A Novel Cargo Delivery System‐AnCar‐ExoLaIMTS Ameliorates Arthritis via Specifically Targeting Pro‐Inflammatory Macrophages

Abstract Macrophages are heterogenic phagocytic cells that play distinct roles in physiological and pathological processes. Targeting different types of macrophages has shown potent therapeutic effects in many diseases. Although many approaches are developed to target anti‐inflammatory macrophages, there are few researches on targeting pro‐inflammatory macrophages, which is partially attributed to their non‐s pecificity phagocytosis of extracellular substances. In this study, a novel recombinant protein is constructed that can be anchored on an exosome membrane with the purpose of targeting pro‐inflammatory macrophages via antigen recognition, which is named AnCar‐ExoLaIMTS. The data indicate that the phagocytosis efficiencies of pro‐inflammatory macrophages for different AnCar‐ExoLaIMTS show obvious differences. The AnCar‐ExoLaIMTS3 has the best targeting ability for pro‐inflammatory macrophages in vitro and in vivo. Mechanically, AnCar‐ExoLaIMTS3 can specifically recognize the leucine‐rich repeat domain of the TLR4 receptor, and then enter into pro‐inflammatory macrophages via the TLR4‐mediated receptor endocytosis pathway. Moreover, AnCar‐ExoLaIMTS3 can efficiently deliver therapeutic cargo to pro‐inflammatory macrophages and inhibit the synovial inflammatory response via downregulation of HIF‐1α level, thus ameliorating the severity of arthritis in vivo. Collectively, the work established a novel gene/drug delivery system that can specifically target pro‐inflammatory macrophages, which may be beneficial for the treatments of arthritis and other inflammatory diseases.


Introduction
Macrophages play a crucial role in innate immune responses, contributing to many physiological and pathological processes including tissue injuries and repair. [1]Following injury, pro-inflammatory macrophages (classically activated macrophages) begin to infiltrate in injured tissue, contributing to clearing the damaged cells and bacteria.Subsequently, the damaged tissue will go through a repairing process accompanied by anti-inflammatory macrophages (alternatively activated macrophages).The imbalance between pro-inflammatory and anti-inflammatory macrophages will lead to dysregulated tissue repair or even chronic inflammation. [2]Targeting different types of macrophages has shown potent therapeutic effects in multiple diseases.Although multiple approaches have been developed to target anti-inflammatory macrophages, there are few researches on targeting pro-inflammatory macrophages partially because of their non-specificity phagocytosis.
Synovial macrophages are the major immune cells in joint microenvironment that actively participate in the repairing of articular injuries.Previous studies have demonstrated that there are a large amount of pro-inflammatory macrophages in inflammatory arthritis, which play important roles in the pathological damage of joints by secreting a variety of cytokines. [3]Liu et al.
found that the ratio of pro-inflammatory macrophages to anti-inflammatory macrophages in synovial fluid was higher in inflammatory arthritis patients compared to the health group. [4]Meanwhile, the inflammatory cytokines secreted by pro-inflammatory macrophages, IL-1 and TNF-, could promote osteoclastogenesis, stimulating the secretion of MMPs and cytokines. [5]In contrast, inhibiting the hyperactivation of pro-inflammatory macrophages or raising the ratio of antiinflammatory to pro-inflammatory macrophages could remarkably relieve the symptoms of inflammatory arthritis and ameliorate progress of disease. [6]However, the current treatment targeting synovial pro-inflammatory macrophages lacks specificity and has potential off-target impacts on other types of articular cells including chondrocytes, synovial fibroblasts, anti-inflammatory macrophages and endotheliocytes.Therefore, it is necessary to develop novel drug delivery systems to specifically target synovial pro-inflammatory macrophages, aiming to improve the therapeutic effect of anti-inflammatory molecules against arthritis.
Exosomes are membrane-bound vehicles with sizes ranging from 30 to 150 nm, which were released from cells through membrane fusion between multivesicular bodies and the plasma membrane. [7]As a carrier of DNA, RNA and protein, exosome functions as mediator for intercellular communication and substance exchange. [8]Due to its excellent immune compatibility and high organotropism, exosome is regarded as a novel potent drug delivery platform for the treatment of many diseases.In recent years, there have been increasing studies focusing on the targeting ability of exosomes for specific types of cells.Kanuma et al. and Wood et al. have modified N segment of exosome membrane protein CD163 or LAMP2 by expression-specific antigen sequences, which mediates the uptake of engineering exosomes by targeted cells. [9]For macrophages, Gowri et al. reported that exosomes with expression of IL-4 receptor on their membrane inhibit tumor growth by targeting tumor-associated macrophages. [10]Furthermore, the exosomes, which were engineered to express molecules on exosome membrane targeting scavenger receptor class A, can ameliorate joint inflammation. [11]lthough these engineered exosomes have exhibited certain therapeutic effects, they also have several limitations including complex preparation process, presence of potentially harmful chemical species and lack of targeting specificity for subset populations of macrophages, especially pro-inflammatory macrophages. [12]herefore, novel delivery methods targeting pro-inflammatory macrophages based on exosomes are needed to enhance the therapeutic effects for arthritis.
Here, we construct a novel recombinant protein that could be anchored on exosome membranes with the purpose of targeting pro-inflammatory macrophages.Just like the Chimeric Antigen Receptor (CAR), this recombination protein consists of four major components including 1) N-terminal antigen domain that recognizes Toll-like Receptor 4 (TLR4), 2) flexible linker region, 3) transmembrane region of LAMP2B, 4) C-terminal GFP.Moreover, this recombination protein would help exosomes to specifically recognize and regulate the targeted cells (pro-inflammatory macrophages), which is similar to the function of CAR in engineered cells.Different from the CAR, this recombinant protein is mainly located on the bilayer membrane of exosome and does not have the function of signal transduction.Based on the above, we named this recombinant protein as AnCar-LaIMTS (analogous chimeric antigen pseudo-receptor with LAMP2B sequence that targets pro-inflammatory macrophages).In this study, AnCar-LaIMTS is used to modify exosomes deriving from seeding cells with purpose of delivering cargo specifically to proinflammatory macrophages, which is named AnCar-Exo LaIMTS .Our data indicate that the phagocytosis efficiencies of proinflammatory macrophages for different AnCar-Exo LaIMTS show obvious differences.In the existing candidate range, the AnCar-Exo LaIMTS3 has the best targeting ability to pro-inflammatory macrophages in vitro and in vivo.Moreover, our bioinformatics analysis revealed the existence of the AnCar-LaIMTS3 amino acid sequence within the Leucine-rich repeat flightlessinteracting protein 1, which is recognized for its interaction with the Leucine-rich repeat domain located in the extracellular portion of TLR4.Through the application of TLR4 neutralizing antibodies or by employing TLR4 −/− mice, we found that this high-efficiency targeting of pro-inflammatory macrophages is achieved through TLR4-mediated endocytosis.In addition, AnCar-Exo LaIMTS3 loading with HIF-1-siRNA (AnCar-Exo LaIMTS3 -siRNA-HIF-1) could efficiently deliver cargo (HIF-1-siRNA) into pro-inflammatory macrophages.Through intra-articular administration of AnCar-Exo LaIMTS -siRNA-HIF-1, the expression of HIF-1 in pro-inflammatory macrophages of synovium is significantly downregulated, while the level of HIF-1 is rarely influenced in other types of cells.On the contrary, the exosomes carrying HIF-1-siRNA without AnCar-LaIMTS randomly enter into multiple types of cells with limited interference effects on HIF-1 in pro-inflammatory macrophages of synovium (Figure 1).In brief, we first found that pro-inflammatory macrophages have selective endocytosis on engineered exosomes modified by different AnCar-LaIMTS.AnCar-Exo LaIMTS3 can improve the targeting ability of the pro-inflammatory macrophages and promote the efficient enrichment of therapeutic cargo in target cells through TLR4.We believe that this novel delivery system of AnCar-Exo LaIMTS will be beneficial for the treatment of arthritis or other inflammatory diseases.

