Cell-derived extracellular vesicles as a platform to ...

Expert Review of Molecular Diagnostics Downloaded from informahealthcare.com by 150.241.253.98 on 05/07/15 For personal use only.

Review

Cell-derived extracellular vesicles as a platform to identify low-invasive disease biomarkers Expert Rev. Mol. Diagn. Early online, 1–17 (2015)

Esperanza Gonza´lez1 and Juan Manuel Falco!n-Pe!rez*1,2 1 Metabolomics Unit and Platform, CIC bioGUNE, CIBERehd, Technology Park of Bizkaia, Derio, Bizkaia, Spain 2 IKERBASQUE, Basque Foundation for Science, 48011, Bilbao, Spain *Author for correspondence: Tel.: +34 944 061 319 Fax: +34 944 061 301 [email protected]

Biomarkers are of great importance for prediction, diagnosis and monitoring the progression and therapeutic success of a disease. Whole body fluids, such as blood or urine, constitute the main desired biological source to identify these markers, mostly due to the minimally invasive procedures used to collect them. An additional benefit of studying these biological fluids that has been demonstrated by many different groups is that they contain cell-released extracellular vesicles, carrying a cargo of lipids, proteins and nucleic acids that reflects cell/ tissue origin and, remarkably, cellular status. In this review, the information obtained from the characterization of this body fluid compartment in human samples is discussed in the context of its usefulness as diagnostic resource for several pathologies, including cancer, inflammatory, vascular and metabolic diseases. The review shows the great variety of methods used for this purpose as well as the different types of molecules that could serve as specific or common disease markers. KEYWORDS: biomarker . body fluids . circulating vesicles . diagnosis . exosomes . extracellular vesicles . human samples

.

microparticles

.

microvesicles

All mammalian cells are highly compartmentalized in membrane-limiting organelles that include mitochondria, endoplasmic reticulum, trans-Golgi network, peroxisomes, different types of endosomes and lysosomes, among others. All these organelles fill the limited intracellular space. The formation and maintenance of all these membrane-limiting compartments require a highly dynamic and regulated intracellular vesicular trafficking that is essential for the proper functioning of the cell. Given the complexity and tight regulation of the membranous compartments, it is conceivable that cellular stress conditions such as deregulation of signaling pathways (e.g., cancer) or infective processes alter the normal vesicular trafficking. The accumulation of endogenous toxic compounds, exposure to xenobiotics or even steric impediments resulting from the formation of protein aggregates or accumulation of lipid drops could also alter the normal vesicular trafficking. This abnormal vesicular trafficking may affect the

informahealthcare.com

10.1586/14737159.2015.1043272

regular cargoes–lipids, surface and cytosolic/ luminal proteins, metabolites and nucleic acids – of extracellular vesicles (EVs). EVs are part of the vesicular trafficking and constitute an important compartment for intercellular communication. It has been shown that in normal conditions, the EVs serve as vehicles to remove unwanted material in a safe way during differentiation processes (e.g., reticulocyte maturation) and to inter-communicate the cells in a paracrine and long-distance manner [1,2]. These vesicles have been detected in many different body fluids. Thus, the isolation and characterization of these vesicles from body fluids could be informative about stressful conditions affecting the cell function or about how the differentiation/formation of a determined cell type or tissue is progressing. Also, the signals that the cells are sending through these vesicles provide information about what type of challenges they are exposed to and how they are responding [3].

