Extracellular vesicles (EVs) are vehicles that transfer signal transduction molecules in intercellular communications and can carry multiple functional molecules, including nucleic acids, proteins, and lipids, to exert their functions. Meanwhile, natural membrane endows EVs with multiple advantages, including lower immunogenicity (1, 2), homologous targeting capacity (3), a long circulation time, and access to biological barriers (4, 5), which are difficult to achieve with artificial materials. Inspired by these brilliant properties, EVs are regarded as a new generation of delivery system. The fabrication of engineered EVs loaded with multiple bioactive molecules or modified with targeting ligands to generate therapeutic effects at the desired site has become an attractive strategy for disease therapies (5, 6).
For fabrication of engineered EVs, multiple methods for therapeutic molecule loading were used, including mainly electroporation (5), precomplexation, transfection (7, 8), saponin permeabilization (4), and sonication (4, 9). However, the existing approaches still suffer from considerable limitations in the following aspects: (i) The above approaches are incapable of removing the residual components unrelated to the therapeutic aims, which may present latent threats (10). (ii) EVs initiate the downstream signaling only after the bioactive molecules in EVs are released in recipient cells. EVs internalized by recipient cells can be entrapped by the lysosomal system, which degrades EV contents and markedly compromises their therapeutic efficiency (11). (iii) Surface biochemical conjugation to introduce targeting properties to EVs may destroy their structures and functions (12). Therefore, the use of appropriate original EVs with natural biological functions and targeting properties, as well as efficient fabrication strategies, are required to address these issues.
The natural biological functions and targeting properties of EVs are mainly dependent on their types (such as exosomes and microvesicles) and their parental cells (13, 14). Apoptotic bodies (ABs) are EVs secreted by cells undergoing apoptosis and are generated by outward blebbing of the plasma membrane (15, 16). There is corroborative evidence that engulfment of apoptotic cells can regulate the immune response to activate the anti-inflammatory pathways (17, 18). A similar consequence can also be observed during the engulfment of the apoptotic cell/vesicle membrane, which may be linked closely to the presence of phosphatidylserine at the surface (18). Moreover, because of the properties of the membrane, ABs released by some kinds of cells, such as leukocytes, may have the ability to target inflammatory regions (19). Therefore, the natural membrane of ABs derived from T cells confers ABs the ability to target inflammatory regions and to modulate inflammatory processes, indicating that ABs could be used as optimal EVs for inflammation modulation.
In this work, we constructed chimeric apoptotic bodies (cABs) functionalized with a natural membrane and a modular mesoporous silica nanoparticle (MSN) delivery system for inflammatory regulation. To remove the residual components of ABs and maintain their targeting and modulation properties, their membranes were collected and subsequently fused with MSNs. Before fusion, MSNs were preloaded with anti-inflammatory cargos [microRNA-21 (miR-21) or curcumin (Cur)] and further modified with various stimuli-responsive molecules to avoid cargo leakage before arriving at the designated environment, which was expected to perform accurate cargo release in target cells (Fig. 1). The resulting engineered EVs inherited the advantages of both natural AB membrane and manufactured nanomaterials, which could actively target macrophages at inflammation sites to regulate the macrophage phenotype. Both the in vitro and in vivo results showed that our engineered EVs are able to modulate cutaneous inflammation, promote regeneration, and ameliorate the inflammatory bowel diseases, demonstrating that this method may be an efficient strategy to engineer EVs in a modularized way for various biomedical applications.
Fabrication and characterization of cABs
Excessive inflammation gives rise to secondary or progressive damage (20). In inflammation progression and resolution, macrophages represent the predominant scavengers and effector cells (21), indicating that functional modulation of macrophages is an efficient therapeutic strategy. Given that macrophages engulf apoptotic cells/vesicles due to the presence of phosphatidylserine on the surface, we hypothesized that fusion of the AB membranes of activated T cells to MSNs would allow the potential inflammatory targeting ability of T cells and the inflammatory regulation properties of the AB membrane to enhance the modular therapeutic molecule delivery system.
We first investigated the biological properties of the AB membrane. After activation, the activated T cells were larger than the unactivated T cells (fig. S1A) and showed higher CD25 expression levels (fig. S1B), as assessed by flow cytometry. We then isolated ABs derived from activated T cells with differential centrifugation after apoptosis induction as previously described (16). Scanning electron microscopy (SEM) showed that the size of the ABs was approximately 1 μm (Fig. 2A), which was consistent with the dynamic light scattering (DLS) data (Fig. 2B). We then detected the expression of Annexin V and C1q in ABs by using a previously described fluorescence staining method (16) (Fig. 2C) and flow cytometry (fig. S1C). ABs had the same membrane marker, CD3, as their source cells and high level of cleaved caspase-3, which indicated the successful apoptosis induction (Fig. 2D). To acquire the intact AB membrane, membrane derivation was achieved by emptying ABs of their intracellular contents by a combination of hypotonic lysis, sonication, and differential centrifugation (22, 23). Nuclear contents were hardly found in PKH67-labeled AB ghosts, indicating the thorough removal of contents from ABs (fig. S1D). Here, we finished the construction of the shells of the cABs, which were also called AB ghosts (Fig. 2E, b).
Concurrently, MSNs, as drug delivery carriers, were prepared by the classical hexadecyl trimethyl ammonium bromide (CTAB)–templated, base-catalyzed sol-gel method (24). According to previous analogous studies (22, 23), to ensure complete membrane coating onto the MSNs, we chose 1:2 as the nanoparticle-to-membrane ratio in the subsequent application. To evaluate the chimeric efficiency, we prelabeled the AB ghosts with PKH67 before fusion, and then the residual fluorescence intensity in the supernatants was examined after centrifugation. The chimeric efficiency was calculated to be nearly 94.91% based on the negligible fluorescence in the supernatants (fig. S1E).
Transmission electron microscopy (TEM) images directly showed the designed core-shell structure of the cABs (Fig. 2E, d), the chemical composition of which was further verified by TEM–energy-dispersive x-ray spectra (EDS; fig. S1F). As shown in fig. S1F, obvious Si and W signals were observed in the core and shell, respectively, after negative staining with phosphotungstic acid (a typical reagent used to stain natural membranes), indicating the successful fusion of MSNs and AB ghosts. Further information about the core-shell structure formation was provided by DLS (Fig. 2F) and zeta potential determination (Fig. 2G). The size of the MSNs increased from 220 nm to approximately 260 nm after coating with the AB membrane. Moreover, the zeta potential of MSNs changed from −25.3 ± 0.4 mV to −17.7 ± 0.4 mV due to AB membrane camouflage, which approached the zeta potential value of pure AB membrane (−16.6 ± 0.6 mV). The completeness of the resulting cABs was also investigated by confocal fluorescence microscopy, and the red MSNs and green membranes exhibited a high degree of colocalization, indicating the complete fusion of the MSNs and AB ghosts (Fig. 2H). According to a previous study (23), the function of EVs mainly relies on the membrane proteins. Therefore, membrane functionality of cABs was evaluated by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) with Coomassie blue staining to analyze the overall protein profiles of the AB ghosts, MSNs, and cABs. The results indicated that the protein composition of cABs was quite similar to that of AB ghosts (Fig. 2I). Major membrane proteins, including CD3, CD8A, integrin-β3, CD44, and CD11b, also showed similar expression levels between AB ghosts and cABs (Fig. 2I). These findings indicated that the membrane composition was retained during cABs synthesis and demonstrated the transfer of the natural membrane to MSNs.
To determine the short-term biological stability of cABs, MSNs and cABs were stored in either water or 1× phosphate-buffered saline (PBS) (pH 7.4), and their sizes were measured over time. The cABs, unlike the MSNs, exhibited a uniform size for 7 days in both solutions. In contrast, the bare MSN cores, which were only stable in water, immediately aggregated in PBS to form 1- to 2-μm particles (Fig. 2J). Regarding the long-term storage stability, cABs showed nearly identical sizes both before freeze drying and after resuspension (Fig. 2J). In addition, cABs displayed little change in absorbance after incubation with 100% serum, while bare cores exhibited an evident increase (Fig. 2K). To a large extent, the therapeutic effect of EVs relies on the in vivo lifetime. We detected cABs retention in the circulation after several time points. These data showed that cABs displayed a relatively higher signal than MSNs at every time point, even though the signal of cABs could be detected 8 hours after injection (Fig. 2L). Collectively, these results provide strong evidence for the successful construction of cABs, which maintained the original morphology of natural EVs and exhibited a similar protein profile as that of natural ABs and excellent biological stability, demonstrating that cABs inherited the structure and function of biomembranes.
