm. The overall patch dimensions are 12 × 12 mm. After completing the modeling, save the 3D structure in STL format. Preparation of microneedle mold Import the designed STL file into the two-photon printer to fabricate the microneedle positive mold using two-photon laser stereolithography technology. The material used was HTL-type 3D printing photopolymer resin. Subsequently, the main structure is immersed in silicone material to prepare a microneedle negative mold. Carefully peel away the main structure to obtain the reverse-replicated microneedle negative mold. Preparation of drug-loaded hydrogel microneedle GS hydrogel microneedle were prepared using the infusion centrifugation method. After uniformly dispersing QUE-Lipo-TAT nanoparticles, add them to the GS hydrogel solution and sonicate in a water bath for 3 min to promote loading and fusion. Inject the drug-loaded hydrogel uniformly into the microneedle array cavity of the silicone mold and process by centrifugation (4000 rpm, 37 °C, 5 min). After UV light curing, carefully remove the microneedle from the mold and store them at 4 °C. Characterization of drug-loaded hydrogel microneedle Microscopic structure of microneedle Observe the surface structural features of microneedle using a scanning electron microscope. Mechanical properties testing Evaluate the mechanical properties of drug-loaded microneedle using UTM-510. Secure the microneedle patch to the bottom platform and conduct mechanical property testing perpendicular to the platform. The probe descent speed was 1.2 mm/min, and testing ceased when a sustained pressure of 600 N was reached. In vivo dissolution experiment of microneedle GS MN loaded with QUE-Lipo-TAT were implanted into the hairless dorsal skin of rats. Samples were retrieved at predetermined time points (0, 2, 5, 10, 15, 20, 25, 30 min) for microscopic observation of microneedle dissolution. Plot the dissolution kinetics curve using the MN residual height ratio (Y = H/H₀, where H₀ was the initial microneedle length and H was the residual length) at each time point (X) as the parameter. Cytological experiments Cell culture and processing An RA environment was established in FLS using tumor necrosis factor-α (TNF-α). The treatment groups were administered QUE, QUE-Lipo, and QUE-Lipo-TAT, respectively. To overexpress caspase-8 in FLS, we pre-transfected the caspase-8 plasmid (Genechem, China) 48 h prior to drug stimulation. Each group performed routine cell culture. Western blot assay Collect cell samples and supernatants from each group, add lysis buffer to extract proteins, and determine protein concentration using the BCA method. Proteins were separated by SDS-PAGE and transferred to PVDF membranes. After blocking the membrane with blocking buffer, incubate overnight at 4 °C with primary antibodies. Primary antibodies include anti-caspase-8 (Abclonal, A0215), anti-caspase-3 (Abclonal, A11319), anti-GSDME (Abclonal, A7432), anti-GSDME-N (Abcam, ab215191), and anti-RIPK3 (Abclonal, A5431). The membrane was washed three times with TBST and incubated with the secondary antibody at room temperature for 1 h. After washing three times with TBST, protein bands were visualized using a Western blot detection system. Detection of pyroptosis using Hoechst 33,342/PI double staining Each group of cell slides was first washed three times by agitation with PBS. Following fixation with 4% paraformaldehyde at room temperature for 15 min, slides were washed three times by agitation with PBS. Hoechst 33,342 solution was added dropwise under light-protected conditions, incubated at room temperature in the dark for 5 min, and then washed three times by agitation with PBS. Continue adding PI solution dropwise under dark conditions. Incubate at room temperature in the dark for 20 min. Wash three times with PBS by gentle shaking. After air-drying, mount with a coverslip containing an anti-fluorescence quencher. Observe, count, and capture images under a fluorescence microscope. Throughout the process, ensure the dye remains protected from light to prevent fluorescence quenching. Enzyme-Linked immunosorbent assay (ELISA) IL-1β and IL-18 levels in the cell culture supernatants of each group were measured using a sandwich ELISA kit. After reaction termination, the optical density (OD) values for each group were determined by enzyme-linked immunosorbent assay at 450 nm within 10 min. Plot a standard curve and calculated the actual concentrations of IL-1β and IL-18 in the samples. Detection of LDH Collection of cell culture supernatants from each group, followed by LDH activity detection according to the LDH assay kit instructions. Cell viability assay Cell viability was determined for each group using the Cell Counting Kit-8 (CCK-8). Cells were uniformly seeded in a 96-well plate at a density of 6 × 10³ cells/well, and after the cells adhered, each group was treated with corresponding interventions. Subsequently, the cells were continuously cultured for 24 h, and then the CCK-8 assay was performed. Measured the optical density of each group at 450 nm using a microplate reader. Cell viability was calculated as follows:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\text{Cell\:viability\:(\%)\:=\:}\frac{{OD}_{s}-{OD}_{b}}{{OD}_{c}-{OD}_{b}}\times\:100\%$$\end{document} ODs represents the absorbance of cells incubated in the drug-treated group, ODc denotes the absorbance of cells cultured in the blank group, and ODb indicates the absorbance of the single medium. Flow cytometry Cells from each group were stained using the Annexin V-PE/FITC detection kit and analyzed by flow cytometry. After co-incubating cells with Cy5.5-labeled drugs, unbound drugs were removed by washing. Cells were collected and resuspended, and fluorescence signals in the Cy5.5 channel were detected. The mean fluorescence intensity (MFI) was used to analyze the level of drug uptake by cells. Transwell assay The migration of synovial cells was observed using the Transwell method. Suspended 5.0 × 10⁴ cells in serum-free DMEM medium and seeded into the upper chamber. Add reagents for each treatment group to the lower chamber. After 24 h of incubation, wash the upper chamber with PBS. Then fix the cells with 4% paraformaldehyde and stain with 0.1% crystal violet. Finally, observe under a microscope and capture representative images of migrating cells. Migration scratch assay Cells were seeded into 24-well plates and cultured until they were fully adherent and confluent. A horizontal scratch was made in the center of each well using the tip of a sterile pipette. After three PBS washes, interventions were performed according to each treatment group. Microscopic observations were conducted at different time points (0, 12, and 24 h), and the scratch area was analyzed using ImageJ software. Hemolysis assay Collect 0.5 mL fresh whole blood from healthy SD rats and prepare a 2% (v/v) red blood cell suspension. Treat the suspension with QUE-Lipo, QUE-Lipo-TAT, and GS extract + QUE-Lipo-TAT, respectively. Use saline as the negative control (0% hemolysis) and sterile water as the positive control (100% hemolysis). After incubation at 37 °C for 3 h, centrifuge the treated samples and measure the absorbance of the supernatant. Hemolysis rate = (Asample - Anegative control)/(Apositive control - Anegative control) × 100%. A represents the absorbance value. Hemolysis rate < 5% indicates the material possesses good biocompatibility. Animal experiments Establishing the CIA rat model Prior to conducting the animal experiments, all procedures were formally authorized by Fujian Medical University’s Institutional Animal Care and Ethics Committee, and the corresponding approval document number is IACUCFJMU2024-0229. A CIA rat model was established using a double-immune approach. Immune-grade bovine type II collagen solution (2 mg/mL) and complete Freund’s adjuvant solution (4 mg/mL) were mixed into an emulsion using a homogenizer under an ice bath. For the primary immunization, the emulsion mixture was injected intradermally at three random sites on the tail and back of rats. A booster immunization was performed by injecting the same emulsion mixture on day 7. All animals were housed under pathogen-free conditions and subjected to a standard feeding cycle. Intracorporeal treatment protocols Following successful establishment of the CIA rat model, rats were randomly assigned to the following five groups (n = 8): control group (HC), model group (CIA), QUE-Lipo-TAT Gel group, QUE MN group, and QUE-Lipo-TAT MN group. Starting from Day 21, QUE-Lipo-TAT Gel, QUE MN, and QUE-Lipo-TAT MN were applied topically to the reddened and swollen areas of the rat ankle joints every 3 days for a total of 7 applications. Body weight and right hind paw thickness of each group of rats were recorded at predetermined intervals, and clinical scores for all four limbs were documented according to the scoring criteria. Distribution of drugs in the body Cy5.5-QUE-Lipo-TAT Gel, Cy5.5-QUE MN, and Cy5.5-QUE-Lipo-TAT MN were applied topically to the reddened skin at the ankle joints of CIA rats. Fluorescence images were acquired using an IVIS imaging system at 3, 6, 12, 24, and 48 h post-administration. Assessment of pro-inflammatory cytokine levels in peripheral blood On day 42, collect peripheral blood from the posterior orbital vein of rats. After collecting the supernatant serum, detect the levels of IL-1β and IL-18 in rat serum according to the instructions of the ELISA kit. Histopathological analysis and immunohistochemical staining On day 42, the rats were sacrificed to collect the hind ankle joints. The joints were fixed in 4% paraformaldehyde solution, followed by decalcification in 15% (w/v) disodium ethylenediaminetetraacetate (EDTA-2Na) solution for three weeks. After paraffin embedding, hematoxylin-eosin (H&E) staining, Safranin-O Fast-Green staining, and Masson staining were performed. For further immunohistochemical staining, specific primary