Skip to main content
NCBI home page
ACS Omega logoLink to ACS Omega
. 2020 Jul 1;5(27):16602–16611. doi: 10.1021/acsomega.0c01441

Mixed-Dimensional Heterostructure Material-Based SERS for Trace Level Identification of Breast Cancer-Derived Exosomes

Avijit Pramanik †, Justin Mayer ‡, Shamily Patibandla †, Kaelin Gates †, Ye Gao †, Dalephine Davis †, Ram Seshadri ‡, Paresh Chandra Ray †,*
PMCID: PMC7364584  PMID: 32685826

Abstract

graphic file with name ao0c01441_0005.jpg

Raman spectroscopy has capability for fingerprint molecular identification with high sensitivity if weak Raman scattering signal can be enhanced by several orders of magnitudes. Herein, we report a heterostructure-based surface-enhanced Raman spectroscopy (SERS) platform using 2D graphene oxide (GO) and 0D plasmonic gold nanostar (GNS), with capability of Raman enhancement factor (EF) in the range of ∼1010 via light–matter and matter–matter interactions. The current manuscript reveals huge Raman enhancement for heterostructure materials occurring via both electromagnetic enhancement mechanism though plasmonic GNS nanoparticle (EF ∼107) and chemical enhancement mechanism through 2D-GO material (EF ∼102). Finite-difference time-domain (FDTD) simulation data and experimental investigation indicate that GNS allows light to be concentrated into nanoscale “hotspots” formed on the heterostructure surface, which significantly enhanced Raman efficiency via a plasmon–exciton light coupling process. Notably, we have shown that mixed-dimensional heterostructure-based SERS can be used for tracking of cancer-derived exosomes from triple-negative breast cancer and HER2(+) breast cancer with a limit of detection (LOD) of 3.8 × 102 exosomes/mL for TNBC-derived exosomes and 4.4 × 102 exosomes/mL for HER2(+) breast cancer-derived exosomes.

1. Introduction

Raman spectroscopy has the ability to provide fingerprint-type identification, which is very important for medical and forensic investigation, as well as other fields.1−5 The main disadvantage for Raman spectroscopy to be used for fingerprint recognition is that Raman scattering is very week where only 1 out of 10 million photons is scattered.4−9 To overcome this, several groups are working on surface-enhanced Raman scattering (SERS) spectroscopy, where the Raman intensity can be enhanced by several orders of magnitude (106 to 1014).6−15 In case of SERS, huge enhancement of Raman intensity for adsorbed molecules on the surface occurs via the electromagnetic mechanism (EM) and the chemical mechanism (CM).6−16 As we and others have reported, the EM exploits the excitation of localized surface-plasmon resonances (LSPR) in plasmonic zero-dimensional (0D) nanostructured materials such as Au, Ag, and Cu.6,11−20 Because of the strong light–matter interaction via plasmon-excitation coupling, the EM enhancement factors can be on the order of ∼106 to 107.21−27 On the other hand, the CM has capability to enhance the Raman signal by ∼101 to 103 times based on dipole–dipole interactions or charge-transfer resonances between the surface and Raman active molecules.24−31 We and others have reported that two-dimensional (2D) nanomaterials such as graphene,11,12,14,19−25 transition-metal dichalcogenides (TMD),15,17,26,27 hexagonal boron nitride (h-BN),26,27 and others26,27 can show CM enhancement higher than 102 when the laser excitation can be resonant to charge-transfer transitions.

Herein, we report huge Raman enhancements from mixed-dimensional heterostructure platform using 2D graphene oxide (GO) and plasmonic gold nanostar (GNS), which has capability for several orders of magnitude Raman enhancement via light–matter and matter–matter interactions. In our design, 2D-GO has been used been used for the SERS enhancement via the chemical enhancement mechanism.14,15,18−20 On the other hand, anisotropic GNS has been used for huge SERS enhancement via the plasmonic enhancement mechanism and “lightning rod effect”.8−14,29 Over the past few years, several groups have designed GO/GNS hybrid nanocomposites for SERS.19,23−25 Most of the synthetic procedure has been used is the direct growth of GNS on GO sheets, where controlling the amount of GNS on GO is difficult. It is now well understood that electromagnetic “hotspots” are the most important parameter to achieve huge SERS enhancement.1−10 To generate huge amount electromagnetic hotspots, we synthesized GNS using 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer. After that, we have performed controlled attachment of GNS on the 2D-GO surface for developing hotspots. Reported data show that because of the formation of good amount electromagnetic hotspots, we have obtained Raman EF is in the range of ∼1010 from GNS-attached GO-based heterostructure material, which is higher than the reported EF data for GO/GNA assembly.19,23−25

