3.1. Proteins Involved in Cancer Progression Detected in CM-derived Ectosomes by LC–MS/MS
In the present study, ectosomes were isolated from conditioned media of in vitro cultured four CM cell lines representing different stages of the disease. Sequential centrifugation with the final step at 18,000× g
was applied to pellet ectosomes. The range between 16,000 and 20,000× g
is considered to be sufficient for successful isolation of ectosomes, but insufficient for pelleting exosomes (at least approx. 100,000× g
is required). Similar centrifugal forces can also be applied to pellet apoptotic bodies; however, viable melanoma cells in culture should not release apoptotic bodies if apoptotic processes are not over-induced [13
]. Since sequential centrifugation is based on the differences in the size of isolated particles, it is biased by the size overlap between EV subpopulations. Therefore, isolated EV populations should rather be analyzed in terms of their relative depletion or enrichment. Herein, we showed, using three independent methods, i.e., TEM, NTA and Western Blot, that isolated EV samples were highly enriched in ectosomes. The majority of isolated vesicles were within the predefined size range for ectosomes, depleted of exosomal markers (CD63 and Hsp70), and enriched in ARF6, a protein marker confirming plasma membrane origin of EVs.
Obtained ectosome samples from primary (WM115, WM793) and metastatic (WM266-4, WM1205Lu) CM cell lines were next analyzed using the shotgun nanoLC–MS/MS approach to profile their protein content. GO analysis for CM ectosomes revealed the presence of proteins involved in cell proliferation, migration, escape from apoptosis, angiogenesis, etc., and most proteins belonging to these groups were more abundant in ectosomes released by metastatic CM cell lines. Moreover, the functional tests performed in the present study have shown that ectosomes stimulated proliferation and migratory properties of recipient melanoma cells, most likely due to the ectosomal transfer of different cancer-promoting molecules.
Regarding the altered adhesion and motility of CM cells, the action of integrins and cadherins has been widely studied. Increased expression of integrin αvβ3 correlated with the progression of melanoma from radial to vertical growth phase [14
], while expression of integrins α2β1 and α3β1 was increased in metastatic cells compared to primary ones [15
]. Additionally, the restored expression of E-cadherin was found to inhibit melanoma cell invasion by decreasing the expression of β3 integrins and MUC18 receptor [17
]. In the present study, all the aforementioned integrin subunits were identified in each ectosome sample together with α4, α5, α6 integrin subunits and MUC18. Other integrin subunits, i.e., α1, β5 and β8, were present in three samples, whereas α9 integrin subunit was found only in ectosomes derived from metastatic WM266-4 cells.
Dissemination of cancer cells is also highly dependent on the activity of matrix metalloproteinases (MMPs). CM ectosomes have already been shown to facilitate the transfer of MMP-1, MMP-2 and MMP-9 and their endogenous activator CD147 [6
]. In the present study, the number of proteins involved in matrix disassembly was higher in ectosomes derived from primary CM cells; however, MMP-14 was the only MMP identified in each sample. In contrast, CD147 was present in all samples, suggesting that CM ectosomes may not always transfer MMPs, but have a regulatory role towards the activity of MMPs that are already present.
Noticeably, several markers of epithelial–mesenchymal transition (EMT) were present in CM ectosomes including N-cadherin and vimentin. Overexpression of N-cadherin in primary melanomas and the loss of E-cadherin expression in primary melanomas and metastatic melanomas correlated with worse overall survival of CM patients [18
]. The absence of E-cadherin and the presence of N-cadherin in CM ectosomes may reflect their potential to induce or promote EMT in recipient cells and such properties have already been shown for exosomes released by bladder cancer cells [19
] and exosomes from plasma of breast cancer patients [20
]. Additionally, Qendro et al. [4
] demonstrated that the pattern of vimentin expression in exosomes can help predict tumor aggressiveness in different subtypes of melanoma.