The inflammatory targeting capacity of cABs in vitro
According to previous studies, the inflamed endothelium highly expressed P-selectin and intercellular adhesion molecule–1 (ICAM-1), which are responsible for activated leukocytes recruitment (25). The activated leukocytes in circulation could migrate to the inflammatory region through CD44 or Mac-1 recognition with the corresponding ligands in inflamed endothelium. Hence, we speculated that cABs might inherit the target-homing abilities of T cells to the inflammatory region through CD44 or Mac-1 recognition. Undoubtedly, the similar CD44 and Mac-1 protein levels were observed between the T cells and ABs (Fig. 3A). Stimulation by lipopolysaccharide (LPS) enhanced the expression of CD44 and Mac-1 on the surface of T cells and secreted ABs, which further promoted migration to the inflammatory region (fig. S2A). To investigate the targeting ability to inflammation sites, we treated human umbilical vein endothelial cells (HUVECs) with tumor necrosis factor–α (TNF-α) to simulate the inflamed vessels in vivo as previously described (26), and untreated HUVECs served as a control. Then, rhodamine B (RhB)–labeled MSNs and cABs were added to the culture system of HUVECs. The results showed that the fluorescence intensity of cABs was higher than that of MSNs in activated HUVECs, suggesting the specific binding of cABs to inflamed HUVECs (Fig. 3B). Moreover, the binding of cABs to activated HUVECs was evidently higher than that in untreated HUVECs, which was ascribed to the increased expression of corresponding ligands of CD44 and Mac-1 in HUVECs, as demonstrated by the increasing green fluorescence intensity (Fig. 3B). These results illustrated that cABs inherited the inflammation tropism of activated T cells. To further verify the specific binding mechanism of cABs, fibroblast-derived–chimeric apoptotic bodies (FB-cABs) were applied as a control, since they had similar structures and properties but minimal adhesion proteins, such as CD44 and Mac-1 (fig. S2B). The fluorescence images revealed that the binding of cABs to activated HUVECs was significantly higher than that of FB-cABs and was similar to that of ABs (Fig. 3B and fig. S2C), further revealing that the binding of cABs to the inflamed endothelium was attributed to the specific interactions between CD44 and Mac-1 on the T-AB membrane and the P-selectin and ICAM-1 overexpressed on activated HUVECs. These results further demonstrated the ability of cABs to target inflammatory sites conferred by their source cell membrane.
Specific engulfment of cABs by macrophages in vitro
Macrophages, the specialized phagocytes in nearly all tissues throughout the body, are responsible for debris clearance and cytokine release (21). It is well accepted that macrophages internalize apoptotic cells specifically and promptly (17, 18). First, we characterized the F4/80 and CD11b double-positive mouse bone marrow–derived macrophages (BMDMs) with flow cytometry (fig. S3A) and immunofluorescence staining (fig. S3B). To investigate the cellular uptake of cABs, we incubated macrophages with PKH26-AB ghosts, RhB-MSNs, and RhB-cABs, and PBS served as control. There was a significant red fluorescence signal in the AB ghosts and cABs groups compared to that in the control and MSN groups (Fig. 3C), implying the specific uptake of cABs by macrophages. To verify the reliable internalization of cABs by macrophages, confocal microscopy measurements varying positions on the Z axis was carried out (fig. S3C). Then, we examined the time- and concentration-dependent cell uptake of cABs by macrophages (Fig. 3D and fig. S3, D to G). As expected, the RhB fluorescence intensity increased with the incubation time and concentration of cABs (fig. S3, D and F). The findings were also emphasized by the flow cytometry results (fig. S3, E and G). To further verify the specific uptake of cABs by macrophages, fibroblasts (FBs), bone marrow–derived mesenchymal stem cells (BMMSCs), and T cells were chosen as the control recipient cells. The results showed that macrophages treated with RhB-labeled cABs showed markedly higher uptake than FBs, BMMSCs, and T cells under the same conditions (Fig. 3E and fig. S3H). Likewise, the flow cytometry analysis demonstrated that there was little detectable fluorescence signal in control recipient cells (fig. S3I). We further examined the uptake of cABs in a coculture system including macrophages (not fluorescence labeled) and FBs (PKH67, green). The confocal microscopy analysis indicated that there was nearly no uptake of cABs (red) by PKH67-labeled FBs (green), which emphasized the specific uptake of cABs by macrophages (Fig. 3F). Therefore, these data demonstrated that cABs inherited the natural properties of ABs, which specifically targeted macrophages.
The inflammatory regulation ability of cABs in vitro
Previous studies have elucidated that engulfment of apoptotic cells promotes macrophage transformation toward the M2 phenotype (18, 27). Here, to explore the potential effect of apoptotic membrane on macrophage polarization in vitro, BMDMs were treated with LPS to induce inflammation, followed by the addition of AB ghosts, cABs, MSNs, and PBS to the culture system, with unstimulated macrophages as a control. The phenotype of macrophages visualized by confocal microscopy showed that LPS increased the M1-like subpopulation [inducible nitric oxide synthase–positive (iNOS+)], which was blunted by treatment with AB ghosts or cABs (Fig. 3G). Meanwhile, the CD206 expression level (M2 marker) was enhanced in the AB ghosts and cABs groups and was unchanged in the MSN group (Fig. 3G). Administration of AB ghosts or cABs reduced the protein expression level of iNOS and significantly increased the levels of CD206 and arginase-1 (Arg1), which are the main markers of M2 macrophages (Fig. 3H). After 24 hours of incubation, we detected the levels of pro- and anti-inflammatory factors in the supernatants. Apparently, LPS vigorously enhanced the release of TNF-α and interleukin-6 (IL-6), which showed levels that were much higher than those in the control group. The addition of AB ghosts abolished the cytokine storms and promoted the secretion of anti-inflammatory factors, transforming growth factor–β (TGF-β), and IL-10, which was also observed in the cABs group (Fig. 3I). Collectively, these results demonstrated that the apoptotic membrane endowed cABs with the inflammatory regulation ability to enhance M2 polarization in vitro.
Construction of the modular cargo loading and intracellular release system
After determining the targeting and inflammation regulating abilities of the shells of cABs, we constructed a modular delivery system as core to transfer different functional molecules like natural EVs. It has been widely reported that EVs play a pivotal role in cellular functional regulation through the transfer of microRNAs or other molecules (15, 28); thus, functional microRNAs and small-molecule drugs were used as cargos for further study.
Curcumin (Cur) is a natural polyphenol that exhibits excellent antioxidant and anti-inflammatory properties and plays a key role in the inflammation progression as a potential regulator of macrophage polarization (29). However, because of its hydrophobic property, curcumin has a poor solubility, which becomes an impediment in its bioavailability, efficacy, and further application. miR-21 is a widely reported immunoregulatory small noncoding RNA (30, 31). In particular, several studies have confirmed the contribution of miR-21 to the resolution of inflammation via its regulation of the phenotype of macrophages (30). Hence, miR-21 and curcumin, the proven effective anti-inflammatory microRNA and small-molecule drug, were chosen as the model cargos to realize the efficient delivery of nucleic acids and small-molecule compounds for inflammation regulation. Here, we established a modular MSN delivery system with intracellular release properties fitting multiple therapeutic cargos to avoid the unexpected cargo leakage, which could be able to achieve accurate, safe, and effective regulation of inflammation assisted by the AB membrane with targeting and inflammation regulating abilities (Fig. 4A).
Specifically, miR-21 binds to the positively charged MSNs via electrostatic interaction, in which the positively charged dimethylamine (DMA) group is linked to the MSNs by S─S bonds (fig. S4A, a). Thus, the high concentration of glutathione (GSH) in the target cell cytoplasm would break the S─S bonds after the internalization of cABs, causing rapid release of miR-21 and resulting in the subsequent biological responses (Fig. 4A, a). Similarly, curcumin is loaded in the pores of MSNs by physical adsorption and then encapsulated by polyester to prevent leakage during nanoparticle circulation (fig. S4A, b). After targeting inflammation site and being engulfed by macrophages, the polyester could be degraded by esterase in the cells to trigger curcumin release (Fig. 4A, b). The calculated cargo encapsulation efficiency of MSNs at different cargo/MSN weight ratios are listed in table S1. The encapsulation efficiency of both miR-21 and curcumin reached 92%, which demonstrated the excellent loading capacity of MSNs for various cargos.
Moreover, the chemical components of MSNsmiR-21 were determined by x-ray photoelectron spectroscopy (XPS), which showed obvious signals of nitrogen and sulfur, indicating the successful grafting of DMA on MSNs with S─S bonds for loading and GSH-triggered release of miR-21 (Fig. 4B). The fabrication process was also verified by infrared spectrometry (Fig. 4C) and zeta potential measurements (Fig. 4D), which reflected the obvious characteristic vibrations of carbon-nitrogen bonds (C─N) in the DMA group and the gradual surface charge changes of the MSNs in accordance with the electronic properties of certain function groups. The surface charge of MSNs+ changed from positive to negative [similar to the surface charge of MSNs modified with sulfhydryl groups (MSNs-SH)] after incubation in PBS solution containing GSH, indicating the GSH-mediated cleavage of the S─S bonds (Fig. 4D). To directly demonstrate the GSH-triggered miR-21 release, MSNsmiR-21 were incubated under two different conditions (PBS with and without GSH) for 7 days. Figure 4E shows that the final proportion of released miR-21 after 7 days in the control group was less than 20%. In contrast, the release of miR-21 markedly increased over time in the presence of GSH, finally reaching 90% after 7 days. After characterizing the formation and function of the miR-21–loaded MSNs, they were further fused with AB membranes to produce cABsmiR-21 using the same approach as fabrication of pure cABs without cargos that were mentioned above. No significant differences could be observed in the membrane-coating percentage, size, and zeta potential after miRNA loading compared with RNA-free cABs (fig. S4, B to D). Confocal fluorescence microscopy analysis was first used to investigate the delivery and release of miR-21 by cABsmiR-21 at the cellular level. As shown in Fig. 4F, free miR-21 was unable to enter the cytoplasm. However, cABs could effectively realize the intracellular delivery of miR-21, which was followed by gradual miR-21 release triggered by GSH in the cytoplasm.
