The Phenotypic Effects of Exosomes Secreted from Distinct Cellular Sources: a Comparative Study Based on miRNA Composition
Exosomes are nano-sized vesicles composed of lipids, proteins, and nucleic acids. Their molecular landscape is diverse, and exosomes derived from different cell types have distinct biological activities. Since exosomes are now being utilized as delivery vehicles for exogenous therapeutic cargoes, their intrinsic properties and biological effects must be understood. We performed miRNA profiling and found substantial differences in the miRNA landscape of prostate cancer (PC3) and human embryonic kidney (HEK) 293 exosomes with little correlation in abundance of common miRNAs (R2 = 0.16). Using a systems-level bioinformatics approach, the most abundant miRNAs in PC3 exosomes but not HEK exosomes were predicted to significantly modulate integrin signaling, with integrin-β3 loss inducing macrophage M2 polarization. PC3 but not HEK exosomes downregulated integrin-β3 expression levels by 70%. There was a dose-dependent polarization of RAW 264.7 macrophages toward an M2 phenotype when treated with PC3-derived exosomes but not HEK-derived exosomes. Conversely, HEK exosomes, widely utilized as delivery vehicles, were predicted to target cadherin signaling, with experimental validation showing a significant increase in the migratory potential of MCF7 breast cancer cells treated with HEK exosomes. Even widely utilized exosomes are unlikely to be inert, and their intrinsic activity ought to be assessed before therapeutic deployment.
KEY WORDSexosomes miRNA profiling phenotypic effects M2 polarization cadherin signaling
Exosomes are nano-sized vesicles secreted by many cells and are composed of a lipid bilayer, membrane proteins, and a hydrophilic core. They are highly versatile carriers of hydrophilic and hydrophobic molecules, proteins, and nucleic acids. While the phenotypic effects of exosomes can be modulated by loading with different therapeutic cargoes, the effects of their intrinsic properties, such as their protein, lipid, or nucleic acid complement, are often overlooked. However, these properties and their effects are of critical importance for clinical translation of exosomes as therapeutics. Furthermore, the overall action of exosomes most likely results from the combined effects of all rather than just a single component (1).
An intriguing aspect of exosome biology is their ability to transfer genetic materials to recipient cells in the form of non-coding RNAs, especially microRNAs (miRNAs) (2). A single miRNA can modulate the expression of hundreds of genes; in total, 30–70% of human genes are predicted miRNA targets through base-pairing of their “seed” sequences (3). Exosomes used to be viewed as a ‘garbage disposal’ system of unwanted cellular proteins and RNAs, but the current view is that exosomes act as natural delivery vehicles for nucleic acids and bioactive proteins that cells use to achieve coherent inter-cellular communication with recipient cells (1,4).
Exosomes seem to have evolved to deliver these bio-messages in vivo. They are relatively stable and protect their cargo from enzymatic degradation (5). Their surface lipids and integral membrane proteins can direct exosome uptake into specific cells (6,7). For therapeutic purposes, exosomes have been supplemented with application-specific cargoes across many disciplines including cardiology, neurology, and oncology (1). Generally, RNA interference strategies such as siRNA or miRNAs are used (8, 9, 10, 11, 12, 13, 14), but exosomes have also been loaded with small molecules (15,16). Additionally, exosomes have been genetically engineered to express specific targeting surface peptides to control their biodistribution and cellular uptake (8,10,14).
Current exosome manipulations do not completely alter their intrinsic properties. Given the heterogeneity of exosomes and extracellular vesicles and current limitations in characterizing subpopulations of released vesicles, total control over intrinsic exosome activity remains a challenge (1). This is an especially important consideration if the intrinsic biology of the exosomes is undesirable.
