Magnetic and Folate Functionalization Enables ... - ACS Publications

Dec 22, 2016 - In this targeted drug delivery system, magnetic targeting provides high local ... (26) The folate receptor (FR) is a glycosylphosphatid...
0 downloads 0 Views 2MB Size
Magnetic and Folate Functionalization Enables Rapid Isolation and Enhanced TumorTargeting of Cell-Derived Microvesicles Wei Zhang,†,‡,⊥ Zi-Li Yu,†,⊥ Min Wu,§ Jian-Gang Ren,† Hou-Fu Xia,† Guo-Liang Sa,† Jun-Yi Zhu,† Dai-Wen Pang,§ Yi-Fang Zhao,*,‡ and Gang Chen*,†,‡ †

The State Key Laboratory Breeding Base of Basic Science of Stomatology (Hubei-MOST) and Key Laboratory of Oral Biomedicine of Ministry of Education, School and Hospital of Stomatology, Wuhan University, Wuhan 430079, P. R. China ‡ Department of Oral and Maxillofacial Surgery, School and Hospital of Stomatology, Wuhan University, Wuhan 430079, P. R. China § Key Laboratory of Analytical Chemistry for Biology and Medicine (Ministry of Education), College of Chemistry and Molecular Sciences, and the Institute for Advanced Studies, Key Laboratory of Analytical Chemistry for Biology and Medicine, Wuhan University, 299 Bayi Road, Wuhan 430072, P. R. China S Supporting Information *

ABSTRACT: Cell-derived microvesicles (MVs), which are biogenic nanosized membrane-bound vesicles that convey bioactive molecules between cells, have recently received attention for use as natural therapeutic platforms. However, the medical applications of MV-based delivery platforms are limited by the lack of effective methods for the efficient isolation of MVs and the convenient tuning of their targeting properties. Herein, we report the development of magnetic and folate (FA)-modified MVs based on a donor cell-assisted membrane modification strategy. MVs inherit the membrane properties of their donor cells, which allows them to be modified with the biotin and FA on their own membrane. By conjugating with streptavidin-modified iron oxide nanoparticles (SA-IONPs), the MVs can be conveniently, efficiently, and rapidly isolated from the supernatant of their donor cells using magnetic activated sorting. Moreover, the conjugated magnetic nanoparticles and FA confer magnetic and ligand targeting activities on the MVs. Then, the MVs were transformed into antitumor delivery platforms by directly loading doxorubicin via electroporation. The modified MVs exhibited significantly enhanced antitumor efficacy both in vitro and in vivo. Taken together, this study provides an efficient and convenient strategy for the simultaneous isolation of cell-derived MVs and transformation into targeted drug delivery nanovectors, thus facilitating the development of natural therapeutic nanoplatforms. KEYWORDS: microvesicle, magnetic modification, drug delivery, cancer therapy, dual targeting

C

cancer sites, nanocarriers can passively accumulate in solid tumors.2,5 Moreover, the active tumor-targeting properties of a nanocarrier delivery system have been established using various modifications.6,7 These nanocarriers also protect their cargo from biodegradation during their journey to the targeted tissues, which results in increased in vivo stability and an extended half-time in circulation that largely facilitate their accumulation at tumor sites.8 Cell-derived microvesicles (MVs) are nanosized (100−1000 nm in diameter) membrane-bound vesicles that are secreted by nearly all cell types via direct budding from the plasma membrane,9,10 and these MVs can transport cargos including

ancer is the world’s most devastating disease and is responsible for several million deaths per year.1 Chemotherapy is one of the most widely used approaches for cancer treatment. However, the limited delivery efficiency to cancer sites reduces the therapeutic efficacy of chemotherapeutic drugs. Moreover, a major drawback of conventional chemotherapy is the nonspecific distribution of chemotherapeutic drugs in normal tissues,2 which damages healthy cells and results in severe systemic toxicity and undesirable side effects.3,4 Therefore, the development of a reliable, biofriendly, and targeted drug delivery system is highly desirable for cancer chemotherapy with improved delivery efficiency, which will result in the enhanced therapeutic outcomes and reduced side effects. Nanocarrier-based drug delivery systems have been rapidly developed in recent years. Due to the enhanced permeability and retention (EPR) effect caused by the leaky vasculature and poor lymphatic drainage in © 2016 American Chemical Society