Likewise, we modified the curcumin-loaded MSNs with polyester as a smart cap to realize the intracellular release of curcumin. The formation of the polyester cap was investigated by TEM, which showed an obvious organic layer around the MSNs (Fig. 4G, a and b). To quantify this procedure, thermogravimetric analysis (TGA) was used, showing that 7.45 weight % (wt %) of polyester cap was formed on MSNs (Fig. 4H). As expected, polyester encapsulation caused an obvious decrease in the pore size from 4 to 0 nm, indicating that the cargo had been effectively encapsulated in these pores (Fig. 4I). The esterase responsiveness of MSNsCur was also investigated by TEM, TGA, and determination of the pore size distribution [Fig. 4, G (c), H, and I]. As seen from the images, the MSNCur turned to smooth MSN without obvious organic layer after coincubation with esterase, which presented only a 1.80 wt % mass difference compared with pure MSN and the reopening of the 4-nm pores, indicating that the majority of the polyester cap could be degraded by esterase to achieve on-demand drug release. The release profile of curcumin from MSNsCur was further detected by determination of the fluorescence spectrum, which showed that the release was highly dependent on esterase (Fig. 4J). Curcumin was markedly released from MSN-Cur without polyester layer and was hardly released from MSN-Cur-polyester in the absence of esterase. With the addition of esterase, curcumin was released gradually from MSN-Cur-polyester. The as-synthesized curcumin-loaded MSNs with polyester were further used as core to prepare cABsCur using the same approach as fabrication of pure cABs without cargos (fig. S4, B to D), which facilitated the targeted delivery and intracellular release of curcumin (Fig. 4K). All these results demonstrated that MSNs had the distinctive loading capability and operability, and modular cABs could be used for different therapies by changing the cores to match various cargos.
cABs promoted M2 polarization of macrophages via miR-21/curcumin delivery
We further assessed the therapeutic potential of cABs loaded with miR-21 or curcumin, and we hypothesized that the natural membrane and therapeutic molecules would have synergistic effects on inflammatory macrophages. We first verified the bioactivity of miR-21 after intracellular delivery. Free miR-21, cABsnull, MSNsmiR-21, and cABsmiR-21 were added to the culture system for 2 hours of coincubation, and then the miR-21 expression level in macrophages was detected by reverse transcription polymerase chain reaction (RT-PCR). The results showed that the miR-21 expression level was distinctly up-regulated in the cABsmiR-21 group (Fig. 5A), indicating the efficient delivery and release of miR-21 by cABsmiR-21. In addition, cABsnull also increased the expression level of miR-21, although this increase showed no significance, which was consistent with previous findings that the immunoregulatory effect of apoptotic cells on macrophages was mediated by miR-21 (31). In addition, the miR-21 level was increased slightly but not significantly in MSNsmiR-21 group due to the deficiency of targeted ability and the protection of natural membrane. To evaluate the function of the delivered miR-21, we detected two representative downstream targets of miR-21, programmed cell death 4 (PDCD4), and phosphatase and tensin homolog deleted on chromosome ten (PTEN), by Western blotting (31). The results showed that the protein expression levels of PDCD4 and PTEN were significantly decreased in the cABsmiR-21 group compared with the other groups (Fig. 5B). These data indicated that the natural AB membrane efficiently mediated functional miR-21 delivery by cABs.
To further evaluate the inflammatory regulation ability of cABsmiR-21, LPS was added to BMDMs to induce inflammation before the administration of cABsmiR-21. Then, cABsnull, free miR-21, MSNsmiR-21, and cABsmiR-21 were added to the culture system, and unstimulated macrophages were served as a control. We then detected the M1 and M2 phenotype distribution in every group. Analysis by confocal microscopy indicated that the proportion of the M1 subpopulation was significantly reduced, and the proportion of the M2 subpopulation was markedly increased in the cABsmiR-21 group (Fig. 5C). Meanwhile, the regulatory effect of cABsmiR-21 on M2 polarization was greater than that of cABsnull, which also exerted a moderate effect to promote M2 polarization, indicating the synergistic effects on inflammation generated by AB membrane and the delivered miR-21. MSNsmiR-21 also resulted in a mild effect on inflammatory macrophages to a much lesser extent than cABsmiR-21 due to the deficiency of the targeting ability. Western blot analysis resulted in findings similar to confocal microscopy (Fig. 5D). In addition, the levels of multiple cytokines released by the treated macrophages were examined by enzyme-linked immunosorbent assay (ELISA). The results showed that cABsmiR-21 reduced the levels of proinflammatory factors (TNF-α and IL-6) and distinctly elevated the levels of anti-inflammatory factors (TGF-β and IL-10; Fig. 5E).
Analogous findings were observed for treatment with cABsCur. The subpopulation distribution displayed that the M2 subpopulation increased, and the M1 subpopulation decreased significantly due to cABsCur treatment (Fig. 5F), which is consistent with the tendency observed in the Western blot results (Fig. 5G). Likewise, the addition of cABsCur lessened the release of TNF-α and IL-6 and elevated the levels of TGF-β and IL-10 (Fig. 5H). Collectively, these results demonstrated that cABs could amplify their anti-inflammatory effect via miR-21/curcumin delivery, and the specificity of cABs as chimeric EVs opens an avenue for the therapy of inflammation-related diseases.
cABs ameliorate cutaneous inflammation and promote regeneration
To investigate the effect of cABs on inflammation in vivo, a cutaneous inflammatory wound model was constructed by generating full-thickness wounds in mice. First, the inflammation site–targeting capability of cABs was evaluated by the in vivo imaging system. The 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindotricarbocyanine iodide (DiR)–labeled MSNs, AB ghosts, cABs, and PBS (control) were injected into mice via the tail vein on day 1 after operation to detect the biodistribution in vivo. The AB ghosts and cABs groups exhibited stronger intensity around the defect than the bare MSNs (Fig. 6A), indicating that the accumulation of nanoparticles at inflammation sites was mainly derived from the specific bioconjugation rather than the enhanced permeation and retention effect. As expected, the enhanced targeting capacity of cABs to inflammatory wound sites was also observed after administration over time (fig. S5A). In line with the aforementioned cell target capacity, it was further observed that cABs were engulfed by macrophages specifically in vivo (Fig. 6B), and the uptake of cABs by macrophages increased over time (fig. S5B). Overall, these results strongly confirmed the dual-target capability of cABs in vivo.
To demonstrate whether cABs could promote skin regeneration, mice were intravenously injected with different agents every 2 days for a total of five times starting on day 1 after operation, and the wound area was monitored at various time points, PBS was used in the blank control group (Fig. 6C). The wound closure rate in mice injected with cABsmiR-21 or cABsCur was significantly accelerated compared to that in the cABsnull or PBS group, notably suggesting the potent therapeutic effect (Fig. 6D). Hematoxylin and eosin (H&E) staining of the dissected skin tissues showed that cABsmiR-21 and cABsCur injection significantly enhanced skin healing compared to that observed in the PBS or cABsnull groups (Fig. 6E). There were well-organized and stratified neo-epithelium in the center of the defect in the cABsmiR-21 and cABsCur groups, with massive collagen deposition, increased formation of hair follicles and sebaceous glands, and loss of inflammatory cell infiltration, which was seldom observed in the PBS group. Mechanistically, cABsmiR-21 and cABsCur promoted wound healing by converting macrophages to an anti-inflammatory phenotype to resolve local inflammation, which was marked by a significant increase in the CD206+ staining rate and a reduction in the iNOS+ staining rate (Fig. 6F); this was verified by the Western blot analysis (Fig. 6G). In addition, we applied Mer proto-oncogene tyrosine kinase (MerTK), another marker of the proresolution macrophage phenotype, to label macrophages and showed the similar results that treatment with cABsmiR-21 or cABsCur increased the number of MerTK+ cells (fig. S6A).
