Microneedle patches deliver targeted GLP-1RAs-loaded nanoparticles for the treatment of rheumatoid arthritis

Abstract

Despite the strong immune-modulatory benefits, there has been a lag in using glucagon-like peptide-1 receptor agonists (GLP-1RAs) in rheumatoid arthritis (RA) due to inefficient articular delivery and poorly understood direct cartilage protection. In this paper, we addressed these critical issues through the mechanism-informed design of a hierarchical drug delivery system. We elucidated previously unnoted ways in which GLP-1RAs curbed RA pathology via direct suppression of the cGAS-STING signaling pathway to inhibit chondrocyte ferroptosis and concomitantly induced macrophage M2 polarization. To turn this mechanistic insight into a targeted therapy, we developed cartilage-targeting silk fibroin nanoparticles modified with a type II collagen-binding peptide WYRGRL for localized GLP-1RAs delivery (GLP-1RAs@NPs-WYRGRL). These GLP-1RAs@NPs-WYRGRL were subsequently loaded into dissolving microneedles (MNs) patches, giving a composite MNs/GLP-1RAs@NPs-WYRGRL system. This system enabled painless transdermal delivery with marked accumulation in arthritic joints, which effectively inhibited chondrocyte ferroptosis via the aforementioned mechanism and also promoted macrophage M2 polarization, thereby attenuating RA progression. This study not only found that the cGAS-STING-ferroptosis pathway was a druggable target in joints, but also pioneered a new flexible delivery platform that resolved the key limitation of low local drug concentrations in joints after systemic administration, and offered a cartilage-protecting perspective for RA.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04261-0.

Introduction

Rheumatoid arthritis (RA) is a kind of chronic systemic autoimmune disorder. Its pathological features include persistent synovial inflammation, progressive degeneration of joint cartilage, and bone erosion [1]. The clinical manifestations are characterized by symmetrical polyarthritis, with associated joint swelling and pain, that may gradually progress to joint deformities, primarily involving the small joints of the hands and feet. The worldwide prevalence of RA is predicted to be about 0.5% to 1% [2]. Current management strategies mainly involve the use of conventional synthetic disease-modifying antirheumatic drugs (csDMARDs), biological DMARDs, and targeted synthetic DMARDs combined with glucocorticoids to alleviate inflammation and pain [3]. However, these therapeutic agents tend to have associated systemic adverse effects, including increased susceptibility to infections, hepatotoxicity, nephrotoxicity, and other complications [4, 5]. Furthermore, challenges such as low drug concentration at the local pathological site after systemic administration, limited target specificity and short drug half-life persist [6]. Hence, there is a great need for more innovative drugs as well as treatment approaches that offer improved target selectivity and are amenable to localized delivery, particularly to small joints, to ultimately enhance therapeutic outcomes in patients with RA.

Recently, glucagon-like peptide-1 receptor agonists (GLP-1RAs) have gradually become a first-line treatment for type 2 diabetes [7]. Beyond known benefits regarding glycemic control, weight loss, and cardiorenal protection [8, 9], GLP-1RAs are powerful anti-inflammatory mediators and articular cartilage protectors [10–12]. It should also be noted that in recent large-scale clinical trials there were signs of good treatment potential for GLP-1RAs in osteoarthritis (OA) [13, 14]. Mechanistically, GLP-1R activation alleviates OA by fostering chondrocyte anabolic metabolism while inhibiting catabolism, safeguarding chondrocytes from endoplasmic reticulum stress and apoptosis, and displaying anti-inflammatory and antioxidant properties [15–17]. Furthermore, evidence indicates that GLP-1RAs inhibit inflammation in fibroblast-like synoviocytes (FLS) and attenuate RA progression [18], suggesting their potential therapeutic utility in rheumatic disorders [19]. However, the precise molecular mechanisms underlying GLP-1RA action in RA, particularly those mediating cartilage protection, remain incompletely elucidated. Moreover, effective delivery of therapeutic drugs to the joint site following systemic administration remains highly challenging, severely limiting their clinical efficacy. Consequently, although GLP-1RAs hold considerable promise for RA treatment, their clinical translation is constrained by both unresolved mechanistic questions and systemic administration failing to achieve therapeutic intra-articular concentrations constrain their application.

The cGAS-STING pathway was initially defined as a “cytoplasmic dsDNA sensor–proinflammatory response” axis that triggers the production of type I interferons and a broad range of other cytokines [20]. Accumulating evidence has firmly established that it is also a prominent driving force behind the pathogenesis of RA, which has been well documented. The release of mitochondrial DNA (mtDNA) activates the cGAS-STING signaling axis, which then triggers macrophages in RA to secrete tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), and interleukin-6 (IL-6) [21]. Furthermore, cytoplasmic dsDNA levels are significantly increased in FLS from RA patients, causing upregulation of genes such as IL-6, CXCL10, TNF-α, MMP-13, IL-1β, and IL-8 via a cGAS-dependent mechanism that also disrupts joint homeostasis [22]. However, the role of the cGAS-STING pathway in directly regulating chondrocyte degeneration in RA remains poorly understood as of now. Recent studies have shown that ferroptosis in chondrocytes aggravates the progress of RA [23], and the cGAS-STING pathway has been proven to positively regulate the ferroptosis in various diseases [24–26]. Moreover, although some studies indicate that GLP-1RAs can rescue diabetic vascular endothelial damage via suppressing STING signaling [27], it remains unclear whether they exert an intervention effect in delaying RA progression by targeting the underlying cGAS-STING-ferroptosis pathway. Therefore, elucidating whether the cGAS-STING-ferroptosis pathway is involved in RA-associated inflammation and cartilage damage and determining whether GLP-1RAs provide multi-target protection to RA by inhibiting this pathway, may provide novel theoretical insights and therapeutic strategies for RA management.

Silk fibroin (SF) has gained much attention because of its remarkable mechanical strength, excellent biocompatibility, and controllable degradation profile [28]. Theoretically, nanoparticles (NPs) encapsulating GLP-1RAs can significantly extend the drug’s half-life and promote accumulation of the drug in the joint cartilage, thereby maintaining a high therapeutic concentration for an extended period and improving treatment efficacy. However, conventional silk fibroin NPs lack sufficient targeting specificity toward articular cartilage. To address this issue, the type II collagen-binding peptide WYRGRL, which exhibits high affinity for cartilage extracellular matrix components, can be employed to functionalize drug-loaded NPs [29]. This functionalization enables cartilage-targeted delivery, allowing for selective enrichment of GLP-1RAs at diseased joint sites and facilitating precise therapeutic intervention. Furthermore, conventional administration routes for RA, including oral intake and systemic injection, are often hindered by low bioavailability and poor patient compliance [30]. Microneedles (MNs), as a minimally invasive, painless, and user-friendly transdermal delivery platform, offer a promising alternative for localized therapy in RA [31]. Notably, RA lesions frequently affect small joints such as the metacarpophalangeal and metatarsophalangeal joints [32]. Direct intra-articular injection is difficult, and periarticular injection provides only temporary relief which is also transient. Therefore, a novel composite system integrating soluble microneedles with targeted nanoparticles—termed a nanoparticle-microneedle delivery platform—holds significant potential for effective and sustained treatment of RA.

Here, we present a novel mechanism-to-therapy strategy designed to overcome the dual barriers of an unclear mechanism and inefficient delivery. We demonstrate that the therapeutic potential of GLP-1RAs in RA is mediated by their dual targeting of two key pathological pathways: suppressing a previously unidentified cGAS-STING-ferroptosis axis in chondrocytes and reprogramming synovial macrophages toward an M2 reparative phenotype, which collectively helps maintain joint homeostasis. To achieve targeted delivery of GLP-1RAs to affected joints, we developed a cartilage-targeting nanoparticle system (GLP-1RAs@NPs-WYRGRL) and integrated it into a methacrylated hyaluronic acid-based microneedle patch (MNs/GLP-1RAs@NPs-WYRGRL). This composite drug delivery system thereby enhances drug stability, enables sustained and targeted delivery, and facilitates localized intra-articular therapy. In summary, this study proposes an innovative therapeutic strategy that not only provides novel insights into the clinical potential and pharmacological mechanisms of GLP-1RAs in RA but also establishes a new cartilage-protective insight that challenges the prevailing synovium-centric approach to RA therapy.

Methods and materials

Reagents and antibodies

2’,3’-cGAMP (cat. no. S7904) was purchased from Selleck (Shanghai, China). Type I collagenase (cat. no. MI00645) was purchased from Mishushengwu (Xi'an, China). Type II collagenase (cat. no. A004174-0001) was purchased from Sangon Biotech (Shanghai, China). GLP-1RAs (cat. no. AG-CP3-0034-M005) were purchased from AdipoGen Life Sciences (San Diego, USA). Delivector™ Avertin (cat. no. DW3120) was purchased from Dowobio (Shanghai, China). The Hematoxylin-Eosin (HE) Stain Kit (cat. no. G1120), modified Safranin-O and Fast Green Stain Kit (cat. no. G1371), and Neutral Balsam (cat. no. G8590) were purchased from Solabio (Beijing, China). Rhodamine B isothiocyanate (RBITC; cat. no. GY418) was obtained from Goyoo Biotech (Nanjing, China). The type II collagen-targeting peptide (WYRGRL) was purchased from Tanshtech (Guangzhou, China). Anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (cat. no. 60004-1-Ig), anti-ACSL4 (cat. no. 22401-1-AP), anti-GPX4 (cat. no. 67763-1-Ig), anti-aggrecan (ACAN) (cat. no. 13880-1-AP), anti-collagen type I (COL1) (cat. no. 66761-1-Ig), anti-collagen type II (COL2) (cat. no. 28459-1-AP), anti-STING (cat. no. 19851-1-AP), anti-IL-6(cat. no. 21865-1-AP), anti-TNF-α (cat. no. 17590-1-AP), Alexa Fluor 594-conjugated goat anti-mouse IgG (cat. no. SA00013-3), Alexa Fluor 488-conjugated goat anti-rabbit IgG (cat. no. SA00013-2), CoraLite^® Plus 647-anti-CD43 (cat. no. CL594-65237), HRP-conjugated Goat anti-Rabbit IgG (cat. no. SA00001-2), and HRP-conjugated Goat anti-Mouse IgG (cat. no. SA00001-1) antibodies were obtained from Proteintech (Wuhan, China). Anti-GLP-1R (cat. no. A8547) was obtained from ABclonal (Wuhan, China). Anti-cGAS (cat. no. sc-515777) and anti-TNF-α (cat. no. sc-52746) antibodies were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). Goat anti-Rabbit IgG (Alexa Fluor^® 594) (cat. no. ab150080) and Goat anti-Mouse IgG (Alexa Fluor^® 488) (cat. no. ab150113) antibodies were purchased from Abcam (Cambridge, MA, USA). PerCP/Cyanine5.5-conjugated anti-CD45 (cat. no. 202220), FITC-conjugated anti-CD11b (cat. no. 201805), PE-conjugated anti-CD68 (cat. no.201003), APC-conjugated anti-CD86 (cat. no. 200315), and PE/Cyanine7-conjugated anti-CD206 (cat. no. 141719) antibodies were purchased from BioLegend (San Diego, CA, USA).