Design and Construction of AnCar-LaIMTS Plasmids
The major components of AnCar-LaIMTS were composed of four elements (Figure 2A), including element 1 (exosome membrane anchoring sequence), element 2 (pro-inflammatory macrophage targeting sequence), element 3 (exosome tracing sequence), element 4 (flexible linker region sequence).For element 1, the sequence coding LAMP2B was used to help AnCar-LaIMTS anchor on exosome membrane. [13]For element 2, sequences derived from a phage display library, which were proposed to recognize pro-inflammatory macrophages in previous study (29-32), were fused with the N terminus of LAMP2B with flexible linker (element 4).Moreover, GFP coding sequence in element 3 was fused with exosomes' intracellular segment (C terminus of LAMP2B) so as to track exosomes through GFP fluorescence.To screen and identify the optimal AnCar, ten targeting sequences were used to code element 2, which came out with 10 different AnCar-LaIMTSs named as AnCar-LaIMTS1 to AnCar-LaIMTS10, while the random sequence in element 2 was named as AnCar-LaRANDOM.Subsequently, all AnCar-LaIMTSs' coding sequences were inserted into PLVXpuro plasmids for lentivirus production (Figure 2B).To test the accuracy of plasmid construction, the eleven plasmids were digested at XhoI and BamHI sites.The electrophoresis results showed that all plasmids produced a fragment of ≈1200 bp, which was consistent with the design size (Figure 2C).The sequences of element 2 of all AnCar-LaIMTSs were further identified by PCR product sequencing as shown in Figure 2D.These data revealed that the constructed plasmids encoded the exact targeting peptides as designed.Overall, we have successfully designed and constructed 11 plasmids that could efficiently encode 11 different AnCar-LaIMTS recombination proteins (AnCar-LaIMTS1∼AnCar-LaIMTS10 and AnCar-LaIMTSRANDOM).
Subsequently, these recombinant plasmids were individually transfected into HEK 293T cells through lentivirus to acquire respective cell lines (Figure 2E).As shown in Figure 2F, the GFP green signals of eleven stable cell lines can be observed under fluorescent microscopy.Moreover, we detected the expression of fusion protein by Western Blot.Data showed that both LAMP2B antibody and GFP antibody could detect the expected fusing protein (130 kDa protein band), while HEK 293T cells and HEK 293T cells transfected with GFP plasmid didn't show fusing protein band, demonstrating that AnCar-LaIMTS proteins could be successfully synthesized in cells (Figure 2G).Collectively, we constructed 11 stable cell lines that could express the corresponding recombinant proteins (AnCar-LaIMTS1 to AnCar-LaIMTS10, AnCar-LaRANDOM).

Characteristics of AnCar-LaIMTS Modified Exosomes
To obtain the corresponding AnCar-modified exosomes from the above 11 stable cell lines (Figure 3A), Ultracentrifugation was used to isolate exosomes from supernatant as reported previously. [7]In brief, the separated supernatants were centrifuged at 300, 2000, and 10 000 g for 10 min.Later, they were centrifuged at 1 00 000 g for 90 min after being filtered through a 0.22 μm filter.Finally, the exosomes were collected and named as AnCar-Exo LaCTRL , AnCar-Exo LaRANDOM , and AnCar-Exo LaIMTS1 to AnCar-Exo LaIMTS10 (Figure 3B).To confirm the success of isolating exosomes from supernatant, aforementioned exosomes were subjected to NanoSight detection.As shown in Figure 3C-F, eleven kinds of AnCar-modified exosomes as well as AnCar-Exo LaCTRL presented classic characteristics of exosomes, such as size distribution and exosome structure.The number of exosomes with the diameters between 30 and 200 nm ranged from 1.23 × 10 11 to 1.92 × 10 11 particles/ml, and the main particle sizes of exosomes were within 99.8 nm-158 nm with average particle size between 123.6 nm -182.1 nm (Figure 3D), suggesting that there were no significant differences in diameters between the AnCar-modified exosomes and native exosomes.Besides, we observed exosome-like structures under transmission electron microscope.It could be seen from Figure 3D that the separated exosome-like vesicles from supernatant displayed typical saucerbilayer membrane, which is a characteristic of exosomes.Moreover, the fusion protein AnCar and surface markers of exosomes including CD63, CD81, and TSG101 could be detected in exosomes (Figure 3E).As shown in Figure 3F, GFP was successfully loaded into AnCar-Exo LaRANDOM , AnCar-Exo LaIMTS1 and AnCar-Exo LaIMTS10 as detected by ImageStreamX compared with AnCar-Exo LaCTRL .Taken together, AnCar-modified exosomes have been successfully separated from 11 engineered stable cell lines by ultracentrifugation with classic characteristics of exosomes including size, structure and contents.

Screening and Identification of the Optimal AnCar-Exo LaIMTS Specifically Targeting Pro-Inflammatory Macrophages In Vitro
In order to screen and identify the optimal AnCar-Exo LaIMTS that targets pro-inflammatory macrophages with high efficiency, confocal microscopy was used to observe the uptake of exosomes by pro-inflammatory macrophages.RAW 264.7 cells were stimulated with 100 ng ml −1 LPS for 24 h to induce proinflammatory transition.AnCar-Exo LaIMTS1 to AnCar-Exo LaIMTS10 , AnCar-Exo LaRANDOM were then added into culture medium and co-cultured with pro-inflammatory macrophages for 12 h (Figure 4A).As shown in Figure 4B, AnCar-Exo LaIMTS1 to AnCar-Exo LaIMTS10 as well as AnCar-Exo LaRANDOM could enter into pro-inflammatory macrophages with variable targeting abilities.Interestingly, AnCar-Exo LaIMTS3 had the best affinity toward pro-inflammatory macrophages, whose fluorescence intensity was higher than that of other AnCar-modified exosomes.Furthermore, 12 h after the incubation of those   AnCar-Exosomes (AnCar-Exo LaIMTS1 to AnCar-Exo LaIMTS10 , AnCar-Exo LaRANDOM ) with pro-inflammatory bone marrowderived macrophages (BMDM) induced by LPS and INF- (Figure 4C,E), AnCar-Exo LaIMTS3 showed the highest uptake by pro-inflammatory BMDM (Figure 4D,F), whose fluorescence intensity was also higher than that of other AnCar-modified exosomes.Above all, AnCar-Exo LaIMTS3 was an optimal AnCar-Exo LaIMTS that can specifically target proinflammatory macrophages in vitro compared with other AnCar-Exo LaIMTS .
We further evaluated the targeting ability of AnCar-Exo LaIMTS3 toward other subtypes of macrophages.BMDM was induced in different subtypes of macrophages.Each subtype was incubated with AnCar-Exo LaIMTS3 for 12 h and then examined by  4G).AnCar-Exo LaIMTS3 could enter into other subtypes of macrophages, but the fluorescence intensity in M and anti-inflammatory macrophages was mild and showed significant decreases (Figure 4H-M).In addition, the data from flow cytometry revealed that the percentage of CD86 + GFP + cells (representing pro-inflammatory macrophages) ingesting AnCar-Exo LaIMTS3 was 27.1% in LPS and INF- stimulated BMDM, while it was 1.08% for CD206 + GFP + cells (represent anti-inflammatory macrophages) in IL-4 stimulated BMDM (Figure 4N-Q).These data demonstrated that AnCar-Exo LaIMTS3 can be largely enriched in pro-inflammatory macrophages compared with M or anti-inflammatory macrophages.In addition, the primary chondrocytes and synovial fibroblasts were also tested for the targeting ability of AnCar-Exo LaIMTS3 .Data of Figure S1, (Supporting Information) indicated that only very few chondrocytes and synovial fibroblasts showed green fluorescence signals after incubated with AnCar-Exo LaIMTS3 for 12 h.In brief, AnCar-Exo LaIMTS3 is the most optimal AnCar-Exo LaIMTS targeting pro-inflammatory macrophages.