In general, cell proliferation and renewal play a crucial role in tissue regeneration. In line with this, we found that cABsmiR-21 and cABsCur significantly enhanced Ki-67 expression levels according to fluorescence imaging (Fig. 6H). After 12 days of treatment, the cutaneous wounds in each group were closed to a large extent, and the original defects were covered with neo-epidermis. However, the thickness and differentiation status of the neo-epidermis differed among the groups. Keratinocytes are an essential contributor to epithelial integrity and resistance to infection and damage (32). Thus, the levels of the keratinocyte marker cytokeratin 14 (KRT14) were measured to assess the reepithelization in the neoskin. The fluorescence intensity in the PBS group was obviously decreased compared to that in normal skin, indicating poor skin regeneration due to the deficiency of mature keratinocytes. However, the administration of cABsmiR-21 or cABsCur could facilitate the expression of KRT14, which was similar to that in normal skin (Fig. 6I). In addition, histological examination of the major organs revealed no noticeable organ damage 12 days after cABs treatment, and no obvious reduction was observed in the body weight of mice during the study, indicating no potential in vivo toxicity caused by cABs (fig. S7, A and B).
Together, these results demonstrated that cABs could convert the macrophages toward an anti-inflammatory phenotype via the transfer of miR-21 or curcumin, which accelerated the resolution of inflammation, cell proliferation, and reepithelialization and lead to improved healing in cutaneous wounds, with collagen deposition and formation of subcutaneous appendages. The biocompatibility of cABs was also confirmed in animal models during the wound healing process.
Administration of cABs ameliorates inflammatory bowel diseases in a colitis model
To further investigate the therapeutic effects of cABs on inflammation, we constructed another acute inflammatory model, the dextran sulfate sodium (DSS)–induced colitis model, according to the previous protocol (33). We first assessed the targeting ability of cABs to inflammation sites by intravenous injection of DiR-labeled MSNs, AB ghosts, and cABs into mice after DSS administration, with PBS served as the control. The colon in the cABs group exhibited a stronger fluorescence intensity compared to that in the bare MSN group (Fig. 7A), indicating that fusion with the AB membrane facilitated the accumulation of cABs in inflammatory colons. More details regarding the accumulation process are shown in fig. S5C. As seen from this figure, the cABs began to appear in the target tissue at 3 hours and then reached a peak at 24 hours (fig. S5C). The further specific internalization of cABs by macrophages in colon tissue was also investigated, which indicated a gradual increase in cellular uptake over time (Fig. 7B and fig. S5D), demonstrating the inherent capacity of cABs to target the inflammation region and macrophages.
To determine the therapeutic effects of cABs on colitis, various agents were suspended in 100-μl solution and injected intravenously at days 3, 5, 7, and 9 of DSS administration. On day 10, the serum, colons, and organs were collected for the subsequent examinations (Fig. 7C). The PBS group exhibited more extreme body weight loss and a higher disease activity index (DAI) than the other groups, which were mitigated in the cABsmiR-21 group (Fig. 7, D and E). Treatment with cABsCur also reduced the extreme loss in body weight and exacerbation of the disease. Meanwhile, DSS administration resulted in the distinct shortening of the colon accompanied by severe hyperemia and ulceration. The therapeutic effects exerted by cABsmiR-21 and cABsCur partially recovered the colon length and relieved the severe diarrhea, the presence of fecal blood, and rectal prolapse exhibited by the PBS group (Fig. 7F and fig. S7C). Histologically, after DSS administration, the colon in PBS group exhibited enormous transmural inflammatory cell infiltration, crypt damage, and loss of mucosal epithelium integrity, which was abrogated by the treatment of cABsmiR-21 or cABsCur that resulted in the reduction of histological activity index (Fig. 7G). Moreover, the colon tissue in cABsmiR-21– or cABsCur-treated groups retained the normal crypt and intestinal gland structure and showed reduced inflammatory infiltration (Fig. 7G). In addition, the concentrations of TNF-α, IL-6, TGF-β, and IL-10 in serum were analyzed by ELISA kits. Treatment of cABsmiR-21 or cABsCur markedly reduced the elevated levels of TNF-α and IL-6 mainly released by inflammatory macrophages and simultaneously increased the concentrations of TGF-β and IL-10, which contributed to inflammation resolution and tissue repair, compared with the effects in the other groups (Fig. 7H). Mechanistically, the fluorescence images showed that the administration of cABsmiR-21 or cABsCur promoted the conversion of macrophages to the M2 phenotype in favor of tissue homeostasis (Fig. 7I and fig. S6B), which is consistent with the Western blot results (Fig. 7J). Histological examination of the major organs exhibited a little inflammatory cell infiltration in the spleen of the PBS group owing to the DSS-induced inflammatory lesion (fig. S7D). The other groups exhibited no noticeable organ damage 10 days after cABs treatment, emphasizing no potential in vivo toxicity caused by cABs. Together, these results demonstrated that cABs could ameliorate the severity and destructive effects of inflammatory bowel diseases via shuttling miR-21 or curcumin that modified the polarization status of macrophages.
Biocompatibility of cABs
To assess the safety of cABs in vivo, the weights and blood of mice were analyzed after the administration of cABs loaded with different cargos. As shown in fig. S8A, no obvious reduction was observed in the body weight of cABsmiR-21-/cABsCur-treated mice over time. Meanwhile, the proportions of red blood cells (RBCs), white blood cells (WBCs), and platelets (PLTs) exhibited no significant differences between the different groups (fig. S8B). The functional indicators of RBCs [hemoglobin (HGB), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), and mean corpuscular hemoglobin concentration (MCHC)] also showed no significant differences between the PBS group and the treated groups (fig. S8C). Moreover, the safety of cABs was evaluated by performing a serum biochemical test. According to the results, treatment of cABsmiR-21/cABsCur did not lead to significant changes in hepatorenal function (fig. S8D). Histological examination revealed no noticeable acute organ damage after cABs treatment, suggesting the lack of in vivo toxicity caused by cABs (fig. S8E). In addition, the cytotoxicity test demonstrated that cABs had no obvious cytotoxic effects on macrophages and FBs (fig. S8F). The cABs also had excellent blood compatibility, evidenced by the low hemolytic index (fig. S8G). Collectively, these results provide the strong evidence about the excellent biocompatibility of cABs platform for delivery of therapeutics.
Recently, given the discovery of the biological functions of EVs, EVs have received extensive attention as new therapeutic frontiers (15, 28). Despite the rapid development in modification of EVs for the design of personalized drug delivery carriers (4, 5), the core challenges have not yet been overcome because of the inadequate understanding of EVs, their components, and the underlying mechanisms. Meanwhile, the safety and standardization deserve to be further explored, including the removal of irrelevant content. To some degree, modification methods would change the intrinsic properties of EVs (4, 12), which increases the risk and uncertainty associated with the treatment. In addition, the expression and bioactivity of functional molecules in EVs depend on multiple factors, including the source cell type, the purity and status of the EVs, and the methods used for isolation, purification, and storage (2). An uncontrolled sorting and encapsulating process generates heterogeneity in EVs intended for therapeutic application. The release mode cannot satisfy the original purpose in the target site/cell. By biological/abiological means, our work established a chimeric platform enhanced by superiority of the biological functions of natural EVs and the modular framework of the smart delivery system. Compared with those in the previous study (23), the modular delivery pattern in this study could be precisely tuned to transfer nucleic acids or small-molecule drugs specifically, which provides a way to design strategies and applications of therapeutic EVs for the treatment of multiple diseases.
In general, products from cells undergoing apoptosis are regarded as wastes or harmful substances (15). However, for the past decades, ABs have been confirmed to contribute to signal transduction and homeostasis regulation (16). It has been reported that apoptosis is the initial process of regeneration (34), but it is unknown whether ABs could contribute to regeneration. According to previous studies (17, 31), apoptotic cells can convert the phenotype of macrophages to achieve the resolution of inflammation after clearance by macrophages, which promotes the transition from the inflammation stage to the stage of tissue repair and regeneration. We also discovered the same effects after the administration of cABs to inflammatory macrophages, which demonstrated the targeting ability and immunoregulatory effects exerted by cABs. This mechanism could be extended to produce the targeted regulation of inflammation by using engineered approaches and provides insight into the strategy of AB-driven targeted delivery systems, which informs and guides the development of engineered ABs. The inherent ability of ABs to target macrophages provides an additional advantage for the engineered design of targeted delivery methods to regulate inflammation and can potentially be generalized to a variety of EV types for these applications. Collectively, we first applied the ABs to promote regeneration, which broadened the application of ABs and engineered EVs and implied the extensive potential for the use of vesicles with particular origins.
