Background

Exosomes are extracellular vesicles of 30–100 nm diameter and are secreted by most eukaryotic cells [1]. They are secreted from various types of cells and contribute to homeostasis by participating in intercellular interactions [2,3,4]. Conventionally, exosome research has focused on surface markers or their internally encapsulated cargo. CD9 and CD63 are verified exosome-specific markers [5]. In addition, the components of the internal cargo (lipids, nucleic acids, metabolites) have been characterized [5,

Rheumatoid arthritis

RA is a chronic disease that causes severe pain, swelling, stiffness, and functional loss of joints [17,18,19]. Although the exact pathogenesis of this disease is unknown, RA onset is reportedly caused by various risk factors, including genetic, epigenetic, dust inhalation, microbiota, and lifestyle [20,21,22]. Immune regulatory factors related to nuclear factor kappa B (NF-κB) stimulation, activation, and functional differentiation of T cells are believed to be closely related to RA. Human leukocyte antigen (HLA)-DBR1 reportedly has the most robust relation with RA among genetic factors. The effects of environmental factors are apparent when investigated at the genetic level. For example, smoking and silica inhalation significantly increase the incidence rate of RA in individuals with susceptibility to HLA-DR4 alleles.

Although RA is a systemic autoimmune disease, symptoms usually occur with severe joint inflammation [23]. Thus, targeting the inflammatory disease microenvironment and cell populations is crucial for treating RA [24]. In synovial tissue, highly proliferative immune cells, vascular endothelial cells, and fibroblast-like synoviocytes (FLSs) establish an inflammatory microenvironment and release various factors such as TNF-α, IL-1, IL-6, matrix metalloproteinases (MMPs), and autoantibodies [25,26,27,28,29]. These components facilitate the development of hyperinflammatory microenvironment. In addition, systemic impaired immune functions and an imbalanced proportion of immune cells, such as in the helper T cell and regulatory T cell ratio, further aggravates RA pathology. The cell types closely related to RA pathogenesis include FLSs, macrophages, B cells, T cells, and dendritic cells (DCs) (Fig. 1).

Fig. 1
figure 1

A The modulation of various kinds of cells in rheumatoid arthritis-microenvironment by engineered exosomes. B Immunosuppression effects on macrophages, dendritic cells (DCs), and T cells. FLS, fibroblast-like synoviocytes; Th1, T-helper 1 cells; Th17, T-helper 17 cells; Treg, Regulatory T cells

FLS, a significant producer of inflammatory cytokines at the synovial site, is primarily responsible for RA initiation [30, 31]. FLSs constitute a 2–3-layer network in normal synovium. However, in the RA pathological condition, an inflammatory response is induced by continuous FLS proliferation and immune cell accumulation, which causes synovial hyperplasia [32]. Activated FLS converts its network to a pannus-like multi-layer structure in RA lesions. This structure can be identified as a distinct synovium lining hypertrophic lesion [33]. During RA progression, MMP-13 overexpression mediates the destruction of bone and cartilage structures. Multiple studies have reported that MMP-13 production is increased in cytokine-stimulated FLSs [34, 35]. Activated FLSs act as effector cells that induce the activation of inflammatory T cells by expressing IL-7 and IL-15, which possess functional similarities with IL-2 [36, 37]. However, synovial macrophages secrete only IL-15, whereas synovial T cells do not secrete either IL-7 or IL-15. In addition, FLS secrete TNF-α similar to other synovial cells such as lymphocytes, macrophages, and endothelial cells [38]. Therefore, cytokines secreted by synovial immune cells and the distinct cytokine profile of FLS contribute to the RA microenvironment.

Activated macrophages secrete IL-1, IL-6, and TNF-α, and they can be induced by various cell subsets, such as FLSs, monocytes, and helper T cells [39]. The cytokines present in the synovial fluid induce severe inflammation in RA [40, 41]. Initial erosion of RA tissue gives way to synovial hyperplasia [42, 43]. Activated macrophages predominantly accumulate at the borderline of pannus-like multi-layer structure and cartilage along with FLSs [33]. In particular, TNF-α and IL-1 induce FLS activation, producing tissue-destructive MMP [44]. TNF-α inhibition by antibodies leads to a considerable reduction in IL-1, IL-6, IL-8, and GM-CSF. Therefore, TNF-α blockade is an excellent therapeutic strategy to suppress inflammation in RA patients.

CD4+ T cells are considered central cellular mediators involved in RA pathogenesis [45]. Among CD4+ T cells, helper T cells and regulatory T cells are significantly related to RA. Helper T cells, characterized by CD4 expression on the cell surface, promote inflammation by activating osteoclasts, FLSs, chondrocytes, and immune cells via various inflammatory cytokines. For example, helper T cells secrete IL-17 to activate osteoclast, which causes bone destruction [46]. Similarly, IFN-γ contributes to macrophage polarization to inflammatory M1 phenotype [47]. In addition, cell surface signaling by membrane proteins such as CD69 and CD11 induces the secretion of MMP and other effector molecules by FLS and chondrocytes, which in turn promotes inflammatory responses and causes joint erosion [23]. Another subset of CD4+ T cells is regulatory T cells (Tregs), CD4+CD25+FOXP3+ cells [48]. Unlike helper T cells, Tregs are considered effective suppressors of RA [49]. Although many Treg cells are present in the synovial fluid of RA patients, they are functionally deficient. This suggests an immune suppressive role of ‘normal’ Tregs in RA [50]. Thus, synovial Tregs in RA patients often exhibit reduced regulatory capacity, which might be attributed to the severe inflammatory condition of the RA microenvironment [51].