Received: August 21, 2016 Accepted: December 22, 2016 Published: December 22, 2016 277

DOI: 10.1021/acsnano.6b05630 ACS Nano 2017, 11, 277−290

Article

www.acsnano.org

Article

ACS Nano

approved by FDA for medical applications, can self-assemble into the phospholipid bilayer.25 Based on this rationale, we propose a donor cell-assisted membrane modification strategy to obtain biotin and folate (FA)-modified MVs, which can be further magnetically functionalized by streptavidin conjugated iron oxide nanoparticles (SA-IONPs). To realize this strategy, donor cells for generating MVs were cultured in medium containing biotin-functionalized DSPE-PEG (DSPE-PEG-Biotin) and FA-functionalized DSPE-PEG (DSPE-PEG-FA). Because DSPE-PEG-Biotin and DSPE-PEG-FA in the culture medium can self-assemble into the plasma membrane, the membrane of donor cells will be modified by biotin and FA. MVs are formed by direct budding from the plasma membrane, thus MVs modified with biotin and FA on their membrane surface (FA/biotin-MVs) can be conveniently obtained. After conjugated with SA-IONPs, the FA/biotin-MVs were allowed to be separated from culture medium in magnetic field. The harvested MVs were then loaded with doxorubicin (DOX) using electroporation and exhibited excellent tumor targeting and therapy properties in a xenograft tumor model. In summary, we here successfully constructed a potential tumortargeted delivery system from cell-derived MVs based on donor-cell assisted membrane modification (Scheme 1): (1)

proteins, RNAs and DNAs from parent cells to recipient cells either locally or at a distance via circulation.11,12 These results have inspired researchers to transform these natural vesicles into drug delivery systems. MVs have multiple advantages compared to traditional artificially synthesized nanovectors. First, the phospholipid bilayer membrane of the MVs not only serves as a natural barrier to protect cargos from degradation during circulation but also improves the cellular internalization of the encapsulated drugs by direct fusion with target cell plasma membrane.13 Second, MVs have the potential to cross physiologic barriers and thus deliver drugs directly to the brain due to its natural origin.14,15 Third, supported by several previous studies including that from our group,16,17 the natural MVs possess intrinsic tumor targeting properties without any modification. Fourth, MVs possess biologic stabilization properties in circulation due to their nonartificial origin.18 Most importantly, MVs are biosafe because they are nearly nonimmunogenic when used autologously, which has been confirmed by several clinic trials. In contrast, the potential immunogenicity and toxicity of artificial delivery systems are inevitable.19 The rapid development of MVs as nanovectors is creating excellent prospects for their potential medical applications. However, problems have arisen. The lack of convenient and rapid methods for scalable isolation as well as the deficiencies in drug loading and unsatisfactory tumor-targeting capability significantly hamper the transformation of MVs into drug delivery vehicles.20 Magnetic nanoparticles combined with an external magnetic field have been used to modify the targeting properties of MVs, and this approach is known as magnetic drug targeting (MDT).21,22 By incubating magnetic materials with the parent cells, Silva and colleagues have loaded superparamagnetic nanoparticles into extracellular vesicles that can be easily concentrated into targeted tissues via an external magnetic field.22 Magnetic nanoparticles that are encapsulated in MVs can also be used as a contrast agent for magnetic resonance imaging (MRI) and hyperthermia treatment.23,24 In addition, the off-targeting effects of receptormediated targeting can be reduced by the large accumulation of vehicles at the tumor sites with the assistance of magnetic attraction. However, the uptake of magnetic nanoparticles and enclosure of these internalized magnetic nanoparticles into released MVs are nonselective and uncontrollable. Therefore, this “indirect encapsulation” strategy may easily result in low efficiency, unsatisfactory yield, and uneven outcomes, limiting its clinical applications.16 Immuno-magnetic isolation based on specific binding of the membrane markers of MVs has been developed to isolate natural vesicles20 and exhibits several advantages (i.e., simplicity and rapidity) compared with ultracentrifugation. However, this method is seriously hindered by the absence of well-defined MV markers, which may lead to a low yield of MVs due to the lack of general markers. Therefore, the development of a biofriendly, high-efficient, and universal strategy to overcome the previously mentioned disadvantages, especially for the scalable production of engineered MVs, is highly desirable. Herein, a separation and targeting strategy was developed to facilitate the transformation of MVs into targeted drug delivery system. It is well-known that the cellular plasma membrane consists mainly of phospholipids that assemble into a stable and sheet-like bilayer. A previous study has reported that the amphiphilic molecule 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-poly(ethylene glycol) (DSPE-PEG), which has been