While it is now appreciated that exosomes display a wide range of biological effects, the underlying mechanisms are not fully understood. Cancer-derived exosomes can exert malignant phenotypes on recipient cells including promoting tumor cell proliferation, immune suppression or subversion, angiogenesis, and preparation of pre-metastatic niches (17, 18, 19, 20, 21, 22, 23, 24, 25). These effects are in stark contrast to the effects reported for exosomes from non-cancerous cells. Mesenchymal stem cell (MSC)-derived exosomes are mostly reported to offer tissue-regenerative effects in a host of settings ranging from protection after cerebral and myocardial ischemia to anti-fibrotic effects and restorative functions in damaged livers and kidneys (26, 27, 28, 29, 30, 31, 32). In the quest for scalable production of therapeutic exosomes, many researchers are utilizing easy-to-manipulate cells such as human embryonic kidney (HEK) 293 assuming that they represent a neutral phenotype (1). However, since HEK 293 represent a transformed cell line, it is not readily apparent if this assumption is valid (33).
The objective of this study was to determine how miRNA composition affects the biological functions of exosomes derived from different cellular sources using bioinformatics approaches to predict their biological effects and experiments to validate predictions. We compared the phenotypic effects of exosomes derived from two distinct cell types: the commonly utilized HEK 293 cells and PC3 prostate cancer cells. We first performed miRNA profiling and used bioinformatics analyses to predict which pathways are targeted by the miRNA cargoes of these two cell types. In vitro internalization assays were then used to determine the rate and extent of exosome uptake into various cell types before testing our in silico predictions in representative assays, including their effects on cell migration and macrophage polarization. Because the biological source of exosomes imparts them with intrinsic functional activities, this study emphasizes the importance of characterizing exosomes before clinical use.
MATERIALS AND METHODS
Prostate cancer cells (PC3) and human embryonic kidney cells (HEK) cells were purchased from ATCC. Forty-eight hours before exosome isolation, PC3 and HEK cells were cultured in exosome-free medium. To isolate exosomes, cell culture supernatants were collected and centrifuged at 2000×g for 30 min to remove cells and debris. Total exosome isolation reagent (Invitrogen, Carlsbad, CA) was added to the supernatant and incubated overnight at 4°C. To collect the exosomes, the mixture was centrifuged at 10,000×g for 1 h at 4°C and the exosome pellet was re-suspended in phosphate buffered saline (PBS), pH 7.4.
Nanoparticle Tracking Analysis
Exosome sizing and enumeration was determined by nanoparticle tracking analysis (NTA). Exosome samples were serially diluted in PBS. Dilutions corresponding to 10–500 μl of culture medium were generally in the range of the Nanosight LM10 instrument (Malvern Instruments, Malvern, UK) of 0.5 and 20E8 particles per milliliter. Each sample was tracked for 30 s using a camera setting of 16. Analysis was performed with the software version 2.3 using a gain of 6.0 and a threshold of 11 for all samples.
Zeta Potential Analysis
Exosome samples with a concentration of 2 μg/ml in PBS were diluted 1:50 in filtered deionized water and analyzed on the NanoBrook Omni. Each HEK exosome and PC3 exosome sample was read a total of three times using the phase analysis light scattering method to obtain the exosome zeta potential. Samples were allowed to stabilize for 5 min and reads were set to 30 s. The analysis was performed in triplicate.
Western Blot Analysis
Twenty-five micrograms exosomes by protein were incubated with 30 μl anti-human CD63 beads (Thermo Fisher Scientific, Waltham, MA). Exosomes were bound to the beads overnight at 4°C using a combination of tumbling and orbital shaking (650 rpm) to prevent settling. Exosome-bound beads were cleaned by magnetic separation according to the manufacturer’s instructions. Exosome beads were re-suspended in 10 μl of RIPA buffer and sonicated for 10 min. Exosomes were held at 4°C for an additional 15 min in RIPA buffer for complete lysis before being combined with 4 μl of 4× LDS buffer. Samples were heated to 95°C for 5 min and then separated on a 4–12% Bis-Tris NuPage gel using SDS-MOPS buffer. The transfer onto PVDF membrane was performed at 100 V for 60 min. Isolates were then confirmed as CD63+ by western blotting using 1:1000 BD anti-human CD63 (556019; Becton Dickinson, Franklin Lakes, NJ) overnight at 4°C.
