Article Cite This: Anal. Chem. XXXX, XXX, XXX−XXX
pubs.acs.org/ac
Diagnosis of Invasive Nonfunctional Pituitary Adenomas by Serum Extracellular Vesicles Huayi Wang,†,§,∥ Kelin Chen,‡ Zhijun Yang,‡ Wenzhe Li,†,∥,⊥ Chen Wang,†,∥ Guojun Zhang,*,‡ Ling Zhu,*,†,∥ Pinan Liu,*,‡ and Yanlian Yang*,†,∥
Downloaded via IDAHO STATE UNIV on July 19, 2019 at 03:22:05 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
†
CAS Key Laboratory of Standardization and Measurement for Nanotechnology, CAS Key Laboratory of Biological Effects of Nanomaterials and Nanosafety, CAS Center for Excellence in Nanoscience, National Center for Nanoscience and Technology, Beijing 100190, China ‡ Beijing TianTan Hospital, Capital Medical University, Beijing 100070, China § Department of Chemistry, Tsinghua University, Beijing, 100084, China ∥ University of Chinese Academy of Sciences, 19 A Yuquan Road, Shijingshan District, Beijing 100049, China S Supporting Information *
ABSTRACT: The invasiveness evaluation of nonfunctional pituitary adenoma (NFPAs) is crucial for the prediction of the malignant potential and for making surgical plans of NFPAs. Current invasiveness evaluation of NFPAs is based on neuroimaging, which can hardly predict the invasive potential and dynamically monitor disease progress. Here we used microbead-assisted flow cytometry to detect and analyze the serum extracellular vesicles (EVs) from 30 NFPAs patients (15 invasive and 15 noninvasive). Lower expressions of folate receptor 1 (FOLR1) and epithelial cell adhesion molecule (EpCAM) were found in serum EVs from the invasive NFPAs patients compared to the noninvasive ones [area under the curve (AUC) of 0.94 for FOLR1 and 0.88 for EpCAM]. Meanwhile, increased mRNA expression of vimentin and N-cadherin, two mesenchymal markers, was found in serum EVs from the invasive NFPAs patients compared to the noninvasive ones. Consistent results were observed in the tumor tissue that invasive NFPAs have lower expression of the epithelial markers while higher expression of the mesenchymal markers. These results suggested the possible role of epithelial−mesenchymal transition (EMT) in the invasiveness of NFPAs. Pituitary tumor transforming gene 1 (PTTG1) mRNA in serum EVs was also found to be an indicator for invasive NFPAs and is related with EMT. These results provide a method for the blood-based diagnosis and invasiveness evaluation of NFPAs and would be beneficial to the diagnosis, prognosis prediction, and surgical risk evaluation of NFPAs.
P
important information to understand the mechanisms of the invasiveness of NFPAs. However, blood-based diagnostic methods for the diagnosis of invasive NFPAs are scarce. Researchers have reported the existence of circulating tumor cells (CTCs) in pituitary adenoma (PA) patients;4 however, the small amount of the rare tumor cells in the blood can hardly reflect the complete landscape of the heterogeneity of the tumor. A valid biomarker in the blood is yet to be established to determine the invasiveness of NFPAs. Extracellular vesicles (EVs) are endosome-derived membrane vesicles with sizes from 30 to 150 nm.5 They are secreted by almost all kinds of cells and are present in various body fluids such as the blood, the urine, and the cerebrospinal fluid.6 EVs contain distinctive membrane proteins, cargo proteins, messenger RNA (mRNA), and microRNA (miRNA) depending on their parental origin.7 The diversity
ituitary adenomas account for about 10% of intracranial tumors, and nonfunctional pituitary adenomas (NFPAs) are a big group among them.1 Some patients may suffer blindness or hypopituitarism because of regional compression with the growth of the tumor.2 More than 40% of NFPAs are invasive. The invasive NFPAs may invade the dura, the bone, and the cavernous sinuses, and some researchers also regard it as a character of malignant potential.3 Moreover, adenoma tissue can be too infiltrative to be completely removed, especially when the invasiveness cannot be evaluated prior to or during the surgery. Current diagnosis of invasive NFPAs is primarily based on imaging techniques which identify the size, or shape, and boundary of NFPAs.1 The morphological changes tracked by the neuroimaging techniques, however, are limited in precisely predicting the invasive potential of NFPAs and the dynamic monitoring of the disease progress. Molecular analysis to understand the mechanisms of the invasiveness of NFPAs is also lacking. Liquid biopsy, in this regard, provides an opportunity for the invasiveness evaluation and prediction and the dynamic monitoring of NFPAs. It can also provide © XXXX American Chemical Society
