A MoS2–MWCNT based fluorometric nanosensor for exosome detection and quantification

Circulating exosomes in body fluids are involved in many diseases and have important roles in pathophysiological processes. Specifically, they have emerged as a promising new class of biomarkers in cancer diagnosis and prognosis because of their high concentration and availability in a variety of biological fluids. The ability to quantitatively detect and characterize these nano-sized vesicles is crucial to make use of exosomes as a reliable biomarker for clinical applications. However, current methods are mostly technically challenging and time-consuming which prevents them from being adopted in clinical practice. In this work, we have developed a rapid sensitive platform for exosome detection and quantification by employing MoS2–multiwall carbon nanotubes as a fluorescence quenching material. This exosome biosensor shows a sensitive and selective biomarker detection. Using this MoS2–MWCNT based fluorometric nanosensor to analyze exosomes derived from MCF-7 breast cancer cells, we found that CD63 expression could be measured based on the retrieved fluorescence of the fluorophore with a good linear response range of 0–15% v/v. In addition, this nanosensing technique is able to quantify exosomes with different surface biomarker expressions and has revealed that exosomes secreted from MCF-7 breast cancer cells have a higher CD24 expression compared to CD63 and CD81.


Introduction
Exosomes are nano-sized (30-150 nm), cup-shaped 1 membranebound structures present inside large multivesicular endosomes 2 and released by most eukaryotic cells. 3 In recent years, extracellular vesicles (EVs), especially exosomes, have received increasing attention due to their cargo of protein, RNA, and DNA from their origin cell. 4 It is currently believed that cells use EVs as a means of extracellular communication and exchange of proteins, lipids and nucleic acids. 5 Recent studies further revealed that many biological uids (including saliva, lymph, urine, milk, blood, synovial and amniotic uids) contain vesicles which are released by early-stage tumors and carry various tumor markers. [6][7][8] These micro-and nano-vesicles are potentially measurable in easily accessible body uids and can be useful as a diagnostic and prognostic biomarker for early detection of many diseases and disorders, mainly, as essential indicators of the state and progression of cancer. 9 According to previous studies, exosomes originating from tumors are an invaluable source of cancer biomarkers. 10 The most common exosome marker target proteins are tetraspanins including CD9, CD63, and CD81. Tumor-associated markers are epithelial cell adhesion molecules (EpCAM), IGF-1R a units (a-IGF-1R), CA125, CD41b, and E-cadherin. [11][12][13] In addition to tetraspanins and tumor associated markers, there are many other promising biomarkers which have important application in medical elds. For instance, CD24 is a biomarker expressed in breast cancer solid tumor and in hematological malignancies. The expression of this protein in ovarian cancer, non-small cell lung cancer, and prostate cancer has also been widely studied and it has been shown that CD24 is associated with an adverse prognosis in these cancers. 14,15 In another study, it has been reported that there is a higher quantity of total exosomes in cancer patients compared with healthy volunteers as exosomes are secreted in larger amounts during carcinogenesis. 16 Hence, achieving an accurate technique to analyze these small vesicles from a heterogeneous biological uid which contains many other biomolecules with similar physical characteristics is essential for both diagnosis and therapy purposes. 17,18 Biomarkers are commonly analyzed by using immunological methods such as enzyme linked immunosorbent assays (ELISA) or Western blot (WB) which are expensive and lengthy assays. 19 Current protocols for exosome concentration measurement are still very semi-quantitative and mainly rely on procedures of total protein concentration or nanoparticle tracking analysis. In most of these current techniques, purication and isolation procedures such as ultracentrifugation are essential before concentration measurement which is a time-consuming process. 4,16 Hence, rapid and sensitive detection of exosomes is still a challenge. Accordingly, a reliable and sensitive detection of exosomes originating from cancer cells is needed for early diagnosis and prognosis of cancer. 20,21 Two-dimensional nanomaterials such as graphene oxide and transition metal dichalcogenides (TMDs) have been widely used as nano-quenchers in biomolecule detection in recent years. 22,23 The mechanism of detection with these nano-quenchers is based on uorescence resonance energy transfer (FRET), in which the uorophore (donor) transfers energy to the nanoquencher (acceptor) and provides an "on-off" sensor to detect the biomolecules of interest. 24,25 MoS 2 -MWCNT is a threedimensional hierarchical nanostructure, which has onedimensional MWCNT backbones with two-dimensional MoS 2 nanosheets grown on the surface of MWCNTs and also possesses partially standing branch features. 26 Preferably, the growth of MoS 2 nanosheets on the outer layer of MWCNTs provides an even larger area for adsorbing the target molecule compared to other 2D materials. 27 Based on immunoelectron microscopy studies, a variety of endocytic membranes have tetraspanin proteins (CD63, CD81, and CD9) which have been identied as the most abundant exosomal markers among various types of exosomes. 28,29 It is therefore assumed that the number of CD63 positive exosomes represents their total number. Herein, we have developed a method using MoS 2 -MWCNT and the CD63 antibody to quantify the exosome concentration in biological samples. To the best of our knowledge, this is the rst study to apply MoS 2 -MWCNT in exosome analysis using its capacity of biomolecule adsorbing and uorescence quenching. CD63 protein as the most common biomarker in all types of exosomes is chosen to identify exosomes in this study. Fluorescently labeled anti-CD63 is adsorbed on the surface of MoS 2 -MWCNT; however their interaction is still not fully understood. 30 The total exosome concentration was measured by using a nanoparticle tracking analysis (NTA) system. In addition, the application of this method for measuring the abundance of other types of exosomal biomarkers has been further studied to distinguish subpopulations of exosomes based on the concentration of biomarkers of interest.

