Such inhibitors include imipramine, GW4869, cambinol, and spiroepoxide as depicted in Figure 2. GW4869, cambinol, and spiroepoxide selectively inhibit neutral sphingomyelinases (nSMase), more specifically neutral sphingomyelinase 2 (nSMase2); whereas imipramine targets aSMase. Each inhibitor may have a different specific mechanism, but they have all been proven to reduce exosome secretion, to varying levels of efficiency. For example, imipramine, an aSMase inhibitor, was found to block the release of exosomes and MVs from glial cells (Bianco et al., 2009). aSMases catalyze the hydrolysis of sphingomyelin to ceramide, which is vital to the release of exosomes and the formation of MVs. When MVs are shed, aSMases are translocated to the cell membrane; once internalized in cells, imipramine acts to proteolytically degrade aSMase. aSMase then detaches from the cell membrane (Catalano and O’driscoll, 2020). A study by Kosgodage et al. (2017) proved that 25 μM imipramine lessened total EV release (both exosomes and MVs) by 77% in PC3 (prostate cancer) cells although the separate effects of imipramine on exosomes and on MVs was not tested (Kosgodage et al., 2017). Imipramine may also be used to inhibit macropinocytosis.

nSMase2 inhibitors block ceramide-mediated exosome biogenesis. GW4869 is a frequently used inhibitor which targets nSMase2 to prevent the ceramide-modulated inward budding of MVBs and the subsequent release of exosomes from MVBs (Catalano and O’driscoll, 2020). GW4869 was used to suppress chemoresistance in colorectal, pancreatic and ovarian cancer cells by reducing exosome secretion (Hu et al., 2015; Cao et al., 2017; Richards et al., 2017). Cambinol, a lipid molecule, binds to the active site of nSMase2 (Figuera-Losada et al., 2015). While its specific nSMase2 inhibitory effect on cancer-derived EVs has not been studied, on the basis of its additional inhibition of NAD-dependent deacetylase activity of cell survival enzymes, it has been found to have antitumorigenic effects on lung cancer cells from the cell line NCI-H460 as well as sensitize those cells to chemotherapy drugs such as DNA-damaging etoposide and tubulin-targeting paclitaxel (Heltweg et al., 2006). Spiroepoxide specifically and non-competitively inhibits nSMase2 (Phuyal et al., 2014) although its ability to diminish exosome secretion in cancer cells has not been documented. GW4869 and spiroepoxide have both been found to be non-toxic to cells (Phuyal et al., 2014). It must be noted that while nSMase2 inhibitors have been proven to reduce exosome release in cancerous cell lines as well as some non-cancerous cell types such as HEK293 cells, a human kidney cell line (Kosaka et al., 2010), they do not affect exosome secretion at all in other cell types such as PC3, a prostate adenocarcinoma cell line (Phuyal et al., 2014). The efficiency of EV inhibitors varies between different cell lines and cell types.

These include bisindolylmaleimide-I and d-pantethine. Bisindolylmaleimide-I, an inhibitor targeting the ATP-binding site of various protein kinase C isoforms, hinders the release of calcium and the externalization of PS, which are known steps in the mechanism driving the release of MVs (Kosgodage et al., 2017). Stratton et al. (2015) detailed that bisindolylmaleimide-I inhibited MV release in PC3 cells by 75% compared to the control, despite intracellular calcium concentrations being increased using the sublytic membrane attack complex (MAC) (Stratton et al., 2015). Meanwhile, d-pantethine, a derivative of vitamin B5, is needed in the production of coenzyme A. It strongly reduces the total levels of cholesterol in cells via its lipid-metabolizing activity (Kosgodage et al., 2017). It also diminishes the shedding of MVs (Kosgodage et al., 2017). Similarly to bisindolylmaleimide-I, it inhibits the translocation of PS (Martin et al., 2016) and therefore prevents microvesiculation. However, d-pantethine severely reduces cell viability by nearly 80% in PC3 cells (Kosgodage et al., 2017).