Scheme 1. Schematic Illustration of the Design of FA/IONPMVs-DOXa

a

After the biotin and FA modification of donor cell (macrophage) membrane, the cell-derived FA and SA-IONP-modified MVs (FA/ IONP-MVs) were separated within a magnetic field. Then, FA/IONPMVs-DOX were generated by loading of the chemotherapeutic drug DOX into FA/IONP-MVs via electroporation. The targeting efficacy and antitumor effects of FA/IONP-MVs-DOX were estimated in a tumor-bearing mouse model.

Macrophages as donor cells are incubated in culture medium containing DSPE-PEG-Biotin and DSPE-PEG-FA to generate biotin and FA-modified MVs; (2) SA-IONPs are added to the supernatant for magnetic isolation of donor cell-derived MVs; (3) after loading with DOX, FA/IONP-MVs-DOX are created and used for in vitro and in vivo tumor targeting and therapy assays. In this targeted drug delivery system, magnetic targeting provides high local nanocarries concentration at the site of cancer, which is vital to receptor-mediated drug uptake by cancer cells and account for better therapeutic outcomes. 278

DOI: 10.1021/acsnano.6b05630 ACS Nano 2017, 11, 277−290

Article

ACS Nano

Figure 1. Characterization of FA/Biotin-MVs. (A) Quantification of FA and biotin on the membrane of unmodified MVs, DSPE-PEG-FAmodified MVs (FA-MVs), DSPE-PEG-Biotin-modified MVs (Biotin-MVs), and DSPE-PEG-FA/Biotin-modified MVs (FA/Biotin-MVs) was performed by flow cytometry. (B) Fluorescence photographs showed the colocalization of FA and biotin signals in FA/Biotin-MVs. (C) Hydrodynamic diameter and (D) zeta potential of the unmodified MVs, FA-MVs, Biotin-MVs, and FA/Biotin-MVs were detected by DLS.

RESULTS AND DISCUSSION Biotin and FA Modification of Macrophage Membrane. Based on the advantages of magnetic modification in MV separation and further utilization, we decided to engineer MVs with this attractive characteristic. Recently, we developed a method with high efficiency, reliable reproducibility, and excellent biocompatibility to modify MVs with biotin using a donor cell-assisted membrane modification strategy. In addition, we successfully tracked cell-derived MVs in vivo by coupling streptavidin-conjugated quantum dots (SA-QDs).16 Here, we proposed biotinylated MVs for use as a platform for efficient conjugation of natural MVs with SA-IONPs based on the high affinity between biotin and SA. This process would endow the MVs with a magnetic property. Moreover, by using the same strategy, the MVs could be equipped with other targeting biomolecules. Folate (folic acid, FA) was modified on the surface of MVs. FA is a non-immunogenic water-soluble B vitamin, which is stable over a broad range of temperatures and pH values.26 FA can induce cellular internalization through the endocytotic pathway when it binds to its receptor on the cell surface.26 The folate receptor (FR) is a glycosylphosphatidylinositol anchored cell surface receptor that is overexpressed in various cancer types, such as cervical, ovarian, breast, lung, and brain tumors. However, the expression of FR is limited in normal tissues and cells.27 Thus, we proposed that FA modification could endow the MVs with more precise tumor targeting property. To obtain the modified MVs, functionalized