Materials
Anti-human CD63 PE (R-phycoerythrin) and CD24 FITC (uorescein isothiocyanate), monoclonal antibodies (Abs), were purchased from eBioscience, Inc. CD81 monoclonal antibody (M38) FITC and Total Exosome Isolation Reagent were obtained from Thermo Fisher Scientic, USA. L-MWCNT-60100 (Shenzhen Nanotech Port Co., Ltd) was used in the synthesis of a MoS 2 -MWCNT nanostructure according to previous work. 26,31 Fluorescence spectra were recorded by using a Tecan microplate reader innite M200 at the excitation wavelengths of 488 and 494 nm based on the used uorophores. Phosphate buffered saline, 1X solution (Fisher Scientic, Inc.) and ultrapure water obtained from a Millipore ltration system were used as the dissolving solutions. All the reagents applied in the breast cancer cell (MCF-7) culture including Dulbecco's modication of Eagle's medium (DMEM), 10% fetal bovine serum (FBS), and antibiotics including penicillin and streptomycin were purchased from Thermo Fisher Scientic, USA and the MCF-7 cell line was purchased from the American Type Culture Collection (ATCC cat. no. HB-72). All NTA analyses were performed by using a ZetaView system, Particle Metrix GmbH.

Characterization
The morphology of the MoS 2 -MWCNT nanostructure and its absorbance spectrum were obtained by using a scanning electron microscope (JEOL JSM-7600F) and a UV-vis spectrophotometer (Tecan microplate reader innite m200), respectively. A Zetasizer Nano Z system (Malvern Instruments Ltd.) was used to measure the zeta potential of the MoS 2 -MWCNT nanostructure.

MoS 2 -MWCNT synthesis
The solvothermal method was applied to synthesize the MoS 2 -MWCNT nanostructure according to the protocol developed in a previous study. 31 Briey, 220 mg (NH 4 ) 2 MoS 4 powder as the single reactant and 100 mg MWCNTs were mixed with 30 mL N,N-dimethylformamide (DMF) as the solvent. A uniformly dispersed solution was obtained aer sonication and then placed in an autoclave (200 C) for 10 hours to form MoS 2 -MWCNT composites. Aer purication and at least 5 times washing with DI water, the MoS 2 -MWCNT nanostructure was obtained with a Mo/C ratio of 1 : 5. The stock solution of MoS 2 -MWCNT is very stable and does not need to be maintained under special lab conditions. It was stored at room temperature and sonicated for one hour before each experiment.

Sample preparation
The MCF-7 cells were sub-cultured every 48 or 72 hours to obtain 80-90% conuence under 37 C, 5% (v/v) CO 2 in a humidied incubator. During sub-culturing, 0.25% trypsin-EDTA solution was used for dissociation of cell monolayers. Aer achieving 80-90% conuence, the cell culture supernatant was collected and centrifuged at 2000g for 30 min before ltering through a 0.22 mm lter. To establish an assay protocol for specic exosome detection and quantication, we rst used a commercial total exosome isolation reagent according to the manufacture's protocol and puried the exosomes through a precipitation method. Aer precipitation of exosomes as a pellet, it was resuspended in PBS solution and was kept at 4 C for immediate use, or at À20 C for long-term storage. Then we characterized the isolated exosomes by using a NTA system to measure the concentration and size distribution.
To measure the quenching ability of the MoS 2 -MWCNT nanocomposites and obtain the optimum concentration of the nanoquencher, a series of different concentrations of MoS 2 -MWCNT nanocomposites were prepared in PBS solution and 2% v/v anti-CD63-PE (with 0.5 mg mL À1 nal concentration) was added to each sample.