Inflammation is common in several diseases, including cancer, obesity, and trauma. Excessive and chronic persistent inflammation can cause tissue lesions, dysfunction, and even disability (20). Previous studies have demonstrated multiple treatments using EVs to target inflammation sites, but there still exist several limitations. On one hand, the display of ligands, antibodies, or aptamers on the EV surface by chemical conjugation increased the immunogenicity and toxicity of EVs (5, 12). On the other hand, the off-target effects made it difficult to maintain a high drug concentration in the local inflammation region for a long term (35). Increasing the dose via systemic administration would produce side effects and even toxicity. Therefore, it is noteworthy that the development of an alternative promising anti-inflammatory therapeutic strategy is imperative for inflammatory diseases. This study takes advantage of the natural chemotactic ability of T cells to target inflammation sites and the specific uptake of ABs by macrophages to realize the dual-target capability, which provides functionalities that would otherwise be difficult to realize via conventional means to regulate inflammation precisely. The uptake pathway of EVs may be very likely specific, which explains the feasibility and popularity of the proposed engineered EVs for targeted drug delivery (3, 36). The design of cABs compensated for the low accumulation in the inflammation region through the interaction of the adhesion molecules on the surface of EVs and targeted cells. In addition, strong evidence was found in this study that there was negligible in vivo toxicity caused by cABs, which emphasized their excellent biocompatibility and safe application, and this may also partially benefit from the lipid bilayer of the natural vesicle. Here, the integration of vesicles membrane with diverse biological properties and modular delivery carriers has been confirmed to be a favorable and effective construction method of engineered EVs for inflammation regulation.
In this study, our emphasis was mainly on macrophages in the inflammation region. Whether cABs exert the immunoregulatory effects on other immune cells still needs to be investigated. For further translation to clinical application, it is valuable to compare cABs with standard clinical treatments, such as corticosteroids (37). Corticosteroids are potent formulations that induce M2 polarization and enhance proresolution responses, while several studies have suggested that side effects associated with systemic administration limit their clinical application (35). Therefore, the superiority of engineered EVs derived from natural membrane advances research on bioinspired EVs for targeted therapy in the future, and more designs and strategies are imperative to cope with clinical demands. To construct the engineered EVs, the natural membrane extraction and MSN manipulation in this study could be simply performed. The low cost of raw materials, simple synthesis, stable performance, and high yield make MSN an ideal core for convenient production of our cABs (38). Furthermore, silica-based materials are also known for their biocompatibility and have been approved by the U.S. Food and Drug Administration, generating a bright future to translate cABs developed in this study for clinical applications (38, 39). Although cABs are a promising candidate for disease therapy, there are still plenty of challenges to overcome from laboratory to bedside. First, additional improvements, such as standardization of the manipulation, characterization, and examination procedures, are required for batch-to-batch manufacturing and clinical use (40). Other aspects that need to be solved include the standard cell culture conditions and vesicle isolation methods. More details on the pharmacokinetic and pharmacodynamic profiles will benefit the clinical dose predictions. Furthermore, more research on the biogenesis, function, and mechanisms of EVs is needed to strengthen the theoretical foundation of the engineering approaches (28). Collectively, although there is still a long way to go before successful clinical application, the engineered EVs herald a new era of the targeted delivery of therapeutics in response to clinical demands, which sheds light on the development of smart bionanomaterials.
MATERIALS AND METHODS
LPS (L2880), PKH26/67, Hoechst 33342, NH3·H2O (AR, 25% ~ 28%), CTAB, ammonium persulfate (APS), N,N,N,N-tetramethylethylenediamine (TEMED), (3-mercaptopropyl)-trimethoxysilane, 2,2′-dithiodipyridine, and 2-dimethylaminoethanethiol hydrochloride were obtained from Sigma-Aldrich (St. Louis, MO, USA). Staurosporine (#9953S) was purchased from Cell Signaling Technology (Boston, MA, USA). Recombinant murine macrophage colony-stimulating factor (M-CSF) (315-02) and recombinant human TNF-α (300-01A) were purchased from PeproTech (Rocky Hill, USA). DiR (D12731) was purchased from Invitrogen (USA). RhB was purchased from Kemiou Chemical Reagent Corporation (Tianjin, China). RPMI-1640 medium, Dulbecco’s modified Eagle medium (DMEM), α-modified Eagle medium (α-MEM), trypsin-EDTA, dispase, type I collagenase, fetal bovine serum (FBS), and PBS were obtained from Gibco (USA). ACK (Ammonium-Chloride-Potassium) lysis buffer (C3702), radioimmunoprecipitation assay lysis buffer (P0013B), and Coomassie Blue Fast Staining Solution (P0017) were obtained from Beyotime (Shanghai, China). A BCA protein assay kit (PA115) was obtained from TIANGEN (Beijing, China). Curcumin (HY-N0005) was purchased from MedChemExpress (Shanghai, China). The mouse miR-21-5p (miR-21) (uagcuuaucagacugauguuga) and miR-21-5p-Cy5 were constructed by and purchased from RiBoBio (Guangzhou, China).
The anti-mouse CD3ε antibody (100314) and Allophycocyanin (APC)–conjugated anti-mouse CD25 antibody (102012) were purchased from BioLegend. The anti-CD28 antibody (16-0281-85) and fluorescein isothiocyanate (FITC)–conjugated anti-CD11b antibody (11-0112) were purchased from eBioscience. The anti-MerTK monoclonal antibody (14-5751-82) was purchased from Invitrogen. Phycoerythrin (PE)–conjugated anti-F4/80 antibody (ab218761), anti-F4/80 antibody (ab6640), anti-CD44 antibody (ab189524), anti-CD11b (ab133357), anti–integrin-β3 antibody (ab75872), anti–ICAM-1 antibody (ab171123), anti-mannose receptor antibody (ab125028), anti-mannose receptor antibody (ab195191), anti-PTEN antibody (ab44712), anti–cytokeratin 14 antibody (ab181595), anti–Ki-67 antibody (ab15580), and anti-iNOS antibody (ab209027) were purchased from Abcam. The anti-mouse C1q antibody (CL7501F) was obtained from CEDARLANE. The anti-CD3 antibody (sc-20047) was obtained from Santa Cruz Biotechnology. The anti-CD8A antibody (PAB16612) was obtained from Abnova. The anti–P-selectin antibody (3633R) was purchased from BioVision. The anti–caspase-3 antibody (#9662 s), anti-PDCD4 antibody (#9535), anti-iNOS antibody (#13120 s), and anti-Arg1 antibody (#93668) were obtained from Cell Signaling Technology. The anti–β-tubulin antibody (CW0098) and anti–glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibody (CW0100) were purchased from CwBio. The secondary antibodies [peroxidase AffiniPure goat anti-mouse immunoglobulin G (IgG), 115-035-003; peroxidase AffiniPure goat anti-rabbit IgG, 111-035-003; FITC-conjugated anti-rat secondary antibody, 112-095-003] were all purchased from Jackson ImmunoResearch.
Primary T cells were derived from the splenocytes of C57BL/6 mice, which were maintained in RPMI-1640 medium supplemented with 20% FBS and 1% penicillin/streptomycin (Invitrogen, USA). The plate was seeded with anti-mouse CD3ε antibody at 2 μg/ml in 1× PBS (37°C, 4 hours). Then, the spleen was crushed and filtered to form a single-cell suspension. After erythrocyte lysis for 10 min, the cells were centrifuged, resuspended, and filtered through a 100-μm cell strainer (Biologix) to remove clumps, and then the cells were incubated in a six-well plate with anti-CD28 antibody (2 μg/ml) for T cell activation. After 24 hours of incubation, the T cells were activated. LPS was added to the culture system at 1 μg/ml for 12 hours to stimulate T cells. The unactivated and activated T cells were determined by the microscope imaging system (Olympus, Japan). T cells were collected and incubated with APC-conjugated anti-mouse CD25 antibody at 37°C for 30 min and then were analyzed by flow cytometry (Beckman Coulter, USA).
Primary macrophages were derived from bone marrow–derived monocytes of C57BL/6 mice which were maintained in DMEM (high glucose) supplemented with 10% FBS and 1% penicillin/streptomycin. The femurs and tibias were dissected from the mice. Then, the bone marrow was rinsed with 1× PBS, which was followed by erythrocyte lysis in ACK lysis buffer. After centrifugation, the cells were resuspended in media with M-CSF (20 ng/ml) to induce the maturation of macrophages. The BMDMs could be used in subsequent experiments for at least 7 days after induction. The BMDMs were collected after detachment with 0.05% trypsin-EDTA and centrifugation at 800 rpm for 5 min, and then the BMDMs were stained with FITC-conjugated anti-CD11b and PE-conjugated anti-F4/80 antibodies. The induction of mature macrophages was evaluated by flow cytometry (Beckman Coulter, USA). Fluorescence staining was carried out by incubation with anti-F4/80 antibody at 4°C overnight. After incubating cells with the FITC-conjugated anti-rat secondary antibody at 37°C for 30 min, the nuclei were counterstained with Hoechst 33342. Fluorescence imaging was performed by using confocal laser scanning microscopy (CLSM) (Nikon, Japan).