! 2015 Informa UK Ltd

ISSN 1473-7159

1


Review

´ lez & Falco!n-Pe!rez Gonza

Purification, detection & characterization

Purification protocols for EVs vary depending on the biological fluid of origin, but generally involve serial centrifugation at low speed to remove cells and cellular debris, followed by ultracentrifugation at 100,000 g to pellet the vesicles [4,5]. Alternatively, EVs can be isolated by immunocapture or size exclusion chromatography [5–7]. The latter, along with methods based on ultrafiltration have attracted great attention due to their capacity of separating the vesicular fraction from traditional contaminants such as protein aggregates and lipoproteins [8,9]. Filtration and microfluidic approaches have also been developed [10,11], although they have yet to be widely adopted. Recently, different precipitation reagents has become commercially available and although the purity of the isolated material is not yet well defined, the methods offer the advantage of requiring small volume of sample and simplified equipment [12,13]. The morphology and size of EVs were first characterized by electron microscopy, and further characterization was traditionally performed using biochemical methods such as immunoblotting, mass spectrometry, 2-D difference gel electrophoresis and microarray analysis [5], although atomic force microscopy [14], dynamic light scattering [15] and Raman tweezers microspectroscopy [16] have also been started to be used. The Vesiclepedia, ExoCarta and EVpedia databases provide a comprehensive record of proteins, RNA and lipid profiles [17–19] by compiling the data coming from the characterization of EVs in different scenarios. These databases constitute a great resource for scientific community in the field. On the other hand, detection and quantification of EVs relies upon indirect methods such as immunoblotting of protein markers, enzymatic activities, protein quantitation, fluorescent labeling of the vesicles or antibody-specific beadcoupled-approaches [5,20,21]. Recently, other approaches based on physical properties are being used, such as qNano and nanoparticle tracking analysis. They have been demonstrated to allow both characterization of size as well as direct quantification of vesicles [22,23]. In addition, the recently implemented high-resolution flow cytometry is making possible the direct analysis of different EV subpopulations, providing qualitative as well as quantitative measurements [20]. EVs from a biomarker perspective

EVs include endosome-derived exosomes, plasma membranederived microparticles (also named microvesicles, shedding vesicles or ectosomes) and apoptotic bodies or blebs, with the latter ones released specifically by cells undergoing apoptosis [24,25]. For years, different terms have been adopted in order to identify and better describe the different types of EVs, and this has generated confusion to deal with the literature. Nevertheless, the scientific community involved in the field is trying to solve the problem by arriving at a consensus about this important aspect [26]. Nowadays, the isolation of pure preparations of each type of EVs implicates a laborious procedure with different steps. This leads to additional experimental variability to that already doi: 10.1586/14737159.2015.1043272

generated by the unavoidable genetic and lifestyle factors and, consequently, increases the rate of false positives in human comparative studies. Also, it cannot be claimed that there is a unique method to isolate EVs or the one that assures complete purity. Given that, it has been proposed that when publishing the methods used for EV isolation, a minimal set of proofs should be provided. Characterization should include a representative protein profile, and data of the morphology and the size distribution [27]. Thus, although EVs can be differentiated by the combination of their physical and biochemical properties [5,20,21,28], the study of all of these types of vesicles as mixed populations in the initial phases of the identification of biomarkers for a determined pathology could be perfectly valuable. However, it is important to highlight that the definition based on size, floatation density and composition of the biomarker-containing vesicle should be further pursued because it would provide relevant information about additional companion markers and reveal the mechanisms involved in the pathogenesis. This added knowledge will offer novel opportunities for therapeutic interventions. Many features argue positively for the potential of EVs as a promising source for new molecular markers. They can be identified and isolated on the basis of their properties regarding size, density, lipid, protein and nucleic acid composition. Moreover, they carry specific markers from the cell of origin and their composition depends on the stimulation and microenvironment of the donor cells and, importantly, their analysis can be performed non-invasively since they can be detected in many body fluids. The aim of this review is to put together the reported candidate markers associated with EVs in human body fluids, hoping that any of these would get into the clinics in future. EVs in human body fluids in disease Circulating EVs in tumoral pathologies Platelet-derived microparticles

The majority of the microparticles found in the blood are derived from platelets and they have not only been extensively studied in thrombotic and vascular diseases [29] but also in cancer patients. Thus, elevated platelet microparticles have been found in 10% of patients with chronic myeloproliferative syndromes such us chronic myelogenous leukemia, polycythemia vera, chronic myelofibrosis,and essential thrombocytopenia [30]. Platelet-derived microparticles have been also studied in nonhematological malignancies. In hormone-refractory prostate cancer (PCa) patients, Helley et al. analyzed the levels of platelet-derived microparticles in whole blood by flow cytometry using an antibody against CD41a, which is a plateletspecific protein. In spite of the limited cohort (43 patients), a significant impact of platelet-derived microparticles on survival was evidenced. A higher CD41a-positive microparticle value correlated with shorter survival [31]. Using a similar approach in gastric cancer, Kim et al. [32] have shown that the levels of CD41a-positive microparticles were better predictors of metastasis than the serum levels of VEGF, IL-6 and RANTES Expert Rev. Mol. Diagn.