DCs are one of the most professional antigen-presenting cells (APCs) [52]. DCs prime T cells by presenting antigenic peptides on their surface complexed with MHC [53]. Along with MHC class II, co-stimulatory molecules such as CD80 and CD86 facilitate interaction and priming of T cells [54]. Auto-reactive helper T cells primed by DCs contribute to the inflammatory microenvironment by secreting pro-inflammatory cytokines, such as IL-12 and IL-23 [55].

Like other immune cells, B cells are also abnormally increased in RA [56]. During RA initiation, B cells produce autoantigen, TNF-α, and IL-6 [57]. Among them, IL-6 contributes to forming an inflammatory microenvironment in the synovial membrane by promoting monocyte differentiation. In addition, B cells act as APCs in RA to induce helper T cell activation by presenting autoantigens to CD4+ T cells [58]. As the disease progress, B cells infiltrate synovial tissue and induce a sustained and destructive autoimmune response.

The synovial microenvironment facilitates optimal antigen processing and presentation by antigen-presenting cells for T cell activation [59]. Studies have revealed that the interaction of activated T cells with monocytes induces the secretion of various inflammatory cytokines and chemokines [60,61,62]. Hence, the interaction of immune cells and inflammatory factors in the inflamed synovial microenvironment plays a pivotal role in RA pathogenesis and progression.

Exosome

Biogenesis of exosomes

Exosomes are nano-sized extracellular vesicles secreted by most eukaryotic cells. Unlike other small extracellular vesicles such as apoptotic bodies and microvesicles, exosomes are primarily generated via cell membrane fusion of exosome-containing endosomes. Exosomes typically have a hydrodynamic size of 30–150 nm in diameter, which can be characterized via nanoparticle trafficking analysis method. The physicochemical properties of exosomes can be characterized through isolation from conditioned cell culture media using ultracentrifugation, size exclusion chromatography, or immunoaffinity-based isolation kits [63]. Exosomes express CD9 and CD63 as surface markers and encapsulate exosome biogenesis-related factors, such as TSG101, ALIX, and Rab27a.

Various studies have recently studied exosome biogenesis, and the accumulated research results provide evidence about exosome biogenesis [1, 64, 65]. Exosomes are generated via direct inward budding of the phospholipid bilayer within the MVB derived from the endosomes and then secreted out of the cell [64]. Exosome secretion occurs through an endosomal sorting complex responsible for transport (ESCRT) protein-dependent or ESCRT-independent mechanism. ESCRT-independent exosome biogenesis showed that sphingolipid ceramide is crucial [66]. This study has revealed that the inhibition of neutral sphingomyelinases reduces the number of exosomes released. In addition, cells can produce CD63-positive MVEs regardless of the depletion of the four subunits of ESCRT [67]. However, several ESCRT proteins are involved in exosome formation and induce ESCRT-dependent exosome biogenesis. For example, the reticulocyte transferrin receptor reportedly interacts with ESCRT accessory protein Alix during exosome formation [68]. Subsequently, Alix contributes to exosome biogenesis and intraluminal vesicle formation by recruiting ESCRT-III complex protein [69].

Various mechanisms determine the contents of the cargo inside the exosome, and as a representative example, proteins such as tetraspanin participate in protein loading in the exosome. Tetraspanin-rich microdomains contribute to the differentiation of receptors and signaling molecules within the exosome plasma membrane [70]. In addition, tetraspanins such as CD9 and CD81 are known to form microdomains and bind to specific proteins; this function serves as a sorting mechanism during the generation of exosomes [70, 71]. Rab27 is an essential protein involved in exosome secretion after biogenesis. Rab27 silencing reduces exosome secretion without changing exosome cargo protein composition [72]. Conversely, inhibiting the proteasomal degradation of Rab27 tends to increase exosome secretion [73].

Various cargo molecules, including DNAs, RNAs, and lipids, are present inside the exosome [74]. Cargos are considered to play an essential role in intercellular communication. In contrast to the previous assumption that exosomes are just cell debris, recent studies have revealed that they are primarily involved in normal biological processes [75]. The exosomes mediate local communication around the recipient cell, and they circulate along the blood vessel after being secreted and accumulate at the target site, acting as mediators in cell-to-cell interactions [76, 77]. Exosomes delivered to the recipient cell enter the recipient cell through fusion, receptor-mediated uptake, and internalization for cell-to-cell communication. A previous study reported that mRNA could be encapsulated in the exosome of the parent cell and translated after being taken up by the recipient cell [7]. Studies on the development of delivery systems have explored the applicability of exosomes to treat diseases [78]. Previous exosome studies reveal that inhibiting exosome biogenesis induces cell apoptosis [64]. Many studies have revealed that exosomes play a crucial role in the progression of intractable diseases, and efforts to understand the exosomes related to the diseases are continuing.