antibodies were used to detect the expression levels of caspase-8, caspase-3, and GSDME in the synovial tissue surrounding the rat ankle joints; CD90 staining was conducted to observe the abundance change of RA-FLS. The tissue sections were observed and photographed under an inverted fluorescence microscope. Biosafety analysis Blood samples were collected from the orbital venous plexus of rats. After serum separation, the levels of ALT, AST, blood urea nitrogen (BUN), and serum creatinine (CREA) in samples from each group were detected using an automatic biochemical analyzer. Rats were sacrificed by excessive carbon dioxide inhalation. The heart, liver, spleen, lungs, and kidneys were harvested, fixed in 4% paraformaldehyde solution, embedded in paraffin, and then subjected to H&E staining. Statistical analysis All data were expressed as mean ± standard deviation (SD). Differences between two groups were analyzed using a two-tailed Student’s t-test. For comparisons among multiple groups, one-way analysis of variance (ANOVA) was applied. A p-value less than 0.05 was considered statistically significant. Results Characterization of drug-loaded nanoparticle formulations The synthesized QUE-Lipo-TAT nanoparticle solution was clear in appearance, presenting a pale yellow and fluid state, and exhibited the Tyndall effect (Fig. 2A). From the UV-visible absorption spectrum, it can be seen that the maximum absorption wavelength of QUE in the QUE-Lipo-TAT sample is λmax = 354 nm, and the standard curve obtained between absorbance and concentration is y = 36.548x − 0.0991 (Fig. 2B).The calculated DL% and EE% of the QUE-Lipo-TAT nanoparticles are 6.4% and 97.4%, respectively. Average particle size, PDI, and zeta potential serve as critical parameters that play a vital role in the stable construction of nanoparticle dispersion systems. As shown in Fig. 2C-D, the average particle size of QUE-Lipo nanoparticles was 122.02 ± 3.84 nm, with a mean PDI of 0.167 ± 0.011 and a mean zeta potential of −23.74 ± 1.03 mV. The average particle size of QUE-Lipo-TAT nanoparticles was 130.69 ± 6.31 nm, with a mean PDI of 0.173 ± 0.013 and a mean zeta potential of −15.74 ± 0.815 mV. There is no significant difference in particle size and PDI between these two nanoparticles, but their zeta potentials exhibit distinct differences. This is attributed to the positive charge carried by TAT, indicating successful modification of DSPE-PEG2K-TAT onto the surface of QUE-Lipo nanoparticles. The specific dynamic light scattering data for the drug-loaded liposomes are shown in Table S1, and the detailed phase diagrams of the zeta potential are presented in Figure S1-S3. Transmission electron micrographs showed that QUE-Lipo (Fig. 2E) and QUE-Lipo-TAT (Fig. 2F) were generally distributed in a spherical shape, presenting a typical liposome morphology. The microscopic images were consistent with the particle size distribution range determined by DLS measurement. The modification of TAT increased the complexity of the liposome surface structure. Additionally, during long-term storage, QUE-Lipo-TAT showed no significant aggregation or degradation, demonstrating excellent stability. As shown in Fig. 2G, the FTIR spectrum of QUE exhibits characteristic spectral bands indicative of its molecular structure: the O-H stretching vibration was observed at 3293 cm⁻¹, the C = O aryl ketonic stretching absorption was at 1663 cm⁻¹, and the C-O-C antisymmetrical and symmetrical stretching vibrations were in the range of 1245–1000 cm⁻¹ [46]. The O-H stretching vibrations of QUE-Lipo and QUE-Lipo-TAT nanoparticles were observed at 3348 cm⁻¹ and 3339 cm⁻¹, respectively. Compared with QUE, the O-H peaks of the nanoparticles become broader and shift to higher wavenumbers. This may be caused by the formation of hydrogen bonds between QUE and the lipid molecules of the carrier. QUE-Lipo and QUE-Lipo-TAT exhibited C–H stretching vibrations at 2933 cm⁻¹ and 2925 cm⁻¹, respectively. This indicates that the drug is encapsulated in the liposome spheres. The C = O band of the aryl ketonic group in QUE (originally at 1663 cm⁻¹) shifted to higher wavenumbers, appearing at 1734 cm⁻¹ in QUE-Lipo and 1739 cm⁻¹ in QUE-Lipo-TAT. The stretching vibration of the C = O bond in the nanoparticles exhibits a broader peak and a higher wavenumber, which may be related to a decrease in the electron delocalization of the C = O bond. This further indicates enhanced hydrogen bonding or hydrophobic interaction between QUE and lipid molecules [47]. At the C-O-C bond in the 1000–1245 cm⁻¹ range, the characteristic peak of the liposomal nanoparticles is broader than that of free QUE, which strengthens the evidence that QUE has been successfully loaded into the liposomes. Furthermore, the intensity of the characteristic peaks of QUE-Lipo-TAT at each corresponding position is higher than that of QUE-Lipo. This indicates that TAT can be stably modified on the surface of QUE liposomal nanoparticles. In summary, these results demonstrate that QUE is successfully and stably encapsulated within the liposomal nanoparticles, while TAT is stably modified onto the nanoparticle surface. As can be seen from the DSC analysis in Fig. 2H, the endothermic peak temperature and enthalpy of QUE are 121.8 °C and 83.2 J/g, respectively, and this endothermic peak may correspond to its polymorphic transition or the process of losing bound water. The endothermic peak temperature and enthalpy of QUE-Lipo are 96.2 °C and 66.7 J/g, respectively. Compared with free QUE, the change in endothermic behavior indicates that the thermal properties of QUE have changed due to the interaction between QUE and liposomes, which also shows that QUE is encapsulated in liposomes in an amorphous form. The endothermic peak temperature and enthalpy of QUE-Lipo-TAT are 97.2 °C and 119.0 J/g, respectively. Compared with QUE-Lipo, the endothermic peak temperature of QUE-Lipo-TAT is similar, but the enthalpy increases significantly. This may be because TAT modification changes the surface structure and interaction mode of drug - loaded liposomes, thus leading to changes in the energy required for the thermal process. These results indicate that QUE-Lipo-TAT is successfully prepared and has good stability. Fig. 2Characterization of nanoparticles. (A) Tyndall effect of the QUE-Lipo-TAT nanoparticle solution; (B) UV-visible absorption spectrum curve of QUE-Lipo-TAT; (C) particle size measurement curves of QUE-Lipo and QUE-Lipo-TAT; (D) bar chart of the particle size, PDI, and zeta potential of the nanoparticles; (E) TEM image of QUE-Lipo; (F) TEM image of QUE-Lipo-TAT; (G) FTIR spectra of QUE and the nanoparticles; (H) DSC analysis chart of QUE and the nanoparticles. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 Characterization of nanoparticles. (A) Tyndall effect of the QUE-Lipo-TAT nanoparticle solution; (B) UV-visible absorption spectrum curve of QUE-Lipo-TAT; (C) particle size measurement curves of QUE-Lipo and QUE-Lipo-TAT; (D) bar chart of the particle size, PDI, and zeta potential of the nanoparticles; (E) TEM image of QUE-Lipo; (F) TEM image of QUE-Lipo-TAT; (G) FTIR spectra of QUE and the nanoparticles; (H) DSC analysis chart of QUE and the nanoparticles. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 Characterization of the double-crosslinked hydrogel As shown in Fig. 3A, this study employed dual crosslinking with GelMA and SilMA to obtain GS hydrogels, which were then used to load quercetin nanoparticles. As shown in Fig. 3B, the GS hydrogel exhibits a fluid state and transforms into a gel-like state after 2 min of light curing. This simple and rapid reaction facilitates clinical application. The frequent movement of knee, ankle, and wrist joints often leads to reduced adhesive strength of the patch, causing it to eventually fall off. Therefore, the patch must possess excellent adhesion and a level of flexibility compatible with skin tissue. When adhered to fingertips, the GS hydrogel exhibits excellent adhesion properties. Compression of the GS hydrogel causes no noticeable deformation, demonstrating good toughness. Figure 3C shows microstructural morphology images of GelMA, SilMA, and GS hydrogels after freeze-drying. It can be observed that incorporating SilMA into GelMA enhances the regularity of hydrogel pores, resulting in a highly porous honeycomb structure. Swelling properties are crucial in tissue engineering applications, as they directly affect the diffusion and mobility of solutes and mechanical performance. As shown in Fig. 3D, the GelMA (G), SilMA (S), and GelMA-SilMA (GS) hydrogels reached swelling equilibrium within 12 h, 8 h, and 10 h, respectively, with swelling ratios of 297.58 ± 15.47%, 88.63 ± 12.90%, and 191.87 ± 13.74%, respectively. The swelling rate of the composite hydrogel formed by adding SilMA to GelMA decreases, which may be attributed to SilMA reducing the water content in the solute [48]. Hydrogels used as tissue engineering scaffolds must possess appropriate biodegradability and degradation rates. As shown in the degradation curve in Fig. 3E, GelMA exhibits a high degradation rate, SilMA shows a low degradation rate, and GS lies between the two. This is because GelMA contains sequences recognizable by collagenase, making it prone to degradation and thus leading to excessively fast degradation [49]. In contrast, SilMA lacks such sensitive domains; therefore, the degradation time of the GS hydrogel immersed in the type II collagenase solution is longer than that of the GelMA hydrogel, which can appropriately extend the degradation duration [37]. This indicates that SilMA in the GS hydrogel can slow down the rapid degradation of GelMA, thereby forming a tightly connected network with a certain timeliness. This network can maintain a moist environment in the skin for a relatively long time and possess sufficient degradability. As shown in the mechanical test results in Fig. 3F-G, the GS hydrogel (tested immediately after photocrosslinking) exhibited an average tensile strength of 0.432 N, a stiffness of 0.05 N/mm, and a Young’s modulus of 288.482 KPa. After being stored at 37 °C for 48 h, the GS hydrogel had an average tensile strength of 18.359 N, a stiffness of 71.770 N/mm, and a Young’s modulus of 328.72 MPa. These results indicate that the stiffness and hardness of the GS hydrogel increased significantly after 48 h of storage. In conclusion, the GS hydrogel stored for 48 h can provide a favorable texture foundation for penetrating the skin and an appropriate degradation duration for the further preparation of bionic microneedle to be inserted into the skin. Fig. 3Characterization of the double-crosslinked GS hydrogel. (A) Schematic diagram of preparing the double-crosslinked GS hydrogel loaded with QUE-Lipo-TAT; (B) gelation process, adhesiveness, and compressive deformation of the GS hydrogel; (C) SEM images of GelMA, SilMA, and the GS hydrogel; (D) swelling curve; (E) degradation curve; (F) stress-strain curve of the GS hydrogel; (G) tensile test for the stiffness and Young’s modulus of the GS hydrogel. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 Characterization of the double-crosslinked GS hydrogel. (A) Schematic diagram of preparing the double-crosslinked GS hydrogel loaded with QUE-Lipo-TAT; (B) gelation process, adhesiveness, and compressive deformation of the GS hydrogel; (C) SEM images of GelMA, SilMA, and the GS hydrogel; (D) swelling curve; (E) degradation curve; (F) stress-strain curve of the GS hydrogel; (G) tensile test for the stiffness and Young’s modulus of the GS hydrogel. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 Preparation and characterization of biomimetic and architecture-inspired microneedle Inspired by bee stingers and the architectural structure of China’s Oriental Pearl Tower, this study designed a bionic, architecture-inspired double-sphere microneedle structure (Figure S4). The drug-loaded microneedle patches were prepared using the micromolding method. A microneedle model was designed using 123D Design. Microneedle patches were fabricated via 3D printing with resin materials, and silicone microneedle molds were obtained using the reverse molding technique. The GS hydrogel loaded with nanoparticles was introduced into the silicone molds, followed by ultraviolet photocuring. After demolding, the designed microneedle patches were obtained (Fig. 4A). Figure 4B shows the overall view of the resin microneedle positive mold, the structure of a single microneedle, and the multi-microneedle array, which are consistent with the designed single microneedle and microneedle patch. Figure S5 shows the silicone microneedle mold, which has good elasticity and allows easy demolding of the hydrogel. As shown in Fig. 4C, the hydrogel microneedle patch after demolding is presented; when the patch is bent, its good flexibility can be observed. Figure 4D shows a magnified view of the hydrogel microneedle array, which exhibits uniform alignment and intact structure. The SEM image in Fig. 4E shows a uniformly arranged microneedle array with a needle height of 950 μm. The diameters of the small sphere, large sphere, and base are 200 μm, 300 μm, and 350 μm, respectively. A buffer ring is positioned between the large sphere and the base, with a needle tip spacing of 850 μm. To ensure microneedle can effectively deliver drugs into the skin, pressure testing of drug-loaded hydrogel microneedle is required. As shown in Fig. 4F, the MN patch stored at 37 °C for 24 h deformed under a displacement of 250 μm, with a maximum load-bearing capacity of 0.588 N/needle. In contrast, the MN patch after 48 h of storage deformed under a displacement of 350 μm, and their maximum load-bearing capacity reached 0.955 N/needle. Interestingly, 250 μm and 350 μm correspond to the distance from the needle tip to the center of the top of the small sphere and the distance from the needle tip to the center of the bottom of the small sphere in the Oriental Pearl Tower structure, respectively. The threshold required for microneedle to penetrate the skin is 0.098 N/needle; therefore, the microneedle stored for 48 h can easily penetrate the stratum corneum and enter the skin structure. GelMA can provide adhesiveness. However, microneedle patch using GelMA alone degrade rapidly after being implanted in the body. SilMA can make up for the poor mechanical properties and rapid degradability of GelMA, and the two can be uniformly cross-linked together. Figure 4G shows the dissolution of drug-loaded microneedle patches in rat skin. The results indicate that the microneedle patches loaded with nanoparticles almost completely lost their height within 30 min, turning into a colloidal state and entering the skin. They exhibit an appropriate morphological dissolution duration, and the hydrogel patch base can be removed after 30 min of application. Fig. 4Design and characterization of hydrogel microneedle. (A) Flowchart of the 3D-designed drug-loaded bionic and architecture-mimetic hydrogel microneedle; (B) positive mold of microneedle prepared by a 3D printer; (C) Bending state of the hydrogel microneedle patch; (D) Enlarged view of the hydrogel patch; (E) SEM images of the hydrogel microneedle; (F) pressure-displacement curve; (G) Dissolution images and dissolution curves of hydrogel microneedle at different time points after in vivo insertion Design and characterization of hydrogel microneedle. (A) Flowchart of the 3D-designed drug-loaded bionic and architecture-mimetic hydrogel microneedle; (B) positive mold of microneedle prepared by a 3D printer; (C) Bending state of the hydrogel microneedle patch; (D) Enlarged view of the hydrogel patch; (E) SEM images of the hydrogel microneedle; (F) pressure-displacement curve; (G) Dissolution images and dissolution curves of hydrogel microneedle at different time points after in vivo insertion TNF-α mediates pyroptosis in RA-FLS by activating the caspase-8/caspase-3/GSDME signaling pathway As shown in Fig. 5A, with increasing TNF-α concentrations from 0 to 100 ng/ml, the expression levels of caspase-8, caspase-3, GSDME, and GSDME-N significantly increased in RA-FLS. ELISA detection revealed that the expression levels of IL-1β and IL-18 also increased with rising TNF-α concentrations (Fig. 5B). In conclusion, TNF-α at a concentration of 100 ng/ml significantly increased the expression levels of pyroptosis-related proteins, including caspase-8, caspase-3, GSDME, GSDME-N, IL-1β, and IL-18, in RA-FLS. The flow cytogram in Fig. 5C shows that in the absence of TNF-α stimulation, cells predominantly cluster in the third quadrant, indicating that most cells are in a healthy state (Annexin V-, PI-) with a low proportion undergoing pyroptosis. Under stimulation with 50 ng/mL TNF-α, the proportion of cells undergoing pyroptosis significantly increased. A marked rise in early pyroptotic cells (Annexin V+, PI-) was observed, alongside a subset of cells progressing to late pyroptosis (Annexin V+, PI+). This indicates that TNF-α has initiated a relatively potent cell death signaling pathway. As TNF-α concentration further increased to 100 ng/mL, the proportion of RA-FLS undergoing pyroptosis again significantly rose, with more cells entering the late pyroptosis (Annexin V+, PI+) zone. This indicates that TNF-α at this concentration greatly promotes the process of pyroptosis. In summary, TNF-α significantly induces pyroptosis in RA-FLS, and the proportion of pyroptotic cells progressively increases with rising TNF-α concentrations. This further demonstrates that TNF-α can induce cell death and lysis in RA-FLS via the pyroptosis pathway, thereby promoting a cascade of inflammatory responses in RA. The fluorescence image of the Hoechst/PI staining in Fig. 5D reveals that the number of PI-positive cells (cells undergoing pyroptosis) in RA-FLS significantly increases with rising TNF-α concentrations, particularly at 100 ng/mL where the proportion of pyroptotic cells reaches its peak. This indicates that TNF-α effectively induces pyroptosis in RA-FLS, and this effect increases with higher concentrations of TNF-α. Figure 5E TEM images show FLS without TNF-α stimulation maintaining normal morphology and structure. Fig. 5(A) Representative immunoblots and bar charts of caspase-8, caspase-3, GSDME, and GSDME-N in groups treated with different concentrations of TNF-α; (B) ELISA detection of IL-1β and IL-18; (C) flow cytometry analysis plots; (D) Hoechst 33,342/PI double staining; (E) TEM images of cells; (F) Representative immunoblots and bar charts of caspase-8, caspase-3, GSDME, and GSDME-N in FLS with caspase-8 overexpression; (G) Immunoblots and bar charts of the rescue assay. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 (A) Representative immunoblots and bar charts of caspase-8, caspase-3, GSDME, and GSDME-N in groups treated with different concentrations of TNF-α; (B) ELISA detection of IL-1β and IL-18; (C) flow cytometry analysis plots; (D) Hoechst 33,342/PI double staining; (E) TEM images of cells; (F) Representative immunoblots and bar charts of caspase-8, caspase-3, GSDME, and GSDME-N in FLS with caspase-8 overexpression; (G) Immunoblots and bar charts of the rescue assay. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 Below are RA-FLS stimulated by TNF-α, showing obvious swelling, irregular cell contours, formation of numerous bubbles, cell membrane rupture, and release of cell debris into the extracellular environment. All these are typical ultrastructural features of pyroptosis. The above results indicate that TNF-α, as a key factor in the pathogenesis of RA, promotes pyroptosis in RA-FLS and enhances inflammatory responses by activating the caspase-8/caspase-3/GSDME pathway. Validation of the caspase-8/caspase-3/GSDME signaling pathway in RA-FLS Figure 5F results show that RA-FLS with OE-caspase-8 constructed by plasmid transfection can effectively overexpress caspase-8 (P < 0.0001). After caspase-8 overexpression, the expressions of caspase-3, GSDME and GSDME-N also increased significantly. This further supports the important role of caspase-8 in the pyroptosis pathway. In addition, the activation of caspase-3 and the increase of GSDME-N, a cleavage product of GSDME, indicate that caspase-8 induces pyroptosis by activating caspase-3 and then initiating the cleavage of GSDME. The Z-IETD-FMK inhibitor specifically inhibits caspase-8 and significantly inhibits the occurrence of this process. This indicates that caspase-8 is located upstream of the caspase-3/GSDME pathway and is involved in the initiation of pyroptosis. The rescue experiment is shown in Fig. 5G. The first group consists of RA-FLS stimulated by TNF-α, and the protein expressions of caspase-3, GSDME and GSDME-N in these cells are upregulated. The second group comprises TNF-α-stimulated RA-FLS transfected with the OE-caspase-8 plasmid, and it shows that the protein expression levels of caspase-3, GSDME, and GSDME-N are significantly higher than those in the first group (P < 0.0001). The third group consists of RA-FLS treated with TNF-α after the addition of Z-DEVD-FMK, a specific inhibitor of caspase-3. The expressions of caspase-3, GSDME and GSDME-N in these cells are significantly inhibited (P < 0.0001). The fourth group involves TNF-α treatment of OE-caspase-8 RA-FLS after the addition of Z-DEVD-FMK. It shows that the protein expression levels of caspase-3, GSDME and GSDME-N are significantly lower than those in the second group but significantly higher than those in the third group (P < 0.0001). The rescue experiment shows that caspase-8 promotes the transmission of the caspase-3/GSDME signaling pathway by activating caspase-3. It verifies the upstream role of caspase-8 in the caspase-3/GSDME signaling pathway and confirms the existence of the caspase-8/caspase-3/GSDME pyroptosis signaling pathway in RA-FLS. Quercetin inhibits pyroptosis of RA-FLS through the caspase-8/caspase-3/GSDME axis, and QUE-Lipo-TAT exerts a stronger inhibitory effect Figure 6A shows that compared with the control group, the TNF-α-stimulated group has significantly higher expressions of caspase-8, caspase-3, GSDME and GSDME-N (P < 0.0001). Compared with the TNF-α group, the QUE-treated group can significantly reduce the expressions of caspase-8, caspase-3, GSDME and GSDME-N to inhibit pyroptosis (P < 0.0001). Treatment with the drug-free Lipo-TAT alone does not show an inhibitory effect on pyroptotic proteins (P > 0.05). When QUE-Lipo-TAT treats TNF-α-stimulated RA-FLS, it can significantly reduce the expression levels of caspase-8, caspase-3, GSDME and GSDME-N (P < 0.0001). Moreover, compared with the QUE-treated group, QUE-Lipo-TAT has a stronger ability to reduce the protein expressions of caspase-8, caspase-3, GSDME and GSDME-N (P < 0.0001). These results indicate that the nanoparticle encapsulation of QUE-Lipo-TAT can significantly enhance the biological effect of QUE in RA-FLS and effectively inhibit the activation of the pyroptosis signaling pathway. ELISA detection shows that the expression levels of IL-1β and IL-18 in each group are consistent with the expression trends of caspase-8, caspase-3, GSDME and GSDME-N in the corresponding groups. Compared with QUE, QUE-Lipo-TAT exerts a stronger inhibitory effect on the secretion of IL-1β and IL-18 by RA-FLS (P < 0.0001) (Fig. 6B). Figure 6C shows the results of LDH assay. Compared with the control group, the LDH release in the TNF-α group is significantly increased, which indicates that TNF-α impairs the membrane integrity of RA-FLS, increases cell membrane permeability, and thereby induces LDH release (P < 0.0001). Compared with the TNF-α group, QUE treatment can significantly reduce the level of LDH, which indicates that QUE effectively inhibits the occurrence of cell membrane damage during pyroptosis (P < 0.0001). In addition, the LDH release in the QUE-Lipo-TAT treatment group is significantly lower than that in the QUE group (P < 0.0001). This indicates that QUE-Lipo-TAT can effectively enhance the uptake of QUE by RA-FLS, exert a stronger cytoprotective effect, and effectively reduce cell membrane damage. The CCK-8 assay results of each group at different time points (6 h, 12 h, 24 h, 36 h, 48 h, 72 h) show that compared with the control group, the cell viability of the TNF-α group (RA group) is significantly reduced, which indicates that TNF-α stimulation causes damage and death of RA-FLS. Treatment with QUE and QUE-Lipo-TAT can significantly increase cell viability, and QUE-Lipo-TAT exhibits a stronger effect than QUE. This suggests that QUE and QUE-Lipo-TAT can effectively protect RA-FLS from death, and QUE-Lipo-TAT further enhances its anti-pyroptosis effect by improving the intracellular delivery and bioavailability of QUE. In addition, the relative difference trend among the groups remains consistent over time, which further verifies that QUE-Lipo-TAT can effectively and sustainably promote the viability recovery of RA-FLS (Fig. 6D). Figure 6E shows the flow cytometry analysis results, indicating that after TNF-α treatment, the proportion of Annexin V-FITC/PI double-positive RA-FLS increases significantly (P < 0.0001), which suggests that TNF-α can effectively induce pyroptosis. QUE and QUE-Lipo-TAT significantly reduce the proportion of Annexin V-FITC/PI double-positive cells (P < 0.0001), and QUE-Lipo-TAT exerts a stronger effect than QUE (P < 0.0001). This indicates that they have obvious cytoprotective effects, and QUE-Lipo-TAT more effectively inhibits pyroptosis and protects cells from damage than QUE. The transwell assay results in Fig. 6F show that compared with the control group, the migration ability of RA-FLS in the TNF-α group is significantly enhanced (P < 0.0001), indicating that TNF-α can promote the migration of RA-FLS. Both QUE and QUE-Lipo-TAT can significantly reduce the migration ability of the cells (P < 0.0001), and QUE-Lipo-TAT exhibits a stronger effect than QUE (P < 0.01). These results indicate that QUE-Lipo-TAT can more effectively inhibit the migration ability of RA-FLS in an inflammatory environment by enhancing the intracellular uptake of QUE. Figure 6G shows that compared with the control group, the scratch wound healing rate of RA-FLS treated with TNF-α is significantly increased at the 12 h and 24 h observation time points (P < 0.0001), indicating that TNF-α stimulation promotes the migration of RA-FLS. After treatment with QUE and QUE-Lipo-TAT, the scratch wound healing rate of RA-FLS is significantly lower than that in the TNF-α group (P < 0.0001), and QUE-Lipo-TAT exhibits a stronger effect than QUE (P < 0.0001). This indicates that both QUE and QUE-Lipo-TAT can effectively inhibit TNF-α-induced migration of RA-FLS, and QUE-Lipo-TAT significantly enhances the ability to inhibit cell migration by improving the bioavailability of QUE. The immunofluorescence staining results in Fig. 7A-C show that, compared with the control group, the fluorescence signals of caspase-8, caspase-3, and GSDME in RA-FLS of the TNF-α group are strong, with their expressions significantly increased. This indicates that TNF-α can effectively induce the expression of pyroptosis-related proteins in RA-FLS to promote pyroptosis. In the QUE and QUE-Lipo-TAT treatment groups, the fluorescence signals of caspase-8, caspase-3, and GSDME are significantly weakened, and the inhibitory effect of QUE-Lipo-TAT is significantly stronger than that of QUE. The results indicate that both QUE and QUE-Lipo-TAT can significantly reduce the levels of proteins related to pyroptosis, and QUE-Lipo-TAT more significantly downregulates the expressions of caspase-8, caspase-3, and GSDME compared with QUE to inhibit the pyroptosis process, thereby blocking the vicious cycle between inflammation and cell death. Fig. 6(A) Representative immunoblots and bar charts of caspase-8, caspase-3, GSDME, and GSDME-N in each group; (B) ELISA detection of IL-1β and IL-18; (C) LDH assay; (D) CCK-8 cell viability assay; (E) Flow cytometry analysis plots; (F) Transwell assay; (G) Wound healing assay. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 (A) Representative immunoblots and bar charts of caspase-8, caspase-3, GSDME, and GSDME-N in each group; (B) ELISA detection of IL-1β and IL-18; (C) LDH assay; (D) CCK-8 cell viability assay; (E) Flow cytometry analysis plots; (F) Transwell assay; (G) Wound healing assay. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 Fig. 7Immunofluorescence staining of proteins. (A-C) From top to bottom, the images correspond to the immunofluorescence staining of caspase-8, caspase-3, and GSDME in each group, respectively Immunofluorescence staining of proteins. (A-C) From top to bottom, the images correspond to the immunofluorescence staining of caspase-8, caspase-3, and GSDME in each group, respectively Cell-penetrating peptide can enhance the inhibitory effect of nanoparticles on pyroptosis in RA-FLS This study further investigates whether TAT modification enhances the cellular transduction and endocytosis of QUE-Lipo. Figure 8A shows that compared with the TNF-α group, the QUE-Lipo and QUE-Lipo-TAT groups significantly reduce the expressions of caspase-8, caspase-3, GSDME, and GSDME-N in RA-FLS (P < 0.0001), indicating that nanoparticles can effectively inhibit the pyroptosis pathway activated by TNF-α. Compared with QUE-Lipo, QUE-Lipo-TAT exhibits a significantly enhanced inhibitory effect on caspase-8, caspase-3, GSDME, and GSDME-N (P < 0.0001). This indicates that TAT modification can significantly improve the cellular transduction and endocytosis of QUE-Lipo, enhance the inhibitory effect of QUE on pyroptosis, and increase the therapeutic potential of nanoparticles. The ELISA results showed that compared with the TNF-α group, QUE-Lipo and QUE-Lipo-TAT significantly reduced the expression of IL-1β and IL-18 (P < 0.0001), and QUE-Lipo-TAT exhibited a stronger effect than QUE-Lipo (P < 0.0001) (Fig. 8B). This