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-Exo^LaIMTS . The data indicate that the phagocytosis efficiencies of pro-inflammatory macrophages for different AnCar-Exo^LaIMTS show obvious differences. The AnCar-Exo^LaIMTS3 has the best targeting ability for pro-inflammatory macrophages in vitro and in vivo. Mechanically, AnCar-Exo^LaIMTS3 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-Exo^LaIMTS3 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.

Keywords: HIF-1α; arthritis; exosome.

Figure 1

Schematic illustration of AnCar‐Exo^LaIMTS and the potential application for inflammatory arthritis. A) the design and construction of AnCar recombinant plasmid. The AnCar is composed of four elements, including exosome membrane anchoring sequence, pro‐inflammatory macrophage targeting sequence, exosome tracing sequence, flexible linker region sequence. B) Preparation and production of AnCar‐Exo^LaIMTS3. C) AnCar‐Exo^LaIMTS3 targets pro‐inflammatory macrophages (PIMs) through TLR4. D) Knockdown HIF‐1α in PIMs ameliorates arthritis via AnCar‐Exo^LaIMTS3‐siRNA‐HIF‐1α. AnCar‐Exo^LaIMTS3‐siRNA‐HIF‐1α can efficiently inhibit the expression of HIF‐1α in synovial macrophages, decrease synovial inflammatory response, ameliorate gait abnormalities and pathological change in mice arthritis model.

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 3

Characteristics of AnCar‐LaIMTS modified exosomes. A) The model and structure of AnCar‐Exo^LaIMTS. LAMP2B is fused with pro‐inflammatory macrophage sequence peptide. B) A schematic diagram illustrating isolation of AnCar‐Exo^LaCTRL, AnCar‐Exo^LaIMTS and AnCar‐Exo^LaRANDOM. AnCar‐Exo^LaCTRL was derived from HEK 293T cells. C) Particle size distribution of AnCar^LaCTRL, AnCar‐Exo^LaIMTS and AnCar‐Exo^LaRANDOM detected by Nanosight. D) The morphology of AnCar^LaCTRL, AnCar‐Exo^LaIMTS and AnCar‐Exo^LaRANDOM measured by transmission electron microscope. Scale bars, 100 nm. E) Western Blot analysis of fusion protein in exosomes. Antibodies were anti‐CD63, anti‐CD81, anti‐TSG101 and anti‐LAMP2B. F) GFP fluorescence in exosomes detection by ImageStreamX.

Figure 3

Characteristics of AnCar‐LaIMTS modified exosomes. A) The model and structure of AnCar‐Exo^LaIMTS. LAMP2B is fused with pro‐inflammatory macrophage sequence peptide. B) A schematic diagram illustrating isolation of AnCar‐Exo^LaCTRL, AnCar‐Exo^LaIMTS and AnCar‐Exo^LaRANDOM. AnCar‐Exo^LaCTRL was derived from HEK 293T cells. C) Particle size distribution of AnCar^LaCTRL, AnCar‐Exo^LaIMTS and AnCar‐Exo^LaRANDOM detected by Nanosight. D) The morphology of AnCar^LaCTRL, AnCar‐Exo^LaIMTS and AnCar‐Exo^LaRANDOM measured by transmission electron microscope. Scale bars, 100 nm. E) Western Blot analysis of fusion protein in exosomes. Antibodies were anti‐CD63, anti‐CD81, anti‐TSG101 and anti‐LAMP2B. F) GFP fluorescence in exosomes detection by ImageStreamX.

Figure 4

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 × 10⁸ 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‐γ‐activated BMDM were incubated with AnCar‐Exo^LaIMTS1 to AnCar‐Exo^LaIMTS10, AnCar‐Exo^LaRANDOM (1 × 10⁸ 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 Mφ, 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 Mφ, pro‐inflammatory macrophage and anti‐inflammatory macrophage. Scale bars, 10 µm. K) Quantification of GFP fluorescence intensities in AnCar‐Exo^LaIMTS3treated Mφ, 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 Mφ or between anti‐inflammatory macrophage and Mφ. N) Flow Cytometry shows the ratio of CD86⁺GFP⁺ in LPS/INF‐γ‐activated or IL‐4‐activated BMDM. O) Flow Cytometry shows the ratio of CD206⁺GFP⁺ in LPS/INF‐γ‐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.

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.

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.

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 8

Identify the cellular uptake mechanisms of AnCar‐Exo^LaIMTS3 targeting pro‐inflammatory macrophage. A) Potential patterns of pro‐inflammatory macrophage uptaking exosomes. B) Quantification of AnCar‐Exo^LaIMTS3 fluorescence intensities in pro‐inflammatory macrophage after phagocytosis (inhibitor: Wortmannin, Wor) and macropinocytosis (inhibitor: Amiloride, Ami) were inhibited. The fluorescence intensities were calculated under three high‐magnification fields with image J. N = 3 independent biological replicates, ***P < 0.001, ****P < 0.0001. C) Quantification of AnCar‐Exo^LaIMTS3 fluorescence intensities in pro‐inflammatory macrophage after the caveolin‐mediated endocytosis (inhibitor: Dynasore, Dyn) and clathrin‐mediated endocytosis (inhibitor: Chlorpromazine, Chl) were inhibited. The fluorescence intensities were calculated under three high‐magnification fields with image J. N = 3 independent biological replicates, ****P < 0.0001. D) Quantification of AnCar‐Exo^LaIMTS3 fluorescence intensities in pro‐inflammatory macrophage after phagocytosis, macropinocytosis, caveolin‐mediated endocytosis and clathrin‐mediated endocytosis were inhibited. The fluorescence intensities were calculated under three high‐magnification fields with image J. N = 3 independent biological replicates, ****P < 0.0001. E) Representative images of pro‐inflammatory macrophages uptake AnCar‐Exo^LaIMTS3 when phagocytosis and macropinocytosis were inhibited. Scale bars, 20 µm. F) Representative images of pro‐inflammatory macrophages uptake AnCar‐Exo^LaIMTS3 when the caveolin‐mediated endocytosis and clathrin‐mediated endocytosis were inhibited. Scale bars, 20 µm. G) Representative images of pro‐inflammatory macrophages uptake AnCar‐Exo^LaIMTS3 when phagocytosis, macropinocytosis, caveolin‐mediated endocytosis and clathrin‐mediated endocytosis were inhibited. 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 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 pro‐inflammatory 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

Targeting HIF‐1α 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‐1α conditional deletion in the macrophages. B) Genotype identification using PCR. C) The strategy of inducing arthritis by injection collagenase and phenotype between HIF‐1α^flox/flox and HIF‐1α^flox/flox‐LysM^cre transgenic mice. n = 5 per group. D) A schematic diagram illustrating targeting HIF‐1α 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‐1α (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‐1α⁺ 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.