As a proof of concept, we have demonstrated that GO-GNS mixed-dimensional heterostructure-based ultrasensitive SERS is versatile for trace level fingerprint identification of exosomes derived from triple-negative breast cancer and HER2(+) breast cancer. Triple-negative breast cancer (TNBC), which lacks estrogen receptors (ER), progesterone receptors (PR), and hormone epidermal growth factor receptor 2 (HER2) protein, is usually highly aggressive and difficult to treat.32−35 Several recent reports show that exosomes can be minimally invasive biomarkers for accessing the stage for TNBC and other types of cancer.36−45 Exosomes are well-documented cell-derived vesicles which are loaded with variety of proteins, micro RNA, noncoding RNAs, and DNA.40−49 Because exosomes can be obtained easily from biological fluids in clinics, they have been considered as minimally invasive cancer biomarkers.38−43

Recently, SERS has been used as noninvasive assay for EV study.35,39−49 For this purpose, SERS nanotags have been used as an alternative label to fluorescent dyes for exosome detection.35,39−49 For instance, Zhang et al. reported35 the design of magnetic beads and Raman reporter-based SERS for rapid capture of exosome using antibody-conjugated magnetic beads and profiling using antigen-targeting SERS nanotags. Their results show that the SERS assay is a sensitive approach for the detection of small EVs, with a limit of detection (LOD) of 2.3 × 106 exosomes/mL. Similarly, Tian et al.44 have reported exosome identification using gold nanostar@Raman reporter@nanoshell-based SERS with an LOD of 2.7 × 104 exosomes/mL. Furthermore, Kwizera et al.47 have reported the design of a miniaturized chip platform using gold nanorods coated with QSY21 Raman reporters for phenotype profiling of small EVs, with an LOD of 2.0 × 106 exosomes/mL. For SERS-based label-free identification of exosomes, in the current manuscript, we have used 2D–0D heterostructure-based SERS for tracking of cancer-derived exosomes from TNBC and other types of breast cancer. Reported data show that heterostructure-based SERS has capability for label-free fingerprint identification of exosomes, with an LOD of 3.8 × 102 exosomes/mL for TNBC-derived exosomes and 4.4 × 102 exosomes/mL for HER2(+) breast cancer-derived exosomes. The observed very good LOD is mainly due to the formation of huge number electromagnetic hotspots on SERS substrate developed by us.

2. Materials and Methods

2.1. Design and Characterization of GO-GNS Multidimensional Heterostructure

As shown in Scheme 1A–C, mixed-dimensional heterostructures were synthesized using a three-step process. As shown in Scheme 1A, in the first step, we have synthesized GNS using tetrachloro Au (III), silver nitrate, sodium hydroxide, and HEPES buffer, as we and other groups reported before.8−10,29 In our synthetic procedure, HEPES has been used as precise shape-directing agent.29 To enhance stability as well as for developing biocompatible materials, GNSs were coated with SH-PEG2000–NH2. Experimental details for the synthesis of PEG-coated GNS have been reported in the Supporting Information. After purification via centrifugation for 1 h at 3500 rpm, PEG-coated GNSs were characterized by tunneling electron microscopy (TEM), X-ray diffraction (XRD), and absorption technique.14−17Figure 1A shows the TEM image of freshly prepared PEG-coated GNS. Figure S1A shows the TEM image of freshly prepared GNS without PEG. From TEM data, we can find that the size of bare GNS is ∼35 ± 10 nm and the size of PEG-coated GNS is ∼50 ± 10 nm. The excitation spectra from GNS, as reported in Figure 1G, shows a strong plasmon band with peak maxima for localized surface plasmon resonance (LSPR) around 580 nm. Figure S2 shows the XRD spectra from PEG-coated GNS which shows the presence (111), (200), (220), (3111) reflection for GNS.

Scheme 1. Mixed-Dimensional Heterostructure Synthesized Using a Three-Step Process.