EVs for low-invasive biomarker identification

protein. In the study, plasma samples from 29 healthy individuals and 109 patients with different stages of gastric cancer were analyzed for their platelet microparticle levels using flow cytometry and plasma levels of VEGF, IL-6 and RANTES by ELISA. The plasma levels of CD41a-positive microparticles, IL-6 and RANTES in patients with gastric cancer were significantly higher than in healthy controls, correlating with the severity of the disease. Interestingly, the study also found that the levels of plasma CD41a-positive microparticles showed more than 90% of diagnostic accuracy in the prediction of distant metastasis [32]. Circulating EVs & hepatocarcinoma

Brodsky et al. investigated by flow cytometry the dynamics of circulating microparticles (300–1200 nm) in the plasma samples from liver transplant patients with hepatocellular carcinoma (HCC) arising in the background of hepatitis C [33]. They were able to analyze different subpopulations of EVs, including apoptotic bodies (stained with annexin-V), hepatic (stained with anti-CPS1) and endothelial (stained with antiCD144) circulating microparticles. Their results indicated that the levels of the hepatic CPS1-positive population of microparticles correlated directly with the size of the tumor. In addition, the levels of circulating microparticles were significantly elevated on the first day after liver transplantation surgery. Then, the levels of microparticles decreased, reaching the levels observed in healthy people, within 2 weeks after surgery. They presented evidences that the proportions of apoptotic bodies, and hepatic and endothelial microparticles correlated well with the clinical outcome of the patients. Although exosomes were excluded due to the purification method based on centrifugation at 10,000 g, and a reduced number of patients were analyzed, this study supports the fact that the levels of microparticles and apoptotic bodies and their proportions would be good markers of the clinical outcome and the viability of transplanted organs. Being one step ahead, predictive biomarkers for the recurrence of HCC would have great benefit in the selection of treatment options. Sugimachi et al. [34] have identified a specific miRNA (miR), miR-718, in exosomes from serum of patients with recurrent HCC after liver transplantation. When validated in a separate cohort of 59 individuals, it showed that decreased expression of miR-718 was related to HCC aggressiveness. They also identified HOXB8 as a potential target gene of miR-718, whose upregulation was associated with poor prognosis. Circulating EpCAM+ & CD63+ EVs

Some proposed vesicle-associated markers such as epithelial cell adhesion molecule (EpCAM) protein are altered by different tumoral diseases. Support for EpCAM as a tumor maker was obtained in a study that focused on patients with lung cancer, which reported significant differences in plasma circulating material of less than 200 nm in size that includes exosomes [35]. In the study, Rabinowits et al. evaluated the plasma circulating levels of exosomes carrying EpCAM protein in 27 patients informahealthcare.com