Evaluating the Distribution and Potential Toxicity of AnCar-Exo LaIMTS3 In Vivo
Next, we evaluated the distribution and potential toxicity of AnCar-Exo LaIMTS3 in vivo, which is important for its safe application in future.AnCar-Exo LaCTRL and AnCar-Exo LaIMTS3 were labelled by DID dye and delivered into collagenase-inducedarthritis mice through systemic administration.Then the distribution of exosomes was detected by Vilber Lourmat at different time points (0, 1, 3, 6, 12, 24, and 48 h) (Figure 5A).As shown in Figure 5B,C, DID fluorescence could be detected in mice injected with AnCar-Exo LaCTRL and AnCar-Exo LaIMTS3 , but not in those with PBS administration.Moreover, the DID fluorescence intensity in mice with AnCar-Exo LaCTRL or AnCar-Exo LaIMTS3 gradually increased after administration and achieved the peak intensity at 3 h.It was subsequently decreased and almost disappeared at 48 h, suggesting that AnCar-Exo LaCTRL and AnCar-Exo LaIMTS3 had similar metabolic rates in vivo.Furthermore, we explored the distribution of exosomes in different organs of mice at 3 h after systemic administration.As shown in Figure 5D-G, the fluorescence of AnCar-Exo LaCTRL and AnCar-Exo LaIMTS3 was mainly detected in liver, spleen and kidney.Interestingly, the fluorescence intensities of AnCar-Exo LaIMTS3 in spleen were stronger than that of AnCar-Exo LaCTRL , while it didn't show significant differences between those two groups in liver and kidney.In addition, hematoxylin-eosin staining revealed that AnCar-Exo LaIMTS3 didn't result in pathological injuries such as cell death and fibrosis in heart, liver, spleen, lung and kidney, suggesting that AnCar-Exo LaIMTS3 didn't induce remarkable toxicity in these organs (Figure 5H; Figure S2, Supporting Information).In brief, AnCar-Exo LaIMTS3 is mainly distributed in liver, spleen and kidney with no significant toxicity in mice.

AnCar-Exo LaIMTS3 Could Specifically Target Pro-Inflammatory Macrophages in Arthritic Mice
There was a large amount of activated macrophages in synovial tissues of inflammatory arthritis, which greatly contributed to joint injury.Our previous studies indicate that targeting these synovial macrophages is a potent strategy for the treatment of arthritis. [14]Then, we further evaluated whether AnCar-Exo LaIMTS3 could specifically target synovial macrophages in inflammatory arthritis.First, we observed the distribution of intra-articularly injected AnCar-Exo LaIMTS3 by Vilber Lourmat at different time points following injection (0, 1, 3, 6, 12, 24, 48, and 72 h) (Figure 6A).As shown in Figure 6B, AnCar-Exo LaIMTS3 , but not the AnCar-Exo LaCTRL , was significantly enriched in inflammatory joints.Statistically, the fluorescence intensity in AnCar-Exo LaIMTS3 treatment group was notably higher than that of AnCar-Exo LaCTRL (Figure 6C).In addition, the intensity of AnCar-Exo LaIMTS3 reached the peak value at 1 h and almost disappeared at 72 h, while the intensity in AnCar-Exo LaCTRL nearly disappeared at 48 h.These results indicated that AnCar-Exo LaIMTS3 could be enriched in inflammatory joints with higher intensity and longer retention time than that of AnCar-Exo LaCTRL .
Next, we detected the targeting ability of AnCar-Exo LaIMTS3 for synovial macrophages in arthritic mice by in vivo imaging.Briefly, tdTomato-LysM Cre mice were used to label synovial macrophages, [15] which were further polarized into proinflammatory macrophages after collagen-induced arthritis, then the targeting ability of systematically administrated AnCar-Exo LaCTRL and AnCar-Exo LaIMTS3 toward pro-inflammatory synovial macrophages were identified by two-photon microscopy in vivo (Figure 6D).As shown in Figure 6E and Movie S1 (Supporting Information), some of the green fluorescence (AnCar-Exo LaCTRL ) gradually entered the synovial tissue of arthritic mice, hardly seen any exosomes co-localized with synovial macrophages by the time 2 h after exosome administration.On the contrary, when arthritic mice were treated with AnCar-Exo LaIMTS3 for 1 h, there were significantly more exosomes, residing on synovial tissue, which was largely co-localized with macrophages (Figure 6F and Movie S1, Supporting Information).All the above data indicated that AnCar-Exo LaIMTS3 could cause pro-inflammatory macrophages in arthritic joints.
At last, we observed the co-localization of AnCar-Exo LaIMTS3 with pro-inflammatory macrophages after their systemic administration in joint.As shown in Figure S3A,B (Supporting Information), after mice were intra-articularly treated with AnCar-Exo LaIMTS3 and AnCar-Exo LaCTRL ,the DID fluorescence of AnCar-Exo LaIMTS3 group, but not AnCar-Exo LaCTRL group, was significantly increased in synovium and largely colocalized with F4/80 + iNOS + positive cells.The ratio of DID positive cells in F4/80+iNOS+ cells was significantly higher than that in F4/80 + iNOS − cells (none pro-inflammatory macrophages) in AnCar-Exo LaIMTS3 group, while it presented no difference in AnCar-Exo LaCTRL group (Figure S4A-D, Supporting Information).These results suggest that pro-inflammatory macrophages are the major subtype of synovial macrophages in collagenaseinduced-arthritis.While AnCar-Exo LaIMTS3 could be largely enriched in these pro-inflammatory macrophages.Collectively, the synovial pro-inflammatory macrophages were the main targeted cells of AnCar-Exo LaIMTS3 in the collagenase-induced-arthritis.

AnCar-Exo LaIMTS3 Could Specifically Target Pro-Inflammatory Macrophage in Caudal fin Injury Model of Zebrafish
Next, we test the targeting ability of AnCar-Exo LaIMTS3 toward pro-inflammatory macrophages that were recruited in the injury site of zebrafish caudal fin. [16]First, Mpeg1:GAL4 and UAS: NTR-mCherry lineages were crossed to obtain the Mpeg1:GAL4/UAS:NTR-mCherry transgenic zebrafishes(Figure 7A), which often were used to trace Screening and identification of the optimal AnCar-Exo LaIMTS that targets pro-inflammatory macrophage in vitro.A) A schematic diagram illustrating that pro-inflammatory macrophages derived from LPS-activated RAW 264.7 cells were incubated with AnCar-Exo LaIMTS1 to AnCar-Exo LaIMTS10 , AnCar-Exo LaRANDOM (1 × 10 8 particles) for 12 h.B) The uptake of exosomes by pro-inflammatory macrophages derived from RAW 264.7.Immunofluorescence showed F4/80 (gray), iNOS (red) and GFP (green) in exosome-treated RAW 264.7 cells.Scale bars, 10 μm.C) A schematic diagram illustrating that how BMDM was separated from bone barrow and pro-inflammatory macrophages derived from LPS/INF--activated BMDM were incubated with AnCar-Exo LaIMTS1 to AnCar-Exo LaIMTS10 , AnCar-Exo LaRANDOM (1 × 10 8 particles) for 12 h.Scale bars, 10 μm.D) The uptake of exosomes by pro-inflammatory macrophages induced from BMDM.Immunofluorescence showed F4/80 (sky blue), iNOS (red) and GFP (green) in exosome-treated BMDM cells.Scale bars, 10 μm.E) Quantification of GFP fluorescence intensities in exosome-treated RAW 264.7 cells.GFP fluorescence intensities were calculated under three high-magnification fields with image J. N = 3 independent biological replicates, *P < 0.05, **P < 0.01 and ***P < 0.001.F) Quantification of GFP fluorescence intensities in exosome-treated BMDM cells.GFP fluorescence intensities were calculated under three high-magnification fields with image J. N = 3 independent biological replicates, *P < 0.05, **P < 0.01 and ***P < 0.001.G) A schematic diagram illustrating that how M, pro-inflammatory macrophage and anti-inflammatory macrophage were induced from BMDM and detected by Immunofluorescence or Flow Cytometry after treated with AnCar-Exo LaIMTS3 .H, I and J) Immunofluorescence showed the uptake of AnCar-Exo LaIMTS3 by M, pro-inflammatory macrophage and anti-inflammatory macrophage.Scale bars, 10 μm.K) Quantification of GFP fluorescence intensities in AnCar-Exo LaIMTS3 treated M, pro-inflammatory macrophage and anti-inflammatory macrophage.GFP fluorescence intensities were calculated under three high-magnification fields with image J. N = 3 independent biological replicates, *P < 0.05, **P < 0.01 and ***P < 0.001.L, M) Flow Cytometry showed the ratio of relative fluorescence intensity between pro-inflammatory macrophage and M or between anti-inflammatory macrophage and M.N) Flow Cytometry shows the ratio of CD86 + GFP + in LPS/INF--activated or IL-4-activated BMDM.O) Flow Cytometry shows the ratio of CD206 + GFP + in LPS/INF--activated or IL-4-activated BMDM.All values were displayed as the way of mean ± SD.One-way analysis of variance (ANOVA) was used to evaluate the difference among groups.
16c] Through fluorescence tracking in vivo, we observed macrophages with red fluorescent signals in those transgenic zebrafishes (Figure 7B).Subsequently, the recruitment of pro-inflammatory macrophages was detected by two-photon microscopy at early stage after injury of caudal fin in transgenic zebrafish (Figure 7C).As shown in Figure 7D, the red fluorescence labeled macrophages were gradually recruited into the injured site during 7 h after injury of caudal fin, which indicated that a lot of pro-inflammatory macrophages were accumulated at early stage after injury.Then, we observed the green fluorescence of AnCar-Exo LaIMTS3 in the above pro-inflammatory macrophages through intracardiac injection (Figure 7E).As shown in Figure S5 (Supporting Information), AnCar-Exo LaIMTS3 could enter into blood circulation system and reached the injured site of caudal fin in zebrafish.Moreover, the representative images in Figure 7F showed that the green fluorescence of AnCar-Exo LaIMTS3 had more co-localization with red fluorescence compared to AnCar-Exo CTRL .The statistical analysis also showed similar results (Figure 7G), suggesting that AnCar-Exo LaIMTS3 showed higher targeting ability toward macrophages.To sum up, our data demonstrate that the AnCar-Exo LaIMTS3 could specifically target pro-inflammatory macrophages in caudal fin injury model of zebrafish.