Scheme 1

Open in a new tab

(A) Synthetic path has been used for the development of (polyethylene glycol) PEG-coated GNS from tetrachloro Au (III), in the presence HEPES buffer. (B) Synthetic path used for the development of graphene oxide from graphite. (C) Synthetic procedure used to develop mixed dimensional heterostructures from PEG-coated GNS and GO.

Figure 1.

Figure 1

Open in a new tab

(A) TEM image shows the morphology of PEG-coated GNS. Inserted picture shows a high-resolution image which indicates that the size is ∼50 ± 10 nm. (B) SEM image shows the morphology of 2D-GO. (C) TEM image shows the morphology of mixed-dimensional heterostructures based on GO and GNS. Inserted picture shows a high-resolution image. (D) SEM image shows the morphology of mixed-dimensional heterostructure-based on GO and GNS. Inserted picture shows a high-resolution image. (E) EDX spectra from mixed-dimensional heterostructure, which indicates the presence of Au, C and O. (F) XRD spectra from mixed dimensional heterostructure, which shows the presence of (002) reflection for GO and (111), (200), (220), (3111) reflection for GNS. (G) Extinction spectra from PEG-coated GNS, GO, and mixed-dimensional heterostructure. (H) Raman spectra from only GO and GO-GNS-based mixed-dimensional heterostructure.

In the second step, we have developed graphene oxide (GO) using modified Hummer’s method, as we and others have reported before.12,14,15,18−22 For this purpose, we have used sodium nitrate, concentrated sulfuric acid, and potassium permanganate as strong oxidizing agents to synthesize graphene oxide from graphite. Experimental details have been reported in the Supporting Information. Finally, H2O2 (50%) was added to the mixture till there was no gas generated. The mixture turned from almost black to yellow. After purification, we have used scanning electron microscopy (SEM) and absorption spectroscopy technique to characterize 2D-GO. Figure 2B shows the SEM image of GO flakes synthesized from graphite using the Hammers method. As shown in Figure 2G, we have not observed any strong absorption peak from GO in the visible region.

Figure 2.

Figure 2

Open in a new tab

(A) Plot shows how the Raman profile from 4-ATP varies in the presence of Si/SiO2, GNS, 2D-GO, and mixed-dimensional heterostructure surfaces. (B) Plot shows how Raman profile from Rh-6G varies in the presence of Si/SiO2, GNS, 2D-GO, and mixed-dimensional heterostructure surfaces. (C) Bianalyte SERS spectra from a mixture of 4-ATP (∼1.4 × 10–9 M) and Rh-6G (∼1.4 × 10–9 M) in the presence of a mixed-dimensional heterostructure surface. We have compared the bianalyte spectra with single-analyte SERS spectra for 4-ATP (∼8.2 × 10–10 M) and Rh-6G (∼8.2 × 10–10 M) separately. (D) Plot shows how Raman intensity at 1507 cm–1 for Rh-6G varies with the concentration in the presence of a mixed-dimensional heterostructure surface. Each data represents the average of four separate experiments. Error bars represent the standard deviation of measurements. (E) Plot shows how Raman intensity at 1078 cm–1 for 4-ATP varies with the concentration in the presence of a mixed-dimensional heterostructure surface. Each data represents the average of four separate experiments. Error bars represent the standard deviation of measurements. (F) FDTD simulation data show how the (|E|2) profile from a sharp branch of GNS varies with the distance for the dimer. TEM image for single three spike-based GNS particle structure has been shown in (F1), which can be simplified as a triangular shape. TEM image for another single three spike-based GNS particle structure has been shown in (F2), which also can be simplified as a triangular shape. (F3) shows FDTD simulation data, where we have taken a triangular structure for our calculation.

Because of the lack of plasmon band in the visible region, Raman enhancement from 2D-GO surface can be mainly due to chemical enhancement. In the next step, we have developed mixed-dimensional heterostructures using GO and PEG-coated GNS. As shown in Scheme 1C, amino functionalized GNSs were covalently linked with graphene oxide nanosheets using 1-ethyl-3-[3-dimethylaminopropyl]carbodiimide (EDC)—N-hydroxysuccinimide (NHS) coupling chemistry.11−16 Experimental details have been reported in the Supporting Information. At the end, the unreacted GO and GNS were removed by centrifugation at 10,000 rpm for 30 min, followed by decantation, and then, the pellet was resuspended in nanopure water for further use. As reported in Figure 1, the morphologies of GNS-attached Graphene oxide-based mixed dimensional heterostructures were characterized by UV–vis, Raman, XRD, TEM, SEM, and energy-dispersive X-ray spectroscopy (EDX) techniques.14−17