Inhibitors targeting ATP-binding cassette (ABC) transporters include glyburide (glibenclamide) and indomethacin. Glyburide, a drug for diabetes, inhibits the ATP-sensitive K⁺ channel of an ABC transporter participating in the release of MVs (Catalano and O’driscoll, 2020). Glyburide non-specifically interacts with the SUR receptor (an ABC transporter) and other proteins such as ABCA1, which regulates cellular cholesterol and phospholipid concentrations (Liu and Tang, 2012). Since cholesterol is vital to the release of both MVs and exosomes, glyburide should be able to have both exosome-inhibiting and MV-inhibiting abilities. Even so, Kosgodage et al. (2017) showed that glyburide did not affect MV biogenesis in PC3 and MCF-7 cells (Kosgodage et al., 2017) and its effect on exosomes was not tested in the study. In comparison, indomethacin, an anti-inflammatory drug, specifically downregulates the transcription of ABCA3, which aids in the transport of lipids (Aung et al., 2011). This explains why indomethacin was able to act as a chemosensitizer for lymphoma cell lines SU-DHL-4, OCl-Ly1 and OCl-Ly3 (Koch et al., 2016). Prior to treatment with indomethacin, the cells encapsulated chemotherapy drugs doxorubicin and pixantrone in exosomes which were then effluxed from cells (Koch et al., 2016). The application of 10 μM indomethacin to the lymphoma cells allowed the drugs to accumulate within their nuclei and exert cytostatic effects (Koch et al., 2016). Although there are many ABC transporter inhibitors such as pelitinib and vatalanib have been identified to sensitize resistant cancer cell lines, e.g., lung and colon cancer cell lines successfully (To et al., 2015a, b), but the study on the effect of these inhibitors in overcoming EV-mediated drug resistance is limited.

Simvastatin, a HMG-CoA reductase inhibitor, prevents the synthesis of cholesterol; since cholesterol is an integral constituent in endosomal membranes, this reduced exosome secretion in epithelial cells and monocytes by 40% (Kulshreshtha et al., 2019). However, simvastatin has not yet been tested with cancerous cell lines. Simvastatin (but not GW4869) was further shown to decrease intracellular concentrations of proteins associated with exosomes such as ALIX, CD63, and CD81 (Kulshreshtha et al., 2019). Simvastatin also inhibits endocytosis (specifically, CIE) dose-dependently (Costa Verdera et al., 2017).

Cytochalasin D is a fungal toxin which binds to actin filament edges, preventing actin polymerization (Catalano and O’driscoll, 2020). Actin polymerization is vital in microvesiculation as well as the trafficking of MVBs toward the cell membrane; thus, cytochalasin D can inhibit EV release. It was proven that HeLa (human cervical adenocarcinoma), Panc 1 (pancreatic carcinoma), PC3, and A293 (human embryonic kidney epithelial) cells secreted exosomes carrying survivin, an anti-apoptotic protein (Khan et al., 2011). It was shown that treatment with cytochalasin D treatment decreased the release of exosomes in those cell lines, therefore reducing the concentration of survivin in the tumor environment (Khan et al., 2011). Cytochalasin D may be capable of inhibiting macropinocytosis, as macropinocytosis also involves actin-dependent pathways (Costa Verdera et al., 2017). However, its effect is not specific in targeting endocytosis mechanisms and its efficacy between different cell lines vary (Dutta and Donaldson, 2012).

Calpeptin is currently the most well-characterized calpain inhibitor. Calpains are cytosolic proteases and their activated forms participate in cellular activities including cell cycle progression, cell motility and cytoskeleton remodeling (Jorfi et al., 2015). Calpains are deregulated in cancer cells and thus play a role in cancer progression, apoptosis and migration. As per Figure 3, calpains promote the shedding of MVs through cytoskeletal rearrangement; thus, calpain inhibitors such as calpeptin can reduce the amount of MV shedding by cells, and they reduce cell proliferation as well. Jorfi et al. (2015) studied the chemosensitizing effect of calpeptin inhibition on the PC3 cell line, where the level of cell proliferation and apoptosis due to docetaxel treatment increased substantially in comparison to PC3 cells which did not receive calpeptin (Jorfi et al., 2015).