Eight micrograms exosomes (quantified by protein) were combined with 30 μl of anti-human CD9 beads (Thermo Fisher Scientific). Exosomes were bound to the beads overnight at 4°C using a combination of tumbling and orbital shaking (650 rpm) to prevent settling. Exosome-bound beads were cleaned by magnetic separation according to the manufacturer’s instructions. Beads were then re-suspended to 100 μl in 1%-BSA-PBS and incubated with 1.5 μl of anti-human CD63-FITC (Miltenyi Biotech). Exosomes were stained for 1 h at room temperature with orbital shaking (450 rpm) in a glass test tube to prevent settling. Beads were analyzed on the MacsQuant® Analyzer 10 flow cytometer.
Exosomes from HEK and PC3 cells were collected and RNA isolated using Trizol reagent. miRNA profiling was performed using the NanoString platform (NanoString Technologies, Seattle, WA) as per the manufacturer’s instructions. nSolver was used to analyze the miRNA profiling data. miRNA reads were normalized to positive controls and the top 100 reads. Negative controls were used to determine the threshold for background. To assess reproducibility, duplicate biological samples were profiled.
miRNA Pathway and Network Analyses
The top 28 miRNAs identified in HEK exosomes and the top 18 miRNAs identified in PC3 exosomes were present at > 1% of the total read counts of all exosomal miRNAs and thus used for further analysis. Messenger RNA targets of the abundant miRNAs were identified with Diana (microT-CDS v5.0) using a filter threshold of greater than 0.7. Targeted pathways were analyzed by submitting the Diana predicted gene-targets to PANTHER v12.0 “statistical over-representation test” against the annotation data set “PANTHER pathway.” Statistically significantly targeted pathways were considered p < 0.05.
Exosomal Uptake and Flow Cytometry
To assess exosome uptake into HEK and PC3 cells, 50,000 MDA-MB-231, HUVEC, PC3, and RAW 264.7 cells plated in exosome-free media (DMEM supplemented with 10% exosome-free serum) were incubated with 2 μg of Bodipy-TR-labeled exosomes (Cat# D7540, Thermo Fisher Scientific). Stained exosomes were cleaned on Exosome Spin Columns (MW3000) according to the manufacturer’s instructions (Cat# 4484449, Thermo Fisher Scientific). Incubation periods of 0.5, 1, 2, and 4 h were used. Cells were detached by trypsin digestion and thoroughly washed with citric acid buffer to remove surface-bound but non-internalized exosomes using the method described in (34). Cellular uptake was analyzed using the MACSQuant® Analyzer (Miltenyi Biotech).
Macrophage Polarization Study
RAW 264.7 macrophages (3 × 106 cells) were plated onto a T-25 flask with DMEM and 100 ng/mL LPS for 6 h at 37°C. Cells were then detached with TrypLE Express Enzyme (Cat#: 2605036, Gibco, Thermo Fisher Scientific). RAW 264.7 cells were plated at 1.6 × 105cells/well of a 24-well plate with new medium. 4 μg, 16 μg and 64 μg/ml HEK and PC3 exosomes were added. After 48 h, RNA was collected according to the TRIzol protocol and cDNA was obtained using the NEB ProtoScript II synthesis kit. A three-step qPCR protocol using PerfeCTa SYBR Green SuperMix (Cat#:95056-02 K, Quantabio, Beverly, MA) was conducted utilizing GAPDH as a housekeeping gene. INOS (M1 marker), (FP 5′-TTTGCTTCCATGCTAATGCGAAAG-3′, RP: 5′-GCTCTGTTGAGGTCTAAAGGCTCCG-3′); TNF-α (M1 Marker), (FP: 5′-ACGGCATGGATCTCAAAGAC-3′, RP: 5′-AGATAGCAAATCGGCTGA-3′); arginase-1 (M2 marker): (FP: 5′-CAGAAGAATGGAAGAGTCAG-3′, RP: 5′-CAGATATGCAGGGAGTCACC-3′); MGL2 (M2 marker), (FP: 5′-AGCGGGAAGAGAAAAACCAG-3′, RP: 5′-AGATGACCACCAGTAGCAGGAG-3′); GAPDH, (FP: 5′-CAAGGTCATCCATGACAACTTTG-3′ RP: 5′-GTCCACCACCCTGTTGCTGTAG-3′).