In RA, exosomes significantly contribute to the inflammatory environment of the joints. Numerous exosomes exist in various biological fluids, including synovial fluid [79]. These exosomes are secreted from multiple cell populations such as FLSs, T cells, and B cells [80,81,82,83]. They trigger inflammation by delivering damage-associated molecular patterns [84]. In addition, the exosome’s cytokines, lipids, and microRNAs aggravate RA’s inflammation [85]. Exosomes also play critical roles in disease progression in various chronic diseases. For example, cancer exosomes deliver oncogenic cargos, which makes tumors a favorable microenvironment. They also overexpress PD-L1 on their surface, thereby resisting immune checkpoint inhibitors. Disease progression can be delayed, and the therapeutic effect of conventional therapies can be improved by targeting disease-favorable exosomes. Therefore, exosomes are potential biomarkers in chronic diseases and have the potential to develop therapeutic approaches [86].

Engineering

Growth in the knowledge of the intrinsic biological functions of exosomes has led to the development of various exosome engineering strategies with biomedical applications. Considering these inherent biological functions of exosomes, bare-state exosomes from various types of cells were utilized to treat intractable diseases, including cancer, RA, multiple sclerosis, and diabetes [87,88,89]. Although exosomes have emerged as a replacement for cell-based therapy owing to their remarkable therapeutic efficacy, the regulation of exosomal compartments has been attempted to maximize their effects. Although exosomes possess homing effect, which can be recruited to parent cells, several engineering strategies have been proposed that disadvantages such as low targetability and poor distribution have still been issued. The engineering strategies can be classified into two categories: modification before and after the isolation of exosomes (Fig. 2).

Fig. 2
figure 2

Exosome engineering via pre-isolation modification and post-isolation modification

Pre-isolation modification

Exosomes can achieve the distinct properties of parent cells by loading the same protein, RNA, and DNA with origin cells as well as by expressing the membrane protein of secreted cells in the exosomal surface [1]. Accordingly, based on exosome biogenesis, the engineering methods for modulating parent cells are developed. The focus of the research includes pre-conditioning methods for altering cell characteristics using various cytokines or stimuli, endogenous modulation for regulating genetic parts, cell membrane modification for exosomal membrane surface engineering, and pre-treatment for cargo modification via the utilization of exosome biogenesis pathway (Fig. 3).

Fig. 3
figure 3

Schematic illustration of pre-isolation modification methods. A Exosomes from target ligand (avidin) labeled and anti-tumor drugs encapsulated donor cells for further labeling with avidin complex [94], as therapeutics for remyelination [95], and migraine [96]; however, tolerogenic Dex are used in the treatment of several autoimmune diseases, including multiple sclerosis [106]. Unlike M1-derived exosomes, M2-derived exosomes facilitate the reprogramming of M1 to M2, reflecting anti-inflammatory properties. Owing to intrinsic properties similar to that of M2, M2-derived exosomes are used as therapeutics for various diseases, such as RA, wound healing [116]. Theranostic nanoparticles and imaging probes such as fluorescent dyes were labeled onto the azide (N3) functional group via copper-free click chemistry utilizing metabolic glycoengineering using mannosamine. Although this strategy was mainly applied for in vivo tracking of several types of cells [117,118,119] and active-targeting of tumors [120], metabolic glycoengineering or bioorthogonal click chemistry was recently used for in situ one-step strategy to label exosomes, considering the exosome biogenesis pathway [93]. The authors noted that the yield of acquisition of dye-labeled exosomes from synthetic metabolite-treated donor cells was higher than that by applying bioorthogonal click chemistry directly to isolated exosomes, and also it preserved the intrinsic properties of exosomes.

In addition to the fluorescent dye conjugation, a study of exosomal surface-modified exosomes with biocompatible polymer via bioorthogonal chemistry based on metabolic glycoengineering was conducted. Park and group reported PEGylated hyaluronic acid (HA)-coated exosomes to target the CD44-overexpressing cells actively and verified the targetability using the RA and PC3 tumor-bearing mouse models, representative models of CD44-overexpression [121]. Briefly, the authors induced the expression of the N3 group in the cellular membrane of donor cells by pre-treatment with N-azidoacetyl-D-mannosamine (Ac4ManNAz), and the parent cells utilized in this research were cancer cells regarding the proof of concept. Subsequently, dibenzocyclooctyne-terminated PEGylated-HA was conjugated with the N3 group using bioorthogonal copper-free click chemistry. Therefore, the PEGylated-HA functionalized exosomes were isolated using ultracentrifugation, and surface editing was verified using biolayer interferometry (BLI) and confocal imaging. Thereafter, macrophage-targeing dextran sulfate (DS)-coated adipose-derived stem cells-derived exosomes for immunomodulation in RA were suggested by the same group [122]. The strategy for exosomal surface editing has been described previously; however, DS was introduced to target the macrophage scavenger receptor class A (SR-A) of activated macrophages.