Isolation and Identification of Chondrocytes

Two 1–2-week-old Sprague-Dawley (SD) rats were anesthetized and euthanized. The skin of the hind limbs was incised to expose the knee joints. The cartilage tissues from the joint surfaces were carefully harvested using a surgical blade and transferred into 1.5 mL EP tubes. Subsequently, 200–300 µL of type II collagenase solution was added to each tube, and the cartilage tissues were cut into tissue fragments of approximately 1–2 mm³. The tissue fragments were then transferred to 15 mL centrifuge tubes, and type II collagenase solution was added to bring the total volume in the centrifuge tube to approximately 6 mL. The cartilage tissues were digested in a shaking incubator at 37 °C and 80 rpm for 1–2 h. After digestion, the tubes were removed from the incubator, and the suspension was filtered through a 70-µm cell strainer. The filtered suspension was centrifuged at 300 × g, after which the supernatant was discarded. DMEM/F12 complete medium was used to resuspend the cell precipitate. Then transfer the cell suspension to the culture flask and incubate at conditions of 37 °C under 5% CO₂ until cells fully adhered. After expansion and passaging, immunofluorescence and flow cytometry were used to test and sort chondrocytes based on the expression of COL2 and COL1 in the cells.

Isolation and Identification of Macrophages

Two 1- to 2-week-old SD rats were anesthetized and euthanized. The skin of the hind limbs was incised, and the tibiae and femora were carefully dissected and stripped of surrounding soft tissues. The epiphyses were removed from both ends of the bones. Bone marrow cavities were flushed with RPMI-1640 complete medium using a syringe, and the resulting bone marrow suspension was collected. The suspension was filtered through a 40 μm cell strainer and transferred to a cell culture flask for adherent cell culture. After 24 to 48 h of incubation, the cells from the supernatant were collected. Re-suspend the cell precipitate in RPMI-1640 complete medium containing M-CSF for further culture. On the third day, replace the medium with a fresh one containing M-CSF. When the fusion rate of macrophages exceeded 90%, the cells were passaged and then subjected to subsequent experiments.

Real-time quantitative polymerase chain Reaction (RT-qPCR) assay

Chondrocytes or macrophages were subjected to different treatments and were collected by centrifugation. 1 mL of RNAiso Plus reagent (Cat. No. 9109; Takara Bio, Shiga, Japan) was added to the cell precipitate. After thoroughly lysing the cell pellet, 0.2 mL of chloroform was added to the lysate. After phase separation, transfer the upper aqueous phase to an EP tube, then mix with an equal volume of isopropanol. After centrifugation again, it can be observed that the RNA precipitate is located at the bottom of the EP tube. Wash the RNA precipitate with 75% ethanol and then centrifuge it to obtain the RNA precipitate. Finally, add RNase-free water to the RNA precipitation to dissolve it. Then, the cDNA synthesis and amplification were performed. The primer sequences were provided by Sangon Biotech (Shanghai, China) and Tsingke Biotechnology (Chongqing, China).

Western blotting assay

RIPA lysis buffer (cat. No. P0013B; Beyotime Biotechnology, Beijing, China) was added to the cell pellet and lysed on ice for at least 30 min. After ultrasonic treatment, centrifuge the lysate at 12,000 × g for 5–10 min. Collect the supernatants and mix them with 5X loading buffer (cat. No. P0015L; Beyotime Biotechnology, Beijing). Then perform gel electrophoresis, and transfer the proteins to a PVDF membrane. The PVDF membrane was removed and blocked with 5% nonfat milk. After washing with TBST solution, the PVDF membrane was incubated with an appropriate dilution of the primary antibodies overnight at 4 °C. After incubation, wash the PVDF membrane again. Subsequently, incubate the PVDF membrane with the secondary antibody for 1 to 2 h. Finally, chemiluminescence detection was carried out using an ECL detection reagent (cat. No. BG0001; Biogen, Chongqing, China), and the resulting signal was captured and recorded using an imaging system (Bio-Rad).

Histology assay

Hematoxylin-Eosin (HE) staining: Begin by hydrating the tissue sections and washing them with PBS. Then, stain the sections with hematoxylin. After washing with PBS, stain with eosin for 3 min. After washing with PBS, mount the sections using Neutral Balsam and capture images via a microscope (Leica, Germany). Safranin O-Fast Green staining: Begin by hydrating the tissue sections and washing them three times with PBS, each for 5 min. Then, add freshly prepared hematoxylin solution and stain for 5 min. Next, treat with acidic differentiation solution for 15 s. Stain the sections with the Fast Green solution for 5 min. Then, wash with a weak acid solution. After air-drying, immerse the sections in Safranin O staining solution for 5 min, and finally rinse them 1 to 2 times with absolute ethanol to remove residual staining solution. When the sections are completely dry, seal them with Neutral Balsam and observe under a microscope (Leica, Germany) while taking pictures for recording.

Transmission electron microscopy (TEM) assay

Electron microscopy fixative (Cat. No. G1102; Servicebio Biotechnology, Wuhan, China) was added to the cell precipitates treated under different conditions and thoroughly mixed. Centrifuge the cell suspension at 1000×g for 5 min to obtain the cell pellet, which was then stored at 4 °C for overnight fixation. Subsequently, the cell clusters underwent gradient dehydration, resin embedding, ultrathin sectioning, and staining. Finally, the cellular ultrastructure was captured and recorded using a TEM.

Immunofluorescence and immunohistochemical assay

After dewaxing and hydration, the tissue sections were immersed in PBS for 5 min, and then the sample areas were circled using a histochemical pen. A 4% paraformaldehyde solution was dropped onto the sample area for fixation for 20 min. Add trypsin antigen retrieval solution and incubate at 37 °C for 15 min. Endogenous peroxidase blocker and BSA blocking solution were added successively to the tissue sections. Then, add the primary antibody and incubate overnight. After incubation, the sections were washed with PBS. For immunofluorescence staining, the fluorophore-conjugated secondary antibody (mouse or rabbit) was dropped on the sections. After being incubated and washed, the sections were stained with DAPI solution. Finally, take and record the pictures under a microscope (Zeiss, German). For immunohistochemical staining, incubate the samples with HRP-labeled secondary antibody for 30 min. Wash the sections with PBS 2 to 3 times, each time for 5 min. Subsequently, add the DAB solution to sections, and observe the color development under a microscope. After the chromogenic reaction is completed, the slides were sealed with neutral gum before capturing images under a microscope (Zeiss, Germany).

RNA sequencing assay

The articular cartilage tissues from RA rats treated with or without GLP-1RAs were transferred to 1.5 mL EP tubes. 1 mL Trizol reagent was added to lyse the tissues and extract total RNA. The quality and integrity of the total RNA were detected using the NanoDrop ND-1000 (NanoDrop, Wilmington, DE, USA) and Bioanalyzer 2100 (Agilent, CA, USA) bioanalyzers. The concentration > 50 ng/µL and the RNA integrity number (RIN) ≥ 7.0 was set as the quality standard. Then, the samples were rapidly frozen and sent to Biotree (Shanghai, China) for RNA sequencing and bioinformatics analysis.

Flow cytometry assay

The flow cytometry was first used to detect the expression levels of COL2 and COL1 in chondrocytes. A volume of 100 µL of cell suspension was mixed with 2.5 µL of rabbit anti-COL2 antibody and 2.5 µL of mouse anti-COL1 antibody, and then incubated in the dark at 37 °C for 1 to 2 h. After the incubation was completed, the samples were washed twice. Then, 2 µL of Alexa Fluor^® 594-conjugated goat anti-mouse IgG and 2 µL of Alexa Fluor^® 488-conjugated goat anti-rabbit IgG were added and incubated for 1–2 h. After incubation, wash the samples once. Then, the cell precipitates were resuspended with 200 µL of PBS. The percentages of positive cell expressing COL1 or COL2 were determined by flow cytometry. When using flow cytometry to analyze the proportion of macrophages in synovial tissue or neutrophils in the skin tissue of the rat’s foot, a cell suspension of the synovial or skin tissue needs to be obtained. First, cut the tissue into 1–2 mm³ pieces. Then, add 2 mL each of 4 mg/mL type I collagenase and type II collagenase, and digest for 1–2 h at 37 ℃. After digestion, filter and centrifuge to obtain the cell pellet, and resuspend the cell pellet in 100 µL of PBS. Then add approximately 2 µL each of FITC-conjugated anti-CD11b, PE-conjugated anti-CD68, APC-conjugated anti-CD86, and PE-Cy7-conjugated anti-CD206 antibodies to the synovial tissue-derived cell suspension, and incubate the mixture at 37 °C for 30 min to identify the proportion of macrophages. Add approximately 2 µL each of PerCP/Cyanine5.5-conjugated anti-CD45, FITC-conjugated anti-CD11b, and CoraLite^® Plus 647-anti-CD43 antibodies to the skin tissue-derived cell suspension from the rat’s foot, and incubate the mixture at 37 °C for 30 min to identify the proportion of neutrophils. After incubation, the cells are washed once and resuspended in PBS. Flow cytometry was used to detect the proportions of macrophages and neutrophils.

Lipid peroxidation detection assay

After digestion of cells with trypsin, the cells were resuspended in complete culture medium. Centrifuge the cell suspension at 1000 rpm for 5 min to obtain the cell pellet. When performing CellROX™ Deep Red staining, 1 µL of CellROX™ Deep Red reagent was mixed with 9 µL of DMSO. Take 2 µL of this mixture and add it to 1 mL of culture medium. Then, resuspend the cell pellet in the CellROX™ Deep Red staining medium and incubate for 60 min. After incubation, the cell pellet was washed with PBS, and the lipid peroxidation level of the cells was analyzed by flow cytometry.

Reactive oxygen species (ROS) detection assay

For C11-BODIPY staining, the C11-BODIPY reagent was added to the cell suspension to achieve a final concentration of 1 to 10 µM, followed by incubation at 37 °C for 30 to 60 min. After the incubation was completed, the cells were washed three times with PBS solution, and the ROS levels were detected and analyzed by flow cytometry. When performing fluorescence staining to detect intracellular ROS levels, the cells were washed with the washing solution. After discarding the washing solution, the ROS staining solution was added and incubated at 37 °C for 30 to 60 min. After this incubation, wash the cells again with washing solution and then photograph under a microscope (Zeiss, Germany).

Enzyme-linked immunosorbent assay (ELISA)

ELISA assay kits were procured from Elabscience (Wuhan, China) and Enzyme Link Biotechnology (Shanghai, China). Weigh 100 mg of synovial tissue and transfer it into a round-bottom centrifuge tube. Add 500 µL of PBS to the tube. Cut the tissue into 1–2 mm³ fragments, followed by homogenization using a tissue grinder. Following homogenization, the samples were centrifuged at 5,000 × g for 5 min and the resulting supernatants were collected. Retrieve the pre-coated ELISA plate, and add 100 µL of standard solutions at varying concentrations or sample supernatants to designated wells in duplicate. Incubate the plate for 2 h. Following incubation, aspirate the contents and wash the plate five times with washing buffer. Subsequently, 100 µL of biotinylated detection antibody working solution was added to each well, and incubated for 1 h in the dark. Repeat the washing step as described above. Then, add 100 µL of 1× streptavidin-HRP (SA-HRP) working solution to each well, and incubate at 37℃ for 30 min in the dark. After another wash cycle, 50 µL of chromogenic substrate solution A and 50 µL of solution B were add to each well. Then mix gently and incubate for 15 min. Finally, after adding the terminating solution, the absorbance was measured at 450 nm using a microplate reader.

Cell viability, MDA, GSH and iron ion assay

Cell viability, malondialdehyde (MDA), glutathione (GSH), and intracellular iron ion levels were determined using the Cell Counting Kit-8 (cat. no. BG0025; BGBIOTECH, Chongqing, China), GSH colorimetric assay kit (cat. no. E-BC-K030-M; Elabscience, Wuhan, China), MDA colorimetric assay kit (cat. no. E-BC-K028-M; Elabscience), and iron colorimetric assay kit (cat. no. E1042; Applygen, Beijing, China), respectively. Collect chondrocytes from different treatment groups and process according to the manufacturers’ instructions. A standard curve was generated for each assay, and the concentrations of target molecules were calculated accordingly.