Integrin β3 Targeting
To assess if macrophage polarization could be attributed to integrin-β3 targeting, as predicted, 2 × 105 macrophages were plated per well of a 24-well plate. Cells were treated in triplicate with 64.0 μg/ml of HEK or PC3 exosomes. As a positive control, 2 × 105 RAW 264.7 cells were electroporated in triplicate with 0.01 or 10 nM of Let-7b. After 48 h, RNA was collected according to the TRIzol protocol and cDNA was obtained using the NEB ProtoScript II synthesis kit. A three-step qPCR protocol using PerfeCTa SYBR Green SuperMix (Cat#:95056-02 K, Quantabio, Beverly, MA) was conducted utilizing GAPDH as a housekeeping gene. ITGB3: (FP: 5′-TGACATCGAGCAGGTGAAAG-3′, RP: 5′-GAGTAGCAAGGCCAATGAGC-3′).
Cell Viability Assay
To assess the effect of HEK and PC3 exosomes on cellular proliferation or death, PC3 and MDA-MB-231 cells were plated onto 96-well plates at 3000 cells/well and RAW 264.7 cells at 6000 cells/well. 24 h after plating, cells were serum starved for 24 h in DMEM. Then medium was replaced with exosome-free medium and incubated with increasing exosome concentrations for 24 h. After an additional 48 h, MTS substrate (Sigma Aldrich, St. Louis, MO) was added to the wells and incubated for 2 h at 37°C. The formazan dye produced by the viable cells was measured using a SpectraMax i3 plate reader at 490 nm.
Cell Migration Assay
MCF7 breast cancer cells were plated onto a 12-well plate at 300,000 cells per well with 1 ml DMEM with 10% FBS. Cells were treated with 4.0, 16.0, or 64.0 μg/ml of HEK or PC3 exosomes. As a positive control one well was treated with 10 nM of hsa-miR-374a-5p following the RNAiMAX protocol. After 24 h, the media were removed and replaced with DMEM without FBS to serum starve the cells, additional doses of exosomes and miRNA were added and cells were incubated for an additional 24 h. After a total of 48 h under exosome treatment and 24 h of serum starvation cells were detached and plated at 5000 cells per well of a 5-μm 96-well migration assay plate (Millipore ECM512). Cells were plated apically in DMEM and attracted to the basolateral chamber by 10% FBS in DMEM. After 24 h total migrated cells were collected and lysed according to manufacturer’s protocol and detected by CyQuant GR dye. MCF7 cells treated with 64.0 μg/ml of HEK or PC3 exosomes over 48 h were also analyzed for expression of E-cadherin, N-cadherin, and vimentin. RNA was collected according to the TriZol protocol and cDNA was obtained using the NEB ProtoScript II synthesis kit. A three-step qPCR protocol using PerfeCTa SYBR Green SuperMix (Cat#:95056-02 K, Quantabio, Beverly, MA) was conducted utilizing GAPDH as a housekeeping gene. The following primers were used: E-Cadherin, (FP: 5′-TGCCCAGAAAATGAAAAAGG-3′ RP: 5′-GTGTATGTGGCAATGCGTTC-3′), N-cadherin, (FP: 5′-GACAATGCCCCTCAAGTGTT-3′ RP: 5′-CCATTAAGCCGAGTGATGGT-3′), vimentin (FP: 5′-GAGAACTTTGCCGTTGAAGC RP: 5′-GAGAACTTTGCCGTTGAAGC-3′), GAPDH (FP: 5′-CAAGGTCATCCATGACAACTTTG-3′ RP: 5′-GTCCACCACCCTGTTGCTGTAG-3′).
Statistical analyses were performed using GraphPad Prism 7. Multiple comparisons were performed by one-way ANOVA followed by Tukey’s post hoc test. Statistically significant differences were marked with asterisks (*p < 0.05, **p < 0.01, ***p < 0.001).