Review

with lung adenocarcinoma and 9 controls. They observed that in patients, the levels of EpCAM-positive exosomes were on average more than threefold higher than in controls. In addition, this study also revealed that while miRNAs were undetectable in control samples, the EpCAM-positive vesicles isolated from patients carried detectable levels of 12 miRNAs (miR-17-3p, miR-21, miR-106a, miR-146, miR-155, miR-191, miR-192, miR-203, miR-205, miR-210, miR-212, miR-214) that were previously demonstrated to constitute a diagnostic miRNA signature in the tissue for non-small-cell lung cancer [36]. The study supports that the quantization of EpCAM-positive exosomes along with the analysis of the 12 aforementioned miRNAs could serve as a minimally invasive screening test for lung adenocarcinoma. In colon cancer, it has also been reported that the amount of circulating EpCAM-positive exosomes is higher than in healthy individuals [37]. The study was conducted on circulating vesicles of size below 220 nm that were isolated from plasma of 12 healthy individuals and 91 patients who had surgery for colon cancer. In this case, the authors found significant differences between the two populations, with the cancer patients having higher levels of plasma circulating exosomes double positive for CD63 and EpCAM (CD63+EpCAM+) proteins. These increased levels of CD63+EpCAM+ exosomes correlated with clinical–pathological parameters of poor prognosis, such us high histological grade and high carcinogenic embryonic antigen levels. Importantly, colorectal cancer patients with high levels of circulating CD63+EpCAM+ exosomes in plasma showed shorter life survival from the diagnostic time than the patients with low amount of these exosomes. Taken together, the data suggest that levels of EVs may have a role as tumor markers in the diagnosis and prognosis of patients with cancer colon. In a similar scenario, Logozzi et al. focused on circulating exosomes in metastatic melanoma, which is an untreatable cancer lacking reliable and non-invasive markers of disease progression [38]. They have developed an in-house ELISA to analyze CD63 and tumor-associated caveolin-1 proteins in Rab5bpositive plasma circulating vesicles. In the study, exosomes and protein aggregates were purified by differential centrifugation and 0.22 mm filtration from plasma of 90 melanoma patients and 58 healthy donors. The levels of circulating double-positive Rab5b+CD63+ or Rab5b+caveolin1+ exosomes in melanoma patients were significantly higher than in healthy donors. Furthermore, the assay could differentiate melanoma patients from healthy donors using 50 ml of unfractionated plasma samples. These results support that an exosome-specific ELISA may be used for the detection and quantitation of circulating exosomes in melanoma [38] and probably many other cancers. In a study on ovarian cancer, Taylor and Gercel-Taylor also found that the levels of EpCAM-positive serum-circulating vesicles were higher in patients than in control samples [39]. They analyzed the serum samples from 50 women diagnosed with different degrees of serous papillary adenocarcinoma of the ovary, using healthy subjects and patients with benign ovarian disease as control. Their results highlight that the levels of doi: 10.1586/14737159.2015.1043272


Review

´ lez & Falco!n-Pe!rez Gonza

EpCAM-positive vesicles in 2.5 ml of serum correlate with the severity of the degree of cancer. In addition, the analysis of the miRNA content of these vesicles showed that in ovarian cancer patients, the levels of eight miRNAs (miR-21, miR-141, miR-200a, miR-200b, miR-200c, miR-203, miR-205 and miR-214) were significantly more than the levels detected in vesicles derived from benign disease, and these miRNAs were not detected in samples from healthy individuals. However, there was no correlation between the intensities of these miRNAs and the severity of the disease. Several other tumor markers in exosomes have been reported by other studies. Gercel-Taylor et al. determined by nanoparticle-tracking analysis the circulating material directly in the serum samples of eight women diagnosed with stage IIIc papillary serous adenocarcinoma of the ovary [7]. They found a fourfold increase in the level of circulating material that could correspond to EVs or protein aggregates. In addition, in one of the ovarian cancer patients was found an increase in the serum circulating vesicles with a size lower than 200 nm, containing the exosomal marker CD63 along with EpCAM and additional tumoral markers such as CD154 (Fas ligand) and epidermal growth factor receptor class III variant (EGFRvIII). Interestingly, the EGFRvIII variant is a constitutively activated mutant of the wild-type tyrosine kinase EGFR which is present in a substantial proportion of malignant human cancers [40]. Yet, this mutant variant is completely absent from normal tissues, which makes it a reliable tumor marker in the circulating vesicles. In fact, not only the protein but also the transcript of this variant has been shown to be differentially expressed in the circulating vesicles isolated from cancer patients [41]. Skog et al. reported that in 7 out of 25 analyzed patients suffering from glioblastoma, serum circulating vesicles lower than 220 nm contained detectable levels of the transcript EGFRvIII. As control, the authors analyzed serum samples from 30 healthy individuals and a positive detection of the tumor marker was made in none of them. Remarkably, five patients analyzed after extensive resection of the tumor showed undetectable levels of the transcript in circulating vesicles, supporting not only a diagnostic value in the detection of the tumor but also in the follow-up of the disease. Circulating EVs & miR-21