Synthesis of NPs, GLP-1RAs@NPs and GLP-1RAs@NPs-WYRGRL

Weigh 2 g of silk fibroin, obtained from Southwest University, Chongqing, and add it to 40 mL of a 9.3 M LiBr solution to dissolve it completely. Centrifuge at 6000 rpm for 5–10 min, collect the supernatant, and perform dialysis in a dialysis bag with a molecular weight range of 8000–14,000 Daltons for 2–3 days. After dialysis, centrifuge again at 6000 rpm, collect the supernatant and measure the protein concentration. Dilute the obtained silk fibroin solution to 10 mg/mL. Then, 5 mL of acetone solution was added to a 15 mL centrifuge tube and vortexed on a vortexed mixer. Drop 1 mL of 1% silk fibroin solution by mass fraction into 5 mL of acetone solution. Finally, the volume ratio of acetone solution to 10 mg/mL silk fibroin solution is 5:1. After complete the drop, continue to vortex for 30–60 s. Once vortexing is completed, centrifuge at 6000 rpm/min for 10 min. Discard the supernatant and collect the precipitate. Add 5 ml of deionized water to the precipitate and perform ultrasonic treatment on ice. The conditions are as follows: 2 min of total ultrasound time, consisting of cycles of 2 s of sonication followed by 2 s of rest, at an amplitude of 30%. After the ultrasound, the mixture was centrifuged at 10,000 rpm for 5–10 min to obtain the NPs precipitate. Then, 2 ml of deionized water was added to resuspend the precipitate to obtain the nanoparticles (NPs) solution. To prepare GLP-1RAs@NPs, the obtained silk fibroin solution was concentrated to 20 mg/ml. Subsequently, 500 µl of 3 mg/ml GLP-1RAs solution was added to 500 µl of 20 mg/ml silk fibroin solution to make the silk fibroin concentration 1%. Then, the 1% silk fibroin solution mixed with GLP-1RAs was dropped into 5 ml of acetone solution, vortexed, centrifuged, and the NPs precipitate was resuspended with double-distilled water to obtain the GLP-1RAs@NPs solution. To prepare GLP-1RAs@NPs-WYRGRL, take the GLP-1RAs@NPs solution and add 10 mg of EDC and 4 mg of WYRGRL to it. After cross-linking for 0.5 to 1 h on a shaker, 10 mg of NHS was added and the cross-linking was continued overnight on the shaker at room temperature. The next day, the synthesized nanoparticles solution was centrifuged at 10,000 rpm for 5–10 min. The GLP-1RAs@NPs-WYRGRL precipitate, which was coated with GLP-1RAs and cross-linked with WYRGRL, was obtained.

Characterization of NPs, GLP-1RAs@NPs and GLP-1RAs@NPs-WYRGRL

The morphology of NPs, GLP-1RAs@NPs, and GLP-1RAs@NPs-WYRGRL was observed by SEM. Ten µL of the pre-prepared NPs, GLP-1RAs@NPs, and GLP-1RAs@NPs-WYRGRL solutions were diluted to 1 mL. After thorough mixing, 20 µL of the NPs, GLP-1RAs@NPs, and GLP-1RAs@NPs-WYRGRL solutions were dropped onto silicon wafers and left to dry naturally. After gold spraying treatment of the samples, the morphology of NPs, GLP-1RAs@NPs, and GLP-1RAs@NPs-WYRGRL was observed and photographed under a SEM. The zeta potential and particle size of NPs, GLP-1RAs@NPs, and GLP-1RAs@NPs-WYRGRL were detected using a zeta potential and particle size analyzer. When conducting Fourier transform infrared spectroscopy analysis, we took NPs, GLP-1RAs@NPs, and GLP-1RAs@NPs-WYRGRL solutions, centrifuged them to obtain the precipitates, and then freeze-dried the precipitates. Take out NPs, GLP-1RAs@NPs, and GLP-1RAs@NPs-WYRGRL, and grind and press them into tablets in potassium iodide. Finally, analyze the spectra of NPs, GLP-1RAs@NPs, and GLP-1RAs@NPs-WYRGRL in a Fourier transform infrared spectrometer.

Envelopment efficiency and release rate of GLP-1RAs from GLP-1RAs@NPs

High-performance liquid chromatography (HPLC) was performed to measure the encapsulation efficiency and release rate. We took 300 µL of the 6 mg/mL GLP-1RAs solution and mixed it evenly with 1 mL of the 15 mg/mL silk fibroin solution. The GLP-1RAs@NPs were prepared by vortexing in an acetone solution, then centrifuged at 10,000 rpm for 10 min to obtain the GLP-1RAs@NPs precipitate. Subsequently, 300 µL of a 9.3 M LiBr solution was added to completely dissolve the GLP-1RAs@NPs precipitate. The peak area of GLP-1RAs in the dissolved precipitate was detected by HPLC. Based on the peak area of the reference sample (6 mg/mL GLP-1RAs), we calculated the concentration of GLP-1RAs in the dissolved GLP-1RAs@NPs precipitate according to their peak area ratio. Then, the mass m of GLP-1RAs in the GLP-1RAs@NPs precipitate was calculated, and the encapsulation efficiency E was calculated based on the total mass M of GLP-1RAs, with E = m/M * 100%.

Determine the release rate of GLP-1RAs from GLP-1RAs@NPs

Take six 300 µL samples of a 6 mg/mL GLP-1RAs solution, and uniformly mix each sample with 1 mL of a 15 mg/mL silk fibroin solution to prepare GLP-1RAs@NPs solution. Adjust the environment of the GLP-1RAs@NPs solution to a microenvironment similar to that of RA and shake it on a shaker at 37 °C. At 0, 1, 3, 5, 7, and 14 days, the GLP-1RAs@NPs precipitate was obtained by centrifugation. Then, 300 µL of 9.3 M LiBr solution was added to completely dissolve the GLP-1RAs@NPs precipitate. The mass of GLP-1RAs in the GLP-1RAs@NPs precipitate in the LiBr solution at 0 days was determined by HPLC and denoted as m0. The mass of GLP-1RAs in the GLP-1RAs@NPs precipitate at 1, 3, 5, 7, and 14 days was also determined by HPLC and denoted as mx. Therefore, the release rate R at different time points can be calculated as R = (m0 - mx) / m0 * 100%.

Ex vivo organ fluorescence imaging

Take 2 mL of the previously prepared GLP-1RAs@NPs and GLP-1RAs@NPs-WYRGRL solutions respectively, and add 5 µL of RBITC solution to each. Conduct the cross-linking reaction under light-protected conditions for 30 min. After the cross-linking reaction is completed, PBS solution is added for washing and then centrifuge to remove the unbound RBITC, thereby obtaining the RBITC-GLP-1RAs@NPs and RBITC-GLP-1RAs@NPs-WYRGRL precipitates. Rats aged 2 to 4 weeks were selected and evenly divided into three groups. After anesthesia, an equal amount of PBS, RBITC-GLP-1RAs@NPs or RBITC-GLP-1RAs@NPs-WYRGRL was injected into each rat through the tail vein, and fluorescence imaging of the heart, liver, spleen, lungs, kidneys, and both lower limbs was performed on the first and third days using an in vivo imaging system.

To observe the infection of RBITC-GLP-1RAs@NPs or RBITC-GLP-1RAs@NPs-WYRGRL on articular cartilage, rats aged 2 to 4 weeks were taken and evenly divided into three groups. After anesthesia, we injected an equal amount of PBS RBITC-GLP-1RAs@NPs or RBITC-GLP-1RAs@NPs-WYRGRL into multiple articular cartilages of each rat. On the first and third days, frozen sections of the joint tissues were made and the invasion depth and distribution of RBITC-GLP-1RAs@NPs or RBITC-GLP-1RAs@NPs-WYRGRL in the articular cartilage were observed under a fluorescence microscope.

Synthesis of MNs, MNs/GLP-1RAs@NPs or MNs/GLP-1RAs@NPs-WYRGRL

Take 0.5 g of hyaluronic acid and add it to 50 mL of deionized water. Stir it with a magnetic stirrer at 400 to 600 rpm until it is completely dissolved. Add 2 mL of methacrylic anhydride to the hyaluronic acid solution, then continuously drop 4 M NaOH solution to maintain the pH value of the methacrylic anhydride/hyaluronic acid solution between 8 and 9, and stir overnight at 4 °C. The next day, the synthesized HAMA solution was dialyzed in a dialysis bag. The dialysis was carried out for 3 days, with the deionized water being changed every 6 to 12 h. After the dialysis was completed, the HAMA solution was placed in an oven at 37℃ for drying. After 1 to 2 days, a HAMA solution with a concentration of approximately 3% was obtained. Take microneedle (MN) mold, with needle shapes resembling regular quadrangular pyramids, a needle length of 1200 μm, a base edge length of 500 * 500 μm, a needle spacing of 800 μm, and an array of 15 * 15. Immerse it in deionized H₂O, and repeatedly blow the tip of the microneedle with deionized water to ensure it is completely filled with deionized H₂O. After exhausting air from the microneedle tip, take it out and place it on the laboratory bench. Use a 200 µL pipette to remove the excess liquid from the microneedle, but the tip of the microneedle was kept filled with deionized H₂O.

Take 1 mL of the previously prepared GLP-1RAs@NPs or GLP-1RAs@NPs-WYRGRL solution and add it to 2 ml of HAMA solution. Mix well and remove the air bubbles in the liquid to obtain 3 ml of HAMA solution containing GLP-1RAs@NPs or GLP-1RAs@NPs-WYRGRL. Take the HAMA solution or the HAMA solution containing GLP-1RAs@NPs or GLP-1RAs@NPs-WYRGRL and add it to the MN mold with the needle tips filled with deionized water. During this process, the HAMA solution or the HAMA solution containing GLP-1RAs@NPs or GLP-1RAs@NPs-WYRGRL will blend with the deionized water at the needle tips, thereby replacing the deionized water at the needle tips with the HAMA solution containing GLP-1RAs@NPs or GLP-1RAs@NPs-WYRGRL. Place the MN mold filled with liquid in a 4℃ refrigerator for 2–3 days to allow the solution in the MN mold to concentrate to a non-flowing state. Take about 600 µl of the base solution for the MN and add it to the mold. Shake it left and right and front and back to evenly distribute the base solution in the MN mold. Continue to place it in the 4℃ refrigerator for 1–2 days to allow the liquid to evaporate as completely as possible. Dry the microneedles. Take the microneedle mold out of the 4 °C refrigerator and place it in a 37 °C oven to continue drying for 1–2 h until the microneedles are completely dry and solidified. Finally, slowly remove the microneedles from the edges of the mold to obtain Microneedles (MNs), MNs/GLP-1RAs@NPs or MNs/GLP-1RAs@NPs-WYRGRL.