Owing to the sophisticated physicochemical properties of phospholipid bilayer structures of cellular membranes, various functional moieties linked with lipid molecules can be incorporated into the cell membrane surfaces by hydrophobic insertion via physical interaction. Furthermore, this modification strategy can endow the exosomal membrane with functional moieties introduced into cells. Wang et al. amended the donor cells to obtain the dual ligand-modified and drug-encapsulated exosomes [123]. The authors pre-treated 1,2-distearoyl-sn-glycero-phosphoethanolamine (DSPE)-PEG-Biotin to introduce the biotin on cellular membrane surfaces, utilizing the physical interaction between DSPE and lipid bilayer of the cellular membrane. Subsequently, the cells were pre-treated with anti-tumor drugs, assuming that the drug will be packaged in exosomes following uptake into the cytosol. Although the anti-cancer drug paclitaxel (PTX) induced partial apoptosis, the authors noted that PTX was also well-encapsulated in exosomes. To edit the exosomes with dual ligands and enhance the targetability, Wang et al. subsequently treated donor cells with avidin, resulting in a biotin-avidin reaction, and successfully isolated these exosomes utilizing the microfluidic chips created by them. Wang et al. also applied the similar strategy to obtain biotinylated exosomes for further engineering with avidin modified with molecular beacon and poly dopamine (PDA) covered Fe3O4 nanoparticles [65], exosomes from various cells have been widely used to modify the surface of theranostic nanoparticles via diverse strategies such as sonication, extrusion, and incubation. The hydroxychloroquine (HCQ)-loaded zinc sulfide (ZnS) nanoparticles, which act as a photosensitizer, were coated with glioblastoma cells-derived exosomes through thin film hydration followed by extrusion and subsequently modified with iRGD peptide which provided the pH- and redox- responsiveness [142]. This study noted that, owing to the homing ability of exosomes, HCQ@ZnS@eRGD was permeable across the blood–brain barrier and showed enhanced targetability to glioblastoma cells in vivo. In addition to the inorganic nanoparticles, exosome-coated poly(lactic-co-glycolic acid) (PLGA) nanoparticles loaded with 1,1-dioctadecyl-3,3,3,3-tetramethylindotricarbocyaine iodide (DiR) were obtained using microfluidic sonication and exhibited a prolonged blood circulation, suggesting the effects of exosomal surfaces [130].

Even though the endogenous modulation of origin cells facilitated the engineering of exosomal proteins and genes, they can be loaded into isolated exosomes by controlling the loading efficiency from the protein or gene delivery view. Wan and Zhong et al. reported a study related to tissue-specific gene therapy of liver fibrosis using Cas9 ribonucleoprotein (RNP) delivery exosomes [133]. The exosomes were isolated from LX-2 cells through ultracentrifugation, and Cas9 RNP complexes were loaded into exosomes via electroporation with 20% entrapment efficiency. Meanwhile, the human umbilical cord MSC-derived exosomes modified with polypeptide (CAQK peptide) and loaded with the CRISPR/Cas9 components were applied for the immunotherapy of spinal cord injury [143]. The CAQK peptide was chemically conjugated on exosomal surfaces via EDC/NHS chemistry, and CRISPR/Cas9 plasmid was loaded using electroporation methods. In addition to the electroporation strategy, McAndrews et al. used the commercially available transfection reagent, Exo-Fect Exosome Transfection Kit, to encapsulate the CRISPR/Cas9 plasmid DNA into MSC-derived exosomes for oncogenic Kras targeted therapy [152]. Even though several stimuli-sensitive drug delivery systems, which facilitate active and passive targeting, have been evaluated, their clinical application remains to be validated in various experimental and preclinical systems. Lee and Sul evaluated reactive oxygen species-responsive tolerogenic exosomes for RA treatment since ROS is overproduced in the inflamed joints [153].

Ma et al. explored chemical reaction-based novel surface-modification strategy, which can be used in situ to modify the ascorbic acid (AA), an antioxidant, on the exosomal surface to overcome the low loading efficacy of common loading methods [132]. In this study, mesenchymal stem cell-derived exosomes were isolated using ultracentrifugation and incubated with DSPE-PEG-SH followed by incubation with HAuCl4 solution to introduce Au3+ on exosomal surfaces. Because AA can act as a reducing and protective agent for AuNPs growth, AA solutions were incubated with Au3+-introduced exosomes. Subsequently, exosomes with their surface-modified with AA-protected AuNPs (mExo@AA) were obtained and used as eye drop formulation to treat dry eye disease. Its therapeutic effects were achieved by eliciting damage repair, ROS scavenging, and macrophage polarization into the M2 type.

For gene/chemo/photothermal therapy and molecular imaging, a combination system composed of polydopamine (PDA) coated magnetic Fe3O4 nanoparticles and DOX-loaded exosomes was prepared using various exosome engineering strategies, including pre-isolation modification and affinity-based reaction [7) [170].