parent cells must be developed. We here selected macrophages as parent cells because macrophages can be recruited by the cytokines released from tumor tissues,28 and macrophagederived MVs were believed to inherent this intrinsic tumor targeting property. Therefore, THP-1-derived macrophages were incubated with DSPE-PEG-Biotin (0−50 μg/mL) and/or DSPE-PEG-FA (0−5 μg/mL, 1/10 of DSPE-PEG-Biotin) to achieve biotin- and/or FA-functionalized parent cells. No obvious cytotoxicity of DSPE-PEG-Biotin and DSPE-PEG-FA was detected from the cell viability analysis and cell morphology observation (Supporting Information Figures S1 and S2). The time-dependent membrane modification efficiency of biotin and FA was evaluated by flow cytometry (Figure S3), and the results demonstrated that both biotin and FA can be efficiently and simultaneously entrenched in macrophages, yielding 33.1%, 74.3%, 87.1%, 90.5%, 98.2%, 98.5%, and 98.4% double positive macrophages (denoted as FA/Biotin) after incubation for 6 h, 12 h, 24 h, 2 days, 4 days, 6 days, and 8 days, respectively. The excellent modification efficacy was further confirmed by fluorescence microscopy observations (Figure S4). The colocalization of FA and biotin was confirmed using a magnetophoresis assay. As shown in Figure S5, after incubation with SA-IONPs for 30 min, the macrophages moved toward and accumulated around the magnet cylinder (diameter = 1.5 mm, height = 1.0 mm, 0.4 T). In addition, the presence of FA and the cell nucleus were confirmed by the red fluorescence emitted by the FA antibody 279

DOI: 10.1021/acsnano.6b05630 ACS Nano 2017, 11, 277−290

Article

ACS Nano

Figure 2. Magnetic separation and characterization of FA/IONP-MVs. (A) Bright-field and fluorescence images of unmodified MVs, DSPEPEG-FA-modified MVs (FA-MVs), DSPE-PEG-Biotin-modified MVs (Biotin-MVs), and DSPE-PEG-FA/Biotin-modified MVs (FA/BiotinMVs) in the presence of a magnet cylinder (diameter = 1.5 mm, height = 1.0 mm, 0.2 T) after incubation with SA-IONPs. The red fluorescence emission is the signature of FA embedded in the membrane of MVs. (B) Digital photographs showed the operating procedures of the magnetic separation of FA/Biotin-MVs. (1) SA-IONPs used for magnetic separation; (2) culture medium harvested from the DSPE-PEGFA/Bitoin treated macrophages; (3) incubation with SA-IONPs (40 μg/mL) for 30 min; (4) magnetic separation in the presence of magnet (100 × 50 × 20 mm, 0.6 T); (5) SA-IONPs-conjugated FA/Biotin-MVs (FA/IONP-MVs) were sorted on the wall of the bottle; and (6) resuspension of the separated FA/IONP-MVs in PBS. Blue arrows in (4) and (5) indicate the separated FA/IONP-MVs. (C) TEM images of SA-IONPs, MVs, FA-MVs, IONP-MVs, and FA/IONP-MVs and SEM image of SA-IONPs. The black arrows indicate IONPs on the surface of MVs. (D) Distribution of the iron load of FA/IONPs-MVs, which was analyzed according to the results of the magnetophoresis assays. (E) Distribution of the number of the IONPs conjugated on MVs, which was calculated based on the TEM images.

and the blue fluorescence emitted by DAPI, respectively. In addition, we also tested the membrane modification of primary macrophages that were separated from human peripheral blood samples. As shown in Figure S6, after incubation with DSPEPEG-Biotin/FA for 2 days, the data from flow cytometry demonstrated that more than 90% of primary macrophages were double positive for biotin and FA staining. Moreover, primary macrophages successfully modified with biotin and FA were also confirmed under fluorescence microscope. Generation, Purification and Characterization of FA/ Biotin-Modified MVs. Once the macrophages were incubated with DSPE-PEG-Biotin and/or DSPE-PEG-FA for 2 days, the cells were stressed by culture in a serum-free medium for an additional 2 days to trigger the release of MVs. Then, MVs were purified from the conditioned medium using a previously reported differential centrifugation protocol.16 To assess the efficiency of MV membrane modification, the purified MVs were incubated with biotin and FA antibodies for 30 min. After removing the excess free antibodies via sucrose gradient centrifugation, the MVs were analyzed via flow cytometry and observed via fluorescence microscopy. FA and biotin were efficiently and concurrently attached to the MVs, yielding 92% double positive MVs (denoted as FA/Biotin-MVs) (Figure 1A), which was further confirmed by observation using