Characterization of MNs/GLP-1RAs@NPs-WYRGRL

MNs dissolution assay. RBITC was used to label GLP-1RAs@NPs-WYRGRL and prepare MNs/GLP-1RAs@NPs-WYRGRL. The dried MNs/GLP-1RAs@NPs-WYRGRL were cut into a 1 × 6 array and placed in double-distilled water. The dissolution of the microneedles was observed and recorded under a fluorescence microscope at 0, 10, 30, and 60 min. To detect the release rate of GLP-1RAs@NPs-WYRGRL from MN/GLP-1RAs@NPs-WYRGRL, we first measured the absorbance of 2 ml of deionized water containing RBITC-GLP-1RAs@NPs-WYRGRL. Then, we combined an equal amount of RBITC-GLP-1RAs@NPs-WYRGRL with microneedles to prepare MN/RBITC-GLP-1RAs@NPs-WYRGRL, dissolved it in 2 mL of deionized water, and measured the absorbance at 0, 10, 20, 30, 60, and 120 min. Finally, the release rate at different time points was calculated based on the standard curve of RBITC-GLP-1RAs@NPs-WYRGRL absorbance. Mechanical property testing of microneedles. After drying, the MNs, MNs/GLP-1RAs@NPs or MNs/GLP-1RAs@NPs-WYRGRL were cut into 3 × 3 or 4 × 4 microneedle arrays. The total force of the microneedles after deformation was detected in the ElectroForce Biodynamic 5500 instrument and test system, and then the force of each microneedle was calculated. The distribution of GLP-1RAs@NPs-WYRGRL in MNs was observed by scanning electron microscopy. The previously prepared microneedles were dried and fixed on the needle holder. The distribution of GLP-1RAs@NPs-WYRGRL in MN was observed under low and high magnification in the SEM and photographed for record.

Safety evaluation

To analyze whether different treatment methods are toxic to rats, we examined the histological changes in the heart, liver, spleen, lungs and kidney tissues and the blood routine and blood biochemical indicators related to liver and kidney function of different groups of rats at 16 weeks. For blood routine tests, after anesthetizing the rats, the tip of the tail was cut to collect blood. The collected blood was placed in an anticoagulant tube and then tested on the machine for indicators such as WBC, RBC, HBG, and Lymph#. For liver and kidney function tests, about 800 µl of the collected blood was placed in a 1.5 ml EP tube and left at room temperature for 2 h. Then, it was centrifuged at 3000 rpm for 15 min. The supernatant was taken for the detection of liver and kidney function indicators such as ALT, AST, UREA, and CREA. Finally, the rats were anesthetized and sacrificed by cervical dislocation. The heart, liver, spleen, lung and kidney were removed for embedding, sectioning, and HE staining. Then, the sections were observed and photographed under a microscope.

Animal experiments

The adult female Sprague-Dawley (SD) rats (about 160 g, 6 weeks old), SD juvenile rats (50–100 g, about 2–4 weeks old) and SD neonatal rats (20–40 g, about 1–2 weeks old) used in the experiment were purchased from the Laboratory Animal Center of Army Medical University. The animal use and experimental procedures met the requirements of the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Animal Ethics Committee of the Army Medical University (no. AMUWEC20250028).

Establish a collagen-induced arthritis (CIA) rat model, that is, rheumatoid arthritis (RA ) model. Prepare type II bovine collagen at a concentration of 2 mg/mL. Take 5 mL of Freund’s complete adjuvant and add it to 5 mL of bovine type II collagen solution. Transfer the mixed solution into a 50 mL centrifuge tube. Use a handheld small electric mixer with a smooth round head rotor to vortex it on ice at a speed of over 10,000 rpm until an emulsion is formed. The emulsion does not disperse within 1 min after being dropped into water. Primary immunization. Six to eight-week-old female SD rats (about 160 g) were anesthetized. The right plantar surface of the foot was disinfected with 75% alcohol, and then about 200 µl of bovine type II collagen emulsion was injected subcutaneously into the foot sole using a 1 mL syringe. After injection, the injection site was pressed for 1 min. The second immunization was carried out on the 21st day after the first immunization. About 200 µl of bovine type II collagen emulsion was injected into the root of the tail using a 1 mL syringe by multiple-point injection. After the second immunization, the CIA-induced rat model was induced.

To evaluate whether MNs could cause further inflammation and neutrophil aggregation, we conducted flow cytometry and RT-qPCR to detect the migration of neutrophils and the levels of inflammatory factors in the foot skin tissues of rats in the control group, microneedle-removed post-acupuncture group, and microneedle-retained post-acupuncture group, with 4–6 rats in each group. To demonstrate the efficacy of GLP-1RAs in the treatment of RA, rats were divided into NC, RA and RA + GLP-1RAs groups, with 4–6 rats in each group. To demonstrate the superior efficacy of the MNs/GLP-1RAs@NPs-WYRGRL in the treatment of RA, rats were divided into NC group, RA group, RA + NPs group, RA + GLP-1RAs@NPs group, RA + GLP-1RAs@NPs-WYRGRL group, RA + MNs group and RA + MNs/GLP-1RAs@NPs-WYRGRL group, with 4–6 rats in each group. After the CIA-induced rat model was established, treatment was initiated in the 4th week and continued for 1–2 times per week until the 16th week. The clinical symptoms, joint scores of the hind feet and body weight of the rats were evaluated weekly. The dosage and administration of GLP-1RAs for treatment is to draw 50 µl of 6 mg/ml GLP-1RAs and inject it into the sole of each rat’s foot. After injection, press the injection site for 1 min. The dosage and administration of GLP-1RAs@NPs, GLP-1RAs@NPs-WYRGRL or MNs/GLP-1RAs@NPs-WYRGRL were as follows: 50 µl of 6 mg/ml GLP-1RAs was drawn to prepare GLP-1RAs@NPs, GLP-1RAs@NPs-WYRGRL or MNs/GLP-1RAs@NPs-WYRGRL, which was then injected or treated with MNs puncture into the sole of each rat’s foot. After the treatment was completed, the safety of the treatment method was evaluated in rats. Then, the hind paws of the rats were embedded, sectioned, stained with HE and safranin O-fast green (SO/FG), and subjected to immunohistochemical and immunofluorescence staining.

Statistical analysis

All data were presented as mean ± standard deviation. The average fluorescence intensity of fluorescence staining and the ratio of positive area of immunohistochemical staining were analyzed using Image J. Data were statistically analyzed using GraphPad Prism 10.3.0 software (GraphPad Software Inc., CA, USA). The Student’s t-test was used for comparison between two groups, and one-way ANOVA followed by Tukey’s test was used for comparison among multiple groups. Statistical significance was set at p < 0.05.

Results

GLP-1RAs alleviate RA progression by suppressing synovial inflammation and cartilage destruction

Recent studies have demonstrated that GLP-1RAs not only exert significant metabolic effects in the management of diabetes and obesity, but also possess anti-inflammatory properties [33]. However, their application in the treatment of RA remains in the initial exploratory stage. In this study, GLP-1RAs were used to treat RA rats. The results showed that collagen-induced arthritis (CIA) rats exhibited obvious pathological manifestations of RA, such as swelling and deformity of the hind paws and joints; however, after treatment with GLP-1RAs, these symptoms were significantly alleviated (Fig. 1A). In addition, the paw thickness, clinical score and body weight in the RA + GLP-1RAs treatment group were significantly reduced at week 16 compared with the RA group (Fig. 1B). Micro-CT analysis further confirmed that the articular surface of hind paws was severely damaged, accompanied by osteophyte formation in the RA model group. However, after GLP-1RAs treatment, the joint structural damage was significantly alleviated and osteophyte formation was reduced (Fig. 1C). To evaluate the pathological changes in joint tissues, we performed HE staining and SO/FG staining on the joint tissues of rats in the NC group, RA group, and RA + GLP-1RAs group. The results demonstrated that the articular cartilage surface of rats in the RA group was significantly eroded, accompanied by extensive inflammatory cell infiltrations, while the cartilage structure in the RA + GLP-1RAs treatment group was relatively intact, and inflammatory cells were significantly reduced (Fig. 1D-E). Subsequently, the expression levels of GLP-1R and inflammation-related factors in the articular cartilage tissues were detected by RT-qPCR and immunohistochemistry. The results indicated that the expression of GLP-1R was upregulated, while the expression levels of IL-6 and TNF-α decreased in the GLP-1RAs treatment group compared with the RA group (Fig. 1F). The results of immunohistochemical staining showed that the positive staining areas of IL-6, GLP-1R, and TNF-α differed significantly between the RA group and the GLP-1RAs treatment group (Fig. 1G-H). Furthermore, ELISA detection showed that after GLP-1RA treatment, the levels of IL-6, IL-1β, and TNF-α in synovial tissue were significantly reduced (Fig. 1I). In conclusion, GLP-1RAs can effectively inhibit articular cartilage destruction and synovial inflammation, thereby alleviating the progression of RA and its clinical symptoms.

Fig. 1.

The effect of GLP-1 receptor agonists (GLP-1RAs) in alleviating the symptoms of rheumatoid arthritis (RA). (A) Observe the morphological changes in the hind paws of rats in the NC, RA, and RA + GLP-1RAs groups, with 4–6 rats per group. (B) Detect and analyze the thickness and clinical scores of the hind paws and body weight of rats in the above groups. (C) Detect the structural changes of the metatarsophalangeal joints of the hind paws in the NC group, RA group and RA + GLP-1RAs group rats via Micro-CT. (D-E) Evaluate the pathological changes of the joint tissue structure through histological staining. (F) The mRNA expression levels of GLP-1R, IL-6, and TNF-α in the articular cartilage tissue in the RA group and RA + GLP-1RAs group rats were detected by RT-qPCR. (G) Detect the protein expression of GLP-1R, IL-6 and TNF-α in the articular cartilage tissue through immunohistochemical staining. (H) Quantitatively analyze the positive staining area of GLP-1R, TNF-α, and IL-6 in the cartilage tissues of different groups. (I) ELISA was used to detect the levels of inflammatory factors in the synovial tissues of the RA group and the RA + GLP-1RAs group. (ns: not significant, *p < 0.05, **p < 0.01, and ****p < 0.0001)

GLP-1RAs inhibit chondrocyte ferroptosis and alleviate RA-induced joint damage via the cGAS-STING signaling pathway