The TEM and SEM images for mixed-dimensional heterostructures are reported in Figure 1C,D. Reported images clearly show that GNSs formed assembly structure on the surface of graphene oxide. Figure 1E shows the EDX data from the mixed-dimensional heterostructures, which clearly indicates the presence of C, O, and Au. XRD data from the mixed-dimensional heterostructures, as reported in Figure 1F, shows the presence of broad (002) reflection for GO. XRD data also indicates the presence of (111), (200), (220), and (3111) reflection for GNS.14,15,18−22 The excitation spectra from the mixed dimensional heterostructures, as reported in Figure 1G, shows a very broad plasmon band above 630 nm, which is due to the assembly structure of GNS on a 2D-GO surface, as we have observed in TEM and SEM images reported in Figure 1C,D. In the self-assembly of GNS, the interparticle plasmon–plasmon couplings will produce plasmonic hot spots where massive near-field amplifications will occur via light–matter and matter–matter interactions.8−13 The above phenomena will enable highly sensitive SERS detections via mixed-dimensional heterostructures. The Raman spectra from only GO and GO-GNS in the mixed-dimensional heterostructure are reported in Figure 1H. Reported Raman spectra shows the presence of disorder band (D band) at ∼1350 cm–1 caused by the graphite edges and the G band at ∼1590 cm–1 caused by the in-plane vibration of sp2-hybridized carbon atoms.14,15,18−22 We have acquired Raman spectra from 10 to 15 spots and then spectra were processed with a baseline removal and averaging program. The observed enhancement of D and G bands in the presence of GNS in the mixed dimensional heterostructures is mainly due to the plasmon effect from GNS.

2.2. Raman Experimental Details Using GO, GNS, and Mixed-Dimensional Heterostructure Substrate

For the measurement of Raman EF, we have used the confocal Raman system with a laser excitation of 670 nm.15 For Raman data collection, we have used 100× magnification and a numerical aperture of 0.9 for our experiment.15 For Raman data collection, we have used 10 s acquisition time and 5-scan averaging, so that we could achieve a very good signal-to-noise ratio. For Raman measurement from TNBC- and HER2(+) breast cancer cell-derived exosomes, we have used a microscope attached with the confocal Raman microscopy system, to locate cluster of exosomes. Once we locate them, the laser beam was focused on cancer cell-derived exosomes cluster through the microscope object.

2.3. FDTD Simulation for Full-Field Electromagnetic Wave Calculations

In our calculation, the electric field intensities were simulated by using gold particle dimmer with 35 nm size for each. The amplitude of electric field was kept as 1 V/m, and courant number was taken as 0.99. We have used 670 nm wavelength, 0.001 nm mesh resolution, and 4000 fs time for our calculation, as we have reported before.12−16

2.4. Breast Cancer Cell Culture and Exosome Isolation from Cell Culture Medium

Triple-negative breast cancer cell line, MDA-MB-231 cells, and HER2(+) KSBR-3 cells were cultured with a culture medium suggested by ATCC, using the procedure we have reported before.12−17,33 When cancer cells were cultured to 60–70% confluency, we have replaced the standard culture medium with exosome-depleted medium, as reported before.33 After that cancer cells were cultured for an additional 80 h. In the next step, exosome was separated from cell media by differential centrifugation to remove cellular debris, as reported before.33 For this purpose, supernatants were collected from cell lines and centrifuged at 3000g for 25 min to eliminate cells and debris. After that we have centrifuged again at 10,000g for 30 min to eliminate microvesicles, as we have reported before.33 At the end, exosomes derived from MDA-MB-231 and SKBR3 cells separately were ultracentrifugated for 1 h and resuspended in PBS. In the next step, we have characterized exosomes suing TEM, DLS, and western blot.