Received: February 19, 2019 Accepted: July 2, 2019 Published: July 2, 2019 A
DOI: 10.1021/acs.analchem.9b00914 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
SH-SY5Y cells were cultured in condition medium RPMI-1640 containing 10% (v/v) FBS and 1% (v/v) streptomycin− penicillin. The cells were cultured in an environment of 5% CO2 at 37 °C in a culture dish (Corning, 430167) with about 10 mL of condition medium. The cell lines were purchased from the Chinese Center for Type Culture Collection. The condition medium should get replaced as the cells grow to dense big clusters of cells and about 50% of saturation. For cells passage, cells were collected by centrifuging at 1000g for 3 min, abandoning most of the upper condition medium, and then gently blowing them by pipet and dispersing the cells in new condition medium. Serum and Tissue Collection from Patients. This study was authorized by Beijing Tian Tan Hospital, Capital Medical University (KY2014-021-02). Thirty NFPAs patients (15 noninvasive and 15 invasive) were enrolled in the study. Informed consent was obtained from all the participants. The peripheral blood was taken on the day before the surgery, and serum was extracted subsequently by centrifuging at 2500g for 10 min. Tissues were collected during surgery. The serum and tissue were stored at −80 °C, and the experiments were performed within 1 month after the storage. EVs Isolation and Characterization. For cells, the FBSEVs-free condition medium should be prepared in advance. The condition medium, as mentioned above, was ultracentrifuged at 100 000g overnight, and the solids on the wall of the centrifuge tube were abandoned. Collecting medium was filtered with a 0.22 μm filter (Millex-GP) to get rid of the possibility of bacteria contamination. Then, the FBS-EVs-free condition medium is prepared and stored at 4 °C. When the cells reached a saturation of about 50%, the condition medium was replaced with FBS-EVs-free condition medium. After culturing for another 24 h when the cells reached a saturation of about 70−80%, the supernatant medium was collected by centrifuging at 1000g for 10 min to discard all the cells, followed by centrifuging at 2000g for 20 min to discard the big cell debris. After that, the media were filtered through a 0.22 μm filter to deplete all the big vesicles. Eventually, the supernatant was ultracentrifuged at 100 000g for 2 h twice to get the EVs stacked to the wall of the tube, and the EVs were resuspended in 200 μL of PBS and stored at 4 °C for the current needs or at −80 °C for preservation. For patients’ serum, serum went through the same step before ultracentrifugation. After that, the supernatant was first ultracentrifuged at 150 000g for 10 h, and second, 150 000g for 2 h. The EVs were also resuspended in 200 μL of PBS and stored in 4 or −80 °C. Formvar 400-mesh carbon-coated grids were used for TEM (transmission electron microscope) experiment. Amounts of 5 μL of EVs suspension were dropped on the grids and held for 2 min to get the EVs absorbed on the surface, followed by 1% uranyl oxalate solution and ddH2O for 45 s, respectively. After being cleaned by ultrapure water, the grids were silenced for at least 20 min in a cleaning culture dish to dry before use. The experiment was carried on an electronic microscope (Hitachi 7700) at 80 kV. Amounts of 30 μL of EVs suspension were used for BCA (bicinchoninic acid) protein assay. The EVs were first lysed by an ultrasonic machine, followed by lysis buffer (Thermo, NP0001) for 30 min in ice. The BCA experimental procedure was entirely in accordance with the manual of the BCA protein assay kit (Solarbio, PC0020), and the absorbance value at 562 nm was obtained by the microplate reader.