Chloramidine (Cl-amidine) is a calcium chelator which irreversibly binds to peptidylarginine deiminase (PAD), consequently causing post-translational protein deimination. PAD-modulated deimination of actin in the cytoskeleton is needed in order for MVs to form (Catalano and O’driscoll, 2020). Kosgodage et al. (2017) demonstrated the ability of Cl-amidine to reduce the release of exosomes as well as MVs (Kosgodage et al., 2017), showing the involvement of PAD in exosome release. Cl-amidine may inhibit EV uptake as well. It has been shown to target both CIE and CDE (Vercauteren et al., 2010) although this targeting is unspecific (Dutta and Donaldson, 2012).

Another inhibitor, NSC23766, acts on Rac1, a Rho GTPase which is involved in cytoskeletal remodeling (Catalano and O’driscoll, 2020). In cancer cells, activated Rac1 (bound to GTP) contributes to chemoresistance by contributing to angiogenesis and by binding to regulators of apoptosis (such as Bcl-2) which arouses anti-apoptotic responses from the cell (Liu et al., 2019). NSC23766 blocks the activation of Rac1 via guanine nucleotide exchange factors (GEF) Trio and Tiam1 (Catalano and O’driscoll, 2020). While NSC23766 has been shown to strongly reduce MV secretion in non-cancerous human cells (Wang et al., 2017), though it does not appear to have a significant effect when used alone in cancer cells without a second EV inhibitor (Wu et al., 2019). It may also function as an inhibitor of macropinocytosis (Hernandez et al., 2010).

Y27632 is a competitive inhibitor of ROCK1 and ROCK2, cytoskeleton-interacting members of the ROCK family (Catalano and O’driscoll, 2020). Y27632 has been used to demonstrate the relation of ROCK1 and ROCK2 to MV formation (Hartmann et al., 2015). ROCK1 and ROCK2 activate LIMK and MYLK (Li et al., 2012), which affect cofilin and myosin to stimulate cytoskeleton re-organization and the contraction of actin filaments (Kosgodage et al., 2017), both activities of which are integral to the process of microvesiculation. Y27632 competes with ATP to bind ROCK1 and ROCK2 catalytic binding sites (Catalano and O’driscoll, 2020). Y27632 acts significantly on MV-sized vesicles, reducing their secretion by 67% in PC3 cells and decreasing exosome-sized secretion by only 3% (Kosgodage et al., 2017). In a study by Li et al. (2012), tissue transglutaminase immunohistochemical staining of MDA-MB-231 (breast adenocarcinoma), U87 (glioblastoma) and HeLa cells following their treatment with Y27632 showed that MVs were not visible at their surfaces (Li et al., 2012) and the medium conditioned by Y27632-treated MDA-MB-231 and U87 cells contained a significantly lower level of MVs compared to the level prior to Y27632 treatment (Li et al., 2012).

These may include Y27632 (described in section “Inhibitors Targeting Other Cytoskeleton-Related Proteins”), U0126, imatinib, and dasatinib. U0126 specifically and non-competitively inhibits MEK 1 and MEK 2, mitogen-activated protein kinase kinase (MAPKK) protein kinases (Catalano and O’driscoll, 2020), which prevents the activation of ERK, needed for microvesiculation to occur. U0126 is proved to reduce MV secretion (Catalano and O’driscoll, 2020) and when used in combination with gemcitabine for 72 h, increased levels of cell death in chemoresistant Suit-2 cells by 11-fold compared to the control (Becker et al., 2016). Dual treatment of another chemoresistant human pancreatic cancer cell line, MPanc-96, with U0126 and gemcitabine for 72 h also increased the levels of cell death, though not as dramatically as in Suit-2 (Becker et al., 2016). These results show that U0126 is an effective MV inhibitor able to induce chemosensitivity in chemoresistant cell lines, although its efficacy varies between cell lines.