fluorescence microscopy. As shown in Figure 1B, nearly all of the red biotin signals were colocalized with the green FA signals in these MVs. As expected, only biotin or FA signals were detected in the MVs that were purified from the supernatants of macrophages incubated with DSPE-PEG-Biotin or DSPE-PEGFA, respectively. The results from flow cytometry and fluorescence observation also indicated that additional modification of FA did not influence the embedding efficacy of biotin. Good biocompatibility during membrane modification is very important for the potential application of MVs. Therefore, the influence of FA and/or biotin modification on the physical properties and biochemical compositions of the MVs was evaluated. The results from dynamic light scattering (DLS) analysis demonstrated that the hydrodynamic diameter (Figure 1C) and zeta potential (Figure 1D) of the FA-MVs, BiotinMVs, and FA/Biotin-MVs were nearly unchanged compared to the unmodified MVs. Real-time quantitative PCR and Western blot analyses further indicated that MV membrane modification did not influence the mRNA or protein expressions of characteristic molecules in the MVs (Figure S7). These results suggested that the MVs could be efficiently functionalized with biotin and FA in a biofriendly fashion via the membrane modification of donor cells. 280

DOI: 10.1021/acsnano.6b05630 ACS Nano 2017, 11, 277−290

Article

ACS Nano

Figure 3. Biocompatibility of FA/IONP-MVs in vitro and in vivo. (A) Proliferation of HeLa cells after incubation with MVs or FA/IONP-MVs (20 μg/mL, 106 /mL) for 0−72 h. (B) Viability of HeLa cells after exposure to increased concentrations of MVs or FA/IONP-MVs for 24 h. (C) Body weights of mice after intravenously injected with MVs or FA/IONP-MVs (100 μL, 1 mg/mL, 5 × 107/mL). (D) The hepatic and renal functions of mice after injection of MVs or FA/IONP-MVs were determined by blood chemistry tests. A/G, albumin/globin ratio; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BUN, blood urea nitrogen; CRE, creatinine. (E) Whole blood cell analysis of mice after injection of MVs or FA/IONP-MVs. WBC, white blood cell; Lymph, lymphocyte; Mon, monocyte; Gran, granulocyte; RBC, red blood cell; HGB, hemoglobin; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin; PLT, platelets; MPV, mean platelet volume; PDW, platelet distribution width; PCT, plateletcrit. (F) H&E staining of major organs harvested from mice injected with PBS, MVs, and FA/IONP-MVs, respectively.

Magnetic Labeling and Separation of FA/Biotin-MVs. The FA- and biotin-modified MVs (FA/Biotin-MVs) were purified by ultracentrifugation and then incubated with FA antibodies and SA-IONPs. The colocalization of the FA and IONPs was evidenced via a magnetophoresis experiment. As shown in Figure 2A, the FA/IONP-MVs that moved toward and accumulated around the magnet cylinder (diameter = 1.5 mm, height = 1.0 mm, 0.4 T) also exhibited red fluorescence due to emission by the FA antibodies. The results were also confirmed by flow cytometry analysis (Figure S8). Based on the magnetic susceptibility of FA/IONP-MVs, magnetic separation was developed to simplify the MV isolation and facilitate their clinical applications. First, conditioned medium was collected

and centrifuged at 2000 g for 20 min to eliminate cell debris and apoptotic bodies. Next, SA-IONPs were added to the supernatant (Figure 2B). Then, the mixture was incubated at 37 °C for 30 min to allow the specific binding of streptavidin to biotin. The SA-IONP-labeled MVs were sorted in a glass bottle within a magnetic field (0.6 T). After removing the supernatant in the presence of an external magnetic field, the IONP-labeled MVs were collected and then resuspended in PBS (denoted as FA/IONP-MVs). The separated MVs were detected using transmission electron microscopy (TEM), and the results indicated that the spherical vesicles were surrounded by IONPs (i.e., dark spots) (Figure 2C). The magnetically separated MVs exhibited no significant difference in their hydrodynamic 281