3. Results and Discussion

3.1. Determine the Raman Enhancement Factor from Mixed-Dimensional Heterostructures

As we have discussed before, the Raman scattering signal has to be enhanced by several orders of magnitude before it can be used for fingerprint identification application.5−10 As a result, we have measured Raman EF for 2D-GO, GNS, and mixed-dimensional heterostructure using 4-aminothiophenol (4-ATP) and Rh-6G as the Raman probe. For the measurement of Raman EF, we have used confocal Raman system (Horiba Scientific) with a laser excitation of 670 nm and 4–6 mW power.15 For Raman data collection, we have used 100× magnification and a numerical aperture of 0.9 for our experiment.15 Before SERS experiment, mixed-dimensional heterostructure substrates were immersed in the different concentrations of 4-ATP or Rh-6G for 2 h, and that SERS substrate was dried under N2 flow. The distribution of GNS monomers, dimers, or clusters is not uniform throughout the substrate as reported in the TEM and SEM images for mixed-dimensional heterostructures. As a result, to obtain high-quality data, we have acquired Raman spectra from 10 to 15 spots, and then, the spectra were processed with a baseline removal and averaging program. The variation of relative standard deviation (RSD) values at different spots for GO-GNS substrate is ∼10.5% for 4-ATP and ∼13.1% for Rh-6G on the GNS-GO surface. To keep the experimental condition same, for each Raman experiment, we have acquired Raman spectra from 10 to 15 spots for GO, GNC, or GO-GN substrates, and then, the spectra were processed by baseline removal and averaging. Because several Raman probes have Raman active bands near D and G bands of GO, which can interfere with the actual intensity measurement, we have subtracted the Raman spectra of each Raman probe with substrates from Raman spectra of only substrates without probes. In every case, we have acquired Raman spectra from 10 to 15 spots, and then, spectra were processed with a baseline removal and averaging program. Raman spectra from different concentrations 4-ATP on GNS, GO, and mixed dimensional heterostructure surfaces is reported in Figure 2A. Raman peaks from 4-ATP on different surfaces are dominated by a1 and b2 vibrational modes.10−20 As shown in Figure 2A, the observed a1 modes are for ν(C–C + NH2 bend) at ∼1590 cm–1 and ν[ν(C–C) + δ(C–H)] at ∼1078 cm–1.11−18 On the other hand, we have observed b2 modes for CC stretch in Ph ring + NH2 rock at ∼1435 cm–1.11−18 We have also observed another strong Raman peak at 464 cm–1, which is due to ν(C–N) + ν(C–S) + γ(CCC) vibrations.11−18 Because among all the observed Raman bands from 4-ATP, Raman mode at 1078 cm–1 is the strongest, we have used that mode for Raman enhancement factor (EF) calculation for GNS, GO, and mixed-dimensional heterostructure substrates. For this purpose, we have used the following equation.11−20

3.1. 1

where Ihetero is the intensity of a1 vibrational mode from 4-ATP at 1078 cm–1 on the mixed-dimensional heterostructure surface. Because for bulk experiment, Si/SiO2 wafers were used as the normal Raman reference substrate, ISiO2 is the intensity of the same mode in the Si/SiO2 surface. Similarly, NSiO2 is the number of 4-ATP used in the Si/SiO2 surface and Nhetero is the number of 4-ATP used for the mixed-dimensional heterostructure surface. In order to calculate the number of 4-ATP used for the Raman experiment, the area of the surface under the focused illumination was determined. For this purpose, Nhetero is calculated using the following equation.

3.1. 2

where ARaman is the laser spot area used for Raman measurement and ASub is the effective area of the whole substrate where analytes are present. Chetero is the concentration of 4-ATP used for the Raman experiment and NA is the Avogadro number. V is the volume of the 4-ATP used for the Raman signal measurement. Similarly, for bulk experiment, NSiO2 is calculated using the following equation.

3.1. 3

where ARaman is the laser spot area used for bulk Raman measurement and ASub is the effective area of the whole substrate where analytes are present. CSiO2 is the concentration of 4-ATP used for the bulk Raman experiment and NA is the Avogadro number. V is the volume of the 4-ATP used for the Raman signal measurement in the bulk experiment.