Identify the Cellular Uptake Mechanisms of AnCar-Exo LaIMTS3 by Pro-Inflammatory Macrophage
Numerous pathways participate in cellular uptake of exosomes by cells, including phagocytosis, membrane fusion, macropinocytosis, caveolin-mediated endocytosis, and clathrinmediated endocytosis [17] (Figure 8A).To further investigate the cellular uptake mechanisms, we used small-molecule inhibitors of those aforementioned pathways in subsequent experiments.As shown in Figure 8B, the green fluorescence (AnCar-Exo LaIMTS3 ) in pro-inflammatory macrophages slightly attenuated with phagocytosis or macropinocytosis inhibitor.This data revealed that the phagocytosis and macropinocytosis partially participated in cellular uptake of AnCar-Exo LaIMTS3 by pro-inflammatory macrophages.Additionally, the green fluorescence in pro-inflammatory macrophages was notably attenuated with caveolin or clathrin inhibitors.Interestingly, the reduction of AnCar-Exo LaIMTS3 enrichment caused by caveolin or clathrin inhibition surpassed that from phagocytosis and macropinocytosis inhibition (Figure 8C).Furthermore, we inhibited caveolin or clathrin-mediated endocytosis while phagocytosis or macropinocytosis was inhibited.The results displayed a significant reduction in green fluorescence enrichment within pro-inflammatory macrophages, especially when both caveolin and clathrin were inhibited.This showed the crucial role of caveolin or clathrin-mediated endocytosis in cellular up-take mechanism of AnCar-ExoLaIMTS3 toward pro-inflammatory macrophages (Figure 8D).The representative figures was shown in Figure 8E,F.Similarly, Previous studies have reported that Caveolin and clathrin-mediated endocytosis can dramatically enhance the endocytosis efficacy of specific macromolecules, often surpassing phagocytosis and macropinocytosis rates by more than 1000-fold. [18]Therefore, the targeting ability of AnCar-Exo LaIMTS3 toward pro-inflammatory macrophage was mainly achieved through the caveolin and clathrin-mediated endocytosis.

AnCar-Exo LaIMTS3 Targets Pro-Inflammatory Macrophage Mainly Through Toll-like Receptor 4 (TLR4)
It was reported that caveolin-mediated and clathrin-mediated endocytosis were closely associated with receptor-mediated uptake processes. [19]18d] Therefore, we deduced that receptor-mediated endocytosis may play an important role in targeting ability of AnCar-Exo LaIMTS3 toward proinflammatory macrophages.Next, we investigated the potential key receptor which was related to receptor-mediated endocytosis of AnCar-Exo LaIMTS3 by pro-inflammatory macrophage.First, we conducted a comprehensive screen of protein data sourced from the NCBI protein database, specifically focusing on proteins harboring the amino acid sequence aligned with AnCar-LaIMTS3.As shown in Supplementary Materials Excel 1, certain proteins were identified containing the sequence of LPSS-GAA, based on the Max score and E value criteria.Among these candidates, we further focused on the leucine-rich repeat flightless-interacting protein 1(Figure 9A,B), which was a positive regulator of TLR4 in innate inflammatory responses. [20]nterestingly, the extracellular fragments of TLR4 consisted of leucine-rich repeats domain, which could interact with leucinerich repeat flightless-interacting protein 1 [20a,b,21] (Figure 9C).A previous study reported an increase in the expression of TLR4 pro-inflammatory macrophage. [22]Similarly, our data also showed that the red fluorescence-labeled with TLR4 antigen in LPS-treated macrophages exhibited a significant up-regulation compared to the control group, suggesting an increased expression of TLR4 protein in pro-inflammatory macrophages (Figure 9D,E).Then, we observed the colocalization of TLR4 and AnCar-Exo LaIMTS3 in pro-inflammatory macrophages at different time points (0.5, 1, 3, and 6 h) using immunofluorescence assay.As shown in Figure 9F,G, the intensity of colocalized fluorescence (yellow) was gradually increased with incubation time, indicating that AnCar-Exo LaIMTS3 may enter into pro-inflammatory macrophages through TLR4-related endocytosis pathway.To further clarify the role of TLR4 in cellular uptake of AnCar-Exo LaIMTS3 , we isolated peritoneal macrophages   9H, the size of nucleus in TLR4 −/-PM under LPS/IFN- stimuli was smaller than the ones from wild-type mice, suggesting TLR4 was essential for activation of macrophages in this condition.Moreover, the green fluorescence signal of AnCar-Exo LaIMTS3 can hardly be observed in LPS/IFN--treated TLR4 −/-PM.The statistical analysis also revealed that green fluorescence intensity of AnCar-Exo LaIMTS3 was obviously decreased in TLR4 −/-PM compared with control ones (Figure 9I).Moreover, we added TLR4-neutralizing antibody in the cell culture medium to antagonize the TLR4-mediated endocytosis pathway.As shown in Figure the enrichment of green fluorescence within pro-inflammatory macrophages was significantly decreased after the treatment of TLR4-neutralizing antibody, further revealing that the targeting ability of AnCar-Exo LaIMTS3 toward pro-inflammatory macrophages was mainly mediated by the TLR4.In conclusion, TLR4-mediated endocytosis plays a key role in the targeting ability of AnCar-Exo LaIMTS3 toward pro-inflammatory macrophages.