indicates that the nanoparticles can alleviate the inflammatory response by reducing pro-inflammatory cytokines in RA-FLS, and TAT modification can significantly enhance the efficacy of QUE-Lipo. The LDH assay results in Fig. 8C show that compared with the TNF-α group, the LDH content in the QUE-Lipo and QUE-Lipo-TAT groups is significantly reduced (P < 0.0001). This indicates that nanoparticles can effectively inhibit pyroptosis-induced cell membrane rupture and reduce LDH release. The effect of QUE-Lipo-TAT is stronger than that of QUE-Lipo (P < 0.0001), which further confirms the advantage of TAT-modified QUE-Lipo in inhibiting cell damage and pyroptosis. In conclusion, TAT-modified QUE-Lipo can further enhance the cellular uptake of QUE-Lipo, significantly strengthen the inhibitory effect on the pyroptosis pathway, reduce the release of inflammatory factors, alleviate cell damage, and thereby exert a cytoprotective effect. The caspase-3/mitochondrial staining results in Fig. 8D show that compared with the TNF-α group, QUE-Lipo and QUE-Lipo-TAT can significantly reduce the high expression of caspase-3 induced by TNF-α stimulation and increase mitochondrial membrane potential, with QUE-Lipo-TAT exhibiting a stronger effect than QUE - Lipo. This indicates that QUE-Lipo and QUE-Lipo-TAT can protect cells by inhibiting the pyroptosis pathway activated by TNF-α and mitochondrial damage, and TAT modification can significantly enhance the effect of QUE-Lipo. The immunofluorescence staining results in Figs. 8E-G show that compared with the TNF-α group, the fluorescence signals of caspase-8, caspase-3, and GSDME in the QUE-Lipo and QUE-Lipo-TAT groups are significantly weakened, and their protein expressions are significantly reduced, with QUE-Lipo-TAT exhibiting a stronger inhibitory effect. This indicates that TAT-modified QUE-Lipo significantly enhances the uptake of nanoparticles by RA-FLS and more effectively inhibits the expressions of caspase-8, caspase-3, and GSDME to prevent pyroptosis. Fig. 8(A) Representative immunoblots and bar charts of caspase-8, caspase-3, GSDME, and GSDME-N in each group; (B) ELISA detection of IL-1β and IL-18; (C) LDH assay; (D) Staining images of caspase-3 and mitochondria; (E-G) From top to bottom, the images correspond to the immunofluorescence staining of cytoskeleton, caspase-8, caspase-3, and GSDME in each group, respectively; (H) Flow cytometry to measure the mean fluorescence intensity of Cy5.5-QUE-Lipo and Cy5.5-QUE-Lipo-TAT in cells; (I) Fluorescence microscopy images and fluorescence intensity of cellular uptake of Cy5.5-QUE-Lipo and Cy5.5-QUE-Lipo-TAT; (J) In vitro hemolysis assays; (K) Immunoblots and bar charts of the rescue assay. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 (A) Representative immunoblots and bar charts of caspase-8, caspase-3, GSDME, and GSDME-N in each group; (B) ELISA detection of IL-1β and IL-18; (C) LDH assay; (D) Staining images of caspase-3 and mitochondria; (E-G) From top to bottom, the images correspond to the immunofluorescence staining of cytoskeleton, caspase-8, caspase-3, and GSDME in each group, respectively; (H) Flow cytometry to measure the mean fluorescence intensity of Cy5.5-QUE-Lipo and Cy5.5-QUE-Lipo-TAT in cells; (I) Fluorescence microscopy images and fluorescence intensity of cellular uptake of Cy5.5-QUE-Lipo and Cy5.5-QUE-Lipo-TAT; (J) In vitro hemolysis assays; (K) Immunoblots and bar charts of the rescue assay. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 Further quantitative cell uptake studies were conducted. Figure 8H shows that compared to the control group (Cy5.5-PBS), the signal peaks of Cy5.5-QUE-Lipo and Cy5.5-QUE-Lipo-TAT exhibit a significant rightward shift, with the peak of Cy5.5-QUE-Lipo-TAT positioned further to the right than that of Cy5.5-QUE-Lipo. The MFI values for Cy5.5-QUE-Lipo and Cy5.5-QUE-Lipo-TAT were significantly higher than those of the control group, and Cy5.5-QUE-Lipo-TAT exhibited a significantly stronger MFI than Cy5.5-QUE-Lipo (P < 0.0001). This indicates that both Cy5.5-QUE-Lipo and Cy5.5-QUE-Lipo-TAT were uptaken by RA-FLS, with Cy5.5-QUE-Lipo-TAT showing greater uptake. Fluorescent staining in Fig. 8I clearly shows that Cy5.5-QUE-Lipo and Cy5.5-QUE-Lipo-TAT accumulate in the cytoplasm surrounding the cell nucleus, suggesting endocytosis of the liposome nanoparticles by RA-FLS. Moreover, the fluorescence signal of Cy5.5-QUE-Lipo-TAT was significantly stronger than that of Cy5.5-QUE-Lipo (P < 0.0001), indicating that RA-FLS exhibits a higher uptake of Cy5.5-QUE-Lipo-TAT compared to Cy5.5-QUE-Lipo. The above results demonstrate that TAT modification enhances the cellular uptake of QUE-Lipo. Additionally, hemolysis assays were conducted to evaluate the toxic effects of different materials on red blood cells. As shown in Fig. 8J, the hemolysis rates induced by QUE-Lipo, QUE-Lipo-TAT, and GS extract + QUE-Lipo-TAT were 2.33%, 2.03%, and 2.08%, respectively, all below 5% and showing no significant hemolysis. This indicates that the prepared QUE-Lipo, QUE-Lipo-TAT, and GS hydrogels all exhibit excellent biocompatibility. Rescue experiment of QUE-Lipo-TAT inhibiting the caspase-8/caspase-3/GSDME pathway in RA-FLS As shown in Fig. 8K, compared with the TNF-α group, in the second group where QUE-Lipo-TAT was added on the basis of TNF-α stimulation, the protein levels of caspase-8, caspase-3, GSDME, and GSDME-N in RA-FLS were significantly reduced (P < 0.0001). Compared with the TNF-α group, the protein levels of caspase-8, caspase-3, GSDME, and GSDME-N in OE-caspase-8 RA-FLS in the third group were significantly increased (P < 0.0001). Finally, after adding TNF-α and QUE-Lipo-TAT to OE-caspase-8 RA-FLS, the expression levels of caspase-8, caspase-3, GSDME, and GSDME-N proteins were higher than those in the second group but still lower than those in the third group, with statistical differences observed in all cases (P < 0.0001). This rescue experiment indicates that QUE-Lipo-TAT can inhibit the downstream caspase-3/GSDME signaling axis by acting on caspase-8, thereby effectively suppressing pyroptosis in RA-FLS. QUE-Lipo-TAT acts on the pyroptosis pathway via caspase-8 and also targets the necroptosis pathway via RIPK3 Caspase-8 can block necroptosis executed by the RIPK3/MLKL pathway by inhibiting RIPK3, a key factor in necroptosis. After QUE-Lipo-TAT inhibits caspase-8, it may activate the necroptosis pathway. Therefore, we need to further investigate the effect of QUE-Lipo-TAT on the necroptosis pathway mediated by RIPK3. As shown in Fig. 9A, compared with the TNF-α group, after treatment with the caspase-8 inhibitor Z-IETD-FMK, the expressions of caspase-8, caspase-3, and GSDME in RA-FLS were all inhibited, while the expression level of RIPK3 was significantly increased (P < 0.0001). This indicates that inhibiting caspase-8 in RA-FLS promotes the upregulation of RIPK3, thereby leading to necroptosis. Furthermore, after adding QUE-Lipo-TAT to RA-FLS treated with Z-IETD-FMK, the expressions of caspase-8, caspase-3, and GSDME were significantly reduced, and the level of RIPK3 was also significantly decreased (P < 0.0001). The immunofluorescence staining results in Fig. 9B showed that compared with the TNF-α group, the fluorescence signal of RIPK3 protein was significantly enhanced after caspase-8 inhibition, indicating a significant increase in the expression of this protein. QUE-Lipo-TAT could significantly reduce the fluorescence signal of RIPK3, suggesting that the expression of RIPK3 was inhibited. These results demonstrate that QUE-Lipo-TAT can not only block the caspase-8/caspase-3/GSDME pyroptosis pathway by inhibiting caspase-8, but also block the necroptosis pathway by inhibiting RIPK3. Therefore, QUE-Lipo-TAT possesses a potential therapeutic mechanism for regulating multiple cell death pathways in RA-FLS. Fig. 9(A) Representative immunoblots and bar charts of caspase-8, caspase-3, GSDME, and RIPK3 in each group; (B) Immunofluorescence staining images of RIPK3. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 (A) Representative immunoblots and bar charts of caspase-8, caspase-3, GSDME, and RIPK3 in each group; (B) Immunofluorescence staining images of RIPK3. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 In vitro drug release experiments and biological effects As shown in Fig. 10A, there were no significant differences in the expression levels of caspase-8, caspase-3, GSDME, and GSDME-N proteins between the TNF-α group and the GS hydrogel extract group (P > 0.05). There was also no significant difference in the protein levels of caspase-8, caspase-3, GSDME, and GSDME-N between the QUE-Lipo-TAT group and the GS hydrogel extract + QUE-Lipo-TAT group (P > 0.05). Compared with the GS hydrogel extract group, the protein expression of caspase-8, caspase-3, GSDME, and GSDME-N was significantly downregulated in the GS hydrogel extract + QUE-Lipo-TAT group (P < 0.001). This indicates that the GS hydrogel matrix itself does not significantly up- or downregulate the expression of this pyroptosis pathway-related protein in RA-FLS. In the QUE-Lipo-TAT-loaded hydrogel material, only the nanoparticles exert an inhibitory effect on this pathway protein. The ELISA results in Fig. 10B are consistent with the above WB findings, indicating that the GS hydrogel matrix itself has no significant effect on IL-1β and IL-18 expression in RA-FLS, whereas QUE-Lipo-TAT exerts an inhibitory effect on these cytokines (P < 0.0001). The LDH detection in Fig. 10C is consistent with the above results, indicating that the GS hydrogel matrix failed to reduce LDH levels in RA-FLS. Only QUE-Lipo-TAT demonstrated the ability to significantly lower LDH levels (P < 0.0001), thereby effectively inhibiting pyroptosis-induced cell membrane rupture. Further evaluate the in vitro drug release kinetics of both the QUE-Lipo-TAT nanoparticles themselves and the GS hydrogel loaded with QUE-Lipo-TAT. Figure 10D shows that drug release