Profiling Reveals Disparate miRNA Landscapes of HEK and PC3 Exosomes
We next considered that the most abundant miRNAs were likely mediators of the miRNA effects of HEK or PC3 exosomes. We defined high-abundance miRNAs as those representing greater than 1% of the total reads of all mapped miRNAs. High-abundance miRNAs present at greater than 1% of the total reads represented ~ 75% of miRNAs in HEK exosomes (28 total miRNAs) and ~ 80% of miRNAs in PC3 exosomes (18 total miRNAs). Several of the miRNAs found at high read counts in PC3 exosomes but not HEK exosomes or were present at very low numbers. For example, let-7b/i/g-5p, miR-100-5p, miR-222-3p and miR-29a-3p were present at high reads in PC3 but not HEK exosomes. Conversely, the second most abundant miRNA in HEK exosomes, miR-25-3p, was only found at low read counts in PC3 exosomes (Fig. 2h).
In Silico Analysis of HEK and PC3 Exosome miRNA Landscapes
Cell-Specific Uptake of Exosomes into Cells
Effects of Exosomes on Cell Viability
Exosomes from Different Cellular Sources Differentially Modulate Macrophage Polarization
Exosomes from HEK Cells and PC-3 Cells Modulate Cellular Migration
Using exosomes as a delivery vehicle for exogenous therapeutic cargoes represents an exciting new area of pharmaceutical research. Before exosomes can be translated into the clinic, however, their innate biological effects must be understood. Since miRNAs mediate many of the epigenetic “reprogramming” effects of exosomes on recipient cells, we performed miRNA profiling of two distinct cell types: a commonly used exosome-producing cell type (HEK 293 human embryonic kidney cells) versus a cancer cell type (PC3 prostate cancer cells). We then used in silico analyses to predict the biological effects of the most abundant miRNAs packaged in HEK 293 versus PC3 exosomes. To assess the predictability of our bioinformatics analysis, the most significant biological pathways for each type of exosomes were assessed experimentally. Our analyses indicated that PC3 exosomes contained an miRNA landscape that would most significantly target integrin signaling (p < 0.0001). Integrin signaling is known to affect macrophage polarization (44). HEK exosomes were not predicted to target integrin signaling, but were instead predicted to uniquely target cadherin-signaling, which plays a role in epithelial to mesenchymal transition and cellular migration.
With regard to integrin signaling, loss of integrin-β3 in macrophages promotes M2 polarization (45,46). Not only was integrin signaling the most significantly predicted pathway in PC3 exosomes but the most abundant miRNA in PC3 exosomes, let-7a-5p, targets and downregulates integrin-β3 expression (47). This prompted us to assess the effect of PC3 exosomes versus HEK exosomes in affecting M2 macrophage polarization. Confirming our in silico predictions, PC3 exosomes downregulated integrin-β3 to 30% of control. Additionally, PC3 but not HEK exosomes showed a dose-dependent ability to polarize macrophages to an M2 phenotype as determined by the expression patterns of several M2 and M1 markers. Commonly, the metabolic state of macrophages is used to understand if the cells are in an inflammatory M1-state, indicated by the high expression of iNOS and NO metabolism, or a reparative, immune-suppressive M2 state, indicated by the high expression of Arginase-1 and arginine metabolism (48). PC3 exosomes shifted RAW 264.7 macrophages toward an M2 phenotype with an 8.4-fold increase in the Arg1/iNOS ratio compared to HEK exosomes. PC3 exosomes polarized macrophages toward an M2 phenotype over HEK exosomes when additional M1 (TNF-α) and M2 (MGL2) markers were analyzed. Generally, we observed a higher M2/M1 ratio with Arginase-1 than with MGL2. This is in line with a recently published study that showed that Arginase-1 is the highest upregulated gene when macrophages are polarized toward the M2 phenotype (39).