Targeting HIF-1𝜶 in Synovial Macrophages by AnCar-Exo LaIMTS3 Ameliorated the Severity of Inflammatory Arthritis
Previous studies have reported that HIF-1 plays an important role in macrophage activation and deletion of HIF-1 in macrophage could dramatically alleviate inflammatory response in arthritis induced by collagen. [23]Here, we crossed HIF-1 flox/+ mice with the LysM cre transgenic mice to generate mice (HIF-1 flox/flox -LysM cre ) with HIF-1 deletion in the macrophages as well as the corresponding controls (HIF-1 flox/flox ) (Figure 10A,B).Subsequently, these mice were further used for arthritis induction and evaluation.As shown in Figure 8C, the arthritic joint of HIF-1 flox/flox mice displayed moderate damage after intraarticular injection of collagenase for 14 days, which was attenuated in HIF-1 flox/flox -LysM cre mice (Movie S2, Supporting Information).In addition, the synovium of HIF-1 flox/flox mice was thickened with higher synovitis score compared with that of HIF-1 flox/flox -LysM cre mice (Figure S6A,B, Supporting Information), suggesting that HIF-1 of macrophages is a key molecular target for the treatment of inflammatory arthritis.Therefore, we intended to specifically downregulate HIF-1 of pro-inflammatory synovial macrophages through AnCar-Exo LaIMTS3 mediated HIF-1 siRNA delivery, with the purpose of reducing synovial inflammatory response of arthritic mice.We screened three sequences of siRNA-HIF-1 and evaluated the interference efficiency of siRNA-HIF-1 in LPS-activated RAW 264.7 or LPS/IFN-activated BMDM.The optimal siRNA HIF-1-3 was chosen for the subsequent experiments (Figure S7A,B, Supporting Information).Also, we also conducted additional experiments to in-vestigate the kinetics of knock-down rate of the HIF-1-siRNA loaded AnCar-Exo LaIMTS3 .Our data showed that treatment of proinflammatory cells (264.7 RAW cells activated by LPS) with HIF-1-siRNA loaded AnCar-Exo LaIMS3 significantly reduced the level of HIF-1 at 24, 48 and 96 h (Figure S7C, Supporting Information).According to previous studies [11,24] and our kinetics data, the AnCar-Exo LaIMTS3 loaded with siRNA-HIF-1 (AnCar-Exo LaIMTS3 -siRNA-HIF-1) was injected into knee joint of arthritic mice once every three days, and then gait characterization, imaging and histopathological analysis were used to comprehensively evaluate the treatment efficiency of AnCar-Exo LaIMTS3 -siRNA-HIF-1 on collagenase induced-arthritis (Figure 10D).As shown in Figure 10E-H, the Print area, Max contact area, Duty cycle, Swing speed and Max intensity were increased, while Swing was reduced in the group of AnCar-Exo LaIMTS3 -siRNA-HIF-1 compared with PBS treated group (Figure S8A,B, Supporting Information).Besides, the degree of changes in gait-related parameters was higher after treatment by AnCar-Exo LaIMTS3 -siRNA-HIF-1 compared with AnCar-Exo LaCTRL -siRNA-HIF-1.In addition, MRI imaging showed that synovial fluid was increased accompanied by a thickened synovial membrane and increased area of subpatellar fat pad in PBS treated group compared with the Sham group.These changes were dramatically inhibited by the treatment of AnCar-Exo LaIMTS3 -siRNA-HIF-1, while the inhibitory effect was weaker in AnCar-Exo LaCTRL -siRNA-HIF-1 group (Figure 10I; Figure S8C, Supporting Information).Similarly, HE staining also suggests that the increased synovitis score in arthritic mice was more significantly reversed by AnCar-Exo LaIMTS3 -siRNA-HIF-1 than that of AnCar-Exo LaCTRL -siRNA-HIF-1 (Figure 10J,K).Moreover, immunofluorescence staining revealed that the number of F4/80+ cells, as well as percentage of F4/80 + HIF-1 + cells, were more significantly decreased in inflammatory joint injected with AnCar-Exo LaIMTS3 -siRNA-HIF-1 compared to that of AnCar-Exo LaCTRL -siRNA-HIF-1 treated group(Figure 10L-N), indicating the efficient konckdown of HIF-1 in macrophage of inflammatory joint and reduction of infiltration of inflammatory cells in synovium.However, there were no significant changes including osteophyte formation and joint space narrowing among the four groups as revealed by X-ray images at an early stage of arthritis (Figure S9, Supporting Information).In addition, the Safranin O/Fast Green staining exhibited that AnCar-Exo LaIMTS3 -siRNA-HIF-1 didn't influence obviously cartilage destruction at two weeks after collagenase inducedarthritis compared with the control (Figure S10, Supporting Information).These data, together with MRI results, indicated that the inflammatory response of the joint occurred earlier than the changes of bone structure and damage of cartilage at the early stage of collagenase-induced-arthritis, and the regulatory effect of AnCar-Exo LaIMTS3 -siRNA-HIF-1 might be related to inhibition of joint inflammatory response.Moreover, we investigated the joint lesion at the late stage (8 weeks) of collagenase-induced-arthritis using the Safranin O/Fast Green staining.The data showed that Figure 6.AnCar-Exo LaIMTS3 could specifically target synovial macrophage in mice of inflammatory arthritis.A) A schematic diagram illustrating that the C57 mice were induced arthritis by collagenase and intra-articularly injected with AnCar-Exo LaIMTS3 and AnCar-Exo LaCTRL exosomes.B) Quantification of joint DID fluorescence intensities in mice treated with AnCar-Exo LaIMTS3 and AnCar-Exo LaCTRL .All values were displayed as the way of mean ± SD.Two-way analysis of variance (ANOVA) was used to evaluate the difference among groups.N = 10 per group, **P < 0.01, ***P < 0.001.C) Representative images of mice treated with PBS or exosomes through intra-articular injection and detected by Vilber Lourmat at different time points.D) A schematic diagram illustrating establishment of macrophage labelling tdTomato-LysM Cre mice and time-series in vivo imaging of exosomes entering into macrophages as detected by two-photon microscopy.E, F) the uptake of DID labeled AnCar-Exo CTRL and AnCar-Exo LaIMTS3 by tdTomato positive synovial macrophages.
the joint cartilage showed significant deterioration with the highest OARSI score in the arthritis group, which could be partially reversed by the treatment of AnCar-Exo LaIMTS3 -siRNA-HIF-1 (Figure 10O-Q).In brief, targeted delivery of siRNA-HIF-1 to inflammatory macrophages by AnCar-Exo LaIMTS3 can efficiently inhibit the expression of HIF-1 in synovial macrophages, decrease synovial inflammatory response, reduce the lesion of cartilage and ameliorate gait abnormalities in mice arthritis model.