Fig. 7
figure 7

A Schematic illustration of post-isolation modification method for treatment of RA. A Preparation strategy of anti-ROS-CII antibodies loaded polymorphonuclear leukocytes (PMN)-derived extracellular vesicles (EVs). [170]. Reproduced with permission, Copyright 2020, Frontiers. B Schematic illustration of Evs-EGCG and cartilage regeneration. [171] Copyright 2021, Elsevier. C Schematic depicting of BBR-PEVs and RA microenvironment reprogramming [172]. Copyright 2021, Wiley

The polymorphonuclear leukocyte (PMNs)-derived exosomes were isolated by differential centrifugation. Then the mixture of exosomes, dried phospholipids consisting of 1,2-dioleoyl-sn-glycero-3-phospho-L-serine, and antibodies was sonicated for 5 min with three amplitude settings. Several kinds of antibodies, including CII post-translationally modified by oxidants-specific antibody (anti-ROS-CII), viral IL-10 fused anti-ROS-CII (anti-ROS-CII-vIL-10), anti-mouse TNFα antibody (anti-mTNF), and both anti-ROS-CII-vIL-10/anti-mTNF (anti-ROS-CII&mTNF/vIL-10), were enriched in exosomes. The data showed that the anti-ROS-CII incorporated exosomes (EV/anti-ROS-CII) bound explicitly to articular cartilage and were highly retained in the arthritic joints without off-target effects compared to non-modified exosomes. Moreover, the binding ability to both ROS-CII and mTNF of anti-ROS-CII&mTNF/vIL-10 incorporated exosomes revealed that the sonication led to incorporation of more than one antibody. In AIA, the EV/anti-ROS-CII&mTNF/vIL-10, which accumulated in the inflamed knee joints after intravenous injection, attenuated the synovial inflammation by downregulating the genes involved in pro-inflammatory response and cartilage degradation, and upregulating the genes related to anti-inflammatory response and cartilage regeneration. Furthermore, histological analysis revealed that EV/anti-ROS-CII bound to the intermediated and deep cartilage showed chondroprotection effects correlating to the genomic response.

The sonication method allows the small-sized therapeutic molecules to be loaded into exosomes. Song et al. encapsulated the epigallocatechin gallate (EGCG), a small antioxidant molecule, into macrophage-derived exosomes to improve its utility against chondrocytes in physiological conditions (Fig. 7B) [171]. The macrophage-derived exosomes were isolated from RAW cells and subsequently EGCG-loaded exosomes (EVs-EGCG) were obtained. Encapsulation and sustained-release behavior of EGCG were verified using UV–Vis absorbance. The results showed that EVs-EGCG internalized into chondrocytes, attenuated hypoxia-inducible factor 1-alpha (HIF-1α)-mediated inhibition of apoptosis, and augmented type II collagen expression, thereby alleviating RA symptoms in a rat model of CIA.

The encapsulation of the small organic compounds can be made feasible using the incubation method [172, 173]. Ma et al. developed berberine, which is known to suppress the activated macrophages and DCs, -coordinated platelet-derived exosomes (BBR-PEVs) (Fig. 7C) [172]. The PEVs were isolated by centrifugation using ultrafiltration tubes (3 kD), and the incubation of DMSO-dissolved BBR was followed. The high-performance liquid chromatography (HPLC) data showed that the loading efficiency was 6.76% and that 91.9% of BBR was sustained-released within 48 h. The BBR-PEVs selectively accumulated into inflamed joints and reshaped RA microenvironments by reducing the proportion of T cells, macrophages, and DCs, downregulating the activated phenotypes of those cells, and regulating the cytokines to a normal state; these effects collectively resulted in relief of RA symptoms. He et al. loaded curcumin (Curc) into MSC-exosomes (Curc-Exos) using the incubation method [173]. The MSC-exosomes were isolated using Exocib kit, then and mixed with Curc solution. The incubated mixtures were centrifuged by sucrose gradient, and 45 and 60% layers were selected as Curc-Exos with 18 ± 5% loading capacity. The Curc-Exos exhibited acidic pH-triggered faster release, followed by apoptosis of HIG-82 cells, a synovial membrane cell line, and downregulated pro-inflammatory mRNAs, implying their anti-inflammatory effects on FLSs.

Yan et al. utilized macrophage-derived exosomes as targeted drug delivery carriers [174]. They encapsulated dexamethasone sodium phosphate (Dex) into exosomes and modified the exosomal surface with folic acid-polyethylene glycol-cholesterol (FA-PEG-Chol) for active targeting of activated macrophages. The exosomes were isolated by differential centrifugation, and Dex was loaded (wt ratio of Exo: Dex = 1:3) using electroporation in the presence of 80 nM of trehalose to prevent aggregation. The electroporation was performed in a double poring pulse with 200 V for 5 ms and a transfer pulse of five pulses with 20 V for 50 ms at room temperature, followed by incubation in 37 ℃ PBS for 1 h to restore the exosomal membrane. Furthermore, the pre-synthesized FA-PEG-Chol was modified by post-insertion since because the cholesterol could be inserted into the lipid bilayer of the exosome. Exo/Dex and FA-PEG-Chol were mixed (wt ratio 1:5) and incubated at 37 ℃ for 2 h. The HPLC data showed the drug loading efficiency of Exo/Dex and FPC-Exo/Dex was 22.38% and 18.81%, respectively. The release of Dex was triggered by acidic pH, and the cellular uptake into activated macrophages was enhanced after surface modification, suggesting FA-mediated active targeting. In the CIA model, augmented accumulation of the FPC-Exo/Dex in inflamed joints suppressed inflammatory cell infiltration, pro-inflammatory cytokine release, as well as increase in anti-inflammatory cytokine level, which resulted in the alleviation of RA symptoms and mitigated bone erosion. The authors noted that FPC-Exo/Dex hardly manifested hepatotoxicity, suggesting their biocompatibility.