This observation provides not only an important phenotypic characterization of prostate cancer exosomes, but also offers a biological explanation of how prostate cancer may mediate macrophage polarization in vivo. Macrophage polarization toward an M2 phenotype is a prominent and deleterious hallmark of many cancers including prostate cancer. Tumor-associated macrophages (TAMs) that adopt an M2 phenotype promote tumorigenesis through many mechanisms including increased angiogenesis, cancer cell proliferation via growth factor secretion, degradation of the extracellular matrix and release of chemokines, (which promote metastasis), and suppression of an immune response against the tumor (49). Conditioned medium from PC3 cells was previously shown to polarize RAW 264.7 macrophages to M2 (50), and it has also been shown that highly apoptotic RM-1 prostate cancer cells (treated with CoCl2 to achieve > 60% apoptosis) caused M2 polarization of bone marrow macrophages through MFG-E8 (51). The exact mechanisms of M2 polarization via non-apoptotic prostate cancer cells is, however, largely unknown.
Our findings are consistent with recent reports that posit a role for exosomes in polarizing macrophages via their miRNA cargoes. In pancreatic cancer, exosomes have been shown to mediate macrophage M2 polarization, and this polarization was modulated by the exosome miRNA content (52). Further, Saha et al. directly demonstrated that the miRNA cargo of exosomes from alcohol-treated monocytes polarized naive monocytes to M2 macrophages (53). Through miRNA-profiling and phenotyping of prostate cancer exosomes, we now suggest a miRNA-mediated effect of exosomes on macrophage polarization via downregulation of integrin-β3 in prostate cancer.
While it has been shown that cancer-cell derived exosomes exert malignant phenotypes and should probably be avoided as delivery vehicles, it is less clear how the miRNA landscapes of other cell lines impacts their biological activity and use as delivery vehicles. HEK cells, for example, are widely used as producer cells of therapeutic exosomes (1). While researchers are modifying and engineering HEK exosomes for drug delivery, it is challenging to completely remove their miRNA cargo. Thus, the intrinsic biological effects of HEK exosomes mediated by miRNAs would remain and could potentially contribute to undesired side effects.
The miRNA landscape of HEK exosomes indicated that they would target cadherin signaling, a pathway not represented in the predicted targets of PC3 exosome miRNAs. When MCF7 breast cancer cells were treated with HEK or PC3 exosomes, HEK exosomes significantly increased the expression of N-cadherin and Vimentin, two genes which are implicated in the metastatic potential of breast cancer (54,55). Furthermore, there was a trend toward reduction of E-cadherin in MCF7 cells treated with HEK exosomes; loss of E-cadherin is also associated with increases in breast cancer motility (56). Not surprisingly from these effects on cadherin genes, HEK exosomes did in fact exert a dose-dependent increase in the migratory potential of MCF7 cells, up to 2-fold over the control MCF7 cells. It appears that in some settings, such as macrophage polarization, HEK exosomes are relatively benign compared to PC3 exosomes, while in other settings HEK exosomes may cause undesired adverse effects. Conversely, although PC3 exosomes did not alter the expression of the assessed cadherin genes in MCF7, these cancer exosomes did promote the migration of MCF7 cells. PC3 exosomes were predicted to modulate integrin signaling, a pathway that is known to play a major role in the migratory potential of breast cancers (57). Additionally, PC3 exosomes were predicted to target PDGF signaling, which could also increase MCF7 motility (58). Further, PC3 exosomes contain hsa-miR-29b in high abundance, a miRNA that is present at very low read counts in HEK exosomes. In a study specific to MCF7 cells, transfection with hsa-miR-29b resulted in significant increases in the migratory ability of the cells (59).
In this study, exosomes were isolated using a precipitation-based method. Studies have shown that different routes of exosome isolation and RNA extraction can result in differential enrichment of different RNA species. While the composition of different RNA species, such as rRNAs and tRNAs, can vary dramatically when using different processing strategies, Tang et al. have shown that there is a high correlation between miRNA composition isolated by ultracentrifugation and precipitation-based exosome isolation methods (R2 ≥ 0.85) (60). In both commonly used methods, ultracentrifugation-based and precipitation-based isolation, exosomes are treated as bulk isolates. Future studies are warranted to assess if and how RNA composition differ between subpopulations of exosomes.