To further investigate the mechanism underlying the inhibitory effects of GLP-1RAs on RA, RNA sequencing analysis was conducted using the articular cartilage tissues from RA rats treated with or without GLP-1RAs. The results of the volcano plot indicated significant differences in gene expression between the two groups (Fig. S1A). Heatmap analysis revealed that the expression levels of STING were reduced in RA + GLP-1RAs compared to the RA group. Additionally, the expression levels of ACSL4 and Hmox1, which are involved in ferroptosis, were also decreased, whereas the expression of GPX2, a gene known to inhibit ferroptosis, was upregulated in the RA + GLP-1RAs group. Meanwhile, the expression levels of M1 macrophage marker gene CD86 and IL-6, IL-1β, and TNF-α were markedly reduced in the RA + GLP-1RAs group (Fig. S1B). The results of KEGG enrichment analysis showed that the RA group showed significant enrichment of rheumatoid arthritis pathway, TNF signaling pathway, osteoclast differentiation, ferroptosis, and cytosolic DNA sensing pathway (Fig. S1C). GSEA enrichment analysis further demonstrated that the RA group showed significant enrichment and activation in biological processes such as osteoclast differentiation, cellular response to IFN, cellular iron ion homeostasis, and ferric iron binding (Fig. S1D), as well as signaling pathways such as rheumatoid arthritis, glycosaminoglycan biosynthesis chondroitin sulfate, cytosolic DNA sensing pathway, and ferroptosis (Fig. S1E). Based on the above results and our previous findings that inhibiting the cGAS-STING signaling pathway could suppress ferroptosis in skeletal stem cells derived from costal cartilage tissue [34], we hypothesized that GLP-1RAs might inhibit ferroptosis of chondrocytes by suppressing the activation of the cGAS-STING signaling pathway, and promote the synthesis of glycosaminoglycans and chondroitin sulfate, thereby alleviating the development of RA. To verify this hypothesis, we conducted immunohistochemical staining on joint tissues in vivo. The results showed that GLP-1RAs reduced the STING and ACSL4 proteins expression in cartilage tissue (Fig. S2A and S2B). To clearly determine the effects of GLP-1RAs on RA-related signaling pathways and protein expression levels through in vitro cell experiments, we isolated chondrocytes from rat articular cartilage. Flow cytometry results indicated that the proportion of COL2-positive cells among the extracted cells was approximately 86.2%, while the proportion of COL1-positive cells was about 2.42% (Fig. S3A). Immunofluorescence staining results also demonstrated that the isolated and cultured cells highly expressed COL2 and lowly expressed COL1 (Fig. S3B). Therefore, it can be concluded that the isolated cells are chondrocytes. Subsequently, transmission electron microscope (TEM) observations showed that LPS treatment led to mitochondrial shrinkage, increased membrane density and reduced volume, and 2’,3’-cGAMP further aggravated this pathological change, while C-176 or GLP-1RAs could effectively alleviate the above-mentioned mitochondrial damage (Fig. 2A). The levels of oxidative stress were detected by CellROX™ Deep Red staining (Fig. 2B-C) and the reactive oxygen species (ROS) levels were detected via ROS staining (Fig. 2D-E) and BODIPY C11staining (Fig. 2F-G). The results indicated that the oxidative stress and ROS levels induced by LPS and 2’,3’-cGAMP could be significantly inhibited by C-176 or GLP-1RAs. In vitro experiments of western blotting indicated that the expression of GLP-1R in chondrocytes treated with LPS significantly decreased, while the expression levels of STING, ACSL4, and IL-6 significantly increased. Following GLP-1RAs treatment, the expression of GLP-1Rs was significantly up-regulated, and the expression of STING, ACSL4 and IL-6 were effectively inhibited (Fig. 2H-I). To further verify the effect of the cGAS-STING signaling pathway on regulating ferroptosis in chondrocytes, we added 2’,3’-cGAMP, C-176 or GLP-1RAs to the LPS-induced inflammatory group. Western blotting results showed that 2’,3’-cGAMP promoted chondrocyte ferroptosis and the release of TNF-α and IL-6 by up-regulating STING expression, p-TBK1, p-IRF3 and down-regulating GPX4 expression, however, this trend could be reversed by C-176 or GLP-1RAs (Fig. 2J-K). The RT-qPCR results indicated that GLP-1RAs decreased the expression of NF-κB, Type I IFN, ACSL4, IL-1β and TNF-α, and increased the expression of GPX4 (Fig. S4). In conclusion, GLP-1RAs can inhibit chondrocyte ferroptosis and alleviate the progress of RA similar to that of C-176 by inhibiting cGAS-STING signaling pathway activation, thereby reducing ROS production and inhibiting ferroptosis(Fig. 2L).

Fig. 2.

GLP-1RAs had a significant impact on inhibiting ferroptosis in chondrocytes. (A) The mitochondrial damage of cells in the NC group, LPS group, LPS + GLP-1RAs group, 2’,3’-cGAMP group, LPS + 2’,3’-cGAMP group, LPS + 2’,3’-cGAMP + C-176 and LPS + 2’,3’-cGAMP + GLP-1RAs group was detected by TEM. (B-C) The lipid peroxidation levels and statistical analysis of mean fluorescence intensity (MFI) in chondrocytes treated as above. (D-E) Fluorescence staining reactive for oxygen species (ROS) and statistical analysis of MFI in chondrocytes treated with NC group, LPS group, LPS + GLP-1RAs group, 2’,3’-cGAMP group, LPS + 2’,3’-cGAMP group, LPS + 2’,3’-cGAMP + C-176 and LPS + 2’,3’-cGAMP + GLP-1RAs group. (F-G) C11-BODIPY staining and statistical analysis of MFI for ROS levels in different chondrocytes treatment groups. (H-I) Western blotting and quantitative analysis of GLP-1R, STING, ACSL4 and IL-6 proteins in chondrocytes of the NC group, LPS group and LPS + GLP-1RAs group. (J-K) Western blot and quantitative analysis of STING, p-TBK1, TBK1, p-IRF3, IRF3, ACSL4, GPX4, IL-6, and TNF-α proteins in chondrocytes of the LPS group, LPS + 2’,3’-cGAMP group LPS + 2’,3’-cGAMP + C-176 and LPS + 2’,3’-cGAMP + GLP-1RAs group. (L) Flowchart depicting the mechanism by which GLP-1RAs inhibit the generation of reactive oxygen species (ROS) through the suppression of the cGAS-STING pathway, thereby inhibiting ferroptosis. (ns: not significant, *p < 0.05, **p < 0.01, and ***p < 0 0.001)

GLP-1RAs promote the polarization of macrophages toward the M2 phenotype, reshape the immune microenvironment and attenuate inflammation in RA

Further GO enrichment analysis revealed that biological processes such as interleukin-6 production, IL-1β production, TNF-α production, macrophage activation, and macrophage migration were enriched in the RA group compared to the RA + GLP-1RAs group (Fig. S5A). GSEA enrichment analysis also showed that the biological functions such as positive regulation of macrophage cytokine production, positive regulation of IL-1β, IL-6, and TNF-α production, as well as signaling pathways such as Fc gamma r-mediated phagocytosis and TNF signaling pathway, were significantly enriched and activated in the RA group (Fig. S5B), suggesting that GLP-1RAs might have the function of inhibiting macrophage activation in RA (Fig. S5C). To validate this mechanism, the macrophages were treated with NC, LPS, or LPS + GLP-1RAs. Flow cytometry results showed that GLP-1RAs significantly reduced the population of LPS-induced M1 macrophages while increasing the population of M2 macrophages (Fig. 3A-B). Immunofluorescence staining results further confirmed that GLP-1RAs could decrease the proportion of M1 macrophages and increase that of M2 macrophages (Fig. 3C-D). The results of RT-qPCR indicated that GLP-1RAs downregulated the expression of IL-6, TNF-α, and iNOS, and upregulated the expression of IL-10, TGF-β, and Arg-1 (Fig. 3E). ELISA detection results also showed that GLP-1RAs suppressed the release of IL-1β, IL-6 and TNF-α induced by LPS, and promoted the release of IL-10 (Fig. 3F). Collectively, GLP-1RAs play a crucial role in promoting the polarization of macrophages towards the M2 phenotype and alleviating inflammation.

Fig. 3.

Effects of GLP-1RAs on the polarization of macrophages. (A-B) The proportion and statistical analysis of CD68+CD86+ M1-type macrophages and CD68+CD206+ M2-type macrophages in the NC group, LPS group and LPS+GLP-1RAs group. (C-D) Statistical analysis of the ratio of CD86+ and CD206+ cells within the CD68+ cells in the NC group, LPS group and LPS+GLP-1RAs group. (E) Representative detection results of mRNA expression levels of marker genes in macrophages from the NC group, LPS group, and LPS + GLP-1RAs group. (F) Application of ELISA to detect the levels of inflammatory cytokines secreted by macrophages in the NC group, LPS group, and LPS+GLP-1RAs group. (ns: not significant, *p < 0.05, **p < 0.01, ***p <0 .001, and ****p <0 .0001).

Synthesis and characterization of GLP-1RAs@NPs-WYRGRL

To enhance the local concentration, stability, and retention time of GLP-1RAs in the joint area, GLP-1RAs were encapsulated into silk fibroin nanoparticles (GLP-1RAs@NPs) and the surface of GLP-1RAs@NPs was further modified with the WYRGRL peptide, thereby developing GLP-1RAs@NPs-WYRGRL for the treatment of RA (Fig. 4A). The Scanning Electron Microscopy (SEM) results showed that NPs, GLP-1RAs@NPs and GLP-1RAs@NPs-WYRGRL all presented spherical morphologies (Fig. 4B). The zeta potential and particle size analysis results indicated that the zeta potentials of NPs, GLP-1RAs@NPs and GLP-1RAs@NPs-WYRGRL were approximately − 14 mV, -25 mV and + 8 mV, respectively (Fig. 4C), and the particle sizes were approximately 113.3 nm, 145.3 nm and 250.2 nm, respectively (Fig. 4D). The differences of particle sizes were statistically significant (Fig. 4E). Fourier transform infrared spectroscopy (FTIR) analysis revealed that there were differences among NPs, GLP-1RAs@NPs and GLP-1RAs@NPs-WYRGRL in the range of 1500–1700 cm⁻¹ and 3200–3400 cm⁻¹ (Fig. 4F), indicating successful conjugation and structural modification. To analyze the encapsulation efficiency of GLP-1RAs@NPs, high-performance liquid chromatography (HPLC) was used to detect the retention time and area of GLP-1RAs or GLP-1AR@NPs. The results showed that the retention time of GLP-1RAs was approximately 2.313 min, while that of GLP-1RAs@NPs was about 2.385 min (Fig. 4G). Based on the peak area, when the mass ratio of Silk to GLP-1RAs was 5:1 and 10:1, the encapsulation efficiency of GLP-1RAs@NPs was approximately 28% and 33%, respectively, but with no statistically significant (Fig. 4H). Furthermore, the release rate of GLP-1RAs from GLP-1RAs@NPs over 1, 3, 5, 7, and 14 days was also assessed by HPLC. The results indicated that approximately 60% of GLP-1RAs was released by the 14th day (Fig. 4I).

Fig. 4.

Preparation and Characterization of GLP-1RAs@NPs-WYRGRL. (A) Flowchart of the preparation process of GLP-1RAs@NPs-WYRGRL. (B) Morphological characteristics of NPs, GLP-1RAs@NPs and GLP-1RAs@NPs-WYRGRL observed by scanning electron microscopy (SEM). (C-D) Zeta potential and particle size distribution of NPs, GLP-1RAs@NPs and GLP-1RAs@NPs-WYRGRL. (E) Statistical analysis of the particle size of NPs, GLP-1RAs@NPs and GLP-1RAs@NPs-WYRGRL. (F) Infrared spectra of NPs, GLP-1RAs@NPs and GLP-1RAs@NPs-WYRGRL detected by Fourier transform infrared spectroscopy (FTIR). (G) Detection of GLP-1RAs and GLP-1RAs@NPs by high-performance liquid chromatography (HPLC). (H) The encapsulation efficiency of GLP-1RAs in GLP-1RAs@NPs was detected and statistically analyzed when the mass ratio of silk to GLP-1RAs was 5:1 and 10:1 respectively. (I) Statistical analysis of the GLP-1RAs release rate from GLP-1RAs@NPs. (ns: not significant, ****p < 0.0001).

GLP-1RAs@NPs-WYRGRL inhibited ferroptosis in chondrocytes by suppressing the cGAS-STING signaling pathway and promoted macrophage polarization toward the M2 phenotype

In this study, western blotting was first conducted to investigate the ability of GLP-1RAs@NPs-WYRGRL to inhibit ferroptosis in chondrocytes. Western blotting results showed that the expression levels of IL-6, cGAS, STING, and ACSL4 were elevated while GLP-1R expression decreased in chondrocytes treated with LPS compared to the NC group. In contrast, GLP-1RAs@NPs-WYRGRL significantly suppressed the expression of IL-6, cGAS, STING and ACSL4 and enhanced the expression of GLP-1R in chondrocytes induced by LPS compared with the NPs-WYRGRL treatment group (Fig. 5A). CellROX™ Deep Red staining and C11-BODIPY staining results showed that GLP-1RAs@NPs-WYRGRL markedly reduced lipid peroxidation (Fig. 5B) and intracellular ROS levels (Fig. 5C) in LPS-induced chondrocytes compared to the NPs-WYRGRL group. Furthermore, the assessments of cell viability, GSH levels, MDA levels, and iron ions levels indicated that GLP-1RAs@NPs-WYRGRL not only significantly reversed the reduction of cell viability and GSH levels caused by LPS, but also attenuated the elevation of MDA and iron ion levels induced by LPS (Fig. 5D). This study further evaluated the effect of GLP-1RAs@NPs-WYRGRL on macrophage polarization. Rhodamine B Isothiocyanate (RBITC) was used to label NPs-WYRGRL and GLP-1RAs@NPs-WYRGRL for immunofluorescence staining. The results of immunofluorescence staining showed that the NPs-WYRGRL did not significantly increase CD206 expression, whereas the GLP-1RAs@NPs-WYRGRL markedly promoted CD206 expression in macrophages with LPS treatment (Fig. 5E). Flow cytometry analysis revealed a significant decrease in the proportion of M1 macrophages and a concurrent increase in the M2 macrophage population in the GLP-1RAs@NPs-WYRGRL group compared to both the LPS group and the NPs-WYRGRL group (Fig. 5F-G).