As we have discussed before, the distribution of GNS monomers, dimers, or clusters is not uniform. As a result, we have measured the SERS EF values at 15 different spots, and average EF values are reported in this manuscript. Using Raman data for 4-ATP reported in Figure 2A, we have estimated the average Raman EF for 2D-GO is ∼1.8 × 102, which can be attributed to the chemical enhancement mechanism as we have discussed before. The observed Raman enhancement from GO can also be due to the resonance with near- and far-field optical properties.11−22 Similarly, using reported 4-ATP Raman data in Figure 2A, we have estimated the average Raman EF for GNS as ∼ 1.1 × 107. The observed very high EF for GNS can be attributed to strong electromagnetic enhancement mechanism.8−13 It is now well documented that sharp branches on GNS will create the lightning rod effect, which will enhance Raman intensity dramatically.8−13 Part of the huge Raman enhancement in the presence of a GNS surface can also be due to the chemical enhancement effect caused by the charge-transfer process happening on the ATP-GNS surface.11−20

On the other hand, using reported data in Figure 2A, we have estimated the average Raman EF for mixed-dimensional heterostructure to be ∼1.4 × 1010. We have observed excellent Raman EF from the mixed-dimensional heterostructure surface because of the strong electromagnetic as well as strong chemical enhancement capability of heterostructure. As reported in Figure 1C,D, in case of mixed-dimensional heterostructures, GNSs formed dimer and higher aggregates on the GO surface. As a result, the electromagnetic field experienced by the 4-ATP in “hot spot” formed by dimers or higher aggregates is much stronger than the field it will experience in monomer. Because of the above fact, a synergistic mechanism is expected in case of mixed-dimensional heterostructure material. Figure 2E shows how the Raman signal at 1078 cm–1 varies with the concentration of 4-ATP in the presence of a mixed-dimensional heterostructure surface. We have used the following equation to determine the limit of detection (LOD).11−20

3.1. 4

In this equation, σ is the standard deviation of the blank and S is the slope of the calibration curve. Standard deviation of blanks was determined from the baseline noise of SERS substrates without analytes over the range of Raman peak of interest. For this purpose, we have recorded 10–15 spectra from substrate in the absence of the analyte. Using the concentration-dependent data as reported in Figure 2E and eq 4, we have estimated the limit of detection (LOD) to be ∼3.1 × 10–13 M for the 4-ATP Raman probe in the presence of a mixed-dimensional heterostructure material.

To understand how the Raman EF values vary with different Raman probes, we have also performed same set of experiments with the Rh-6G Raman probe. Figure 2B shows the SERS spectra from Rh-6G at different concentrations on the GNS, 2D-GO, and mixed-dimensional heterostructure surface. The observed Raman modes from Rh-6G at 376 cm–1 is due to the N–C–C bending modes of the ethylamine group of the Rh-6G ring. Other observed prominent Raman modes are at 615 cm–1 due to the C–C–C ring in-plane bending mode, 778 cm–1 due to the C–H out-of-plane bending, 1181 cm–1 due to the C–H in-plane bending, 1366 cm–1 due to the C–N stretch, 1507 cm–1 due to the aromatic C–C stretching, and 1603 and 1650 cm–1 due to the C=N stretch, as we and others reported before.8−14 Using eqs 1–3 and reported data in Figure 2B, we have estimated the Raman EF to be ∼3.8 × 1010 for mixed-dimensional heterostructure, ∼2.6 × 107 for GNS, and ∼2.1 × 102 for 2D-GO. Although the Raman EF values vary a little bit for different Raman probes, the orders of magnitude remain the same. In both cases, we have observed ∼103 higher Raman EF from the mixed-dimensional heterostructure surface with respect to GNS, and it is due to the strong electromagnetic as well as chemical enhancement capability of the mixed-dimensional heterostructure surface. Figure 2D shows how the Raman signal at 1507 cm–1 varies with the concentration in the presence of a mixed-dimensional heterostructure surface. Using Figure 2D and eq 4, we have determined the LOD, which is estimated to be ∼2.5 × 10–13 M for a Rh-6G Raman probe in the presence of a mixed-dimensional heterostructure material.

Because we have observed very high sensitivity, to demonstrate that the observed Raman signals is coming from a few molecules, we have performed SERS experiment using a bianalyte Raman probe. Le Ru et al.16 have proposed that simultaneous use of two analyte molecules is necessary to determine whether the observed SERS signals are from the single or few molecules nature. Figure 3C shows the bianalyte SERS spectra from a mixture of 4-ATP (∼1.4 × 10–9 M) and Rh-6G (∼1.4 × 10–9 M) in the presence of a mixed-dimensional heterostructure surface. By comparing the bianalyte spectra with single-analyte SERS spectra for 4-ATP and Rh-6G separately, as reported in Figure 3C, we can conclude that the observed SERS signal from bianalyte is a mixture of SERS spectra from individual Raman probes. The above observation clearly evidenced that the SERS signal comes from a very small number of molecules.