While many of the phenotypic effects of exosomes may be mediated by miRNAs, not all exosomal effects or behaviors are related to miRNAs. Other molecular characterizations such as proteomic and lipidomic studies should also be considered within a complete pipeline of exosome characterization (61). Also, the influence of other non-coding RNAs besides miRNAs, which are also present in exosomes, need further study (1). The ability of PC3 exosomes to target macrophages twice as efficiently as HEK exosomes could be due to differences in protein and lipid composition. For example, it has been shown that exosomes derived from different cells have different membrane proteins that promote specific uptake and internalization into cells (62). Further studies to understand how PC3 exosomes are taken up by macrophages could reveal strategies for engineering macrophage-targeted exosomes or synthetic delivery vehicles.
As exosomes are explored as delivery vehicles, their intrinsic biological activities cannot be ignored. Our data show that PC3 exosomes should likely not be used as a delivery vehicle in anti-neoplastic therapies where M2 macrophage polarization is known to exacerbate cancer progression. The use of HEK exosomes for drug delivery applications should also be critically evaluated, as HEK exosomes significantly modulated the expression of cadherin genes away from an epithelial phenotype and toward a more migratory mesenchymal phenotype.
Here we show that exosomes derived from different cellular sources show different intrinsic biological effects that can at least in part be explained by their miRNA composition. Our data suggest that prostate cancer exosomes modulate macrophage polarization and that HEK exosomes increase the migration of MCF7 cancer cells. These findings highlight that before exosomes are clinically employed as delivery vehicles, it is critical to understand their intrinsic biological activities without assuming inertness.
We acknowledge support by the NIH through awards R01EB023262 and R21EB021454. We thank Dr. Prashant Singh for performing the miRNA profiling for HEK and PC3 cells and exosomes. The authors declare no competing financial interests.
J.N. and S.F. conceived and designed the experiments. S.F. carried out the exosomal characterization, and cellular uptake experiment. S.K. carried out the bioinformatics analysis and MTS assay. C.J.L. performed the macrophage polarization study. M.D. provided M2 polarization insights and reviewed the manuscript. J.N. and S.F. analyzed the data. S.F., and J.N. wrote, reviewed and edited the manuscript. J.N supervised the project. All authors approved the final manuscript.
- 5.Gallo A, Tandon M, Alevizos I, Illei GG. The majority of microRNAs detectable in serum and saliva is concentrated in exosomes. Plos One. 2012;7(3).Google Scholar
- 6.Mulcahy LA, Pink RC, Carter DR. Routes and mechanisms of extracellular vesicle uptake. J Extracell Vesicles 2014;3.Google Scholar
- 29.Zhang Z, Yang J, Yan W, Li Y, Shen Z, Asahara T. Pretreatment of cardiac stem cells with exosomes derived from mesenchymal stem cells enhances myocardial repair. J Am Heart Assoc. 2016;5(1).Google Scholar
- 35.Tominaga N, Hagiwara K, Kosaka N, Honma K, Nakagama H, Ochiya T. RPN2-mediated glycosylation of tetraspanin CD63 regulates breast cancer cell malignancy. Mol Cancer. 2014;13.Google Scholar
- 38.Ohshima H. Encyclopedia of biocolloid and biointerface science. Hoboken, New Jersey: John Wiley & Sons, Inc.; 2016. volumes cm p.Google Scholar
- 44.Cha BH, Shin SR, Leijten J, Li YC, Singh S, Liu JC, et al. Integrin-mediated interactions control macrophage polarization in 3D hydrogels. Adv Healthc Mater. 2017;6(21). https://doi.org/10.1002/adhm.201700289.
- 52.Su MJ, Aldawsari H, Amiji M. Pancreatic cancer cell exosome-mediated macrophage reprogramming and the role of microRNAs 155 and 125b2 transfection using nanoparticle delivery systems. Sci Rep 2016;6:30110. https://doi.org/10.1038/srep30110.
- 53.Saha B, Momen-Heravi F, Kodys K, Szabo G. MicroRNA cargo of extracellular vesicles from alcohol-exposed monocytes signals naive monocytes to differentiate into M2 macrophages with increased phagocytic activity. Eur J Clin Investig. 2016;46:82–3.Google Scholar