FBs were derived from the dermal FBs of C57BL/6 mice, which were maintained in α-MEM (Gibco, USA) supplemented with 10% FBS and 1% penicillin/streptomycin. Briefly, the hairless full-thickness skin was excised from the back of mice, and the subcutaneous fat was removed before the skin samples were cut into small strips. After washing in 1× PBS several times, the skin was incubated with dispase at 4°C overnight. Then, the epidermis was removed, and the dermis was cut into pieces with scissors before digestion with type I collagenase for 90 min at 37°C with shaking. Next, 5 to 10 ml PBS was added to stop the digestion process, and the suspension was centrifuged at 800 rpm for 5 min to concentrate the FBs. The supernatant was removed, and the pellet was resuspended in media.
HUVECs were purchased from American Type Culture Collection and maintained in DMEM (Gibco, USA) supplemented with 10% FBS and 1% penicillin/streptomycin. All cells were incubated at 37°C in a humidified atmosphere with 5% CO2 (Thermo Fisher Scientific). The media were refreshed every third day, and the cells were passaged once they reached 80 to 90% confluence.
AB isolation and characterization
T cells were treated with staurosporine (0.5 μM) for 3 to 4 hours to induce apoptosis. Then, the culture media were collected and centrifuged at 50 g for 5 min to remove the cells and debris. After repeating this twice, the supernatant was further centrifuged at 1000g (10 min) to concentrate the ABs in the pellet. Then, the pellet was suspended with 1× PBS and stored in −80°C for subsequent experiments. The protein concentration was determined using a BCA protein assay kit.
DLS analysis was performed using a Zetasizer Nano ZSE (Malvern, UK). The morphology of the ABs was observed by SEM (Hitachi, Japan). Annexin V–FITC staining was conducted using an Annexin V–FITC/PI apoptosis assay kit (A005-2, 7Sea Biotech, China). ABs were incubated with anti-mouse C1q antibody at 37°C for 30 min. Then, fluorescence imaging was conducted with CLSM (Nikon, Japan), and flow cytometry analysis was also carried out (Beckman Coulter, USA). Western blotting was performed to characterize the protein constitution of the ABs. The protein samples from T cells, apoptotic T cells, and ABs were loaded into the Bio-Rad Electrophoresis System. The proteins in the gel were transferred to polyvinylidene difluoride (PVDF) membranes. After blocking in 5% bovine serum albumin (BSA) solution (DY60105, DIYIBio), the membranes were incubated with primary antibodies (CD3, CD44, caspase-3, and β-tubulin) at 4°C overnight. Then, the membranes were incubated with the corresponding secondary antibodies at room temperature for 1 hour. Films were developed using Western chemiluminescent horseradish peroxidase (HRP) substrate (Millipore) and evaluated with an imaging system (Tanon 5500, Shanghai).
The ABs were subjected to hypotonic treatment by resuspension in a hypotonic lysis buffer consisting of 10 mM tris (pH 7.4), 10 mM MgCl2, and 1 mM phenylmethylsulfonyl fluoride at 4°C for 1 hour, followed by gentle sonication for 5 s (VCX 130 PB, Sonics, USA). After centrifugation at 100g for 10 min to remove the debris, the AB ghosts were concentrated by centrifugation at 10,000g for 10 min and then washed in water at least three times until the intracellular components were removed, which was verified by fluorescence microscopy. Nuclear staining with Hoechst 33342 served as the index for the assessment of the residual components in the PKH67-labeled ABs. The resulting AB ghosts were dispersed in water and stored in 4°C for subsequent experiments.
Synthesis of MSNs
MSNs were prepared by the classical CTAB-templated, base-catalyzed sol-gel method according to a previous work (24). The pH value of 1000-ml deionized water was adjusted to approximately 11 with 52.8 ml of ammonium hydroxide (25 to 28 wt % NH3·H2O). The temperature was raised to 323 K, and then 1.12 g of CTAB was added. After the CTAB was completely dissolved, 5.8 ml of tetraethylorthosilicate (TEOS) was added dropwise with rapid stirring. After 2 hours, the mixture was incubated overnight, centrifuged, and washed thoroughly with distilled water and ethanol. As-synthesized silica nanoparticles were dispersed in ethanol by sonication for 30 min, followed by the addition of 20 ml of 1:1 mixture (v/v) of water and 1,3,5-trimethylbenzene. The mixture was placed in the autoclave and kept at 140°C for 4 days without stirring. The resulting white powder was washed with ethanol and water for five times each. Then, the surfactant templates were removed by extraction using acidic methanol (9 ml of HCl/400 ml of methanol, 36 hours) at 70°C, and then the MSNs were centrifuged, washed several times with ethanol, and dried under a vacuum for 20 hours.
Manipulation of cABs
To fuse the AB ghosts with MSNs, 1 mg of MSNs were mixed with 2 mg of AB ghosts, and then the suspension was sonicated for 2 min intermittently and gently in a bath sonicator (SY25-12, Shengyuan Supersonic, China). Then, the cABs were concentrated after centrifugation at 5000g for 10 min to form a pellet, followed by resuspension in water before storage. The size and zeta potential of the AB ghosts, MSNs, and cABs were measured by DLS (Zetasizer Nano ZSE, Malvern, UK). The morphology of the cABs was determined by TEM (TECNAI Spirit, FEI). A total of 4 μl of the cABs solution at a concentration of 1 mg/ml was deposited onto a carbon-coated 400-square mesh copper grid. Ten minutes after the sample was deposited, the grid was rinsed with 10 drops of water. A drop of 1% phosphotungstic acid was added to the grid to conduct the negative staining. The grid was subsequently dried naturally and visualized using the 120-kV FEI TEM. TEM-EDS were conducted using a field emission TEM (JEOL, Japan).
The membrane loading percentage was confirmed by the detection of the fluorescence intensity. Briefly, PKH67-labeled AB ghosts were mixed with MSNs, followed by sonication in a bath sonicator for 2 min to generate cABs. The suspension before fusion and the supernatant after centrifugation were collected, and the fluorescence intensity was detected using a multimode plate reader (HH3400, PerkinElmer) (excitation = 496 nm/emission = 520 nm). The membrane loading percentage (%) = (the fluorescence intensity before fusion − the fluorescence intensity after fusion)/(the fluorescence intensity before fusion) × 100%.
PKH67-AB ghosts were coated onto the RhB-MSNs, followed by sonication for 2 min. The generated cABs and bare RhB-MSN suspension samples were prepared at 1 mg/ml. Then, the samples were resuspended with 80% glycerine (MP Biomedicals), and 4 μl of sample was added to the coverslip to be observed with a confocal microscope (Nikon, Japan).
Membrane protein retention
Western blotting analysis was performed to identify protein retention during the manipulation of cABs. All samples were prepared at a final protein concentration of 1 mg/ml in SDS-PAGE loading buffer (CW0027S, CwBio). Twenty microliters of sample was loaded into each well of a 10% SDS-polyacrylamide gel in a Bio-Rad Electrophoresis System. Protein staining was accomplished using Coomassie Blue Fast Staining solution, and the gels were destained in deionized water at 4°C overnight before imaging. For Western blot analysis, PVDF membranes were blocked with 5% BSA solution for 1 hour and then incubated with antibodies against CD3, CD8A, CD11b, CD44, and integrin-β3. Films were developed using western chemiluminescent HRP substrate (Millipore) with an imaging system (Tanon 5500, China).
To assess short-term storage stability, MSNs and cABs were resuspended in water and 1× PBS, respectively, to measure the size change by DLS for 1 week. To evaluate the long-term storage stability, the sizes of cABs before lyophilization and after resuspension were measured. Serum stability tests were conducted by resuspending cABs in 100% FBS at a final concentration of 0.1 mg/ml, with water as the control. The absorbance at 560 nm was measured at different time points (1, 5, 10, 15, 30 60, and 90 min) using a multiplate reader (Epoch, BioTek, USA). The change in absorbance reflected the nanoparticle aggregation due to poor stability in serum.
In vivo circulation time assay
To evaluate the in vivo circulation time, 150 μl of RhB-cABs and RhB-MSNs (2 mg/ml) were injected into the normal mice intravenously. Blood was collected at 1, 5, 15, 30 min and 1, 3, and 8 hours following the injection. Twenty-microliter blood samples were diluted with 30 μl of PBS in a 96-well plate, and the fluorescence intensity of RhB was measured using a multimode plate reader (HH3400, PerkinElmer; excitation = 540 nm/emission = 625 nm).
First, the expression level of the adhesion molecules of different samples was explored. Twenty microliters of protein from the groups of T cells, ABs derived from T cell (ABs), T-cABs, and FB-cABs were loaded into each lane. To further evaluate the protein expression levels after LPS stimulation, equal amounts of the protein samples were loaded into each lane to separate the proteins with different molecular weights. The protocol used was the same as that mentioned above. The detected proteins were Mac-1, CD44, and GAPDH, which was used as the internal reference.