Discussion
Macrophages are heterogenic phagocytic cells that play distinct roles in multiple physiological and pathological processes. [25]argeting different types of macrophages has shown potent therapeutic effects in diseases, which has attracted great interest in recent years. [26]TAMs (tumor-associated macrophages) are responsible for the cancer microenvironment, influencing growth and progression of tumors. [27]Weissleder et al. reported that -cyclodextrin nanoparticles controlled tumor growth and prevented tumor relapse by targeting TAMs. [28]Besides, exosomes modified by IL-4 receptor reprogrammed TAMs into pro-inflammatory macrophage and inhibited tumor growth. [10]lthough many approaches have been developed to target alternatively activated macrophages (M2 macrophages), there are few researches on targeting pro-inflammatory macrophages partially because of the non-specificity phagocytosis. [29]Interestingly, our data indicated that the phagocytosis efficiencies of proinflammatory macrophages showed obvious differences among the exosomes modified by different types of AnCar-LaIMTS.In the existing candidate range, the AnCar-Exo LaIMTS3 has the best targeting ability to pro-inflammatory macrophages in vitro and in vivo, indicating that modification of exosomes through biomolecular engineering is a prospective method to target pro-inflammatory macrophages.
Except for targeting ability, efficient loading of cargo into exosomes is also a crucial way to improve their therapeutic effectiveness for diseases.Recently, researchers have shown an increased interest in cargo loading of exosomes by itself.Ryosuke et al engineered surface of exosome membrane by conjugating L7Ae to the C-terminus of CD63 in order to make exosome itself carry more mRNA after the 3′-untranslated region of mRNA is inserted into a C/D box sequence. [30]Thus, it is expected that modification of the C-terminus of AnCar-LaIMTS according to the respective treatment needed will make exosomes selectively load specific cargos and improve their therapeutic effects.In addition, it is available to improve inflammatory responsiveness of exosome release in arthritis.Park et al reported that exosomes modified by a thioketal linker-embedded poly (ethylene glycol) enhanced the cellular uptake of ROS-responsive exosomes after the production of excessive ROS in arthritis. [31]Therefore, combined application of exosomes and biomaterials may further improve the therapeutic effects of exosomes.
Engineered exosomes can enhance targeting efficiency to specific cells by binding to ligand receptors, which could maximize effectiveness of drug delivery. [32]There are various pathways for extracellular macromolecules to enter cells, among which receptor-mediated endocytosis enables selective and efficient uptake of large molecules in engineered exosomes. [17,19]Previous study has reported that exosomes with membrane surface modified by PD-L1 could target cells with high expression of PD-1 receptor, which have therapeutic efficiency in both UC and psoriasis disease. [33]In addition, exosomes engineered by high binding affinity could deliver drugs to HER2 receptor-expressing tumors in vivo. [34]Our data indicate that the expression of TLR4 receptor is increased in pro-inflammatory macrophages which greatly contributed to the targeting ability of AnCar-Exo LaIMTS3 towards pro-inflammatory macrophages.
In conclusion, we construct new engineered exosomes that can efficiently target pro-inflammatory macrophages to exert the effect of targeted treatment of arthritis in this study.Wherein, the optimal AnCar-Exo LaIMTS3 can efficiently deliver therapeutic contents to pro-inflammatory macrophages mainly through TLR4mediated endocytosis, which ameliorates the severity of arthritis without obvious toxicity.This work provides a novel perspective and potential strategy for targeting treatments of arthritis and other inflammatory diseases in the future.
Plasmid Construction: As previous study reported, [13] three repeated pro-inflammatory macrophages were designed targeting sequences (1-10, Table S1, Supporting Information) or random sequences in the extracellular domain between signal peptide and mature peptide of human LAMP2B(NP_054701.1)through a flexible peptide linker (GGGGS) and added GFP tag to the carboxy-terminal region of LAMP2B recombinant protein.The sequences coding recombinant protein were fused into PLVXpuro at the XhoI and BamHI sites.Lentivirus was produced by transfecting HEK293T cells transiently.AnCar sequences were synthesized, and inserted into PLVX-puro vector with XhoI and BamHI restriction sites.For the lentivirus package, HEK-293T cells were transfected with the above plasmids, the psAX2 packaging plasmid and pMD2G envelope plasmid for 48 h to obtain the lentivirus supernatant.
Establishment of Stable Cell Lines: HEK293T cells were cultured in a 6-well plate at the density of 30-50% at day 1.Next, those Lentiviruses with anti-puromycin gene were added according to the MOI of HEK293T (MOI = 5) in the afternoon on day 2.After the density of cells reached over 80%, the complete medium was replaced with fresh complete medium containing 2 μg ml −1 puromycin.When cell death didn't occur in six-in well plate on day 5, the concentration of puromycin was reduced to 0.5 μg ml −1 .Then, the mixed clonal cell lines were digested with 0.05% trypsin and counted.60 cells cultured in 10 ml complete media were seeded in a 96well plate.After 10 days, the stable monoclonal cell line was passaged and expanded until the number of cells was enough for exosome production.
Preparation of Exosomes: When the density reached 80%, cells in four 10-cm dishes were washed with DPBS twice and the medium was exchanged with DMEM/High glucose supplemented with 10% exosomedepleted FBS and 1% penicillin/streptomycin solution.After 24 h, the supernatant was collected for isolation of exosomes.Then, ultracentrifugation was used to obtain exosomes.First, the supernatant was centrifuged at 300 g for 10 min in order to remove the dead cells.Second, the supernatant transferred into a new centrifuge tube was centrifuged at 2000 g for 10 min in order to remove cell debris.Third, the supernatant transferred into a new centrifuge tube was centrifuged at 10 000 g for 10 min to discard microvesicles.Fourth, the supernatant transferred into a new cen-trifuge tube was filtered with a 0.22 μm filter (Steritop; Millipore) to remove apoptotic bodies.Fifth, the supernatant was centrifuged at 1 00 000 g for 90 min to obtain exosomes.Lastly, the exosomes were treated with RIPA supplemented with a protease inhibitor or resuspended with 200 μl DPBS and stored at −80 °C until use.
Exosome Identification: The characteristics of exosomes were evaluated using four approaches, including Nanosight, transmission electron microscope (TEM), Western blot, flow cytometer (ImageStreamX Mark II).For Nansight, the exosomes adequately resuspended with 1 ml PBS were gently added into Nanosight NS300 (Malvern) sample pool.Next, put the laser module back on the pedestal and place a probe with OMEGA thermometer in Copper hole of laser module.Then adjust the suitable screen gain and camera level.When all parameters were ready, measured the samples.Finally, the samples and particle concentration (particles/ml) was counted.For TEM, place 5-10 μl exosomes on the copper screen and settle for 3 min.Next, remove the liquid near the edge by using filter paper.Then, negatively stained with phosphotungstic acid after PBS rinsed.Finally, exsomes were dried at room temperature and examined by TEM (Japan, JEM-1200EX).Western blot was performed according to the previous studies. [38]The antibodies were mouse anti-lamp2 antibody, rabbit anti-TSG101, rabbit anti-CD63, rabbit anti-CD81, and mouse anti--actin.For the flow cytometer, it could detect the exosomes with GFP due to the insertion of fusion protein LAMP2B.Exosomes were isolated from supernatant and measured by amnis Im-ageStream, in which the 100 nm FITC particles were used as the standard control.
The Separation of Primary Cells and Cell Culture: The primary chondrocytes were separated from murine articular cartilage. [39]The following steps have been done in sterile environment: 1) obtained femoral heads, femoral condyles and tibial plateau after removing skin and soft tissues.2) Placed them in 1 × DPBS and washed twice with 1 × DPBS.3) Discarded DPBS and put them into centrifuge tube with 10 ml 0.1% type II collagenase at 37 °C in an incubator under 5%CO 2 overnight.4) The tube was centrifuged at 1000 rpm for 5 min.5) Discarded the supernatant and seed primary chondrocytes in culture dishes with DMEM/F12 supplemented with 10% FBS.
The primary fibroblasts were separated from mouse synovium.The following steps have been done in sterile environment: 1) obtained synovium near the knee joint as much as possible.2) Cut the synovium into 1 mm 3 pieces and digested them with type IV collagenase overnight in 15 ml centrifuge tube placed in shaker at 37 °C.3) Centrifuged at 1000 rpm for 5 min.4) Discarded the supernatant and seed the cells in culture dishes.The severity of articular cartilage damage among groups were evaluated using the OARSI scoring system, maximal and summed scores were calculated, *P < 0.05, ***P < 0.001.All values were displayed as the way of mean ± SD.One-way analysis of variance (ANOVA) was used to evaluate the difference among groups.