DOI: 10.1021/acsnano.6b05630 ACS Nano 2017, 11, 277−290

Article

ACS Nano

Figure 4. Tumor targeting of FA/IONP-MVs in vitro and in vivo. (A) Fluorescence imaging for the uptake of CFSE-labeled MVs and FA/ IONP-MVs by HeLa cells after 2 h incubation. To evaluate the effect of magnetic targeting, the cell culture plates were placed on a permanent magnet (100 × 50 × 20 mm, 0.6 T) during the first hour of incubation (denoted as +MF). Free FA (1 mM) was added to compete with the performance of FA-mediated targeting (denote as +FA). (B) Quantitative analysis of CSFE-labeled MVs and FA/IONP-MVs per cell. (C) NIR fluorescence images of DiR-labeled MVs, IONP-MVs, and FA/IONP-MVs in microcentrifuge tubes. (D) Top: In vivo NIR fluorescence images of mice injected with DiR-labeled MVs, IONP-MVs, and FA/IONP-MVs (100 μL, 1 mg/mL, 5 × 107/mL) at the indicated time points after injection. In IONP-MVs plus MF and FA/IONP-MVs plus MF groups, the permanent magnets (100 × 50 × 20 mm, 0.6 T) were attached to the tumor site. Bottom: Ex vivo fluorescence images of major organs and tumors excised from mice intravenously injected with DiR-labeled MVs, IONP-MVs, and FA/IONP-MVs for 4 h. (E) Quantification analysis of the fluorescence signals measured from the major organs and tumors 4 h after injection with DiR-labeled MVs, IONP-MVs, and FA/IONP-MVs. (F) The mass of iron in the organs and tumors was measured by ICP-AES. *, p < 0.05; **, p < 0.01.

interest, these CD63 positive exosomes were negative for biotin and FA antibody labeling (Figure S10), indicating that, at least in 2 days, the treatment of donor cells with DSPE-PEG-Biotin/ FA only modified the membrane structure of MVs but not exosomes. The iron mass and the number of IONPs that were conjugated on the membrane of FA/Biotin-MVs were analyzed quantitatively according to the results from the magnetophoresis assays (Figure 2D) and TEM observation (Figure 2E), respectively. By monitoring of a single vesicle trajectory toward the magnet and calculating the magnetic mobility, the distribution of iron load in the vesicles was estimated. The

diameter and zeta potential compared to those separated using ultracentrifugation (Figure S9). And the results from real-time quantitative PCR and Western blot analyses indicated that no significant differences were observed between the MVs separated by these two distinct methods (Figure S7). Moreover, whether exosomes, another type of extracellular vesicles, could be functionalized and separated based on the present strategy was explored. The macrophages were treated with DSPE-PEG-Biotin/FA for 2 days, and then exosomes were harvested and detected using flow cytometry based on a CD63Dynabeads-dependent manner as previously reported.29 Of 282

DOI: 10.1021/acsnano.6b05630 ACS Nano 2017, 11, 277−290

Article

ACS Nano

magnetic targeting property that may be useful in further experiments. Biocompatibility of FA/IONP-MVs in Vitro and in Vivo. To assess the biocompatibility of FA/IONP-MVs in vitro, cell viability and proliferation assays were performed. As shown in Figure 3A, the growth activity of HeLa cells was not significantly altered by incubation with FA/IONP-MVs for 72 h, suggesting the absence of a substantial proliferationpromoting effect on the tumor cells. The results of cell viability assays indicated that FA/IONP-MVs exhibited no cytotoxicity when incubated with HeLa cells for 24 h, even at the high concentration of 40 μg/mL (Figure 3B). To further evaluate the biosafety of FA/IONP-MVs in vivo, FA/IONP-MVs were injected into mice via the tail vein. The body weights of the mice from FA/IONP-MVs- and MVstreated groups did not differ significantly from those of the control group (Figure 3C). Three weeks after injection, the mice were euthanized for blood biochemistry and hematology analyses as well as histological examinations. Sex-, age- and weight-matched healthy mice were used as controls. The indicators for liver and kidney functions did not change significantly, suggesting that intravenous injection of FA/ IONP-MVs did not lead to reductions in hepatic or renal functions (Figures 3D and S11). No statistically significant differences in the blood cells, hemoglobin, and platelets indicators were determined from the hematology analyses (Figure 3E and S12). In addition, the heart, liver, spleen, lungs, and kidneys were harvested for H&E staining. Apparent organ injury and inflammation changes were not observed compared with those in the control group, indicating negligible histological toxicity. These results demonstrated that our proposed FA/IONP-MVs were biocompatible as nanocarriers without significant side effects in vitro and in vivo. Cancer Cell Targeting of FA/IONP-MVs in Vitro. The intracellular localization of FA/IONP-MVs and their quantitative cellular uptake in HeLa cells were evaluated using fluorescence microscopy. The HeLa cells were visualized as red fluorescence after cell staining with Cellmask, and the MVs that were labeled with CFSE exhibited green fluorescence. As shown in Figure 4A,B, both the unmodified MVs and FA/ IONP-MVs were readily taken up by HeLa cells after 2 h of incubation, and FA/IONP-MVs exhibited significantly enhanced cell internalization compared to that of the unmodified MVs (16.1 vs 6.4 MVs per cell). This enhancement was abrogated by coincubation of the HeLa cells with 1 mM free FA for competitive inhibition (8.8 vs 16.1 MVs per cell), suggesting that the FA/FR system contributed to cellular internalization of the modified MVs. Moreover, when the HeLa cells were incubated with FA/IONP-MVs and 1 mM free FA within a magnetic field, significantly higher cellular internalization was observed compared to that in the absence of a magnetic field (23.4 vs 8.8 MVs per cell). As expected, the largest increase in the cell uptake was observed for HeLa cells incubated with FA/ IONP-MVs within a magnetic field and without free FA for competition (32.2 MVs per cell), thus demonstrating the superior targeting properties of the FA/IONP-MVs. Moreover, the mass of iron of FA/IONP-MVs taken up by the HeLa cells was measured using an iron assay kit according to the manufacture’s instruction. The mass of iron in the control HeLa cells and the HeLa cells treated with nonfunctionalized MVs was 0.88 and 0.87 pg/cell, respectively. Increased iron mass of 0.51, 0.31, 0.64, 1.03 pg/cell was estimated in the FA/ IONP-MVs, FA/IONP-MVs plus free FA, FA/IONP-MVs plus