of the contents, especially the membrane proteins, provides a rich library of biomarkers which can be applied for disease diagnosis.8,9 Different from the limited numbers of CTCs, EVs are abundant in blood, making them desired biomarkers for the diagnosis of diseases.10 To our knowledge, no EV-based diagnostic method for PA has been developed. The detection of EVs is difficult due to the nanosize of EVs. Strategies have been taken to improve capture efficiency.11 These include nanowires12 and engineered nanointerfaces for isolating and molecule profiling of EVs.13 Magnetic beads and latex beads are other types of materials employed for the assistance of EV detection.14−18 The detection of EVs using traditional flow cytometry has long been challenging as the nanoscale size of EVs exceeded the resolution limit of flow cytometry. By adhering EVs on aldehyde microsized beads and subsequently staining the captured EVs with antibodies specific to their membrane proteins, the EV−microbeads complex can be detected by flow cytometry. Using this method, researchers found GPC1+EVs to be a diagnostic marker for early pancreatic cancer19 and demonstrated that EpCAM+ and HER2+ EVs can be used for the diagnosis and HER2 phenotyping of breast cancer.20 Folate receptor 1 (FOLR1) is a membrane-bound glycoprotein protein that is mainly responsible for the transport of folates into cells.21,22 It is normally overexpressed in epithelial original tumor tissues like ovarian, breast, and colorectal carcinomas.23,24 Both mRNA and protein of FOLR1 have been reported to be overexpressed in tumor tissues of NFPAs compared to functional adenomas (FPAs).25−27 Epithelial cell adhesion molecule (EpCAM), a common epithelial membrane protein applied for tumor diagnosis, is also found to be overexpressed in tumor tissues from NFPAs patients than from FPA ones.28,29 Nonetheless, the expression difference of FOLR1 and EpCAM between invasive and noninvasive patients has not been investigated so far. Here, we establish a method for the diagnosis and invasiveness evaluation of NFPAs based on microbead-assisted flow cytometry. Applying this method, we demonstrate that FOLR1 and EpCAM in EVs can accurately distinguish invasive NFPAs patients from the noninvasive ones with high sensitivity and specificity. We also demonstrate the positive correlation between the invasion of NFPAs and epithelial−mesenchymal transition (EMT). Additionally, mRNA expression of pituitary tumor transforming gene 1 (PTTG1) in serum EVs, a gene that has been considered to promote EMT, can also be a valid marker for invasive NFPAs. We show that the expression of markers in serum EVs is consistent with the one in tumor tissues, suggesting that serum EVs can be a surrogate for the invasive tissue-based assessments and have great potential in clinical application. These results show the clinical significance of serum EVs in the diagnosis of the invasiveness evaluation of NFPAs and will help to reduce the surgical risk and meanwhile improve the prognosis of NFPAs.
■
MATERIALS AND METHODS Cell Lines. The AtT-20 mouse pituitary suspension cell lines were cultured in condition medium RPMI-1640 (Gibco, with NaHCO3 1.5 g/L, glucose 2.5 g/L, and sodium pyruvate 0.11 g/L) containing 10% (v/v) fetal bovine serum (FBS) and 1% (v/v) streptomycin−penicillin. The hBMECs cells were cultured in ECM (ScienCell ECM 1001) condition medium. Hela cells were cultured in DMEM (Gibco) condition medium containing 10% FBS, 1% (v/v) streptomycin−penicillin, and B
DOI: 10.1021/acs.analchem.9b00914 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry Nanoparticle Tracking Analysis. A NanoSight LM14 system (NanoSight Technology, Malvern, U.K.) was used for the measurement of EVs concentration and size. A syringe was used for sample injection, and the measurement was carried out by a high-sensitivity scientific complementary metal oxide semiconductor camera, at camera setting 16 with an acquisition time of 60 s and a detection threshold setting of 7. Between different samples, the chamber was washed using PBS for three times. The data were collected by nanoparticle tracking analysis software (NTA version 2.3; Malvern Instruments, Malvern, U.K.). Flow Cytometry Analysis of EVs. Before the experiment, the beads suspension and EVs suspension should be