Li et al. decorated the surface of M2 macrophage-derived exosomes (EXs) with a polyarginine peptide (R9) [178]. The developed exosome was considered for potential treatment of RA. The engineered exosome exhibited higher cellular uptake and showed an excellent targeting effect. For example, the engineered exosomes showed 176% accumulation at the site of inflammation. This was due to higher Cur@EXs-R9 capture and uptake by inflammatory cells. In addition, the M2 macrophage-derived exosome effectively repolarized the macrophage at the inflamed site, and the authors proposed that Cur@EXs-R9 is a potential treatment option for RA.

Others

Engineering strategies utilizing both pre-isolation and post-isolation modification have been introduced to confer multifunctionality on exosomes and maximize the therapeutic effects in RA. The parent cells are first engineered to exert immunosuppressive effects, and exosomes are isolated from these engineered cells and subjected to either exosomal surface modification or cargo loading.

Lee et al. developed a reactive oxygen species (ROS), one of the RAM stimuli,-responsive tolerogenic dendritic cell-derived exosomes (TKDex) (Fig. 8A) [153].

Fig. 8
figure 8

A Schematic illustration of TKDex and cytokine regulation. B Therapeutic regime of RA treatment by TKDex. C The size of spleen after treatment. D, E Represent cytokine level in blood. F Quantification of regulatory T cell after treatment of TKDex [153]. Reproduced with permission, Copyright 2021, Elsevier. G Schematic illustration of neutrophil-derived exosomes (NEs-Exo). H Illustration for therapeutic regime [175]. Reproduced with permission, Copyright 2021, Elsevier

They hypothesized that PEG containing ROS-responsive moieties on the exosomal surface protected the exosomes in the bloodstream, while facilitating the exposure of exosomes in the inflamed joint, leading to their uptake into DCs, macrophages, or T cells in the RAM to induce the immunomodulation. The authors first induced immunosuppressive tolerogenic dendritic cells (TolDCs) via preconditioning DCs 2.4 cells with 40 ng/mL IL-10 for 24 h. Subsequently, the dendritic cell-derived exosomes (TolDex) were isolated utilizing TFF systems equipped with an ultrafiltration filter capsule (MWCO = 300 kDa). Meanwhile, the ROS-sensitive linkage (DSPE-TK-mPEG) for exosomal surface modification was synthesized by conjugation of thioketal (TK) linker, methoxy PEG, and 1,2-distearoyl-sn-glycero-phosphoethanolamine (DSPE). Following TolDex isolation, the DSPE-TK-mPEG was modified on the surface of TolDex via the hydrophobic insertion method. The hydrophobic insertion was performed by incubating DMSO-dissolved DSPE-TK-mPEG with TolDex. TKDex was internalized into mature DCs (mDCs) activated with LPS, whereas TolDex modified with PEG without TK (PEGDex) showed insignificant cellular uptake, suggesting its stimuli-sensitive behaviors. The in vitro data exhibited that pro-inflammatory factors were downregulated in mDCs and T cell proliferation was suppressed. In CIA, the negligible difference in the joint-to-liver ratio of TKDex and PEGDex implied that TK moiety did not contribute to former’s degree of accumulation in the lesions. Decrease in IL-6 level, increased TGF-β level, and increased proportion of Tregs suggested that TKDex could modulate T cells as well as DCs in inflamed joints. It is noteworthy that the therapeutic activity of Dex are induced in only ROS-overproduced RAM.

Zhang et al. modified ultrasmall Prussian blue nanoparticles (uPB), nanoenzyme, on the surface of neutrophil-derived exosomes to regulate the RAM (Fig. 8G) [175]. Since LFA-1 on the activated neutrophil membrane can bind with activated FLSs, chondrocytes, and macrophages via intracellular activation molecule 1 (ICAM-1), LPS was injected into mice, followed by the isolation of neutrophils. Subsequently, the neutrophil-derived exosomes (NEs-Exo), possessing the LFA-1, were isolated by differential centrifugation. In addition, an antioxidative enzyme mimic, Prussian blue was fabricated into ultrasmall nanoparticles and further anchored to the exosomal surface to relieve the ROS-abundant conditions in inflamed joints. DBCO-sulfo-NHS was conjugated on the exosomal surface in advance, and N3-uPB was linked by copper-free click chemistry. The uPB-Exo showed enhanced uptake into activated FLSs, chondrocytes, and macrophages, as well as ROS scavenging in vitro, indicating the role of LFA-1 in active targeting and activity of nanozymes, respectively. Moreover, uPB-Exo revealed protective effects against oxidative stress and cytokines secreted from apoptotic cells by mimicking SOD-2 and inhibiting apoptosis of activated FLSs, chondrocytes, and macrophages. The authors also validated that apoptosis was related PI3K/AKT signaling pathway. In CIA, both NEs-Exo and uPB-Exo were highly accumulated into paws and penetrated deep into the cartilage. The significant differences in the reduction of TNF-α, IL-1β, and Th1, as well as the increase in Treg count suggested the synergistic effects of uPB with NEs-Exo.