To explore the specific targeted ability of cABs derived from T cell (T-cABs), FB-cABs were fabricated as control using the same protocol. HUVECs were seeded in 24-well plates and cultured until 80% confluence was reached. Then, TNF-α (50 ng/ml) was added to stimulate HUVECs. After stimulation for 6 hours, cells were washed with cold PBS and fixed with 4% paraformaldehyde (PFA) for 1 hour. Then, PKH26-ABs (50 μg/ml), RhB-MSNs, RhB–T-cABs, and RhB–FB-cABs were added to the culture plates at room temperature for 60 s. After treatment, the cells were washed several times with cold PBS and blocked with normal goat serum for 30 min. Subsequently, the cells were stained with P-selectin and ICAM-1. After counterstaining with Hoechst 33342, the cells were imaging with CLSM.
MTT (Methylthiazolyldiphenyl-tetrazolium bromide) assay was adopted to examine the cytotoxicity of cABs in vitro. Briefly, macrophages and FBs were seeded in 96-well plates at a concentration of 103 cell per well. After cell adhesion, MSNs (25 μg/ml) and various concentrations of cABs (0, 10, 25, 50, 100, and 200 μg/ml) were added to the plate. After 24 hours of treatment, cell viability was detected by the standard MTT assay according to the instructions. The optical density at 492 nm was measured with a multiplate reader (Epoch, BioTek, USA). The blood compatibility test was carried out as follows. Briefly, cABs were added to RBC suspensions at the final concentrations of 10, 25, 50, 100, and 200 μg/ml in 0.5 ml of PBS solution. The same amount of RBCs incubated with 0.5 ml of PBS or water was used as negative control or positive control, respectively. The samples were mixed gently, incubated at room temperature for 4, 24, and 48 hours, and centrifuged at 4000 rpm for 5 min. The broken RBCs will release hemoglobin into the supernatant. The absorbance of the supernatants at 450 nm was measured using a multiplate reader.
The percentage of hemolysis was calculated as follows: Hemolysis (%) = (sample absorbance − negative control)/(positive control − negative control) × 100%.
For safety evaluation of cABs in vivo, 100 μl of cABsmiR-21/cABsCur (1 mg/ml) was injected into normal mice every other day for a total of five times (n = 6). At the endpoint, the blood was obtained to carry out the hematology and serum biochemistry tests. The major organs were harvested to conduct H&E staining. Samples from PBS-treated mice were used as controls.
The immunoregulatory effect of the apoptotic membrane in vitro
To explore the immunoregulatory effect mediated by the apoptotic membrane, macrophages were treated with LPS (1 μg/ml) to induce inflammation. Simultaneously, AB ghosts (25 μg/ml), MSNs, and cABs were added to the culture system, and PBS and unstimulated macrophages were prepared in parallel for comparison. After 24 hours, the cells were fixed with 4% PFA at 4°C, followed by blocking with normal goat serum. Subsequently, the cells were incubated with anti-mannose receptor antibody and anti-iNOS antibody at 4°C overnight and stained with Hoechst 33342. The fluorescence imaging was observed by CLSM. The protein level was analyzed by Western blotting. The detected proteins included iNOS, Arg1, and CD206, and GAPDH was used as the internal reference. The levels of TNF-α, IL-6, TGF-β, and IL-10 in supernatants were detected with ELISA kits according to the instructions (NeoBioscience, China).
miR-21 loading and GSH-triggered release from MSNsmiR-21
The loading of miR-21 on the MSNs was achieved via electrostatic interactions. MSNs (50 mg) and (3-mercaptopropyl)-trimethoxysilane (200 μl) were completely dispersed in 20 ml of ethanol by sonication for 10 min (100 W). The solution was stirred overnight under nitrogen at 60°C to obtain MSNs-SH. Next, 225 mg of 2,2′-dithiodipyridine and 145 mg of 2-dimethylaminoethanethiol hydrochloride were added to the above solution to continue the reaction for 12 hours. The mixture was further centrifuged, washed several times with ethanol, and dried under vacuum for 20 hours to obtain the DMA-grafted MSNs with disulfide bonds (MSNs+). Furthermore, 0.06 mg of MSNs+ and 15 nmol of miR-21 were mixed in 60 μl of nuclease-free water at 4°C by repeated aspiration for 1 hour. The mixture was further centrifuged and washed three times with nuclease-free water to obtain MSNsmiR-21. MSNsmiR-21-cy5 were produced by the same method. The encapsulation capacity of MSNs for miR-21 was calculated as follows
The major elements (C, N, O, S, and Si) in the MSNs+ were identified through XPS. The characteristic peaks of the elements S and N are magnified to demonstrate the successful modification of MSN+ (ESCALAB Xi+, Thermo Fisher Scientific). The DLS measurements and infrared spectroscopy were performed by a Malvern Zetasizer Nano Series and Avatar 320 FT-IR spectrometer, respectively, to further investigate the stepwise modification.
In the GSH-triggered release experiment, a certain amount of MSNsmiR-21-cy5 was dispersed in 400 μl of two different solutions (a: nuclease-free water with GSH; b: nuclease-free water without GSH) at room temperature. Subsequently, the supernatant was obtained periodically from the suspension by centrifugation (10,000 rpm, 10 min). The release of miR-21–cy5 from the nanoparticles was determined by detection of the fluorescence intensity in the supernatants using a multimode plate reader (HH3400, PerkinElmer) (excitation = 649 nm/emission = 670 nm). Meanwhile, free miR-21-cy5, PKH67-cABsnull, and PKH67-cABsmiR-21-cy5 were added to the macrophage culture system. After incubation for different times, the cells were counterstained with Hoechst 33342 and observed with CLSM.
Drug encapsulation and intracellular release from MSNsCur
To encapsulate curcumin, 10 mg of MSNs, 1 mg of curcumin, and 10 μl of glyceryl dimethacrylate (GDMA) were completely dispersed in 10 μl of water by sonication for 20 min (100 W). The mixture was placed in a vacuum oven and repeatedly vacuumed (−1 to −0.8 bar) three times for 10 min each time to generate curcumin- and GDMA-coloaded MSNs. Separately, 24 mg of APS was dissolved in 3 ml of water under nitrogen for 20 min, and then 80 μl of TEMED was dissolved in 800 μl of water. After that, 2500 μl of APS solution and 440 μl of TEMED solution were mixed with the curcumin- and GDMA-coloaded MSNs under nitrogen to form a polymerized glyceryl dimethacrylate (PGDMA) cap to prevent curcumin leakage. The solution was stirred at room temperature for 1 hour under nitrogen to complete the polymerization. The resulting nanoparticles were washed with water and dried in a vacuum oven at 50°C overnight. The capacity of MSNs to encapsulate curcumin was calculated as follows
The bare MSNs and MSNs-polyester before and after degradation by esterase were observed by TEM as mentioned above. TGA was also conducted for quantitative analysis of these samples (STA449F5, NETZSCH). The pore sizes of the particles can be measured by a surface area and porosimetry analyzer (V-Sorb 4800, Gold APP). After approximately 100 mg of the powdered sample is dried in advance, the gas is adsorbed by the static capacity method, and the pore diameter–pore volume pattern is determined by performing the pore size test procedure. Intracellular release can be detected by a fluorescence spectrophotometer (Gangdong F-280, China) in vitro. Specifically, 10 mg of MSN-Cur-polyester was dispersed in 10 ml of two different phosphate buffer solutions (a: PBS with esterase; b: PBS without esterase) at 37°C with shaking, and then the supernatant was collected by centrifugation at different time points. The fluorescence intensity of the supernatant was tested to obtain the release result in vitro. Meanwhile, PKH26-cABsCur (25 μg/ml) was added to the macrophage culture system, and PBS was used as a control. After incubation for different times, the cells were counterstained with Hoechst 33342 and then observed with CLSM.
Confirmation of the functional delivery of cargos in vitro
To investigate the successful functional delivery of cargos to macrophages in vitro, macrophages were treated with cABsnull (50 μg/ml), MSNsmiR-21, cABsmiR-21, and an equal amount of free miR-21 for 2 hours. Untreated macrophages were prepared in parallel as a control. The delivery of miR-21 was confirmed by real-time RT-PCR and Western blotting. Total RNA was extracted from macrophages by using TRIzol reagent (Invitrogen, USA) according to the manufacturer’s protocol. After quantification and verification of the purity, 500 ng of total RNA was reverse-transcribed into complementary DNA (cDNA) with a PrimeScript RT reagent kit (RR037A, TaKaRa, Japan). Real-time RT-PCR was performed using cDNA with the SYBR Premix Ex Taq II kit (TaKaRa, Japan) and the Bulge-Loop miRNA qRT-PCR primer set for miR-21 and U6 (RiBoBio, China). Real-time RT-PCR was performed by 10 min of incubation at 95°C followed by 40 cycles of 95°C for 2 s, 60°C for 20 s, and 70°C for 10 s with the CFX96TM Real-time RT-PCR System (Bio-Rad, USA). The qRT-PCR reaction was carried out in a 10-μl reaction volume. U6 was used as the internal control for the quantitation of miR-21. The target molecules of miR-21, PTEN, and PDCD4 were also detected by Western blotting. The steps used were the same as those mentioned above.