Imatinib and dasatinib, medications for leukemia, both inhibit ATP-binding sites of the catalytic sites of bcr-abl tyrosine kinase enzymes (Catalano and O’driscoll, 2020). Phosphorylated receptor tyrosine kinases promote anti-apoptotic activity in cancer-derived exosomes (Song et al., 2016). Dasatinib was developed when myeloid leukemia cells were shown to acquire resistance to imatinib, and dasatinib was proven to be effective on all mutants of the bcr-abl tyrosine kinase enzyme (Mineo et al., 2012). Imatinib and dasatinib reduced exosome biogenesis in the chronic myeloid leukemia cell line K562 by 58 and 56%, respectively (Mineo et al., 2012).

These include manumycin A, tipifarnib, and sulphisoxazole. Manumycin A and tipifarnib both diminish levels of exosome production primarily by targeting Ras farnesyltransferase enzymes, hence inhibiting their activation in order for Ras/Raf/ERK1/2 signaling to proceed, which is part of the ESCRT-dependent pathway of exosome biogenesis (Hanson and Cashikar, 2012; Datta et al., 2018). However, manumycin A also inhibits nSMase2 activity (Canals and Hannun, 2013) and hnRNP H1 (Datta et al., 2017). It has been suggested at low concentrations of manumycin A, nSMase2 inhibition does not occur and that both nSMase2 and Ras farnesyltransferase activity is needed in order for manumycin A to efficiently inhibit exosome biogenesis (Datta et al., 2017). Manumycin A (250 m) was found to inhibit exosome biogenesis by 50, 60, and 65% in prostate cancer cell lines C4-2B, PC3, and 22Rv1, respectively (Datta et al., 2018). Meanwhile, tipifarnib activity was found to reduce the phosphorylation of signal transducers and activators of transcription 3 (STAT3) and ERK, the latter of which is needed in exosome biogenesis, and both of which are needed for the growth of cancer cells (Datta et al., 2018). Tipifarnib (1 μM) was found to inhibit exosome biogenesis by 70% in a modified version of human prostate cancer cell line C4-2B which expressed the exosomal marker CD63 (Datta et al., 2018).

Sulphisoxazole are antibacterial drugs which, by targeting endothelin receptor type A (a G-protein coupled receptor), demonstrates inhibitory activity on components within or relating to the ESCRT-dependent pathway such as ALIX and VPS4B (which aids in ILV formation) as well as some RABs (Catalano and O’driscoll, 2020). It was proven to reduce the secretion of small EVs in breast adenocarcinoma cell lines MCF-7, MCF-10A, and MDA-MB-231 (Im et al., 2019). Most of the small EVs inhibited were likely to be exosomes, since sulphisoxazole did not inhibit the secretion of MVs from the breast cancer cell lines used (Im et al., 2019).

Cannabidiol and SMR peptides are recently discovered to be able to inhibit EV release. Cannabidiol, a phytocannabinoid, has been discovered to inhibit EV release from HEPG2 as well as MDA-MB-231 and PC3 cell lines (Kosgodage et al., 2018). Its effect is dose-dependent and varies across different cancer cell types. It has been found to affect mitochondrial functions via reducing the expression of STAT3 and prohibitin, both of which positively regulate cell proliferation, and hence is a possible agent to sensitize chemoresistant cells to chemotherapy drugs (Kosgodage et al., 2018). Meanwhile, SMR peptides impedes mortalin expression (Huang et al., 2019). Mortalin is present in erythroleukemia K562 cell-derived EVs and has been implicated in EV release, although its precise relation is not known (Pilzer and Fishelson, 2005). SMR peptides blocked EV release in the cell lines MCF-7 and MDA-MB-231 (Huang et al., 2019). It inhibited the growth of the breast cancer cells without affecting their cell viabilities (Huang et al., 2019).