Figure 3.

Figure 3

Open in a new tab

(A) TEM image of exosomes derived from MDA-MB-231 triple-negative breast cancer cells. (B) Raman Spectra from exosomes derived from TNBC cells (105 and 103 cell/mL) on a mixed-dimensional heterostructure surface. (C) Plot shows how SERS intensity at 1605 cm–1 for lipid varies with the concentration of exosomes derived from MDA-MB-231 triple-negative breast cancer cells. Each data represents the average of four separate experiments. Error bars represent the standard deviation of measurements. (D) TEM image of exosomes derived from HER2(+) SKBR3 breast cancer cells. (E) Raman Spectra from exosomes derived from SKBR3 cells (105 and 103 cell/mL) on a mixed-dimensional heterostructure surface. (F) Plot shows how SERS intensity at 1014 cm–1 for tryptophane varies with the concentration of exosomes derived from HER2(+) SKBR3 breast cancer cells. Each data represents the average of four separate experiments. Error bars represent the standard deviation of measurements.

To understand hot spot effects on SERS, we have performed the finite-difference time-domain (FDTD) simulation.11,14,15,30,31 As shown in Figure S1, several GNS structures developed by our group have three spikes and they can be simplified as a triangular structure where three corners can act as spikes for the lighting rod effect, as shown in Figure 2F(1,2). For FDTD calculation, we have used a simple triangular structure with three corners as spikes, as shown in Figure 2F(3). For the simulation, we have used a triangular gold particle with 30 nm length for each side. We have used 670 nm wavelength, 0.001 nm mesh resolution, and 4000 fs time for our calculation, as we have reported before.12−16Figure 2F shows how that the square of field enhancement (|E|2) in sharp branch for GNS varies with distance for dimer. Reported data clearly show that more than 2 orders of magnitudes higher field enhancement in hot spot position than that of an individual particle. Because the Raman EF α|E|4, our FDTD simulation data indicate that there is a possibility of 4 orders of magnitude Raman enhancement in hot spot position on the heterostructure. As we have discussed before, the distribution of GNS monomers, dimers, or clusters is not uniform and as a result, the measured EF value will vary at different spots for the mixed-dimensional heterostructure surface. As a result, the experimental average EF values reported in this manuscript are lower than expected theoretical values calculated by FDTD simulation.

Because adsorption affinity, molecular aggregation under dry condition, and other several factors can alter the estimated EF values from mixed-dimensional heterostructure, we have measured the average EF values on the heterostructure surface using Rh-6G as s Raman probe, when they have been prepared in different batches. Experimental data reported in Figure S3 show that the variation of RSD values for substrates made in different batches is ∼15.5% for Rh-6G. For this measurement, we have performed 10 different spots in each substrate and averaged it. Similarly, we have determined the variation of RSD values for substrates made in different batches for 4-ATP and it is ∼14.1%.

3.2. Mixed-Dimensional Heterostructure-Based SERS for Label-free Identification of TNBC- and HER2(+) Breast Cancer-Derived Exosomes

Inspired by the huge Raman EF, we attempted to explore whether mixed-dimensional heterostructure-based Raman can be used for tracking different cancer-derived exosomes from TNBC and HER2(+) breast cancer cells. Exosomes were derived from triple-negative breast cancer cell line, M.D. Anderson Metastasis Breast cancer (MDA-MB)-231 cells, and HER2(+)-type Sloan-Kettering Breast Cancer (SKBR) 3 cells separately using the procedure we have reported recently.33 For this purpose, cells were cultured first separately in a standard culture medium and then with an exosome-depleted medium. At the end, exosome was separated by centrifugation to eliminate cellular debris and microvesicles.

The TEM image reported in Figure 3A indicates that the size of freshly separated MDA-MB-231 cell-derived exosomes varies between 200 ± 50 nm. Table 1 shows the DLS measurement data which also indicate similar size for TNBC-derived exosomes, which are between 180 ± 60 nm. We have also performed the western blot test, which indicates that CD63 proteins are overexpressed by exosomes which are derived from MDA-MB-231 cells. Similarly, Figure 3D shows the TEM image of SKBR3 cell-derived exosomes, where the size varies between 18 0 ± 60 nm. DLS measurement also indicates similar size for SKBR-3-derived exosomes, which are between 170 ± 60 nm. The western blot test indicates that HER2 is overexpressed by exosomes derived from SKBR3 cells.