from QUE-Lipo-TAT nanoparticles exhibits a time-dependent profile. The initial phase (0–24 h) features a rapid release rate, with 53.53% released at 6 h, 71.14% at 12 h, and 86.21% at 24 h. By 48 h, the release rate reached 94.05% and gradually plateaued. Drug release from the GS hydrogel loaded with QUE-Lipo-TAT exhibited a prolonged time-dependent profile, with release rates of 27.41% at 6 h, 40.85% at 12 h, 56.71% at 24 h, and 80.97% at 120 h. This indicates that GS hydrogels can further delay drug release from nanoparticles and extend the therapeutic time window. Subsequently, the biological effects of the drug release kinetics of the QUE-Lipo-TAT-loaded GS hydrogel were evaluated. As shown in Fig. 10E, incubation of RA-FLS with GS hydrogel loaded with QUE-Lipo-TAT resulted in a progressive decrease in the expression levels of caspase-8, caspase-3, GSDME, and GSDME-N proteins at the 0 h (RA), 6 h, 12 h, and 24 h treatment time points. Figure 10F-G shows that as the culture time increases, the levels of IL-1β, IL-18, and LDH in RA-FLS gradually decrease, exhibiting a differential pattern consistent with the pyroptosis protein expression observed in Western blots. Fluorescent staining in Fig. 10H shows that Cy5.5-QUE-Lipo-TAT exhibits progressively enhanced fluorescence aggregation in the cytoplasm surrounding the nuclei of RA-FLS at the 6 h, 12 h, and 24 h time points (P < 0.0001). These results indicate that QUE-Lipo-TAT can be released from the GS hydrogel and successfully internalized by RA-FLS. Corresponding to the in vitro drug release curve, cellular uptake of QUE-Lipo-TAT progressively increased with extended culture duration. This further suggests that the GS hydrogel enhances the sustained-release properties of drugs within nanoparticles while reducing burst release effects. Fig. 10(A) Representative immunoblots and bar charts of caspase-8, caspase-3, GSDME, and RIPK3 in each group; (B) ELISA detection for IL-1β and IL-18; (C) LDH assay; (D) In vitro drug release curves of QUE-Lipo-TAT nanoparticles and QUE-Lipo-TAT-loaded GS hydrogels; (E) Representative immunoblots and bar charts of caspase-8, caspase-3, GSDME, and RIPK3 at different time points after drug treatment; (F) ELISA detection for IL-1β and IL-18; (G) LDH assay; (H) Fluorescence microscopy images and fluorescence intensity of cellular uptake of Cy5.5-QUE-Lipo-TAT at different time points after drug treatment. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 (A) Representative immunoblots and bar charts of caspase-8, caspase-3, GSDME, and RIPK3 in each group; (B) ELISA detection for IL-1β and IL-18; (C) LDH assay; (D) In vitro drug release curves of QUE-Lipo-TAT nanoparticles and QUE-Lipo-TAT-loaded GS hydrogels; (E) Representative immunoblots and bar charts of caspase-8, caspase-3, GSDME, and RIPK3 at different time points after drug treatment; (F) ELISA detection for IL-1β and IL-18; (G) LDH assay; (H) Fluorescence microscopy images and fluorescence intensity of cellular uptake of Cy5.5-QUE-Lipo-TAT at different time points after drug treatment. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 Dual-crosslinked GS hydrogel microneedle loaded with QUE-Lipo-TAT ameliorate arthritis in CIA rats The experimental protocol for CIA rats is shown in Fig. 11A. During the observation period, the body weight of rats in the QUE-Lipo-TAT MN group was significantly higher than that in the CIA group, and also higher than that in the QUE-Lipo-TAT Gel group and QUE MN group. This indicates that the combination mode of QUE-Lipo-TAT and MN has good safety and a systemic improvement effect (Fig. 11B). As shown in Fig. 11C, the paw thickness of rats in the QUE-Lipo-TAT MN group was significantly lower than that in the CIA group, and the effect was superior to that in both the QUE-Lipo-TAT Gel group and the QUE MN group. This indicates that the combination of QUE-Lipo-TAT and microneedle structure can more effectively treat joint inflammation in RA to alleviate symptoms. In addition, the clinical score data showed a consistent trend with the changes in hind paw thickness, which further supports the significant effect of QUE-Lipo-TAT MN in relieving arthritis symptoms and improving joint function (Fig. 11D). The results of the ELISA assay showed that compared with the control group, the serum levels of IL-1β and IL-18 in CIA rats were significantly increased (P < 0.0001), reflecting the inflammatory state of CIA rats. After treatment with QUE-Lipo-TAT Gel and QUE MN, the levels of IL-1β and IL-18 in CIA rats were significantly decreased (P < 0.0001), and the inhibitory effect of the QUE-Lipo-TAT MN group was significantly better than that of both the QUE-Lipo-TAT Gel group and the QUE MN group. This indicates that QUE-Lipo-TAT MN can more effectively reduce the levels of inflammatory cytokines in CIA rats (Fig. 11E). In conclusion, these results indicate that the dual-crosslinked GS hydrogel microneedle loaded with QUE-Lipo-TAT possess good biosafety, can significantly alleviate the symptoms of CIA rats, improve joint function, and reduce the levels of inflammatory cytokines, thus providing an effective new strategy for the treatment of RA. Fig. 11(A) Schematic diagram of the establishment and treatment of CIA rats; (B-D) Body weight, hind paw swelling degree, and arthritis score of rats in each group; (E) ELISA detection of IL-1β and IL-18 in the serum of rats in each group; (F) Representative fluorescence images of Cy5.5-labeled drugs and nanoparticles in rats; (G) H&E staining and Safranin O staining of ankle joints, and Masson staining of synovial tissue. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 (A) Schematic diagram of the establishment and treatment of CIA rats; (B-D) Body weight, hind paw swelling degree, and arthritis score of rats in each group; (E) ELISA detection of IL-1β and IL-18 in the serum of rats in each group; (F) Representative fluorescence images of Cy5.5-labeled drugs and nanoparticles in rats; (G) H&E staining and Safranin O staining of ankle joints, and Masson staining of synovial tissue. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 In vivo biodistribution of QUE-Lipo-TAT MN In vivo fluorescence imaging was performed using CY5.5-labeled QUE-Lipo-TAT Gel, QUE MN, and QUE-Lipo-TAT MN to evaluate the biodistribution of the payload. As shown in Fig. 11F, at 3 h, 6 h, and 12 h after microneedle administration, the CY5.5-labeled drug in each group rapidly accumulated at the inflamed joint sites of CIA rats, and the fluorescence intensity of the drug label in the QUE-Lipo-TAT MN group was consistently significantly higher than that in the QUE-Lipo-TAT Gel group and QUE MN group. At 24 h and 48 h, the fluorescence signals in the QUE-Lipo-TAT Gel group and QUE MN group almost disappeared, while the QUE-Lipo-TAT MN group still exhibited obvious fluorescence signals, which gradually decreased from 0 h to 48 h. In addition, local dissection revealed that the fluorescence signal was concentrated at the inflamed joint sites, and no fluorescence signal was detected in the heart, liver, spleen, lung, or kidney. These results indicate that the combination of QUE-Lipo-TAT and MN exhibits excellent drug absorption, accumulation effects, and sustained drug release capacity. It can effectively accumulate drugs in situ at inflamed joints, with a clear local therapeutic area, no diffusion to other organs, and thus possesses favorable targeting property, biosafety, and therapeutic potential. QUE-Lipo-TAT MN ameliorates synovial inflammation and cartilage damage in CIA rats by inhibiting the pyroptosis pathway and exhibits biosafety Proliferative synovial tissue is a key pathological feature of RA, while bone and cartilage damage are major adverse consequences. We performed histological analysis on the inflamed joints of CIA rats in each group. As shown by H&E staining in Fig. 11G, compared with the healthy group, the ankle joints of rats in the CIA group showed massive inflammatory cell infiltration, severe synovial hyperplasia, and pannus invasion—presenting the typical pathological features of RA. Compared with the CIA group, the joint conditions in the QUE-Lipo-TAT Gel group and QUE MN group were somewhat improved, but partial inflammatory cell infiltration and synovial hyperplasia were still observed. In contrast, the QUE-Lipo-TAT MN group exhibited a significant improvement in the pathological features of joints and synovium: inflammatory cell infiltration, synovial hyperplasia, and pannus invasion were significantly inhibited, and the surface of articular cartilage was smooth, which was similar to that of the healthy group. Safranin O staining showed that compared with the healthy group, the joint surfaces of CIA rats were rough and their cartilage tissue was severely damaged. Compared with the CIA group, the joints in the QUE-Lipo-TAT Gel group and QUE MN group were somewhat improved, but the normal cartilage structure could not be fully restored. In contrast, the QUE-Lipo-TAT MN group significantly improved joint damage, and the cartilage tissue recovered to a state close to that of the healthy group. Finally, Masson staining was used to evaluate the fibrosis of the joint synovium. Compared with the healthy group, the significant deposition of collagen fibers in the synovial tissue of the CIA group indicated severe synovial fibrosis. Compared with the CIA group, the deposition of collagen fibers in the QUE-Lipo-TAT Gel group and QUE MN group was reduced. In contrast, the QUE-Lipo-TAT MN group significantly decreased the level of collagen fiber deposition in the synovial tissue, indicating that QUE-Lipo-TAT MN can most effectively alleviate synovial fibrosis and significantly relieve synovial inflammation. In conclusion, the ingenious combination of QUE-Lipo-TAT and MN can maximize the reduction of joint inflammatory status, inflammatory cell infiltration, articular cartilage damage, synovial hyperplasia, and fibrosis in CIA rats, demonstrating