results showed that the biotin- and FA-modified MVs contained an iron mass of 0.91 ± 0.35 fg. And the number of IONPs on each MVs was 22.8 ± 8.8, as the iron mass for each particle was approximately 0.04 fg according to the manufacture’s instruction. By counting the IONPs around MVs from TEM observation, the number of IONPs on each MV was 25.5 ± 9.7, which is consistent with the results from magnetophoresis assays. Differential centrifugation is the most common method employed to isolate MVs. However, several key disadvantages cannot be ignored, especially for large-scale isolation, which is required for clinical applications. First, the current separation protocols of MVs that are based on differential centrifugation produce a heterogeneous mixture of apoptotic bodies, MVs and exosomes. Due to the overlap in the vesicle size of exosomes, MVs, and apoptotic bodies, the isolation of pure MVs from exosomes and apoptotic bodies is impossible via differential centrifugation.20 In addition, these components may induce adverse effects and reduce the therapeutic efficacy. Second, protein aggregates could potentially pretend to be MVs and may not be separated from MVs when using the currently available MVs purification techniques.30 The biological performance of protein aggregate-contaminated MVs may differ from that of the purified MVs, which results in heterogeneous outcomes. Third, the high shear force during ultracentrifugation may induce the aggregation and rupture of MVs,20 which may result in a lower yield and decreased delivery efficacy of the MVs. Fourth, ultracentrifuges are required for the ultracentrifugation of MVs, and these instruments are very expensive and consume a great deal of space, which prevents many laboratories from acquiring them. Fifth, differential centrifugation is also tedious and time-consuming. According to the standard protocol for MV isolation, at least 4 h is required. Therefore, immuno-magnetic isolation based on specific binding between the antibodies and the receptors on the surface of extracellular vesicles (EVs) may be suitable. This method relies on the specific biomarkers expressed on the membrane of the EVs, such as exosomes, which have general biomarkers on their surface (e.g., CD63). However, no ubiquitous biomarkers on the MV surface could be used to distinguish them from other EVs. Therefore, based on immunomagnetic isolation, only a portion of the MVs can be separated, which decreases the yield and affects the biological behaviors of the isolated MVs.18 In this study, we introduced membrane modification-mediated magnetic separation of MVs, and this approach possesses several important advantages over ultracentrifugation and immuno-magnetic isolation. Because the magnetic nanospheres were conjugated onto the surface of the MVs, the purified FA/biotin-MVs were easily separated from the culture medium without contamination of the protein aggregate and exosomes, as evidenced directly by the TEM (Figure 2C) and DLS (Figure S9) results that the size range of the magnetically isolated vesicles are from 300 to 1000 nm. In addition, the moderate operation conditions may preserve the original structure of the MVs by preventing aggregation and rupture. Moreover, this simple and convenient strategy consists of only one step and takes