Fig. 5.

The role of GLP-1RAs@NPs-WYRGRL in modulating ferroptosis of chondrocytes and polarization of macrophages. (A) Western blot analysis of the expression levels of GLP-1R, STING, cGAS, ACSL4, and IL-6 in chondrocytes treated with NC, lipopolysaccharide (LPS), LPS + NPs-WYRGRL or LPS + GLP-1RAs@NPs-WYRGRL. (B) Detection and quantitative analysis of lipid peroxidation in chondrocytes treated with NC, LPS, LPS + NPs-WYRGRL, or LPS + GLP-1RAs@NPs-WYRGRL via CellROX™ Deep Red staining. (C) Detection and statistical analysis of intracellular ROS levels under various treatment groups via C11-BODIPY staining. (D) Statistical analysis of cell viability and GSH, MDA and iron ions levels under various treatment groups. (E) Immunofluorescence staining for CD206 in macrophages treated with LPS, RBITC-NPs-WYRGRL or RBITC-GLP-1RAs@NPs-WYRGRL. (F-G) Detection and statistical analysis of the proportions of CD86 + M1-type macrophages and CD206 + M2-type macrophages cells in CD68 + macrophages treated with LPS, LPS + NPs-WYRGRL or LPS + GLP-1RAs@NPs-WYRGRL. (ns: not significant, *p < 0.05, **p < 0.01, and ***p < 0 0.001)

WYRGRL modification enhanced targeted delivery and retention of GLP-1RAs@NPs in cartilage

In this study, RBITC was used to label GLP-1RAs@NPs and GLP-1RAs@NPs-WYRGRL, and the targeting ability of RBITC-GLP-1RAs@NPs and RBITC-GLP-1RAs@NPs-WYRGRL toward chondrocytes and articular cartilage in vivo and in vitro was observed through fluorescence staining (Fig. 6A and C). The results of the study in vitro indicated that, regardless of whether it was on the first or third day, the uptake of GLP-1RAs@NPs-WYRGRL by chondrocytes was significantly higher than that of GLP-1RAs@NPs, and the difference in mean fluorescence intensity (MFI) was statistically significant (Fig. 6B). To verify whether GLP-1RAs@NPs-WYRGRL has the ability to target articular cartilage, GLP-1RAs@NPs and GLP-1RAs@NPs-WYRGRL were labeled with RBITC and injected into the experimental animals via the tail vein (Fig. 6C). The distribution of fluorescence signal in the liver, kidneys, heart, spleen, lungs, and lower limb joints was observed using an in vivo imaging system on the first and third days, respectively. The results showed that compared with the RBITC-GLP-1RAs@NPs group, the fluorescence intensity in the kidney was lower on the first day in the rats injected with RBITC-GLP-1RAs@NPs-WYRGRL, while the fluorescence intensity in the lower limb joints was significantly increased. On the third day, this trend remained consistent, and the fluorescence intensity in the lower limb joints was further increased compared with that on the first day (Fig. 6D). To further evaluate the invasion and the distribution area of RBITC-GLP-1RAs@NPs-WYRGRL in articular cartilage, RBITC-GLP-1RAs@NPs or RBITC-GLP-1RAs@NPs-WYRGRL was injected into the joint cavity, then the fluorescence intensity of articular cartilage tissue sections was observed and calculated on the first and third days (Fig. 6E). The results indicated that, regardless of whether it was on the first day or the third day, the depth of penetration and the area of the RBITC-GLP-1RAs@NPs-WYRGRL group into the articular cartilage were significantly higher than those of the RBITC-GLP-1RAs@NPs group (Fig. 6F). Moreover, with the extension of time, the fluorescence intensity in the cartilage tissue gradually increased (Fig. 6G). In conclusion, GLP-1RAs@NPs-WYRGRL significantly enhances the targeting efficiency of drug-loaded nanoparticles to chondrocytes and articular cartilage, prolongs their retention time in cartilage tissue, and thereby reduces excretion through the kidneys.

Fig. 6.

Detection of the targeting characteristics of GLP-1RAs@NPs-WYRGRL to chondrocytes and cartilage tissues. (A) Schematic diagram of the invasion of RBITC-GLP-1RAs@NPs and RBITC-GLP-1RAs@NPs-WYRGRL to chondrocytes or cartilage tissues. (B) Uptake analysis of RBITC-NPs and RBITC-NPs-WYRGRL by chondrocytes in vitro, along with statistical analysis of their MFI on the first and third days. (C) Flowchart of tail vein injection of RBITC-GLP-1RAs@NPs and RBITC-GLP-1RAs@NPs-WYRGRL. (D) Evaluate the tissue distribution of RBITC-GLP-1RAs@NPs and RBITC-GLP-1RAs@NPs-WYRGRL in the kidney, lung, spleen, liver, heart, and articular cartilage on the first and third days following intravenous administration via tail vein injection. Experiments were performed in three groups of SD rats (n = 6 per group): NC group, and two treatment groups administered RBITC-GLP-1RAs@NPs grou or RBITC-GLP-1RAs@NPs-WYRGRL, respectively. (E) Schematic diagram of intra-articular injection of RBITC-GLP-1RAs@NPs and RBITC-GLP-1RAs@NPs-WYRGRL. (F) Analyze the infiltration and distribution of RBITC-GLP-1RAs@NPs and RBITC-GLP-1RAs@NPs-WYRGRL within articular cartilage on the first and third days via fluorescence detection. Experiments were performed in three groups of SD rats (n = 6 per group): NC group, and two treatment groups administered RBITC-GLP-1RAs@NPs grou or RBITC-GLP-1RAs@NPs-WYRGRL, respectively. (G) Statistical analysis of the MFI of RBITC-GLP-1RAs@NPs and RBITC-GLP-1RAs@NPs-WYRGRL within articular cartilage on the first and third days. (**p < 0.01, ***p < 0.001 and ****p < 0.0001)

Synthesis and characterization of MNs/GLP-1RAs@NPs-WYRGRL

MNs could rapidly dissolve in the skin, achieving local drug release. In this study, a HAMA solution was combined with GLP-1RAs@NPs-WYRGRL to prepare targeted drug-loaded MNs (MNs/GLP-1RAs@NPs-WYRGRL) (Fig. 7A). The results of SEM showed that the HAMA MNs were complete in shape, neatly arranged, and had clear tip structures, indicating the successful preparation of HAMA MNs (Fig. 7B). To verify whether GLP-1RAs@NPs-WYRGRL was successfully loaded into the microneedles, GLP-1RAs@NPs-WYRGRL was labeled with RBITC and fluorescence observation was conducted. The results showed that there were a large number of red fluorescence signals inside the MNs/GLP-1RAs@NPs-WYRGRL compared with MNs (Fig. 7C). SEM further confirmed that a large number of GLP-1RAs@NPs-WYRGRL were visible inside the MNs/GLP-1RAs@NPs-WYRGRL (Fig. 7D). The results of dissolution experiment showed that the MNs/GLP-1RAs@NPs-WYRGRL gradually dissolved in deionized water with the extension of time (Fig. 7E). Mechanical property analysis indicated that when compressed by about 0.65 mm, MNs could withstand approximately 1.85 N of force, while the MNs/GLP-1RAs@NPs-WYRGRL could withstand less pressure, about 1.6 N (Fig. 7F-G). When the compression distance of the needle tip was set to either 0.25–0.5 mm, the tip pressure resistance gradually decreased from the MNs, MNs/GLP-1RAs@NPs to MNs/GLP-1RAs@NPs-WYRGRL groups with statistically significant differences (Fig. 7H). The results of drug release experiments showed that the MNs/GLP-1RAs@NPs-WYRGRL could release approximately 92% of GLP-1RAs@NPs-WYRGRL within 2 h post-administration (Fig. 7I). Additionally, this study further evaluated the penetration ability of the MNs/GLP-1RAs@NPs-WYRGRL into the skin and the tip retention characteristics. By observing the skin morphology and fluorescence signals, it was found that MNs/GLP-1RAs@NPs-WYRGRL could not only effectively puncture the skin but also retain the tip within the skin tissue (Fig. 7J-K). In conclusion, MNs/GLP-1RAs@NPs-WYRGRL can serve as an efficient drug delivery system to achieve sustained GLP-1RAs release at the disease site and prolong the drug effect.

Fig. 7.

Preparation and Characterization of Microneedles (MNs)/GLP-1RAs@NPs-WYRGRL. (A) Flowchart of the preparation process of MNs/GLP-1RAs@NPs-WYRGRL. (B) Morphological characteristics of HAMA MNs were observed by SEM. (C) Fluorescence imaging of MNs and MNs/GLP-1RAs@NPs-WYRGRL. (D) Observe the distribution of GLP-1RAs@NPs-WYRGRL in MNs and MNs/GLP-1RAs@NPs-WYRGR via SEM. (E) Analyze the dissolution of MNs/GLP-1RAs@NPs-WYRGRL in PBS solution at 0, 10, 30, and 60 min. (F-G) Mechanical strength testing and statistical analysis of MNs, MNs/GLP-1RAs@NPs, and MNs/GLP-1RAs@NPs-WYRGRL. (H) Detect the mechanical strength of MNs, MNs/GLP-1RAs@NPs, and MNs/GLP-1RAs@NPs-WYRGRL at compression depths of 0.25 mm and 0.5 mm. (I) Analyze the release rate of GLP-1RAs@NPs-WYRGRL from MNs/GLP-1RAs@NPs-WYRGRL. (J) Transdermal penetration experiments of MNs/GLP-1RAs@NPs-WYRGRL were conducted to evaluate its skin penetration ability. (K) Detected and analyze the microneedle tips distribution of MNs/RBITC-GLP-1RAs@NPs-WYRGRL after transdermal penetration. (ns: not significant, *p < 0.05 and ****p < 0.0001)