homogeneous. Aldehyde beads (Thermo, 4 μm, A37304, with a concentration of 1.3 × 109 beads/mL) were saturated with EVs and thoroughly mixed by pipet and gentle shaking for 1 h at room temperature. An amount of 100 μL of PBS was added subsequently, and the mixture was incubated overnight at 4 °C. After that, 10% (w/w) bovine serum albumin (BSA) (Beijing Biodee Biotechnology) with 100 mM glycine in PBS was used to stop the reaction. A concentration of 2% BSA was used as a solution for washing and follow-up experiments. After washing, first antibodies [folate receptor, purchased from Abcam, dilution 1/200; EpCAM, purchased from Cell Signal Technology (CST), dilution 1/500; E-cadherin, purchased from Abcam, dilution 1/100] were added as the recommended antibody dilutions in their specifications. A total of 2 h of gentle shaking and twice cleaning was necessary before adding the Alexa-647-tagged secondary antibodies (CST, dilution 1/ 1000) for 30 min with rotation at 4 °C. The experiment should be carried in dark as long as the second antibody is involved. After the same rotation and washing, the samples were ready for the flow cytometry analysis. Western Blot Analysis for EVs. The protein concentration of EVs was obtained by BCA protein assay as mentioned above. Equal amounts of total protein were loaded into a 10% Bis−Tris mini gels (Invitrogen); MOPS SDS (Life technologies) was used as running buffer for electrophoresis. The proteins were then transferred into poly(vinylidene fluoride) membranes (PVDF, Invitrogen), followed by FOLR1 (Abcam), EpCAM (CST), CD81 (Santa Cruz), flotillin (Abcam), calnexin (Abcam), and albumin (Abcam) overnight antibody. The secondary antibody with HRP (CST) was added later, and the proteins were visualized by chemiluminescence (Thermo, 34095). Immunogold Labeling and TEM. Amounts of 10 μL of EVs suspension were loaded on 200-mesh grids and allowed to stand for 20 min. Then, 10 μL of 4% paraformaldehyde was added and held for 10 min. Blocking was performed by adding 10 μL of 10% BSA (w/w), and antibodies [(anti-EpCAM (mouse mAb; Cell Signal Technology, dilution, 1/200) and anti-FOLR1 (mouse mab; Abcam, dilution, 1/200)] were added and held for 1 h. Secondary antibody conjugated with 10 nm gold particles (Aurion, BioValley, France) was added after washing by PBS. Then, the grids were fixed with 2.5% glutaraldehyde for 10 min; 1% uranyl acetate was used for negative stain. Quantitative Reverse Transcription Polymerase Chain Reaction (qRT-PCR) for EVs RNA. Amounts of 50 μL of EVs suspension were used for per PCR experiment. TRIzol (Life Technologies) was first added into the suspension for RNA extraction according to the reagent protocol, followed by a Quant Script RT kit (cat. no. KR103-04, Tiangen, China)
and Supernal Premix Plus kit (cat. no, FP205-02, Tiangen, China). Primers for our research are listed as follows: hPTTG1 mRNA: forward primer, F-5′-GCTCTGTTCCTGCCTCAGAT-3′; reverse primer, R-5′-GAGAGGCACTCCACTCAAGG-3′. GADPH mRNA: forward primer, F-5′-GAGAAGGCTGGGGCTCATTT-3′; reverse primer, R-5′-AGTGATGGCATGGACTGTGG-3′. Vimentin mRNA: forward primer, F-5′-GAACGCCAGATGCGTGAAATG-3′; reverse primer, R-5′-CCAGAGGGAGTGAATCCAGATTA-3′. N-Cadherin mRNA: forward primer, F-5′-GAGGAGTCAGTGAAGGAGTCA-3′; reverse primer, R-5′-GGCAAGTTGGAGGGATG-3′. The primers were purchased from Sangon Biotech Company in Shanghai, China. Immunofluorescence of Human Tissue Specimen. Tumor tissue sections were fixed with 4% paraformaldehyde for 10 min, followed by 3% H2O2 for 10 min and 3% Triton X100 for 10 min. Next, samples were blocked with 10% FBS for 30 min at room temperature, followed by incubation with the vimentin antibody (CST), FOLR1 (Abcam), EpCAM (CST) for 2 h, and then followed by Alexa-647-tagged secondary antibodies for half an hour. 4′,6-Diamidino-2-phenylindole (DAPI) was added for nucleus staining. Statistical Analysis. Unpaired t test was used to calculate the difference of the expression of the markers. Receiver operating characteristic (ROC) curves were used to determine the specificity and sensitivity of the markers in differentiating noninvasive and invasive NFPAs. All the statistical analyses were performed in Graphpad Prism. Normalization of the protein expression of EpCAM and Ecadherin, the mRNA expression of vimentin and N-cadherin in