CD63 positive exosome detection
To demonstrate the positive detection of the CD63 exosome, the MoS 2 -MWCNT nanostructure was incubated with 2% v/v (0.5 mg mL À1 ) of anti-CD63-PE at room temperature in a dark environment. Aer 10 minutes, a series of concentrations of exosome solution from 5 to 800 v/v% were added to the quenched solution and incubated at room temperature for 60 minutes (the nal volume of each sample was 50 mL, and the nal concentration of the antibody was 0.1 mg mL À1 ). Then the samples were added to the individual wells on a black 384-well at-bottomed microplate and recovered uorescence of anti-CD63-PE was measured by using a microplate reader for each sample.

Specicity of exosome detection
The specicity of the developed exosome sensing technique was examined by measuring the relative uorescence intensity recovery of anti-CD63-PE, anti-CD81-FITC, and anti-CD24-FITC when mixed with the MCF-7 cell culturing medium. The supernatant of the MCF-7 cell culturing medium was collected aer 48 hours of incubation in a humidied incubator and was centrifuged at 2000g for 30 min to remove dead cells, large proteins and cellular debris as a pellet. Then, it was ltered through a 0.22 mm lter. Before adding the quenched anti-CD63-PE/MoS 2 -MWCNT complex, 500 mL of the total exosome isolation reagent was added to 1 mL of the centrifuged cell media and the exosomes were isolated according to the protocol provided by the manufacture. As the type and amount of dye in each dye-labelled antibody are different, the uorescence recovery rate of the samples was measured with respect to the uorescence of each dye-labelled antibody (0.1 mg mL À1 in PBS).