M2 macrophage-derived exosomes loaded with a plasmid DNA encoding IL-10 (IL-10 pDNA) and synthetic glucocorticoid were developed for macrophage re-polarization in RA (Fig. 9A) [176].

Fig. 9
figure 9

A Schematic illustration of M2-macrophage derived exosomes (M2-Exo/pDNA). B Therapeutic regime of M2-Exo/pDNA and macrophage polarization. C Photographs of paw thickness [176]. Reproduced with permission, Copyright 2022, Elsevier. D Schematic illustration of anti-inflammatory exosomes (AI-Exo). E Photoacoustic images of joint. F skeletal structure of paw joint [177]. Reproduced with permission, Copyright 2022, Royal Society of Chemistry

The authors aimed to inhibit pro-inflammatory factors and promote M1 polarization to M2 macrophages. M2 macrophages (RAW 264.7 cells) were induced with 100 ng/mL IL-4. The pDNA-lipid complexes, comprising earlier synthesized IL-10 pDNA and transfection agent, were incubated with M2 macrophages at 37 ℃ for 3 days. Subsequently, the IL-10 pDNA encapsulated M2-derived exosomes (M2 Exo/pDNA) were isolated using differential centrifugation with ultrafiltration. In brief, supernatant centrifuged at 300 g, 2000 g, and 10,000 g was concentrated using ultrafiltration through MWCO 100 kDa filter, followed by ultracentrifugation at 120,000 g. Lastly, to obtain the betamethasone sodium phosphate (BSP) loaded M2 Exo/pDNA (M2 Exo/pDNA/BSP), the M2 Exo/pDNA and BSP were mixed (wt ratio 1:2) with 80 nM trehalose and electroporated with poring pulse at 200 V, pulsed twice for 5 ms and transfer pulse for five pulses at 20 V for 50 ms at room temperature. M2 Exo/pDNA/BSP showed sustained release of BSP (approximately 98.93% within 20 h). The enhanced uptake of M2 Exo into activated macrophages was attributed to the overexpression of LFA-1 or VLA-4 on M2 Exo, which can bind to ICAM-1 or VCAM-1 on activated macrophages, suggesting the inherent targetability of M2 Exo to M1 macrophages. As expected, the M2 Exo/pDNA/BSP exhibited maximal effects on M1 to M2 polarization compared to free BSP, pDNA/BSP, M2 Exo, M2 Exo/pDNA, and M2 Exo/BSP. In CIA, highly accumulated M2 Exo/pDNA/BSP in inflamed joints regulated pro-and anti-inflammatory cytokines and protected the cartilage, without inducing hepatotoxicity; Thus, this system led to effective co-delivery of genes and drugs without side effects allowing synergy.

Tang et al. utilized an external energy source to improve the therapeutic effects in RA (Fig. 9D) [177]. The authors prepared anti-inflammatory exosomes (AI-Exo) by loading IL-10 into macrophage (M0 type)-derived exosomes and irradiated using non-invasive ultrasound (US) during therapy to enhance blood vessel permeability to accelerate drug uptake into tissues and cells. The AI-Exo was prepared as follows; 1) Exosomes were isolated from RAW 264.7 cells using ExoQuick-TC®ULTRA EV Isolation kit, and 2) IL-10 and exosomes were electroporated (wt ratio 2:15, 600 V for 90 μs, 5 times pulses with 1 s interval). The in vitro data showed that US irradiation enhanced the cellular uptake of AI-Exo into activated macrophages. Similarly, the US augmented the accumulation of AI-Exo into inflamed joints in CIA, alleviating RA symptoms. The mechanistic analysis verified that the therapeutic effect of AI-Exo+US group was attributed to the reduction in IL-6 and TNF-α level and decrease CD86/CD206 ratio in macrophages. The authors noted that IL-10 plays a role in promoting polarization of M1 macrophages to M2 macrophages, and US can enhance the targetability by promoting the activated macrophage phagocytosis of AI-Exo.