the EVs, as well as the expression of the these markers in the tumor tissues from the noninvasive and invasive NFPAs patients was performed according to the following equation: x − xmin r= xmax − xmin (1) where x represents the individual expression level of the marker proteins in the EVs as tested by flow cytometry, the individual mRNA expression fold of the markers in the EVs as tested by qRT-PCR, or the expression level of the marker proteins in the tumor tissues as tested by immunofluorescence staining; xmin represents the minimum value of the protein or mRNA expression level, and xmax represents the maximum value of the protein or mRNA expression level. The ratio of EMT tendency was calculated according to the following equation: REMT =
rvimentin + rN‐cad rEpCAM + rE‐cad
(2)
where r represents the normalized expression of the markers as calculated according to eq 1.
■
RESULTS Establishment of Microbead-Assisted Flow Cytometry Analysis of Membrane Proteins in EVs. We established an approach based on flow cytometry to detect the expression of membrane proteins on EVs. EVs were purified from cell culture or human sera by differential centrifugation as C
DOI: 10.1021/acs.analchem.9b00914 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry previously described.19 Since the nanoscaled size of eoxosmes exceeded the detective limit of flow cytometry, aldehyde latex beads with a diameter of 4 μm were added to EVs suspensions for the enrichment of EVs. Fluorescent antibodies targeting the membrane proteins in EVs were mixed with EVs-bound aldehyde beads. In this way, the EVs-bound microbeads could yield detectable signals for flow cytometry analysis (Scheme 1).
purified from condition medium by differential centrifugation. Transmission electron microscopy and nanoparticle tracking analysis were used to characterize the morphology, size, and concentration of the extracted EVs. TEM images showed a “saucer-like” morphology of the extracted vesicles (Figure 1A), characteristic of the typical morphology of EVs as previously described.33,34 NTA results showed a main diameter of around 122 nm and a concentration of 6.5 × 1011 vesicles per mL of serum (Figure 1B), which is in accordance with a previously reported value of about 3 × 1012 vesicles in 1 mL of serum.35 The EVs were found to be successfully bound to the surface of aldehyde beads as proved by TEM images showing that the aldehyde microbeads were densely covered with EVs (Figure 1C). Flow cytometry analysis showed high expression of FOLR1 in hBMECs and Hela cells, while low expression of FOLR1 in AtT-20 and SH-SY5Y (Figure 2, parts A and B). The expression of FOLR1 in the EVs derived from the cells exhibited the similar trend that hBMECs and Hela cells had high expression of FOLR1 while AtT-20 and SH-SY5Y had low expression of FOLR1 (Figure 2, parts A and B), indicating that the expression of FOLR1 in the cells correlated with the one in the source cells. The expression of FOLR1 in the EVs derived from the pituitary adenoma cells AtT-20 was significantly lower than the one in the EVs from the normal cells hBMECs (hBMECs cells with AtT-20 cells, unpaired t test, ****, P < 0.0001; hBMECs EVs with AtT-20 EVs, unpaired t test, ***, P < 0.001, Figure 2A). These results were confirmed by Western blot analysis showing high expression of FOLR1 in EVs from hBMECs and Hela cells while low expression of FOLR1 in EVs from SH-SY5Y and AtT-20 (Figure 2C). The positive markers flotillin-1 and CD81 and the negative marker calnexin were used to verify the purity of extracted EVs (Figure 2C; uncropped blots in Figure S-3).36 Immunogold TEM also verified the expression of FOLR1 in EVs from hBMECs (Figure 2D). As hBMECs are typical endothelial cells while AtT-20 is a pituitary adenoma cell reported to be invasive due to EMT,37 we ascribed the expression difference between cells to EMT. EMT prompted AtT-20 cells to be invasive and lower expression of FOLR1, and the low expression was also reflected in EVs. These results highlighted the efficiency of microbead-assisted flow cytometry for the analysis of marker membrane proteins in EVs, and the expression differences between normal cell lines and invasive pituitary adenoma cell lines remind us that FOLR1 protein can work as diagnostic biomarkers for invasive pituitary adenoma.