Furthermore, ectosomes facilitate the transfer of proangiogenic factors or up-regulate their expression in endothelial cells [21
]. In the present study, comparable numbers of angiogenesis-related proteins were identified in all ectosome samples, including proteins involved in vascular endothelial growth factor (VEGF)-mediated signaling. Nevertheless, VEGF was not found in any sample and only ectosomes from primary WM793 cells carried VEGF receptor 1 (VEGFR1). In addition, neuropilin 1, another VEGF receptor that integrates proangiogenic signals, was present in ectosomes from metastatic WM1205Lu cells.
Cancer progression is also associated with a prothrombotic state. In the studies by Lima et al. [9
], melanoma-derived ectosomes displayed a greater procoagulant activity (resulting from elevated levels of tissue factor (TF)) than melanocyte-derived ectosomes. A more recent study showed that accumulation of pancreatic cancer-derived ectosomes at the site of thrombosis is mediated by αvβ1 and αvβ3 integrins [24
], both identified in each ectosome sample in the present study. Although we did not identify TF in any CM ectosome sample, other procoagulant molecules such as urokinase plasminogen activator receptor (uPAR), tissue plasminogen activator (tPA) and plasminogen activator inhibitor type 1 (PAI-1) were present in ectosomes from isogenic primary WM793 and metastatic WM1205Lu cells. In addition, PAI-2 was identified in WM793 ectosomes. Its absence in WM1205Lu ectosomes may reflect the primary origin of WM793 cells, since PAI-2 was previously shown to inhibit uPA activity in less invasive tumors in a mice melanoma model [25
Moreover, in the present study, ectosomes of metastatic origin had a higher percentage of proteins associated with negative regulation of the apoptotic process, including annexins or programmed cell death protein 10 (PDCD10). Furthermore, T-cadherin (cadherin-13) was identified in WM793 and WM1205Lu ectosome samples. Bosserhoff et al. [26
] have already demonstrated in mice that the growth of T-cadherin-positive melanoma tumors was diminished in comparison to T-cadherin-negative control, suggesting that loss of T-cadherin desensitizes melanoma cells to apoptosis.
Tumor-derived EVs are associated with suppression of immune response towards transformed cells, for instance, by inducing chemotaxis of blood leukocytes [27
]. In the present study, CM ectosomes carried several molecules involved in leukocyte migration such as L1 cell adhesion molecule (L1CAM) and integrins. In a study by Valenti et al. [28
], CM ectosomes inhibited differentiation of monocytes to antigen-presenting dendritic cells and the remaining monocytes released transforming growth factor β (TGF-β), which inhibited T-cell cytolytic activity. CM ectosomes were also shown to suppress the immune response by vesicle-associated Fas (FasL) and TRAIL ligands [11
]. Our CM ectosome samples did not contain any of the aforementioned molecules; however, multiple proteins involved in T-cell response were identified, including components of major histocompatibility complex I (MHC I). Removal of MHC I molecules via ectosomes may lead to the loss of tumor recognition by cytotoxic T-cells, and may limit the efficacy of cancer immunotherapy.
Cancer cells have also been shown to overexpress different transporter proteins involved in multidrug resistance (MDR) and the efflux of anticancer drugs, i.e., P-glycoprotein (Pgp), multidrug resistance-associated protein 1 (MRP1) and breast cancer resistance protein (BCRP) [29
]. In the present study, MRP1 was present in ectosomes released by three CM cell lines (besides WM793) and previous studies by Walsh et al. [30
] showed an association of higher MRP1 levels in biopsy specimens with aggressiveness and spread of metastatic melanoma.