great potential in the treatment of RA. As shown by the immunohistochemical staining of joint synovial tissue in Fig. 12A, compared with the healthy group, the positive expression areas of caspase-8, caspase-3, and GSDME in the CIA group were significantly upregulated, indicating that these pyroptotic proteins are involved in the pathological process of RA arthritis. Compared with the CIA group, the number of caspase-8/caspase-3/GSDME-positive cells in the QUE-Lipo-TAT Gel group and QUE MN group decreased, but failed to fully return to the level of the healthy group. In contrast, the positive expression areas of caspase-8/caspase-3/GSDME in the QUE-Lipo-TAT MN group were significantly downregulated and close to the level of the healthy group. Furthermore, we performed immunohistochemical staining for CD90, a specific marker of RA-FLS [50].Compared to the healthy group, the CIA group exhibited a large number of CD90-positive cells, indicating substantial proliferation of RA-FLS. Compared to the CIA group, both QUE-Lipo-TAT Gel and QUE MN reduced the number of RA-FLS. The QUE-Lipo-TAT MN group showed the lowest number of CD90-positive cells, most closely resembling the healthy group. This demonstrates that QUE-Lipo-TAT MN significantly inhibits RA-FLS proliferation. In summary, the delivery of QUE-Lipo-TAT via MN can most effectively inhibit the caspase-8/caspase-3/GSDME pyroptosis pathway in RA-FLS, suppress the tumor-like proliferation of RA-FLS, and block the vicious cycle of cell death, tissue inflammation, and uncontrolled expansion of RA-FLS. Thus, it effectively alleviates synovial inflammation and provides a novel therapeutic strategy for the treatment of RA. Fig. 12(A) Immunohistochemical staining images of proteins (caspase-8, caspase-3, GSDME, and CD90) in CIA rats of each group; (B) Blood biochemical indicators (ALT, AST, BUN, and CREA) in CIA rats of each group; (C) H&E staining of the heart, liver, spleen, lung, and kidney in CIA rats of each group. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 (A) Immunohistochemical staining images of proteins (caspase-8, caspase-3, GSDME, and CD90) in CIA rats of each group; (B) Blood biochemical indicators (ALT, AST, BUN, and CREA) in CIA rats of each group; (C) H&E staining of the heart, liver, spleen, lung, and kidney in CIA rats of each group. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 As shown by the blood biochemical tests in Fig. 12B, compared with the healthy group, the levels of ALT, AST, BUN, and CREA in CIA rats were significantly increased, indicating that the liver and kidney functions of CIA rats were impaired. After CIA rats received QUE-Lipo-TAT MN treatment, various indicators of liver and kidney functions were significantly improved and could return to a level close to that of the normal group (P > 0.05). This suggests that QUE-Lipo-TAT MN has good biocompatibility and safety, and can effectively improve the liver and kidney function impairment in CIA rats. As shown by H&E staining in Fig. 12C, no abnormal changes or pathological damage were observed in the heart, liver, spleen, lung, or kidney of the QUE-Lipo-TAT MN group, with no significant difference from the normal group. This further confirms that the use of QUE-Lipo-TAT does not induce significant adverse reactions in vivo and exhibits good biosafety. In summary, QUE-Lipo-TAT MN exhibits excellent biocompatibility and biosafety, and effectively alleviates liver and kidney damage in CIA rats. Discussion Rheumatoid arthritis is a chronic and refractory joint disease characterized by proliferative synovial pannus, with clinical manifestations mainly including synovial inflammatory response and joint tissue damage. As the key component of invasive synovium, RA-FLS plays a crucial role in the pathological process of destructive arthritis and is closely associated with disease activity. In healthy joints, FLS play a crucial role in regulating the composition of the extracellular matrix and synovial fluid by producing lubricants such as hyaluronic acid. FLS also facilitate lubrication and nutrition of the articular cartilage surface by reducing cell and protein adhesion. In RA joints, RA-FLS proliferate extensively to become a key component of invasive synovium, which represents the core pathological mechanism of RA [51]. Dying cells fuel the flame of inflammation. Pyroptosis and necroptosis promote the progression of inflammatory diseases such as RA, while inflammation further accelerates cell death—these two processes are interdependent and inseparable [52]. The molecular mechanisms of pyroptosis include the canonical pathway, non-canonical pathway, and alternative pathway. The canonical pathway is induced by caspase-1, while the non-canonical pathway via caspase-4/−5/−11 [53]. In the alternative pathway, other members of the caspase family can also induce pyroptosis. For example, caspase-8 can induce GSDMC-dependent pyroptosis [54]; caspase-3 can cleave GSDME to generate GSDME-N, which causes cell membrane perforation [55]; in the NLRP3 and AIM2 inflammasome signaling pathways, the pyrin domain of ASC can bind to the death effector domain of pro-caspase-8, activate it, and then cleave downstream caspase-3 to induce pyroptosis [56]. However, no relevant studies on the caspase-8/caspase-3/GSDME pathway in RA have been reported. This study is the first to report the expression of this pathway in RA-FLS and its role in inducing pyroptosis. RA-FLS are considered to have the duality of being both passive responders and active aggressors [57]. RA-FLS undergoing pyroptosis undergo cell lysis and release a large number of proinflammatory cytokines, including IL-1β and IL-18, to promote local inflammatory cascades. This further induces pyroptosis in surrounding normal FLS, triggering a vicious cycle. In this study, it was found that the use of QUE-Lipo-TAT can effectively inhibit caspase-8, thereby blocking the caspase-8/caspase-3/GSDME pathway. This inhibition prevents cell lysis of RA-FLS caused by pyroptosis, reduces the levels of proinflammatory cytokines, and suppresses inflammation to break the vicious cycle. Necroptosis is a type of PCD that is negatively regulated by caspase-8 and dependent on the RIPK3/MLKL pathway. When caspase-8 is inhibited, RIPK3-dependent necroptosis occurs [58]. TNF-tumor necrosis factor receptor 1 (TNFR1) signaling activates RIPK1, and the activated RIPK1 interacts with RIPK3 via sharing the RIP homotypic interaction motifs (RHIM) to form an amyloid structure, thereby activating RIPK3 [59]. RIPK3 recruits and activates MLKL to the necrosome [60]. Subsequently, MLKL localizes to the cell membrane, and its amino-terminal coiled-coil domain forms homotrimers. These homotrimers oligomerize to form pores on the cell membrane, leading to cell lysis and release of intracellular contents. This ultimately results in necroptosis and further triggers an inflammatory response [61]. The spread of inflammatory sources caused by pyroptosis and necroptosis of RA-FLS may be one of the root causes of the occurrence and development of RA. The key link to block this vicious cycle lies in caspase-8; therefore, caspase-8 may be an important target for RA (Fig. 13). This study found that after stimulation with TNF-α, FLS can activate caspase-8 and the downstream caspase-3, which in turn cleaves GSDME to form GSDME-N. GSDME-N then forms pores on the cell surface, promoting pyroptosis and the release of IL-1β and IL-18. When caspase-8 is inhibited, the protein levels in the caspase-8/caspase-3/GSDME pathway decrease, while the expression of RIPK3 is significantly upregulated. When QUE-Lipo-TAT inhibits caspase-8 to block the caspase-8/caspase-3/GSDME pathway (thereby suppressing pyroptosis), it can also inhibit the expression of RIPK3 to block the occurrence of necroptosis, achieving a “two birds with one stone” therapeutic effect. Fig. 13Molecular mechanism diagram of QUE-Lipo-TAT nanoparticles in treating RA-FLS. In RA, after TNF-α is recognized and bound by TNFR1 on the surface of RA-FLS, TNF complex IIb is formed inside the cells. TNF complex IIb further activates caspase-8 and the downstream caspase-3/GSDME pathway, which in turn cleaves GSDME to form GSDME-N. GSDME-N then forms pores on the cell membrane of RA-FLS, leading to pyroptosis and the release of large amounts of pro-inflammatory cytokines. When caspase-8 is inhibited, RIPK3 is activated, which further activates downstream MLKL. MLKL forms cleavage pores on the cell membrane of RA-FLS, resulting in necroptosis. Ultimately, this promotes inflammatory cell death in RA and a vicious cycle of inflammation. QUE-Lipo-TAT nanoparticles inhibit pyroptosis and necroptosis by suppressing caspase-8 and RIPK3, respectively; this interrupts the release of pro-inflammatory cytokines and the vicious cycle of inflammatory death in RA-FLS, thereby treating RA Molecular mechanism diagram of QUE-Lipo-TAT nanoparticles in treating RA-FLS. In RA, after TNF-α is recognized and bound by TNFR1 on the surface of RA-FLS, TNF complex IIb is formed inside the cells. TNF complex IIb further activates caspase-8 and the downstream caspase-3/GSDME pathway, which in turn cleaves GSDME to form GSDME-N. GSDME-N then forms pores on the cell membrane of RA-FLS, leading to pyroptosis and the release of large amounts of pro-inflammatory cytokines. When caspase-8 is inhibited, RIPK3 is activated, which further activates downstream MLKL. MLKL forms cleavage pores on the cell membrane of RA-FLS, resulting in necroptosis. Ultimately, this promotes inflammatory cell death in RA and a vicious cycle of inflammation. QUE-Lipo-TAT nanoparticles inhibit pyroptosis and necroptosis by suppressing caspase-8 and RIPK3, respectively; this interrupts the release of pro-inflammatory cytokines and the vicious cycle of inflammatory death in RA-FLS, thereby treating RA At present, the clinical strategies for RA mainly focus on drug therapy aimed at symptom relief and disease