Scheme 1. Schematic Illustration of Aldehyde Latex Microbeads-Assisted Flow Cytometry Analysis of EVsa
a The EVs were purified from cell culture supernatant or human sera by differential centrifugation. The 4 μm aldehyde latex beads were incubated with EVs extractions for EVs enrichment. The captured EVs were stained with the antibodies specific to the membrane proteins on the EVs for flow cytometry analysis.
We first tested if microbead-assisted flow cytometry was an efficient method for the detection of membrane proteins in EVs. We focused on the expression of FOLR1 in the EVs, as FOLR1 was regarded as a marker for PA.30−32 Four cell lines were chosen for comparison: two brain tumor cell lines, pituitary adenoma cell line AtT-20 and neuroblastoma cell SHSY5Y, the cervical cancer Hela cell, and the normal human brain microvascular endothelial cell (hBMECs). EVs were
Figure 1. Morphology and size characterization of EVs and EV-bound microbeads. (A) Transmission electron microscope (TEM) image of EVs extracted from AtT-20 cells. (B) Nanoparticle tracking analysis (NTA) of the size distribution of EVs extracted from AtT-20 cells. (C) TEM image of EVs adsorbed on aldehyde latex bead. Inset: wide-field TEM image of EVs-binding bead. D
DOI: 10.1021/acs.analchem.9b00914 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
Figure 2. Folate receptor 1 (FOLR1) expression shows consistency between EVs and cells and can be a potential biomarker for the diagnosis of pituitary adenoma. (A) The expression of FOLR1 in the cells and cell-derived EVs (hBMECs cells with AtT-20 cells, unpaired t test, ****, P < 0.0001; hBMECs EVs with AtT-20 EVs, unpaired t test, ***, P < 0.001). (B) Representative flow cytometry analysis of the expression of FOLR1 in the cell-derived EVs. Four cell lines, hBMECs, Hela cells, SH-SY5Y, and AtT-20, were chosen for comparison. (C) Western blot analysis of the expression of FOLR1, flotillin-1, CD81, calnexin in EVs from hBMECs, Hela, SH-SY5Y, and AtT-20 (uncropped blot in Figure S-1). (D) Immunogold TEM analysis of the expression of FOLR1 in EVs from hBMECs cells.
FOLR1 and EpCAM Can Be Biomarkers To Distinguish Invasive NFPAs from Noninvasive NFPAs. After proving the efficacy of microbead-assisted flow cytometry in the analysis of EVs membrane proteins in PA, we applied this method in the diagnosis of invasive NFPAs. The expressions of two marker proteins in EVs isolated from sera of invasive (n = 10) and noninvasive (n = 10) NFPAs patients were investigated: FOLR1, a protein that has been reported to be highly expressed in NFPAs compared to FPA,30,31 and EpCAM, a common epithelial membrane protein used for tumor diagnosis.38,39 All patients enrolled in our experiment are diagnosed according to the Knosp classification and with typical characteristics of invasive or noninvasive pituitary adenoma judged by magnetic resonance imaging (MRI) and computerized tomography (CT) image.