Other molecules, such as miR-21, could also serve as biomarkers for different diseases. Skog et al. reported that miRNA-21, that is known to be over-expressed in glioblastoma tumors [42], was elevated in serum circulating vesicles from patients as compared to controls. In another study, elevated levels of miR-21 were also found in serum circulating material purified by using Exoquick reagent from patients with esophageal squamous cell cancer [43]. In this study, Exoquick-purified material from serum samples of 51 patients with newly diagnosed esophageal squamous cell cancer and 41 patients with benign diseases, including asymptomatic cholecystolithiasis and hernia, was studied. The authors showed that levels of miR-21 in the esophageal squamous cell doi: 10.1586/14737159.2015.1043272

cancer group were significantly higher compared with the levels in the control benign group. In addition, the authors also observed that older age, negative metastasis and clinical disease stage were associated with higher miR-21 expression. In the same line [44], laryngeal squamous cell carcinoma patients with polyps of vocal cords were evaluated for exosomal miR-21 and non-coding RNA HOX transcript antisense RNA (HOTAIR) in serum. It was found that serum exosomal miR-21 and HOTAIR significantly correlated with the clinical parameters of laryngeal squamous cell carcinoma. Interestingly, miR-21 was also examined in serum exosomes from patients with HCC or chronic hepatitis B (CHB) [45]. The expression level of serum exosomal miR-21 was significantly higher in patients with HCC than in those with CHB or healthy volunteers and, high level of miR-21 expression correlated with cirrhosis and advanced tumor. The findings of this work also indicated that miR-21 is enriched in serum exosomes, providing increased sensitivity of detection than the whole serum. Other cancer biomarkers associated with circulating EVs

Other examples of studies on biomarkers associated with circulating EVs include a study performed in gastric cancer, in which the analysis of RNA content of plasma circulating vesicles also evidenced the value of these vesicles as biomarkers [46]. In the study, 37 patients with biopsy-proven gastric cancer at different clinical stages were studied and compared with a control group consisting of 10 healthy subjects. The study concluded that enhanced expression of oncogenic HER-2/neu and MAGE-1 mRNA was found in plasma circulating vesicles isolated from gastric cancer patients. On the other hand, Manterola et al. focused specifically on serum microvesicles for predictors of glioblastoma multiforme [47]. This study included 25 newly diagnosed patients for screening and another 50 patients for validation. The authors found that the expression of one small non-coding RNA (RNU6-1) and two microRNAs (miR-320 and miR-574-3p) were significantly associated with diagnosis of glioblastoma multiforme. In a more recent study, the authors aimed to find plasma exosomal miRNAs for prognosis in castration-resistant prostate cancer [48]. In a screening cohort of 23 castration-resistant prostate cancer patients, they identified two candidate miRNAs that were further evaluated in a follow-up cohort of 100 castrationresistant prostate cancer patients. The authors found that higher levels of miR-1290 and -375 were significantly associated with poor overall survival. Circulating EVs in non-tumoral pathologies Circulating microparticles & immune diseases

Apart from cancer, other pathological conditions have been shown to alter the composition of circulating cell-derived EVs. In particular, elevated levels of circulating microparticles have also been reported for autoimmune diseases including primary Sjo¨gren’s syndrome, rheumatoid arthritis and systemic lupus erythematosus (SLE) [49]. In this latter disease, Nielsen et al. Expert Rev. Mol. Diagn.