Ketotifen (an antihistamine) and dimethyl amiloride (DMA) (a derivative of amiloride, a drug to treat high blood pressure) inhibit exosome release by altering intracellular calcium levels (Khan et al., 2018; Catalano and O’driscoll, 2020). Calcium-dependent pathways are prevalent in exosome release (Khan et al., 2018). Ketotifen and DMA hinder calcium entry into cells. DMA targets calcium channels (Catalano and O’driscoll, 2020) whereas ketotifen inhibits the expression of CDC42, Rac, Rho and MMP-9 (Kim et al., 2014) as indicated in Figure 4. Ketotifen sensitized HeLa cells, as well as MCF-7 and BT549 breast cancer cells, to doxorubicin proportionally to its effect on exosome release inhibition (Khan et al., 2018). DMA was shown to decrease exosome secretion both in vitro and in vivo (Chalmin et al., 2010). An in vitro study using CT26 (mouse colon carcinoma), EL4 (mouse lymphoma) and H23 (human lung adenocarcinoma) cell lines showed that DMA reduced exosome release as per analysis of the culture medium (Chalmin et al., 2010). In vivo results of the study are discussed in section “Targeting EVs in Clinical Applications.” Additionally, as calcium channels are also involved in micropinocytosis, DMA has been shown to inhibit macropinocytosis in mammalian cells, though it has not yet been used in cancer cell lines specifically.

Choosing the appropriate inhibitor to investigate a specific endocytic pathway can be challenging as they may be unspecific or toxic for different cell lines. Inhibitors of CDE include dynasore and ikarugamycin (IKA) as well as older inhibitors such as chlorpromazine, potassium depletion and hypertonic sucrose (Dutta and Donaldson, 2012). While potassium depletion and hypertonic sucrose are still commonly used to study CDE, they are non-specific as they may target CIE as well as CDE (Dutta and Donaldson, 2012). Potassium depletion may result in decreases of protein and DNA synthesis, whereas hypertonic sucrose results in shrinkage of cells and could affect cytoskeletal actin. Chlorpromazine blocks CDE effectively by assembling clathrin and adaptor proteins on endosomal membranes, depleting plasma membranes of clathrin (Vercauteren et al., 2010). It has also targets CIE and was discovered to be toxic to several non-cancerous cell lines (Vercauteren et al., 2010).

Dynasore and IKA are more recent inhibitors. Dynasore is a widely used CDE inhibitor which has been thoroughly characterized. But like the other CDE inhibitors, dynasore too may have non-specific effects. Dynasore non-competitively inhibits GTPase activity of the dynamin proteins dynamin1, dynamin2 and Drp1 (mitochondrial dynamin) in just seconds. Dynamin proteins are necessary for a late step of CDE which involves the production of a clathrin-coated endocytic vesicle (as indicated in Figure 5) and may participate in earlier stages of CDE (Kirchhausen et al., 2008). Dynasore inhibits the endocytosis of exosomes (Dutta and Donaldson, 2012). Dynasore may not only act as a CDE inhibitor: previous studies have shown that dynamin proteins regulate the assembly of actin filaments into bundles (Gu et al., 2010). Hence, it could be hypothesized that dynasore could have some additional effect on exosome release in cells. However, this has not been tested in cancer cells and this hypothesis is dubious since dynasore was shown to be ineffective in inhibiting exocytosis in synaptic vesicles (Newton et al., 2006). In comparison, IKA has been found to exclusively inhibit CDE in H1299, HCC366 and H1437 lung cancer cell lines (Elkin et al., 2016). It targets different receptors involved in CDE according to the cell line. In H1299 cells, it acts on transferrin receptors (TfNR) to inhibit CME (Elkin et al., 2016). 4 μM of IKA was also found to have no significant effect on cell viability of H1299 for the first 8 h; after 48 h, significant levels of apoptosis occurred and only 20% of the cells remained viable (Elkin et al., 2016). At 32 μM of IKA, the concentration of IKA which absolutely inhibits CDE, reductions in cell viability of H1299 were detected within an hour of treatment with IKA and after 8 h of incubation, only 20% of cells remained viable (Elkin et al., 2016). The cytotoxicity of IKA with other cancerous cell lines has not been thoroughly examined. The efficacy, specificity and toxicity of CDE inhibitors may be highly variable between different cell lines and suitable inhibitors must be chosen carefully for various types of research.