Engineering other subpopulations of extracellular vesicles, including matrix-bound nanovesicles (MBV) and MV, was also reported. The MBVs are similar to MV; however, they are anchored to the extracellular matrix (ECM) and identified as functional components of ECM [206]. Crum et al. verified that the MBV isolated from urinary bladder ECM by enzymatic digestion attenuated the inflammation in pristane-induced arthritis via synovial and splenic macrophage polarization from M1 to M2 phenotype [207]. Meanwhile, MV was used for develo** biomimetic nanoparticles. The zeolitic imidazolate framework-8 (ZIF-8), one of the common metal–organic frameworks (MOF), encapsulated the MTX and was camouflaged with macrophage-derived MV and further modified with DSPE-PEG-FA by hydrophobic insertion to enhance the biocompatibility and targetability [208]. In the same group, the Dex-encapsulated and macrophage-derived MV-coated ZIF-8 was also developed for targeted delivery to arthritic joints [209]. The macrophage-derived MV was employed to coat the poly(lactic-co-glycolic acid) (PLGA) nanoparticles, eliciting enhanced RA targeting [210].

Challenges and perspectives

Considering the biological effects of the exosome in cell-to-cell communication, using an exosome as an RA therapeutics is a potential approach for treating RA. In addition to DC and T cells, exosomes derived from various cells, such as NK cells, can enhance each cell's function and modulate the disease microenvironment. Although the etiology of RA has not been precisely elucidated, therapeutic strategies that convert autoimmunity into a state of immune tolerance have been proposed. Because systemic immunosuppression induces undesirable immunodeficiency, the importance of inducing antigen-specific tolerance is emerging to treat autoimmune diseases effectively [211]. Recently, OA or chronic multiple sclerosis was reportedly treated in an antigen-specific manner by inducing tolerogenic DC using nanovaccines utilizing antigen-specificity [212, 213]. Although Rheumavax, an autologous DC modified by exposure to citrullinated peptides, was introduced several years ago and was safe and effective in clinical trials with a single administration, issues related to cell-based therapy remain a challenge. In this perspective, exosomes are non-immunogenic and less toxic than organic/inorganic therapeutic nanoparticles; therefore, they can be used to substitute cell-based therapy. Moreover, DC-derived exosomes can be developed into tolerogenic nanovaccines through various engineering methods. Development of antigen-specific nanovaccines using DC exosomes is a promising strategy for develo** novel RA therapeutics, thereby preventing non-specific immune system downregulation.

Challenges such as low in vivo stability and unspecific uptake of bare exosomes have been identified, the need to utilize engineered exosomes has emerged. Engineered exosomes equipped with targeting moieties or PEGylation for long-term circulation have emerged as excellent alternatives as they overcome low in vivo biostability of bare exosomes.

However, considering that mass production of engineered exosomes is challenging, the issues related to their producibility and cost need to be addressed. Some strategies, such as three-dimensional cultivation using hallow-fiber bioreactor [214], and focused US irradiation [215], can improve the production efficiency; however, other challenges remain. In addition, the exosomes have a complex structure compared to conventional drugs. Therefore, difficulties in their characterization required for drug approval hinders further application of engineered exosomes in clinical settings. For example, problems of animal serum retention in the cell culture medium used for exosome production can be a huddle for the clinical application of exosomes, as well as the limitation of the human application of stem cells.

Exosome-based therapies, encompassing bare and engineered exosomes, offer promising prospects for RA treatment. Even though both strategies have been studied with various types of exosomes, it is crucial to weigh the potential advantages against the limitation of each approach. Bare exosomes possess inherent immunomodulatory properties depending on the origin cell types, and autologous or allogeneic exosomes are biocompatible and have lower immunogenicity, minimizing the risk of immune responses or adverse reactions. Nevertheless, the variability of naturally controlled cargoes might lead to inconsistent therapeutic efficacy. In addition, since the mode of action of bare exosomes' immunomodulatory effects and anti-inflammatory effects are complex and unpredictable, the diverse exosomal content and interaction with various types of cells might lead to undesirable biological effects.

On the other hand, various therapeutic cargo like anti-inflammatory drugs, disease-modifying molecules, and immunomodulatory molecules can be loaded within engineered exosomes, and the loading efficiency can be precisely controlled. The targeting moieties on exosomal surfaces enable exosome delivery at the intended site of action, resulting in enhanced therapeutic efficacy and lower side effects. Furthermore, the engineered exosomes can be personalized by customizing to the needs of individual patients based on the heterogeneity of specific diseases. Since the engineered exosomes have more chance to recognize immune systems than bare exosomes, the engineering strategy should augment the biocompatibility for lowing the immunogenicity.

In addition to using the exosome directly, in situ exosome biogenesis is an emerging strategy. In situ exosome biogenesis modulation is a strategy without the production process of the exosome, which is expensive. For example, the strategy to inhibit the secretion of exosomes derived from the cell population that overproduces exosomes in the disease microenvironment showed promising microenvironment modulation capability in the pre-clinical state. Conversely, strategies to induce the secretion of exosomes from suppressed cell populations within the disease microenvironment can be considered. Because exosomes function as cell avatars, increased secretion of exosomes can be utilized to restore, reinvigorate, or suppress biological functions within the microenvironment of the suppressor cell population.

Conclusion

In conclusion, using exosomes as a potential therapeutic approach for RA is promising due to their non-immunogenicity and less toxicity than other therapeutic nanoparticles. However, issues related to the production and characterization of engineered exosomes need to be addressed. Therefore, further research is required to address these challenges and fully utilize the potential of exosomes in treating RA and other autoimmune diseases.