improvement; however, such therapies are usually accompanied by adverse events [62]. Therefore, it is of great significance to explore novel therapies with high efficacy, low toxicity, convenience, and affordability in the treatment of RA. QUE is a flavonoid compound found in fruits and vegetables, with anti-inflammatory, anti-rheumatic, and hepatoprotective activities, and it can interact with a variety of molecules and targets [63]. Clinical studies have shown that QUE can significantly reduce the levels of pro-inflammatory cytokines in RA, while remarkably improving clinical symptoms such as morning stiffness and pain, and disease activity in RA patients [64]. Basic research has demonstrated that QUE can prevent and treat bone destruction in RA by inhibiting the secretion of pro-inflammatory cytokines, down-regulating lipopolysaccharide-induced cyclooxygenase-2 expression, and suppressing bone resorption [65]; QUE can antagonize the negative effects in arthritis of RA rats, effectively prevent joint damage, and possess hepatoprotective effects [66]; QUE can inhibit the proliferation and invasion of RA-FLS [67]. In terms of molecular mechanisms, caspase-8 can regulate the aggressive behavior of RA-FLS. Silencing caspase-8 in RA-FLS inhibits calpain activity, reduces fibronectin and type II collagen in the extracellular matrix, and suppresses RA-FLS migration. Therefore, caspase-8 is crucial for the adhesion, migration, and invasion of RA-FLS [68]. This study found that QUE-Lipo-TAT nanoparticles can inhibit aggressive behaviors such as migration of RA-FLS by inhibiting caspase-8. However, QUE has extremely low bioavailability, so it is crucial to package, modify QUE and develop effective drug delivery systems for it. Nanoparticle delivery systems are regarded as a key strategy to enhance the transdermal penetration of drugs in TDDS. Among them, liposomes are bilayer spherical vesicular structures composed of amphiphilic lipids. Their phospholipid components can interact with the lipid structure of the stratum corneum to improve the transdermal penetration efficiency of drugs, and they can also be embedded in hydrogels to achieve transdermal drug delivery [69]. Modification of drugs by introducing functional ligands can significantly optimize drug delivery efficiency. The function of CPPs lies in enhancing the cellular internalization capability of nanocarriers [70]. CPPs possess the ability to mediate the entry of nanoparticles into the cytoplasm, which provides an important opportunity for technological innovation in the fields of biomedical research and drug development [71]. Since positively charged nanoparticles tend to generate electrostatic adsorption with negatively charged phospholipid bilayers [72], the surface electrical properties of liposomes can be regulated by conjugating CPPs to promote cellular uptake [73]. In this study, QUE was encapsulated in liposomes, and the surface of the nanoparticles was modified with DSPE-PEG2K-TAT to enhance the ability to fuse with cell membranes and improve the intracellular delivery efficiency of QUE. Compared with QUE alone, the synthesized QUE-Lipo-TAT can more potently inhibit the caspase-8/caspase-3/GSDME pathway to block pyroptosis in RA-FLS, more strongly suppress the migration and disordered proliferation of RA-FLS, and more effectively inhibit joint inflammation, synovial hyperplasia, and bone destruction, thus exhibiting greater therapeutic potential for RA. The skin stratum corneum (SC) serves as the primary and most formidable barrier against the invasion of exogenous substances, with its defensive structure described as a tightly integrated “mortar and brick” model [74]. As a chronic inflammatory disease, RA treatment requires long-term, repeated administration. Frequent direct injections may lead to complications such as disruption of skin barrier function, damage to articular cartilage, and bleeding, and increase the risk of infection, leading to infectious arthritis. Additionally, the convenience and safety of the administration mode directly impact patient compliance, with the pain associated with long-term direct injections leading to reduced patient acceptance [75, 76]. Compared to direct injection, TDDS can ensure transdermal drug delivery efficiency while avoiding damage to skin tissue [77]. The microneedle in the TDDS strategy can evade cutaneous nociceptors by constructing micrometer-scale permeation channels in the stratum corneum, enabling localized targeted delivery of small-molecule drugs [78]. Microneedle systems have garnered significant attention in the field of transdermal drug delivery due to their high safety profile, excellent delivery efficiency, ease of use, and non-invasive, pain-free application [79]. The GS hydrogel microneedle developed in this study offer the advantages of minimally invasive delivery, self-administered medication, excellent mechanical strength, and good biocompatibility. Microneedle penetrate only the SC, avoiding contact with nerve endings, resulting in no obvious pain during the administration process. The patch design enables convenient self-administration by patients. The backing layer can be removed after half an hour, reducing the wear time of the patch. After dissolution, the hydrogel leaves no solid residue, reducing the risk of infection. In summary, this “microinjection” system circumvents the drawbacks of direct injection administration, making it more suitable for long-term home treatment of RA patients. Furthermore, the low bioavailability and rapid systemic clearance of directly injected drugs have prompted researchers to explore novel drug delivery strategies [80]. Hydrogel microneedle is composed of swellable materials. Upon skin penetration, the crosslinked structure at the needle tip swells by absorbing interstitial fluid, subsequently dissolving to form an in situ hydrogel. As the hydrogel is fabricated from biodegradable polymers, it enables sustained drug release [81]. In this study, the GS hydrogel microneedle transformed into a colloidal state within 30 min upon entering the skin. The morphological evolution of this microneedle system progressed from photopolymerized rigid microneedle to absorbing interstitial fluid, expanding, embedding within the dermis, and ultimately dissolving to form an in situ hydrogel. Combined in vitro drug release assays, biological effect studies, and in vivo biodistribution experiments demonstrate that the in situ hydrogel formed within RA joints effectively achieves sustained, localized drug release. This approach prevents burst release, thereby enabling long-term therapeutic efficacy. In summary, this microneedle system demonstrates superiority over direct injection in terms of safe and convenient drug delivery, reduced complications, effective sustained-release of the drug, and enhanced patient compliance. GelMA hydrogel possesses good cell adhesion properties; however, it degrades too quickly, has poor toughness and mechanical structure, and is prone to breakage. In this study, SilMA was incorporated into GelMA to form a double-crosslinked GS hydrogel, which enhanced the mechanical properties of the crosslinked network, including increased toughness, reduced deformation, and improved compressive strength. Compared with single-network GelMA or SilMA, the GS hydrogel exhibited more excellent comprehensive performance in both biological properties and mechanical properties. The bionic Oriental Pearl Tower-shaped microneedle prepared for the first time using GS hydrogel exhibit good fracture resistance and stiffness, and can easily penetrate the stratum corneum. After entering the skin, these microneedle can absorb interstitial fluid to promote the swelling of their double-sphere and single-ring structure. Additionally, the GS hydrogel has good adhesiveness; these two factors together enable the microneedle to be firmly embedded in the skin and prevent them from falling out or being squeezed out at frequently moving joint sites. The GS hydrogel microneedle loaded with QUE-Lipo-TAT can slow down the degradation rate to achieve the effect of sustained drug release. This helps nanoparticles more effectively inhibit pyroptosis and necroptosis, reduce the levels of pro-inflammatory cytokines, further suppress the disordered proliferation of invasive synovium, alleviate bone and cartilage damage, effectively treat CIA rats, and exhibit good biosafety. Conclusion This study delves into the specific role of caspase-8 in RA, proposes a novel pathogenic mechanism of the caspase-8/caspase-3/GSDME pathway in RA, and reveals its key role in RA-FLS cell death and inflammation. We successfully synthesized QUE-Lipo-TAT nanoparticles by encapsulating QUE in liposomes and modifying their surface with DSPE-PEG2K-TAT using nanotechnology. These nanoparticles can act on the caspase-8 and RIPK3 targets of RA-FLS to inhibit pyroptosis and necroptosis respectively, reduce inflammatory cell death and the release of pro-inflammatory cytokines, suppress the abnormal proliferation and migration of RA-FLS, thus achieving a “kill two birds with one stone” therapeutic effect. After adopting a combination of second-generation and third-generation TDDS enhancement strategies to successfully synthesize QUE-Lipo-TAT and verify its effect on RA-FLS, a novel bionic Oriental Pearl Tower-shaped GelMA-SilMA double-crosslinked hydrogel microneedle was further constructed for effective drug delivery and controlled release, so as to alleviate chronic synovial inflammation, synovial fibrosis, and bone and cartilage damage, thereby treating CIA rats. This study provides new targets, new pathways, and a new drug delivery method for the treatment of RA, promotes the development of targeted and individualized therapeutic strategies, and offers new ideas for the precision treatment of RA. Supplementary Information Supplementary Material 1. Supplementary Material 1. Supplementary Material 2. Supplementary Material 2.