Eight-week-old female C57BL/6 mice were purchased from the Animal Center of the Fourth Military Medical University, Xi’an, China. All animal studies were performed by following protocols approved by the Animal Care Committee of the Fourth Military Medical University, China. All mice were maintained under specific pathogen-free conditions with a 12-hour light/12-hour dark cycle. All mice were allowed access to food pellets and tap water ad libitum.
Mouse models of cutaneous wounds and colitis
To evaluate the effects of cABs on wound healing, full-thickness round wounds of equal size (diameter = 1.0 cm) were aseptically generated in the middle of the back. The cutaneous wound mice were randomly grouped (n = 6 in each group) and intravenously injected with PBS (100 μl), cABsnull (100 μg), cABsmiR-21 (100 μg), and cABsCur (100 μg) on days 1, 3, 5, 7, and 9 after wounding. The wound size was measured on days 0, 3, 5, 7, 10, and 12 after injury and calculated with ImageJ. The body weight was recorded every day. At the endpoint, the mice were euthanized, and the injured areas and major organs were harvested for further investigation.
Colitis was induced in C57BL/6 mice by administration of 2.5% (w/v) DSS (MP Biomedicals) in the drinking water for 10 days. In the treated group, cABsnull (100 μg), cABsmiR-21 (100 μg), and cABsCur (100 μg) were suspended in 100 μl of solution and injected intravenously into each group (n = 6) on days 3, 5, 7, and 9 of DSS administration. In the untreated group, mice were injected with an equal amount of PBS, and normal mice served as the control. The body weight was measured, and the DAI was monitored and recorded daily. The DAI was scored according to the following criteria: (i) body weight loss = 0 (no change), 1 (1 to 5%), 2 (5 to 10%), 3 (10 to 20%), and 4 (>20%); (ii) stool consistency or diarrhea = 0 (normal), 1 (slightly soft), 2 (loose), 3 (unformed/mild diarrhea), and 4 (severe diarrhea); (iii) fecal blood = 0 (negative fecal occult blood), 1 (faint positive fecal occult blood), 2 (certain positive fecal occult blood), 3 (moderate rectal bleeding), and 4 (severe rectal bleeding). The DAI is the sum of the scores for body weight loss, stool consistency, and fecal bleeding. At the endpoint, mice were sacrificed; after euthanization, the colons and major organs were dissected carefully for histological experiments, and the colon length was measured.
In vivo biodistribution test
AB ghosts, MSNs, and cABs were prelabeled with DiR, and the labeled nanoparticles were intravenously injected into cutaneous wound mice and colitis mice (300 μg/150 μl per mouse), respectively. PBS was used as a control. At the set time points, the colitis mice were euthanized, and the colons were excised. Then, the cutaneous wound mice (under gaseous anesthesia) and the colons were performed the fluorescence imaging to determine the biodistribution of labeled nanoparticles using an in vivo imaging system (IVIS) at set time points (Xenogen).
Specific uptake of cABs by macrophages
To verify the specific uptake of cABs by macrophages in vitro, macrophages were treated with PKH26-AB ghosts (50 μg/ml), RhB-MSNs, and RhB-cABs for 3 hours, and PBS was used as a control. Then, the cells were fixed with 4% PFA overnight and blocked at 37°C for 30 min. After incubation with anti-F4/80 antibody and FITC–anti-rat secondary antibody as mentioned above, the nuclei were counterstained with Hoechst 33342. Fluorescence imaging was performed by CLSM. Macrophages were also treated with RhB-cABs (25 μg/ml) at different time points. Meanwhile, various concentrations of RhB-cABs were added to the culture system for 12 hours. To explore the cell-specific uptake of cABs, BMMSCs, FBs, and T cells served as controls, and RhB-cABs (10 μg/ml) were added to the culture system for 3 hours. The uptake of RhB-cABs by macrophages was analyzed by CLSM and flow cytometry. In addition, PKH67-labeled FBs were cocultured with the unlabeled macrophages at a ratio of 1:1. After cell adhesion, RhB-cABs (10 μg/ml) were added to the medium for 2 hours. The uptake of RhB-cABs was observed by CLSM.
For observation of the uptake of cABs by macrophages in vivo, 100 μl of PKH26-AB ghosts, RhB-MSNs, and RhB-cABs (1 mg/ml) were injected into the cutaneous wound mice and colitis mice via the tail vein, with PBS as a control. After a set period of time, the mice were euthanized, and the local skin and colon were obtained and fixed with 4% PFA overnight at 4°C. After dehydration and embedding in optimal cutting temperature compound (OCT, Leica), frozen sections were prepared, and slides were stained with anti-F4/80 antibodies, followed by Hoechst staining. The fluorescence imaging was performed by CLSM.
Phenotype switching of macrophages
In vitro measurements showed that macrophages were stimulated with LPS (1 μg/ml) to induce inflammation before cABs treatment for 2 hours. Then, cABsnull (50 μg/ml), MSNsmiR-21, cABsmiR-21, MSNsCur, and cABsCur and equal amounts of miR-21/curcumin were added to the media for 4 hours. After refreshing the media, the cells were incubated for another 20 hours. For the in vivo measurements, the tissues were harvested to prepare frozen sections. After fixation and permeation with 0.1% Triton X-100 (room temperature, 10 min), the cells and tissues were blocked with normal goat serum (37°C, 1 hour) and stained with iNOS, CD206, or MerTK at 4°C overnight, followed by counterstaining with Hoechst for 15 min. Fluorescence imaging was performed with confocal microscopy. The protein levels (iNOS, Arg1, and CD206) were analyzed by Western blotting as mentioned above, and GAPDH or β-tubulin was used as the internal control. The levels of TNF-α, IL-6, TGF-β, and IL-10 in the cell culture supernatants or serum were determined using an ELISA kit according to the manufacturer’s recommended protocol (NeoBioscience, China).
The expression level of cytokeratin 14 was explored in the neoskin tissue surrounding the wound. Briefly, after blocking with normal goat serum, the skin sections were incubated with anti-cytokeratin 14 antibody at 4°C overnight. Then, the skin sections were incubated with the FITC-conjugated secondary antibody at 37°C for 1 hour. Afterward, the sections were counterstained with Hoechst for 15 min. The results were analyzed by CLSM.
In vivo proliferation analysis
To explore the cell proliferation rate in the neoskin, the expression level of Ki-67 was determined in the skin tissue. The skin tissue sections were permeated with 0.1% Triton X-100 for 15 min and blocked with normal goat serum at 37°C for 30 min. After incubation with anti–Ki-67 antibody at 4°C overnight, the sections were incubated with FITC-conjugated secondary antibody at 37°C for 1 hour. Before imaging, the cells were stained with Hoechst for 15 min. The fluorescence imaging was observed by CLSM.
At the endpoint, the cutaneous wound and colitis mice were euthanized, and the renascent skin, colons, and major organs were collected for H&E staining. Images were obtained with an Olympus BX41 Microscope (Olympus, Japan).
The histological score of colitis was graded according to the following criteria: (i) severity of inflammation = 0 (none), 1 (mild), 2 (moderate), 3 (severe); (ii) extent of inflammation = 0 (none), 1 (mucosa), 2 (mucosa and submucosa), 3 (transmural); (iii) crypt damage = 0 (none), 1 (1/3 damaged), 2 (2/3 damaged), 3 (crypt lost, surface epithelium present), and 4 (crypt and surface epithelium lost). The histological score was calculated by adding the scores of the three parameters.
Data are presented as means ± SD of at least triplicate measurements. Statistical analysis was performed by Student’s t test (two-tailed), one-way analysis of variance (ANOVA), or Kruskal-Wallis H test with the statistical software SPSS 19.0 (IBM, USA). Tukey’s post hoc test was used for multiple post hoc comparisons to determine the significance between the groups after one-way ANOVA. The difference between groups was considered statistically significant for *P < 0.05, very significant for **P < 0.01, and the most significant for ***P < 0.001. Graph analysis was performed using GraphPad Prism 7.00 (GraphPad Software, USA).
Acknowledgments: We thank B.Li., W.Liu., and C.Hu. for critical suggestion and guidance as the work was in progress. We are also grateful to X.Qiu. and L.Bao. for technical and editorial assistance. Funding: This work was supported by the National Key Research and Development Program of China (2016YFC1101400 to Y.J.), Young Elite Scientist Sponsorship Program by CAST (2017QNRC001 to Sh.L.), the National Natural Science Foundation of China (31800817 to Si.L. 81670915 to Z.D., 31870970 to J.Z., and 81601606 to X.C.), Natural Science Basic Research Program of Shaanxi (2018JM3026 to Sh.L.), and open fund of the State Key Laboratory of Military Stomatology (2017KA02 to X.C.). Author contributions: G.D., R.T., and X.L. contributed to the study design, data acquisition, and interpretation. P.Y. characterized properties of the materials. Q.Y. and Si.L. performed the animal experiments. J.L., J.Z., and Z.D. contributed to data analysis and interpretation. Y.J., Sh.L., and X.C. developed the concept and supervised experiments. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.