EVs for low-invasive biomarker identification

examined by flow cytometry the presence of several proteins in plasma circulating large microparticles purified from 72 patients with SLE and 29 healthy donors [50]. Among the tested proteins, annexin-V and protein markers of different cell types such as CD42a (specific of platelet), CD45 (specific for leukocytes) and CD146 (specific for endothelial cells) were included. The results revealed that the amount of circulating microparticles in SLE patients was higher than in controls, although this increase was not found for all types of microparticles. The authors also found significantly higher levels of annexin-V negative microparticles in SLE samples, while the annexin-V positive microparticles were significantly decreased. Among the annexin-V positive microparticles, no significant differences were observed regarding their origin [50]. Circulating EVs & hepatic diseases

Several studies have been conducted in order to find biomarkers for hepatic diseases, using EVs as source. Kornek et al. showed that the levels of circulating microparticles released by immune cells could also be used to assess the extent and characteristics of hepatic inflammation in patients with chronic liver disease [51]. The study reports that increased levels of circulating microparticles released from T lymphocytes (CD4+CD8+) correlated with the grade of inflammation in patients with chronic hepatitis C (CHC), which was determined by histological analysis and alanine aminotransferase activity. In patients with nonalcoholic fatty liver or non-alcoholic steatohepatitis (NASH), the levels of microparticles released from invariant natural killer T cells and macrophages/monocytes (CD14+) were significantly increased and correlated with alanine aminotransferase activity and the severity of NASH determined by histology. This study indicates that the quantification and characterization of immune cell-released microparticles could differentiate between patients with non-alcoholic fatty liver or NASH and those with CHC. In another study, protein- or vesicles-associated circulating microRNAs have also been shown to be a useful tool to differentiate and grade various liver injuries including NASH and CHB and CHC [52]. In each case, the authors analyzed the microRNAs that were present in 1 ml of serum sample processed by using the precipitating reagent Exoquick. This reagent mostly excludes circulating free nucleic acids and enriches the sample with nucleic acid-containing protein complexes and EVs. The profiling by microarray of this population of circulating microRNAs in four CHB and 12 NASH patients identified a number of microRNAs that could be useful in liver pathology. So, the pattern of expression of these selected microRNAs was further analyzed in a cohort of 95 CHC, 16 CHB and 20 NASH, and 24 healthy control subjects. The expression of nine miRNAs (miR-1225-5p, miR-1275, miR-638, miR-762, miR-320c, miR-451, miR-1974, miR-1207-5p and miR-1246) allowed categorizing patients as CHC or normal liver with 96% accuracy. The expression pattern of 12 miRNAs (miR-1225-5p, miR-1275, miR-638, miR-762, miR-320c, miR-451, miR-1974, miR-630, miR-1207-5p, miR-720, miR-1246 and informahealthcare.com

Review

miR-486-5p) led to the distinction of CHC, CHB, NASH and normal liver with 87% accuracy. These results showed that miRNA profiling represents a promising alternative to diagnose liver disease. While these results suggest there is great potential and benefit of miRNA profiling, future studies in larger populations of CHC patients are required to fully elucidate the diagnostic potential of serum miRNA expression. In another study on liver condition, plasma CD133 and CD39 microparticles subsets were analyzed [53]. CD133 and CD39 are expressed by hematopoietic stem cells which are mobilized after liver injury, and they target sites of injury, limit vascular inflammation, and boost hepatic regeneration. Concretely, plasma microparticles expressing CD39 can block endothelial activation. In the study, patients with acute (n = 5) and acute on chronic (n = 5) liver injury were enrolled with matched controls (n = 7). Plasma microparticles increased in patients with liver injury and were characterized by significantly higher levels of CD39 (p < 0.05), which could potentially serve as a biomarker of liver failure in the clinic. Other putative biomarkers associated with circulating EVs

EVs have been also found to be promising diagnostic tools in other clinical fields, for example, in pregnancy-associated problems. The diagnostic values of these vesicles for problems that can arise during pregnancy such as preterm delivery were supported by the comparison of serum circulating exosomes. These were purified from 10 ml of blood specimens obtained from three different groups: 27 pregnant women delivering without complications at term (>37 weeks gestation), 19 pregnant women delivering preterm (