The bone marrow-derived macrophages (BMDM) were obtained from bone marrow of adult mice [40] using the following steps in sterile environment: 1) Obtained femur and tibia after removing the muscle around them. 2) Cut base and knee joint of femur and epiphyses of tibia to obtain bone marrow through a 23G needle.3) Washed the bone marrow with 2, 3 ml DMEM/High glucose supplemented with 10% FBS in 10 cm dishes.4) Transferred the supernatant into 15 ml centrifuge tube and centrifuged at 200 g for 5 min under 4 °C.5) Discarded the supernatant and resuspended sediment with complete medium.6) Seed cells in 10 cm dishes.7) Collected the supernatant in 10 cm dishes after 24 h and centrifuge at 1000 rpm.8) Resuspended the sediment with complete medium and seed in 6-well plates supplemented with 10 ng ml −1 M-CSF.9) Changed the medium with fresh complete medium supplemented with 10 ng ml −1 M-CSF at day 3. 10) Evaluated the formation of BMDM.In order to polarize BMDM into pro-inflammatory macrophages, 100 ng ml −1 LPS and 50 ng ml −1 INF-were added in complete medium and incubated for 24 h.To polarize BMDM into anti-inflammatory macrophages, 10 ng ml −1 IL-4 was added in complete medium and incubated for 24 h.
As previous study reported, [2b] the primary peritoneal macrophages were isolated using the following steps: 1) Surgical instruments (scissors, forceps, etc.) were sterilized, and the dissection table was disinfected with ultraviolet light.2) Mice were euthanized by cervical dislocation and immersed in 75% ethanol for 5 min to ensure disinfection.3) The abdominal skin was gently lifted using forceps, a small incision was made, and the abdominal wall was carefully dissected to expose the peritoneum below the xiphoid process, ensuring the integrity of the peritoneal membrane.4) A 10 mL syringe was used to aspirate 1640 medium containing 5% FBS (precooled on ice), and it was carefully injected into the abdominal cavity above the liver to prevent accidental penetration of the intestines.5) ≈5, 6 mL of medium was gently infused into the abdominal cavity, followed by a 15 min abdominal massage to ensure thorough distribution.6) The liquid from the abdominal cavity was aspirated using a syringe and transferred to a 15 mL centrifuge tube, then centrifuged at 1000 rpm for 5 min.7) The collected cells were suspended in 1640 medium supplemented with 10% FBS and 1% antibiotics, and incubated at 37 °C for 2-4 h.8) After incubation, the cells were washed with PBS to remove non-adherent cells, and the remaining adherent cells were considered peritoneal macrophages (with a purity of 85%-95%).
For RAW 264.7 cell line, incubated with 100 ng ml −1 LPS for 24 h to generate pro-inflammatory macrophages or incubated with 20 ng ml −1 IL-4 to obtain anti-inflammatory macrophages.
The Inhibition of Pathways Participating in Exosome Uptake by Cells: To generate pro-inflammatory macrophages, RAW 264.7 cells were incubated with 100 ng ml −1 LPS for 24 h.Then, for phagocytosis and macropinocytosis, Wortmannin (1 μm) and Amiloride (1 mm) were used to treat pro-inflammatory macrophages for 30 min. [41]For clathrin-mediated and caveolin-mediated endocytosis, chlorpromazine (10 μg ml −1 ) was used to treat macrophages for 30 min, while the dynasore (80 μg ml −1 ) was used to treat macrophages for 1 h. [42]Subsequently, AnCar-Exo LaIMTS3 was added for co-incubation for 12 h.Finally, immunofluorescence was performed to observe the enrichment level of AnCar-Exo LaIMTS3 in pro-inflammatory macrophages.
Flow Cytometry: The BMDM was seeded in 35 mm-confocal dishes at the density of 2 × 105 cells with complete medium containing nothing or 100 ng ml −1 LPS and 50 ng ml −1 INF- or 10 ng ml −1 IL-4 for 24 h.The complete medium was then replaced with complete medium containing 10% exosome-depleted FBS and incubated with 1 × 10 8 AnCar-Exo LaIMTS3 for 12 h at 37 °C.Later, Collected the M, pro-inflammatory and anti-inflammatory macrophages and resuspended them with FACS buffer (PBS containing 1% FBS, 1 mM EDTA).Next, prepared for blank control, and Isotype control for the following analysis.Incubated with antibodies against F4/80, CD86, CD206 (1:100, Biolegend) at 4 °C for 30 min.Then, washed with FACS buffer.The cells were centrifuged at 500 g for 5 min.At last, over 50 000 events were collected for analyzing by Flow cytometry (Beckman, CytoFLEX SRT).
Exosome Tracing In Vivo: Exosomes were characterized in mice following systemic or local administration of them.Arthritis was induced by injecting collagenase into knee joint.Exosomes were labeled with DID.For systemic administration, mice were randomly divided into three groups.After injected collagenase at day 7, the arthritis mice were injected with exosomes (1 × 10 9 particles per mouse) according to the dose of previous study. [43]Then, the mice imaging was detected by Fusion FX Edge system (Vilber Lourmat).The signal intensity was recorded after injecting exosomes at 0, 1, 3, 6, 12, 24, and 48 h.Three hours after injection, three mice were sacrificed to observe the distribution of exosomes in major organs.For local administration in particular cavities, mice were also randomly divided into three groups.Seven days after collagenase injection, the arthritic mice were injected with exosomes (1 × 10 9 particles per mouse).Mice were then imaged by Fusion FX Edge system.The signal intensity was recorded after injecting exosomes at 0, 1, 3, 6, 12, 24, 48, and 72 h.The exposure time, focus and position of mice were the same between mice with local or systemic administration.The FusionCapt Ad-vance software was used to analyze the images.Briefly, the interesting regions showing signal intensity were identified.Then, the background of images was balanced and original intensities of fluorescence were recorded.Last, the real intensities of exosomes were obtained by subtracting their background fluorescence from the original intensities.
For tdTomato-LysM Cre mice, the mice were injected with 1 × 10 9 particles into a vail vein after being placed in 10cm-dishes containing 3% lowmelting hydrogel to immobilize the mouse.Then mice were detected by Two-Photon Microscopy from 0 min to 120 min.
16b] First, an injury in the zebrafish caudal fin was induced using a surgical scalpel to create the injury model.After 7 hours, the zebrafish were anesthetized and fixed.Subsequently, a microinjector was used to administer exosomes into the zebrafish heart region.Finally, the injured portion of the zebrafish caudal fin was positioned under a twophoton microscope for observation.
Arthritis Induction and Exosome Injection: Twelve-weeks-old' C57BL/6J male mice (Charles River) were divided into four groups (5 males/group).The approaches to induce arthritis were described in previous study. [43]fter being anesthetized with 0.8% pentobarbital, the mice in sham group had the skin of the right knee joint cut without injecting collagenase.While for arthritis group, the knee joints were fully exposed and injected with 1 unit type VII collagenase (Sigma, C0773) in 10 μl CaCl 2 at day 1 and day 3, respectively.For transgenic mice, the same procedures were applied to experimental groups.
Gait Analysis: Before formal gait analysis experiment, all mice were trained to be familiar with the test environment and can cross the walkway without interruptions or hitches in video-based CatWalk gait analysis system (Netherlands, Noldus).The system was composed of an enclosed walkway, a yellow fluorescent lamp, a high-speed color camera and analysis software.The gait data were recorded by a high-speed color camera.
X-Ray and MRI: All mice were examined by X-ray using Faxitron MX-20 (Faxitron Bioptics, Tucson USA) at 26 kV.For MRI, the mice were anesthetized with 0.8% pentobarbital and scanned by a small animal MRI (Germany, Bruker BioSpec USR 70/20).After the right leg knee joint was fixed with four channels coil, the mice were scanned with T1WI, T2WI.Histology Analysis: All samples of knee joints went through 4% paraformaldehyde fixation, 0.5 m EDTA at pH 7.4, 30% sucrose and optimal cutting temperature (OCT) compound embedding.14a] After being blocked with QuickBlock Blocking Buffer for Immunol Staining (Beyotime, P0260), those sections were incubated at 4 °C overnight with the primary antibody such as F4/80, iNOS, CD206.Then sections were incubated with secondary antibodies conjugated with fluorescence (1:500; Invitrogen/Abcam) at 37 °C for 1 h.Last, DAPI (1:1000; D8417, Sigma-Aldrich) was used to stain the nucleus.Samples were detected by the living cells workstation (Olympus, Tokyo, Japan).
Bioinformatics Analysis of Proteins Containing the Ancar-Laimts3 Amino Acid Sequence: To explore proteins containing the AnCar-LaIMTS3 amino acid sequence, a comprehensive search was conducted through the NCBI protein database.The focus was on identifying proteins with sequences aligned to AnCar-LaIMTS3.Initially, a protein BLAST by inputting the AnCar-LaIMTS3 sequence into the "Enter Query Sequence" field was employed.The search by setting the organism as Mus and selecting "proteinprotein blast" as the program was narrowed down.Subsequently, a range of proteins containing this sequence was identified.The selection by considering the Max score and E value was further refined.This process led to shortlisting the top 100 ranked proteins.
Immunofluorescence: The OCT sections (10 μm thickness) were chosen for Immunofluorescence.All samples of knee joint sequentially went through 4% paraformaldehyde fixation, 0.5 m EDTA at pH 7.4, 30% sucrose and optimal cutting temperature (OCT) compound embedding.14a] After being blocked with QuickBlock Blocking Buffer for Immunol Staining (Beyotime, P0260), sections were incubated at 4 °C overnight with the primary antibody such as F4/80, iNOS, CD206.Sections were then incubated with secondary antibodies conjugated with fluorescence (1:500; Invitrogen/Abcam) at 37 °C for 1 h.Last, DAPI (1:1000; D8417, Sigma-Aldrich) was used to stain the nucleus.Signals in the section were detected by the living cells workstation (Olympus, Tokyo, Japan).
Statistical analysis: All values were displayed as the way of mean ± SD.One-way or two-way analysis of variance (ANOVA) with a Tukey post-hoc test was used to evaluate the difference among groups.P < 0.05 were considered statistically significant (*P < 0.05, **P < 0.01, ***P < 0.005, and ****P < 0.001).The software for Statistical analysis was Graph Pad Prism 9.0 (GraphPad Software Inc., La Jolla, CA, USA).