3.2. Melanoma-Derived Ectosomes as a Source of Potential Disease Biomarkers
Over the years, a number of potential markers for CM have been investigated. However, only a few of them have been recommended for clinical practice, such as lactate dehydrogenase (LDH) or S100B protein, but their poor sensitivity and specificity are major limitations for routine use. The serum levels of S100B, which was found in all CM ectosome samples in the present study, is linked to the tumor burden and reflects a clinical stage of CM. It is not, however, a good indicator of treatment response for patients with stage I, II and III CM, but it is still used as a biomarker for monitoring patients with advanced metastatic disease only [31
]. In the present study, all CM ectosome samples contained also LDH, the elevated serum level of which is used as an independent and highly significant predictor of survival in CM [33
The melanoma biomarker field recently saw advancements by different proteomic strategies. Findeisen et al. [34
] used matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI–TOF–MS) and showed that serum amyloidal protein (SAA) and c-reactive protein (CRP) could serve as prognostic serum biomarkers for early-stage CM patients. The same technique was applied for comparative analysis of CM specific spots after two-dimensional gel electrophoresis that revealed five potential biomarkers, i.e., eukaryotic elongation factor 2, enolase 1, aldolase A, glyceraldehyde-3-phosphate dehydrogenase and heterogeneous nuclear ribonucleoprotein A2/B1 [35
]. A similar study identified six more proteins overexpressed in melanoma cell lines compared to normal melanocytes, i.e., galectin-1, inosine-5′-monophosphate dehydrogenase 2, serine/threonine-protein phosphatase 2A 65 kDa regulatory subunit A α isoform, protein DJ-1, cyclophilin A and cofilin-1 [36
Stable isotope labeling with amino acids in the cell culture (SILAC) before MS analysis provides further advantages in biomarker-focused proteomic studies. The peptide peaks of the differentially labeled samples can be accurately quantified relative to each other to determine the peptide and protein ratios. Using SILAC, Liu et al. [37
] detected differential expression of CUB-domain-containing protein 1 (CDCP1) in plasma membrane proteome of two CM cell lines of low and high metastatic potential. Similar SILAC-based proteomic comparison of primary WM115 and metastatic WM266-4 cell lines indicated changes in cyclophilin A expression related to the disease stage [38
]. Finally, the metastatic potential of CM cells was also correlated by proteomic studies with expression levels of annexin 1 [39
], nucleophosmine- and hepatoma-derived growth factor (HDGF) [40
Only five of the aforementioned proteins (SAA, CRP, CDCP1, HDGF and cyclophilin A) have not been detected in CM ectosomes in the present study, thus proving CM ectosomes as a potential biomarker source. Ectosome content may not only become a source of biomarkers of different disease stages, but may also contribute to differential diagnosis between benign melanocytic lesions and CM. However, while discussing the presence/absence of a particular protein in CM ectosomes, it is necessary to acknowledge the limitations of protein identification by shotgun LC–MS/MS. Such proteomic analysis of unseparated, complex protein mixture allows identification of a much higher number of proteins, which are not lost during gel band or spot excision. On the other hand, signals from underrepresented peptides might be lost among the most abundant proteins. The expression level for particular proteins might have been low and not detected in a given sample replicate in the present study, thus the presence of proteins described as being absent in CM ectosomes cannot be completely ruled out.
Finally, it is important to notice that the number of studies on melanoma EVs is still very limited. Melanoma-derived growth regulatory protein (MIA) and S100B protein were detected in exosomes from serum of CM patients and their quantification presented with diagnostic and prognostic potential towards stage IV of CM [41
]. MIA and S100B were also identified in ectosomes in the present study, thus the population of larger vesicles seems to have not less potential in terms of biomarker discovery research. More recently, Cresticelli et al. [42
] performed proteomic analysis of small and large EVs isolated from melanoma tumors and confirmed that metastatic melanoma tissues contain a mixture of EVs derived from tumor and immune cells. The same group also showed that two mitochondrial inner membrane proteins: cytochrome c oxidase subunit 2 (MT-CO2) and cytochrome c oxidase subunit 6C (COX6c) are enriched in the plasma of melanoma patients as well as in tumor tissues-derived EVs compared to healthy controls [43
]. Studies using human tumor tissues are particularly valuable, since they fully reflect the complexity and cellular interactions present within the tumor microenvironment.
Nevertheless, studies based on the proteomic analyses of EVs only create the premise for the large, independent and multicenter clinical trials that are necessary for the validation of any novel biomarkers. Diagnostic and prognostic potential of elevated numbers of ectosomes, or ectosomes bearing certain molecules, depends on the establishment of proper isolation protocols. To gain valid information for clinical practice, optimal concentrations of uncontaminated vesicle populations with maintained native form and function are required. Our results obtained in the present study are promising and suggest that these investigations should be continued in the future for CM, a disease for which effective biomarkers are still lacking.