Results and discussion
The proposed nano-sensor provides a rapid and sensitive technique to measure exosome concentration and could detect any biomarker of interest. Fig. 1 shows the MoS 2 -MWCNT based uorescence sensing platform for exosome detection. Firstly, anti-CD63-PE is adsorbed on the surface of MoS 2 -MWCNT and its uorescence is quenched off due to the uorescence resonance energy transfer (FRET) from the PEconjugated antibody to the MoS 2 -MWCNT nanostructure. Aer introducing exosomes into this mixture, due to the higher binding force between target exosomes and anti-CD63-PE, a detachment of anti-CD63-PE from MoS 2 -MWCNT causes a rapid uorescence recovery. The morphology and absorbance spectra of the MoS 2 -MWCNT nanostructure are shown in Fig. S1 and S2, † respectively. MoS 2 has a layered structure extended out of cylindrical tubules and provides a much larger surface area to adsorb anti-CD63-PE. It also shows a broad absorbance spectrum from the UV to NIR region. According to the Beer-Lambert law A ¼ (a Â L) Â C, where A is the measured absorbance, a is the extinction coefficient, L is the path length and C is the concentration. The extinction coefficient was calculated to be 102.31 L g À1 cm À1 which is a much higher value compared with that of previously reported 2D layered materials. 32,33 Based on the absorbance measurement of six different MoS 2 -MWCNT concentrations, 500 mg mL À1 shows the highest absorbance value. In addition, according to the quenchability data ( Fig. 2a and b), the MoS 2 -MWCNT sensor shows the total quenching of the uorescence at this concentration, which conrms that the optimum concentration of MoS 2 -MWCNT is 500 mg mL À1 . To assess the quenching ability of MoS 2 -MWCNT, the quenching uorescence data (I 0 /I C ) are plotted versus the concentration of MoS 2 -MWCNT (Fig. 2b). As can be seen, the uorescence quenching data have a nonlinear trend. Therefore, the traditional linear form of the Stern-Volmer (SV) model could not be applied to describe the quenching ability of the MoS 2 -MWCNT nanostructure as a quencher for anti-CD63-PE. The positive deviation at a high concentration of the quencher could be attributed to the simultaneous presence of dynamic (collisional) and static quenching. 34,35 Chen et al. 36 have developed a general nonlinear SV model based on the Sips adsorption isotherm model 37 to describe the uorescence quenching with nonlinear trends that could be written as the following equation: where C is the concentration of the quencher, and I 0 and I C are the uorescence intensity in the absence and presence of the quencher, respectively. K S and K D are the static and dynamic quenching constants, respectively. The exponent n refers to the effective adsorption sites, adsorption strength or the index of affinity heterogeneity. The parameter f refers to the fractional accessible sites for the uorophore in the quencher, which should be between 0 and 1. Table S1 † presents the parameters of the general nonlinear SV model for the data shown in Fig. 2b. The near unity value for parameter f indicates the efficient quenching ability of MoS 2 -MWCNT. This parameter is in good agreement with a high n value (3.474), which could be attributed to the high affinity of anti-CD63-PE to MoS 2 -MWCNT as a quencher. Furthermore, the combination of static and dynamic quenching, giving rise to the nonlinear quenching behavior of MoS 2 -MWCNT, could be attributed to the nonuniform surface with inequivalent adsorption sites existing in MoS 2 -MWCNT. To conrm the adsorption of anti-CD63-PE on the surface of the MoS 2 -MWCNT nanostructure, the zetapotential of MoS 2 -MWCNT was measured before and aer addition of the antibody (Fig. S3 †). The substantial change in the zeta-potential from À30 mV to À13 mV aer the addition of the antibody shows that it is successfully attached to the MoS 2 -MWCNT nanostructure. The original concentration and size distribution of exosomes were measured by using nanoparticle tracking analysis (NTA), which uses the Brownian motion of particles in liquid and characterizes exosomes based on their size and concentration (Fig. 3). The recovered uorescence intensity indicates the increase of recovered uorescence by increasing the concentration of the exosome (Fig. 4a). The calibration curve of recovered uorescence versus concentration of the exosome is shown in Fig. 4b which exhibits a linear region between 0 and 15 v/v% (0-11.1 Â 10 6 particles per mL) concentrations with a detection limit of 2.0408 v/v% (14.8 Â 10 5 particles per mL) based on the 3s rule. Cancer derived exosomes in body uids are available in high concentrations (10 9 -10 12 exosomes per mL   in blood), 38 which are comparable to normal exosome concentration (10 11 exosomes per mL in blood). 8 It is thus conceivable that the application of MoS 2 -MWCNT to exosome detection provides an assay with a sufficient detection sensitivity.
To demonstrate the mechanism of exosome sensing via the MoS 2 -MWCNT sensor, the uorescence spectrum of anti-CD63-PE alone and in the presence of an exosome was recorded (Fig. 5, curves a and b). It showed a strong uorescence emission in both measurements (peak at 575 nm) owing to the presence of PE (R-phycoerythrin), while upon the addition of the nano-quencher, the uorescence intensity of anti-CD63-PE dropped signicantly, which indicates that most of the anti-CD63-PE was adsorbed on the surface of MoS 2 -MWCNT and caused the adequate quenching of the uorescence (Fig. 5, curve c). However, aer adding the target exosome, the uorescence was recovered considerably. This trend shows a high binding rate of anti-CD63-PE and exosomes which prevents placing the PE uorophore at close distances to the nanoquencher and stops FRET occurring between them (Fig. 5, curve d).
The kinetic behavior of anti-CD63-PE and exosome hybridization was studied by monitoring the recovery of uorescence intensity in 60 min aer introducing exosome solution into the quenched anti-CD63-PE/MoS 2 -MWCNT complex (Fig. 6). The preparation time for the quenched complex was less than 5 minutes and uorescence recovered very fast in a few seconds, once the exosome solution was added to the quenched complex. However, the uorescence intensity was not stable in the rst 10 minutes, which could be due to some false interactions between exosomes and anti-CD63-PE. Based on this gure, aer one hour, uorescence was more stable which demonstrates a specic and stable binding to the antigen. Hence, all the uorescence measurements in this study were also done aer one hour to achieve a stable uorescence intensity. Table S2 † presents a comparison of the reported results for the LOD and detection time of exosomes via different methods. The results indicate that the developed uorometric method in this study has an acceptable LOD as well as detection time, and can compete with other methods due to its simple procedure and low equipment cost.
For testing the specicity of this sensor, uorescent dyelabeled anti-CD63, CD81, and CD24 were used as the sensing agents for the detection of specic exosomes derived from MCF-7 cells. According to Fig. 7, the uorescence recovery of CD24 is much higher than that of CD63 and CD81, and it is in good agreement with previous studies which have reported an enhancement in the concentration of CD24 protein in cancer   cells. 6,15 It is therefore concluded that this detection platform is highly sensitive to the concentration of biomarkers at different expression levels and has a great potential application in the detection of specic biomarkers for diseases.

Conclusions
In this study, we developed an easy, fast and efficient sensing platform to detect exosomes with specic biomarkers using MoS 2 -MWCNT nanocomposites. This uorescence-based biosensor has a linear range of 0-15 v/v% (0-11.1 Â 10 6 particles per mL) with a detection limit of 2.0408 v/v% (14.8 Â 10 5 particles per mL) which is lower than the exosome concentration range in body uids 8 and is comparable to other methods (e.g., 5 Â 10 7 particles per mL 39 and 2.2 Â 10 7 exosomes per mL 8 ). Also, we tested the specicity of this sensor which showed a high selectivity for the biomarkers of interest. The robustness of the biosensor provides a mechanism to measure different biomarkers upon choosing the related antibodies. It is therefore not limited to detect only anti-CD63 and has the potential to measure any other biomarker. Furthermore, this technique could be a potential platform for in situ detection of exosomes for early detection and therapy of cancer.

Conflicts of interest
There are no conicts to declare.