Figure 2 .
Figure 2. Design and construction of AnCar-LaIMTS plasmids.A) The major components of AnCar-LaIMTS consisted of four elements.B) Eleven kinds of AnCar-LaIMTS sequences were inserted into PLVX-puro plasmids.C) The results of electrophoresis images.Target genes were digested at XhoI and BamHI sites.AnCar-LaIMTS sequence was labelled by red arrow.D) DNA sequencing analysis.Sequences of DNA were consistent with peptide sequences.E) Eleven kinds of AnCar-LaIMTS plasmids were transfected into HEK 293T cell through lentivirus.F) GFP protein expression in 11 cell lines.The expressed GFP-fusing proteins could be observed green florescence under fluorescent microscopy.Scale bars, 25 μm.G) Western Blot analysis of fusion protein in cells.Antibodies were anti-LAMP2B and anti-GFP.Control means HEK 293T.GFP means HEK 293T transfected with blank GFP protein plasmid.

Figure 5 .
Figure 5. Evaluating the distribution and potential toxicity of AnCar-Exo LaIMTS3 in vivo.A) A schematic diagram illustrating that the C57 mice were induced arthritis by collagenase and injected with AnCar-Exo LaIMTS3 through systemic administration.B) Quantification of whole-body DID fluorescence intensities in mice treated with AnCar-Exo LaIMTS3 and AnCar-Exo LaCTRL .All values were displayed as the way of mean ± SD.Two-way analysis of variance (ANOVA) was used to evaluate the difference among groups.N = 10 per group.C) Representative images of mice treated with PBS or exosomes and detected by Vilber Lourmat at different time points.D-F) DID fluorescence in different organs of mice at 3 h after exosome injection.A: Heart, B: Liver, C: Spleen, D: Lung, E: Left Kidney, F: Right Kidney, E: Left Leg, H: Right Leg.N = 3 per group.G) Quantification of DID fluorescence intensities in different organs between AnCar-Exo LaIMTS3 and AnCar-Exo LaCTRL at 3 h after injection, N = 3 per group.All values were displayed as the way of mean ± SD.One-way analysis of variance (ANOVA) was used to evaluate the difference among groups.***P < 0.001.H) Hematoxylin-eosin staining of liver and spleen at 48 h after injection.N = 3 per group.Scale bars, 100 μm.
www.advancedscience.com(PM) from TLR4 −/− mice, and detected the change of AnCar-Exo LaIMTS3 fluorescence signals in TLR4 −/-PM treated with LPS and IFN-.As indicated in Figure

Figure 7 .
Figure 7. AnCar-Exo LaIMTS3 could specifically target recruited macrophage in zebrafish caudal fin injury model.A) A schematic diagram showing the establishment of macrophage labeling transgenic zebrafish.B) Genotype identification using mCherry protein under fluorescence microscope.C) A schematic diagram showing the caudal fin injury model and macrophage aggregation during 7 h.D) Injury to the caudal fin leads to the aggregation of macrophages at the injured site.E) A schematic diagram illustrating the entry of exosome into macrophage detected by two-photon microscopy.F, G) Two-photon microscopy showed that mCherry-labeled macrophages ingested exosome labelled by DID dye during 5 min.N = 3 independent biological replicates, Scale bars, 50 μm.***P < 0.001.All values were displayed as the way of mean ± SD.One-way analysis of variance (ANOVA) was used to evaluate the difference among groups.

Figure 9 .
Figure 9. AnCar-Exo LaIMTS3 targeted pro-inflammatory macrophage through Toll-like receptor 4 (TLR4).A) The Max score of protein blast in NCBI.B) The E value of protein blast in NCBI.C) A schematic diagram that Leucine-rich repeat flightless-interacting protein 1 binds to Leucine-rich repeat domain.D) Immunofluorescence showed the number of TLR4 activated by LPS increased.Scale bars, 20 μm.E) Quantification of TLR4 fluorescence intensities in macrophages between control and LPS group.The fluorescence intensities were calculated under three high-magnification fields with image J. N = 3 independent biological replicates, **P < 0.01.F) Immunofluorescence showed the co-localization of AnCar-Exo LaIMTS3 and TLR4 gradually increased within 6 hours.Scale bars, 20 μm.G) Quantification of the co-localization fluorescence intensities between AnCar-Exo LaIMTS3 and TLR4 in macrophages between control and LPS group.The fluorescence intensities were calculated under three high-magnification fields with image J. N = 3 independent biological replicates, **P < 0.01, ****P < 0.0001.H) Immunofluorescence showed AnCar-Exo LaIMTS3 in TLR4 −/− and TLR4 +/+ BMDM.Scale bars, 20 μm.I) Quantification of AnCar-Exo LaIMTS3 fluorescence intensities in TLR4 −/− and TLR4 +/+ BMDM.The fluorescence intensities were calculated under three high-magnification fields with image J. N = 3 independent biological replicates, ****P < 0.0001.J) Quantification of AnCar-Exo LaIMTS3 fluorescence intensities in macrophages treated by TLR4 neutralizing antibody.The fluorescence intensities were calculated under three high-magnification fields with image J. N = 3 independent biological replicates, ****P < 0.0001.K) Immunofluorescence showed the enrichment level of exosomes in peritoneal proinflammatory macrophages decreases as the concentration of TLR4-neutralizing antibodies increases, Scale bars, 20 μm.All values were displayed as the way of mean ± SD.One-way analysis of variance (ANOVA) was used to evaluate the difference among groups.

Figure 10 .
Figure 10.Targeting HIF-1 in synovial macrophage by AnCar-Exo LaIMTS3 ameliorated the severity of inflammatory arthritis.A) A schematic diagram illustrating the establishment of transgenic mouse model with HIF-1 conditional deletion in the macrophages.B) Genotype identification using PCR.C) The strategy of inducing arthritis by injection collagenase and phenotype between HIF-1 flox/flox and HIF-1 flox/flox -LysM cre transgenic mice.n = 5 per group.D) A schematic diagram illustrating targeting HIF-1 in synovial macrophage by AnCar-Exo LaIMTS3 to treatment arthritis and detection by gait analysis, medical image and histopathology.E-H) The changes of gait parameters among groups (n = 5 per group), including Print area, Max contact area, Duty cycle, Max intensity.The values of the right hind parameters were displayed, N = 5 per group, *P < 0.05, **P < 0.01.I) Medial, medium and lateral tibial MRI images of right legs among groups.The red dotted line marks the articular capsule tissue.J, K) Synovitis was shown by H&E staining and Synovitis scores after 2 weeks' treatment among groups.N = 5 per group.*P < 0.05.L) Immunofluorescence of F4/80 (green), HIF-1 (red) in synovial tissues after treatment.Scale bars, 10 μm.M) Quantification analysis of F4/80 positive cells among groups.The fluorescence intensities were calculated under three high-magnification fields with image J, N = 3 independent biological replicates, *P < 0.05, ***P < 0.001.N) Quantification analysis of F4/80 + HIF-1 + cells among groups.The fluorescence intensities were calculated under three high-magnification fields with image J, N = 3 independent biological replicates, *P < 0.05, ***P < 0.001.O-Q) Representative images of Safranin O/Fast Green-stained sections of knee joints among groups after 8 weeks' treatment.N = 5 per group.Scale bar, 100 μm.The severity of articular cartilage damage among groups were evaluated using the OARSI scoring system, maximal and summed scores were calculated, *P < 0.05, ***P < 0.001.All values were displayed as the way of mean ± SD.One-way analysis of variance (ANOVA) was used to evaluate the difference among groups.