MNs/GLP-1RAs@NPs-WYRGRL significantly enhanced the therapeutic effect on RA

Although MN patches have several advantages, they can sometimes cause tiny punctures that trigger further migration of inflammatory cells and elicit an inflammatory response. To analyze the influence of this inflammatory response on microneedle therapy, we detected the migration of inflammatory cells by flow cytometry in the NC group, the MN-removed-after-puncture group, and the MN-puncture group, mainly including monocytes and neutrophils (Fig. S6A). The results showed that the ratio of CD45 + CD11b+CD43 + inflammatory cells, which mainly consist of monocytes and neutrophils, induced by MNs or MN-induced puncture injuries increased significantly on day 1, but no significant difference is observed compared with the NC group on days 2 and 3 (Fig. S6B-C). The RT-qPCR results also indicated that the inflammatory factors IL-6 and TNF-α increased significantly on day 1, but there were no significant differences compared to the NC group on days 2 and 3 (Fig. S6D). The results indicate that the damage caused by MNs recovers quickly, and the impact on neutrophils or monocytes in the tissue is not significant after 2–3 days. To evaluate the advantages of MNs/GLP-1RAs@NPs-WYRGRL on treating RA, the RA rats were induced via CIA method and were divided into the NC group, RA group, RA + NPs group, RA + GLP-1RAs@NPs group, RA + GLP-1RAs@NPs-WYRGRL group, RA + MNs group and RA + MNs/GLP-1RAs@NPs-WYRGRL group. In the experimental groups, treatment was administered 1–2 times every 7 days, and the changes in hind paw thickness and clinical scores were recorded during the treatment process (Fig. 8A). After microneedle puncture, clear microneedle puncture marks could be seen on the surface of the hind paws of the rats (Fig. 8B). At the 16th week, Micro-CT, HE staining, SO/FG staining, immunohistochemistry, immunofluorescence staining and flow cytometry analysis were performed. From the 8th week on, the clinical scores of the hind paws in each treatment group gradually decreased, and the thickness of the hind paws gradually reduced. By the 16th week, compared with other treatment groups, the MNs/GLP-1RAs@NPs-WYRGRL group showed the most significant improvement in the clinical scores and thickness of the hind paws compared with the RA group with statistically significant difference. Moreover, there were significant differences in body weight between the NC group and the GLP-1RAs@NPs group. In contrast, no statistically significant differences in body weight were observed between the NC group and the GLP-1RAs@NPs-WYRGRL group or the MNs/GLP-1RAs@NPs-WYRGRL group (Fig. 8C). At the 16th week, the swelling and deformity of the hind paw joints in the MNs/GLP-1RAs@NPs-WYRGRL group of rats were also the most significantly alleviated (Fig. 8D). Micro-CT results showed that the RA group had severe destruction of articular cartilage and obvious osteophyte formation compared with the NC group; the improvement effects of the NPs and MNs treatment groups were not obvious, while the GLP-1RAs@NPs, GLP-1RAs@NPs-WYRGRL and MNs/GLP-1RAs@NPs-WYRGRL groups could effectively alleviate cartilage destruction and osteophyte formation, especially the MNs/GLP-1RAs@NPs-WYRGRL group (Fig. 8E). The results of HE and SO/FG staining also indicated that the MNs/GLP-1RAs@NPs-WYRGRL group had the best inhibitory effect on joint destruction and could effectively restore the structure of articular cartilage (Fig. 8F-G).

Fig. 8.

MNs/GLP-1RAs@NPs-WYRGRL effectively inhibit RA. (A) Schematic diagram of the treatment process for RA with MNs/GLP-1RAs@NPs-WYRGRL. (B) Photographs of the MNs/GLP-1RAs@NPs-WYRGRL puncturing rat skin. (C) Statistical analysis of joint clinical scores, hind paw thickness, and body weight in the NC group, RA group, RA + NPs group, RA + GLP-1RAs@NPs group, RA + GLP-1RAs@NPs-WYRGRL group, RA + MNs group, and RA + MNs/GLP-1RAs@NPs-WYRGRL group from weeks 1 to 16, with 4–6 rats per group. (D) Representative images of hind paw morphology from the NC group, RA group, RA + NPs group, RA + GLP-1RAs@NPs group, RA + GLP-1RAs@NPs-WYRGRL group, RA + MNs group, and RA + MNs/GLP-1RAs@NPs-WYRGRL group. (E) Representative images of Micro-CT imaging evaluations for each experimental group. (F-G) HE staining and Safranin O/Fast Green staining of joint tissues from different treatment groups. (*p < 0.05 and **p < 0.01)

During the treatment of RA, NPs, GLP-1RAs@NPs or GLP-1RAs@NPs-WYRGRL were injected subcutaneously into the skin of the hind paws of rats, or MN and MN/GLP-1RAs@NPs-WYRGRL were used to puncture the skin and left in place. However, it is not clear whether long-term treatment would cause toxic effects in rats. Therefore, in the 16th week, HE staining analysis was conducted on the kidney, lung, spleen, liver, and heart tissues of each group, and liver and kidney function-related indicators were detected. The results showed that there were no significant differences in the tissue structures in the RA group and each treatment group compared with the NC group (Fig. S7A). The blood routine test showed that, at the 16th week, monocyte count (Mon#) or granulocyte count (Gran#) increased significantly in RA group and RA + MNs group with statistically significant differences. White blood cell (WBC) count, lymphocyte count (Lymph#), and red blood cell distribution width (RDW) increased in the RA group, whereas red blood cell (RBC) count, hemoglobin (HGB) concentration, and platelet (PLT) count decreased. However, no statistically significant differences were observed among the groups (Fig. S7B). The liver and kidney function indicators (ALT, AST, UREA and CREA) were slightly elevated in the RA group, RA + NPs group and RA + MN group, but still did not reach a statistically significant level (Fig. S7C). In conclusion, MNs/GLP-1RAs@NPs-WYRGRL did not show obvious toxicity in long-term treatment, indicating its good biological safety and suggesting it as a safe and effective treatment strategy for RA.

MNs/GLP-1RAs@NPs-WYRGRL inhibited chondrocyte ferroptosis and facilitates the repair of articular cartilage effectively

Previous research results of this study had shown that MNs/GLP-1RAs@NPs-WYRGRL could effectively improve articular cartilage injury, but its specific mechanism remained unclear. To clarify the mechanism, immunohistochemical and immunofluorescence staining analyses were conducted on the articular cartilage tissues of each group. The immunohistochemical results showed that the expression of GLP-1R was down-regulated and the expression of STING, ACSL4 and TNF-α was up-regulated in the RA group compared to the NC group. The NPs and MNs groups did not significantly improve these indicators, while the GLP-1RAs@NPs, GLP-1RAs@NPs-WYRGRL and MNs/GLP-1RAs@NPs-WYRGRL groups could significantly up-regulate the expression of GLP-1R and down-regulate the expression of STING, ACSL4 and TNF-α. Among them, the MNs/GLP-1RAs@NPs-WYRGRL group had the most significant improvement (Fig. 9A). Moreover, the percentage of positive staining area showed statistically significant differences among different groups (Fig. 9B). The immunofluorescence staining results further confirmed that compared with the RA group, the MNs/GLP-1RAs@NPs-WYRGRL group could significantly reduce the expression of TNF-α, up-regulate the expression of GPX4, and promote the expression of extracellular matrix components ACAN and COL2 in chondrocytes (Fig. 9C). The MFI also showed statistically significant differences among different groups (Fig. 9D). In conclusion, MNs/GLP-1RAs@NPs-WYRGRL could increase the expression of GLP-1R, inhibit the cGAS-STING signaling pathway, suppress ferroptosis in chondrocytes of articular cartilage, and promote the expression of chondrocyte-related mechanisms such as ACAN and COL2.

Fig. 9.

Analyze and evaluate the efficacy of MNs/GLP-1RAs@NPs-WYRGRL on articular cartilage in rats with RA. (A-B) Immunohistochemical staining to detect the expression of GLP-1RAs, IL-6, ACSL4, and TNF-α in articular cartilage tissues of rats in NC group, RA group, RA+NPs group, RA+GLP-1RAs@NPs group, RA+GLP-1RAs@NPs-WYRGRL group, RA+MNs group, RA+MNs/GLP-1RAs@NPs-WYRGRL group and quantitative analysis of the positive staining area percentage. (C-D) Immunofluorescence staining of ACAN/GPX4 and TNF-α/COL2 in articular cartilage of each experimental group and quantitative analysis of the MFI. (ns: not significant, *p<0.05, **p<0.01, ***p<0.001, and ****p<0.0001).

MNs/GLP-1RAs@NPs-WYRGRL suppressed the progression of RA through promoting the polarization of synovial macrophages to the M2 type

To clarify whether MNs/GLP-1RAs@NPs-WYRGRL could alleviate the inflammatory response of RA by regulating the polarization state of synovial macrophages, this study detected the types of synovial macrophages and the cytokines they secrete in each group. The results of immunofluorescence staining of the synovial tissues indicated that, compared with the NC group, the proportion of CD68 + CD86+ M1 macrophages significantly increased in the RA group, while the proportion of CD68 + CD206+ M2 macrophages did not change significantly. Compared with the RA group, the GLP-1RAs@NPs, GLP-1RAs@NPs-WYRGRL and MNs/GLP-1RAs@NPs-WYRGRL groups were able to significantly reduce the population of M1 macrophages and increase the population of M2 macrophages, among these groups the improvement was most significant in the MNs/GLP-1RAs@NPs-WYRGRL group (Fig. 10A-D). Flow cytometry combined with statistical analysis indicated that compared with the RA group, the CD11b+CD68 + CD86+ M1 macrophages in the MNs/GLP-1RAs@NPs-WYRGRL group were significantly decreased, while the CD11b+CD68 + CD206+ M2 macrophages were significantly increased (Fig. 10E-F), and the M2/M1 ratio was the highest (Fig. 10G). ELISA assay results indicated that the levels of IL-10 and TGF-β in the GLP-1RAs@NPs, GLP-1RAs@NPs-WYRGRL and MNs/GLP-1RAs@NPs-WYRGRL groups were significantly increased, especially in the MNs/GLP-1RAs@NPs-WYRGRL group compared with the RA group. Meanwhile, the levels of IL-1β, IL-6, and TNF-α were significantly decreased, with the most significant down-regulation observed in the MN/GLP-1RAs@NPs-WYRGRL group (Fig. 10H). In conclusion, MNs/GLP-1RAs@NPs-WYRGRL could effectively inhibit RA inflammatory responses by promoting the polarization of synovial macrophages towards the M2 phenotype.

Fig. 10.

Analysis and evaluation of the role of MNs/GLP-1RAs@NPs-WYRGRL in regulating the polarization of synovial macrophages. (A-B) Detect and analyze the positive rate of CD86+ cells in CD68+ cells in the synovial tissues of the joints via immunofluorescence staining. (C-D) Detect and analyze the positive rate of CD206+ cells in CD68+ cells in the synovial tissues of the joints via immunofluorescence staining. (E) Detect the changes in the proportions of CD68+CD86+ M1 macrophages and CD68+CD206+ M2 macrophages in the synovial tissues of the joints of rats in the NC group, RA group, RA + NPs group, RA + GLP-1RAs@NPs group, RA + GLP-1RAs@NPs-WYRGRL group, RA + MNs group and RA + MNs/GLP-1RAs@NPs-WYRGRL group via flow cytometry. (F) Detect and analyze the positive rate of CD68+CD86+ cells or CD68+CD206+ cells in CD11b+ cells in the synovial tissues of the joints via flow cytometry. (G) Statistical analysis of the M2/M1 macrophage ratio in the above-mentioned groups. (H) Detect the expression levels of cytokines IL-10, TGF-β, TNF-α, IL-1β and IL-6 in the synovial tissues in the NC group, RA group, RA + NPs group, RA + GLP-1RAs@NPs group, RA + GLP-1RAs@NPs-WYRGRL group, RA + MNs group, and RA + MNs/GLP-1RAs@NPs-WYRGRL group via ELISA assay. (ns:not significant, *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001)

Discussion

The management of rheumatoid arthritis (RA) has long been dominated by strategies targeting synovial inflammation, often overlooking the pivotal role of chondrocyte dysfunction in joint destruction [35, 36]. Our study demonstrates that direct cartilage protection is not only achievable but central to achieving therapeutic efficacy. We unveil a dual-pathway mechanism wherein GLP-1RAs concurrently halt a novel cGAS-STING-ferroptosis axis in chondrocytes and reprogram synovial macrophages toward an M2 reparative phenotype. More importantly, combined with drug-encapsulating silk fibroin NPs, cartilage-targeting WYRGRL peptide, and sustained-release soluble MNs, we designed and prepared the cartilage-targeting MNs/GLP-1RAs@NPs-WYRGRL which offers a novel and highly translatable strategy for precise, efficient, and low-side-effect therapy in the management of rheumatoid arthritis.