Caveolin-mediated endocytosis (CavME) inhibitors include genistein and methyl-β-cyclodextrin (MβCD) as well as heparin. Genistein is a highly specific, dose-dependent inhibitor of tyrosine kinases such as the EGF receptor kinase, and pp60^v–arc and pplloBn^gag–fes kinases (Costa Verdera et al., 2017). It disrupts the arrangements of actin and prevents the mobilization of dynamin for plasma membranes, two mechanisms vital for CIE (Costa Verdera et al., 2017). Genistein has been reported to be toxic to ovarian tumor cells A2780, CaOV3, ES2, and SK-OV-3 at concentrations above 10 μM (Antosiak et al., 2017). Meanwhile, MβCD removes cholesterol from the plasma membrane, disturbing lipid rafts within it, subsequently diminishing the uptake of exosome-sized vesicles (≤150 nm) (Kosgodage et al., 2017). However, its effect is non-specific as it also inhibits exosome secretion from cells as well as other endocytic pathways such as macropinocytosis and CDE (Dutta and Donaldson, 2012). In the study by Kosgodage et al. (2017), MβCD reduced exosome secretion by 58% compared to the control. MβCD has been documented to severely impair cell morphology especially under incubation times which are longer than 2 h, using concentrations of MβCD higher than 5 mM, which is in the range of concentrations of MβCD typically used to inhibit exocytosis (Kosgodage et al., 2017). MβCD is typically used in combination with lovastatin, an inhibitor of cholesterol synthesis, to prevent cholesterol being synthesized in cells and replacing the cholesterol in the plasma membrane extracted by MβCD (Kosgodage et al., 2017).

Heparin, an anticoagulant, competitively inhibits cancer cell surface receptors which depend on heparin sulfate proteoglycan (HSPG) coreceptors for the uptake of exosomes when HSPGs are present (Sento et al., 2016). An example of a HSPG which is necessary for exosome biogenesis is Syndecan (Sento et al., 2016). Heparin primarily acts on the receptors of treated cells rather than the receptors on treated exosomes, as shown by a study using the urothelial carcinoma cell line SW780 (Franzen et al., 2014). In another study by Christianson et al. (2013), it was observed that exosome uptake in U87 cells was reduced by 55% compared to the untreated control. In addition, exosome uptake was reduced significantly in medium depleted of Ca²⁺ compared to medium consisting Ca²⁺ (Christianson et al., 2013). It was concluded that heparin inhibits exosome internalization in a dose- and charge density-dependent way (Christianson et al., 2013).

5-(N-ethyl-N-isopropyl)amirolide (EIPA) is one chemical derivative of amiloride which has been regularly used to inhibit macropinocytosis. EIPA lowers sub-membranous pH by targeting Na⁺/H⁺ exchangers (Koivusalo et al., 2010), and like NSC23766, disrupts Rac1 activation as well as the assembly of actin, hence reducing the uptake of EVs as demonstrated using HeLa cells (Costa Verdera et al., 2017). While EIPA is regarded as effective as a pharological inhibitor of macropinocytosis, it inflicts “collateral damage” on treated cells with effects on ion transport, intracellular pH and the cytoskeleton (Costa Verdera et al., 2017). Imipramine has been suggested as an alternative; intracellular pH of treated cells has been proven to recover following treatment with imipramine, and imipramine was shown to effectively reduce macropinocytosis activity in a range of cell lines including mammary carcinoma 4T1 cells (Lin et al., 2018).