EVs were extracted from patients’ sera by differential centrifugation. The TEM image revealed that the extracted vesicles exhibited the “saucer-like” morphology that is characteristic of EVs (Figure 3A). NTA showed a main diameter of 179 nm and a concentration of 1.12 × 1011 vesicles per mL of serum, in accordance with a previous report (Figure 3B).10 Bicinchoninic acid assay was applied for determining the protein concentration of EVs. These results indicated the success of isolating EVs from patients’ serum. FOLR1 and EpCAM expressions of EVs were evaluated by the method established above. Flow cytometry was used for determining the expression difference. Flow cytometry analysis showed that the expression of FOLR1 was significantly lower in EVs from invasive NFPAs patients than the ones from noninvasive NFPAs patients (unpaired Student t test, ***, P < 0.001, Figure 3C). Similarly, the expression of EpCAM in EVs was also significantly lower in invasive NFPAs compared to the E
DOI: 10.1021/acs.analchem.9b00914 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
Figure 3. FOLR1+ EVs and EpCAM+ EVs can markedly distinguish invasive NFPAs patients from noninvasive ones. (A) TEM image of EVs extracted from patients’ sera. (B) NTA analysis of the size distribution of EVs extracted from patients’ sera. (C) Flow cytometry analysis of the expression of FOLR1 from EVs of noninvasive (n = 10) and invasive NFPAs patients (n = 10) (***, P < 0.001). (D) Flow cytometry analysis of the EpCAM expression of noninvasive patients (n = 10). (E) Representative flow cytometry histogram of the expression of FOLR1 and EpCAM in EVs derived from one patient. (F) Immunogold TEM analysis of the expression of EpCAM and FOLR1 in serum EVs from a representative noninvasive NFPAs patient. (G) Western blot analysis of the expression of FOLR1 and EpCAM in EVs extracted from noninvasive and invasive PA patients. Flotillin-1 and CD81 were used as positive control, and HSA was used as negative control (uncropped blot in Figure S5). (H) Receiver operating characteristic (ROC) curve of the percentage of FOLR1 and EpCAM EV-bound beads and EV protein concentration (blue line) in distinguishing noninvasive (n = 10) and invasive (n = 10) NFPAs patients. The area under the ROC curve (AUC) is indicated.
noninvasive ones (unpaired Student t test, **, P < 0.01, Figure 3D). Immunogold TEM analysis confirmed the expression of FOLR1 and EpCAM in serum EVs from the noninvasive NFPAs patient (Figure 3, parts E and F). These results were verified with Western blot analysis which also showed lower expression of FOLR1 and EpCAM in invasive NFPAs compared to the noninvasive ones (Figure 3G; uncropped blots in Figure S3). In addition, we observed high expression of two positive markers, CD81 and flotillin-1, and no expression of the negative marker human serum albumin (HSA) in the
isolated EVs, which demonstrated the purity of the isolated EVs from patient sera (Figure 3G). We further performed ROC analysis to evaluate the discriminatory efficacy of FOLR1 and EpCAM in distinguishing invasive and noninvasive NFPAs. We found that both FOLR1 and EpCAM had excellent sensitivity and specificity in discriminating invasive and noninvasive NFPAs [area under curve (AUC) 0.940, 95% confidence interval (CI) 0.8331 for FLOR1, and AUC 0.880, 95% CI 0.7278 for EpCAM; Figure 3H, Table S-1]. As it has been reported that the protein concentration of serum EVs could also be a used as a F
DOI: 10.1021/acs.analchem.9b00914 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
Figure 4. Invasive NFPAs patients exhibited epithelial−mesenchymal transition (EMT). (A) Normalized expression level of N-cadherin mRNA, vimentin mRNA, EpCAM, and E-cadherin in EVs derived from noninvasive (n = 11) and invasive (n = 11) NFPAs patients (vimentin mRNA, unpaired Student t test, *, P < 0.05; EpCAM, unpaired Student t test, **, P < 0.005; E-cadherin, unpaired Student t test, *, P < 0.05). (B) The relative EMT ratio in EVs of the noninvasive and invasive NFPAs patients. The relative EMT ratio was calculated according to the equation in the Materials and Methods, unpaired Student t test, ***, P value