Recently, significant progress has been made in the field of rheumatoid arthritis (RA) drug treatment, mainly including biological agents and targeted synthetic drugs, such as TNF-α inhibitors [37, 38] and specific intracellular signaling pathways inhibitors [39]. Although numerous large-scale clinical studies have fully confirmed that these new drugs can rapidly and persistently alleviate RA symptoms, their safety has always been controversial, particularly regarding potential risks such as venous thromboembolism and infection [40, 41]. Therefore, exploring new therapeutic targets and developing new therapeutic drugs is particularly important for treating RA. GLP-1RAs used for the treatment of type 2 diabetes was gradually expanded to other diseases, such as obesity [42], non-alcoholic fatty liver disease [43], and Alzheimer’s disease [44]. Recent studies have unexpectedly found that it also holds great potential in the treatment of OA [14]. Nevertheless, the efficacy and mechanism of GLP-1RAs in inhibiting RA remained unclear entirely. Here, we had confirmed that GLP-1RAs could significantly inhibit the activation of the cGAS-STING signaling pathway in RA. The cGAS-STING pathway has been confirmed to be closely related to the pathogenesis of many autoimmune diseases in recent years [45], but research on its functional role in RA is still in the initial exploration stage. It had been reported that the accumulation of cytoplasmic double-stranded DNA (dsDNA) and the expression level of cGAS were positively correlated with the severity of synovitis in RA patients [22]. The activation of the cGAS-STING pathway has been confirmed to mediate the progression of synovitis and joint destruction by promoting the migration and invasion of fibroblast-like synoviocytes (FLS) and the production of inflammatory factors [22, 46]. However, this article also analyzes how the cGAS-STING pathway influences RA from the perspective of articular cartilage cells; it is known that this signaling pathway is closely associated with articular cartilage surface destruction during the occurrence and development of RA, which may be related to its regulation of chondrocyte ferroptosis and promotion of inflammatory factor expression, as supported by RNA sequencing, molecular biology experiments, and both in vivo/in vitro experiments. After adding either the cGAS–STING pathway inhibitor C-176 or GLP-1RAs, the cGAS-STING signaling pathway was inhibited effectively accompanied with suppression of chondrocyte ferroptosis, inflammatory factor expression, and osteophyte formation on articular surfaces, and enhancement of cartilage repair, reinforcing a causal link between cGAS–STING activation and chondrocyte ferroptosis in RA. Notably, the therapeutic effects of GLP-1RAs on ferroptosis induced by the activation of cGAS-STING pathway were similar to that of the cGAS-STING pathway inhibitor C-176. Given that prior studies have robustly demonstrated that the cGAS-STING pathway promoted the ferroptosis [24–26], our findings further corroborate its pivotal involvement in chondrocyte ferroptosis during RA pathogenesis. We therefore propose that GLP-1RAs exert chondroprotective effects in RA, at least in part, by suppressing ferroptosis in chondrocytes mediated by cGAS–STING pathway.

The pathogenesis of RA is complex and not yet clear. Previous studies had shown that inhibiting the cGAS-STING pathway in macrophages could effectively reduce the secretion of inflammatory factors TNF-α, IL-1β and IL-6 in the synovial tissue of RA joints [21, 47]. In our research on maintaining synovial homeostasis, we found that macrophages and related inflammatory factors were enriched in RA tissues through GO enrichment analysis. One of the hallmarks of RA is persistent synovitis, which is associated with the local infiltration of various immune cells and excessive inflammatory mediators [48]. Among these infiltrating cells, macrophages in synovial tissue, as an early cell subpopulation to enter the joint and synovial area, are the main contributors of tumor necrosis factor (TNF), profoundly mediating the chronic development of RA [49]. Inhibition of M1 macrophage polarization or driving M2 macrophage polarization were regarded as an effective therapeutic strategy for RA [50]. In this study, the results of flow cytometry, immunofluorescence staining, RT-qPCR and ELISA assays demonstrated that GLP-1RAs could drive macrophage polarization towards the M2 type, thereby inhibiting the continuous development of synovitis. In conclusion, GLP-1RAs promoted macrophage polarization towards the M2 type, improving the inflammatory microenvironment of the synovium, maintaining synovial homeostasis, and inhibiting the progression of RA.

Although it had been confirmed that GLP-1RAs had excellent therapeutic effects in the treatment of RA, how to enrich it in the pathological microenvironment of arthritis and avoid off-target effects of the drug was the main challenge currently faced. With the development of nanotechnology, drug delivery systems targeting the joint area of RA have shown great therapeutic potential [51]. Currently, the targeting objects of these multifunctional drug delivery nanoplatforms are mainly focused on the inflammatory microenvironment of arthritis (such as infiltrating M1-type macrophages and TNF-α) and FLS, aiming to eliminate the adverse effects of these factors [52–54]. However, our research took a different approach, designing drug-loaded nanoparticles targeting articular cartilage from the perspective of protecting chondrocytes and increasing the drug concentration in the joint. Compared with oral administration or the extremely difficult intra-articular injection into small joints, MNs offered a minimally invasive, painless transdermal drug delivery system. They also have significant advantages such as avoiding the first-pass effect of oral administration and improving patient compliance [55]. Therefore, they have not only shone brightly in the fields of dermatology and wound treatment but have also increasingly attracted attention in the treatment of RA recently [56]. In this study, we firstly confirmed that compared with RBITC-GLP-1RAs@NPs, RBITC-GLP-1RAs@NPs-WYRGRL could effectively target tissues with abundant cartilage, such as joints, and could be retained in cartilage tissues more effectively, and thus be excreted by the kidneys less. Secondly, we integrated GLP-1RAs@NPs-WYRGRL into HAMA MNs to innovatively construct a targeted drug-loaded MN delivery system (MNs/GLP-1RAs@NPs-WYRGRL) with articular cartilage targeting and excellent sustained-release performance via combining the advantages of GLP-1RAs@NPs-WYRGRL and MNs. In the vivo study of RA rats, we known that there were significant differences in body weight between the NC group and the GLP-1RAs treatment group, as well as between the NC group and the GLP-1RAs@NPs group. However, there were no significant differences in body weight between the NC group and the GLP-1RAs@NPs-WYRGRL group or the MNs/GLP-1RAs@NPs-WYRGRL group. This indicates that local administration of non-targeted GLP-1RAs or GLP-1RAs@NPs can affect body weight and systemic metabolism; however, the targeted cartilage delivery strategy can minimize systemic metabolic side effects, such as changes in body weight. Moreove, it was found that MNs/GLP-1RAs@NPs-WYRGRL could release GLP-1RAs@NPs-WYRGRL, target articular cartilage, significantly improve the clinical score of rat joints, reduce the thickness of their foot pads, alleviate cartilage damage and osteophyte formation, promote the secretion of extracellular matrix by chondrocytes and the repair of articular cartilage, and inhibit inflammatory responses, ultimately effectively alleviating the occurrence and development of RA. Furthermore, the assessment of biological safety was a crucial step for biomaterials. We have fully verified that the MNs/GLP-1RAs@NPs-WYRGRL did not show obvious toxicity, indicating its excellent biological safety through in vivo experiments. In conclusion, this new targeted therapy strategy provided a brand-new and highly promising solution for precise, efficient and painless treatment of RA.

In this study, the cGAS-STING-ferroptosis axis was first confirmed in chondrocytes in RA, rather than in macrophages or tumor cells as previously reported [57, 58]. Moreover, existing studies mostly focus on the regulation of RA synovial cells (such as FLS) or macrophages by the cGAS-STING pathway, whereas this study, for the first time, clearly demonstrates the central role of the cGAS-STING-ferroptosis axis in RA chondrocytes, and reveals that GLP-1RAs exerted therapeutic effects on RA through a dual mechanism of directly inhibiting this axis and simultaneously promoting the M2 polarization of macrophages. By combining cartilage-targeting nanocarriers with microneedle drug delivery systems, this mechanism has been transformed into a clinically applicable targeted treatment strategy, compensating for the shortcomings of existing studies, which emphasize mechanistic insights and synovial tissue while neglecting translational applicability and cartilage-specific targeting. Although this study had achieved some positive results, the following limitations still exist: The direct molecular targets of GLP-1RAs in regulating the cGAS-STING pathway had not been fully clarified and further verification was needed using molecular biology techniques; FLS, osteoclasts and monocytes in the RA joint cavity also played an indispensable key role in the progression of RA. In the future, single-cell sequencing technology could be used to further analyze the effects of GLP-1RAs on the cGAS-STING pathway and ferroptosis in different cell subpopulations, such as chondrocytes, macrophages, FLS, and neutrophils, in RA joints. The safety and pharmacokinetic characteristics of long-term local application of GLP-1RAs in human joints still need to be further evaluated, which is crucial for its transformation from the laboratory to the clinic.

In this study, it elucidated that GLP-1RAs suppressed the progression of RA through a dual mechanism of inhibiting chondrocytes ferroptosis caused by the cGAS-STING pathway and promoting macrophage polarization towards the M2 type. Moreover, our research findings not only expanded the application prospects of GLP-1RAs in RA diseases, provide a brand-new and pioneering perspective for the precise targeted treatment of RA, but also represent a promising new direction for the future treatment of RA.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

RA

Rheumatoid Arthritis

GLP-1RAs

Glucagon-like peptide-1 receptor agonists

GLP-1R

Glucagon-like peptide-1 receptor

MNs

Microneedles

FLS

Fibroblast-like synoviocytes

HAMA

Hyaluronic Acid Methacrylate

NPs

Nanoparticles

mtDNA

mitochondrial DNA

TNF-α

Tumor necrosis factor-α

IL-1β

Interleukin-1β

IL-6

Interleukin-6

KEGG

Kyoto encyclopedia of genes and genomes

GO

Gene ontology

GSEA

Gene set enrichment analysis

cGAS

Cyclic GMP-AMP synthase

STING

Stimulator of interferon genes

Author contributions

Conceptualization: HZ, LL and MS. Investigation: HZ, LL, SZ, YL and ZW. Validation: XZ, JZ, YZ and CL. Writing-original draft: HZ, ZT and YL. Writing-review & editing: LL and JL. Methodology: HZ, MS, ZW, YL, SL and LL. Software: HZ, YZ and LL. Visualization: All of the authors have read and provided comments on the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China grant 82572992 (T.ZQ), the National Natural Science Foundation of China grant 82472517 (L.CQ), the National Natural Science Foundation of China grant 82203921 (L.SJ), the Guizhou Province Science and Technology Plan Project (Qiankehejichu-ZK[2024]Yiban 132 [Zuli Wang]) (W.ZL), Guizhou Province’s Scientist Workstation Project (Qiankehepingtai KXJZ [2024]004),the National Natural Science Foundation of China (82471385), Guizhou Province Scientific and Technological Innovation Talent Team Project (Qiankeherencai CXTD [2025]048) (S.M), and the Natural Science Foundation of Chongqing CSTB2024NSCQ-MSX0622 (Z.J ).

Data availability

All data are available from the corresponding authors upon reasonable request.

Declarations

Ethics approval and consent to participate

The animal use and experimental procedures met the requirements of the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Animal Ethics Committee of the Army Medical University (no. AMUWEC20250028).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.