Table 1. Size Distribution for Exosomes Derived from MDA-MB-231 and SKBR3 Cells.

exosome sources size measured by DLS (nm) size measured by TEM (nm)
MDA-MB-231 cells 180 ± 60 200 ± 50
SKBR3 cells 170 ± 60 180 ± 60
Open in a new tab

For Raman measurement from TNBC- and HER2(+) breast cancer cell-derived exosomes, we have used a microscope attached with the confocal Raman microscopy system, to locate cluster of exosomes. Once we locate them, the laser beam was focused on cancer cells derived exosomes cluster through the microscope object. To acquire reproducible Raman spectra, we have measured the SERS data from 10 to 15 spots and then we have performed the baseline removal and averaging. Figure 3B shows Raman spectra from TNBC-derived exosomes in the presence of a mixed-dimensional heterostructure. The observed vibrational Raman bands can be assigned to the spectral contributions of mainly lipids, protein, and nucleic acids as reported in Table 2.40−43 As reported in Figure 3B, we have observed several lipid bands, and these are at ∼1605 cm–1 due to the ergosterol, ∼1454 cm–1 due to δ(CH3, CH2) in acyl chain, ∼1260 cm–1 due to δ(=CH2) in acyl chain, and ∼707 cm–1 due to cholesterol.40−43 On the other hand, we have observed one protein peak at ∼880 cm–1 due to tryptophan. Similarly, we have also observed one DNA peak at ∼1520 cm–1, which is due to Purine A, G ring. Figure 3C shows how the lipid Raman band intensity at ∼1605 cm–1 varies with the concentration of TNBC-derived exosomes. Concentration dependent SERS data are reported in Figure 3B,C. From the slope of the concentration dependent curve and using eq 4, we have calculated that the limit of detection (LOD) for TNBC-derived exosome is 3.8 × 102 exosomes/mL.

Table 2. Fingerprint Raman Peaks (cm–1) Observed from Exosomes Derived from MDA-MB-231 and SKBR3 Cancer Cellsa.

exosome from MDA-MB-231 cells exosome from SK-BR3 cells vibrational mode
  1630 tryptophan due to –C=O stretch
1605   lipid band due to the ergosterol
1510   DNA peak due to purine A, G ring
1454   lipid band due to δ(CH3, CH2) in acyl chain
  1388 DNA peak due to NH in-plane deformation
1260   lipid band due to δ(=CH2) in acyl chain
  1198 tyrosine band due to ring deformation
  1101 phenylalanine band due to C–C stretch
1056   lipid band is due to C··O stretch
  1014 tryptophan band due to the ring breathing
970   lipid band due to Phosphate monoester groups
707   lipid band due to cholesterol
Open in a new tab
a

Bands are assigned using reported data.40−43

Figure 3E shows Raman spectra from SK-BR-3 derived exosomes in the presence of mixed-dimensional heterostructure. As reported in Table 2 and Figure 3E, we have observed several protein bands and those are at ∼1634, ∼1019 cm–1 due to the tryptophan, ∼1198, ∼858 cm–1 due to tyrosine, ∼1101 cm–1 due to phenylalanine.40−43 On the other hand, we have observed one lipid peak at ∼1315 cm–1 due to δ(CH2) in acyl chain.40−43 Similarly, we have also observed DNA peaks at ∼1388 cm–1, which is due to pyrimidine and imidazole rings A/G stacking.40−43

Experimentally observed SERS spectra from TNBC-derived exosomes and HER2(+) breast cancer-derived exosomes indicate that there several fingerprint Raman band for each type. As reported in Table 2, lipid bands at ∼1605, ∼1260, ∼1056, and ∼970 cm–1 are unique for TNBC-derived exosomes. Similarly, DNA band at ∼1510 cm–1 is unique for TNBC-derived exosomes. On the other hand, protein bands at ∼1634, ∼1198, ∼1101, and ∼1014 cm–1 are unique for HER2(+) breast cancer cell-derived exosomes. Similarly, DNA bands at ∼1388 cm–1 is unique for HER2(+) breast cancer cell-derived exosomes. Concentration-dependent SERS data are reported in Figure 3E,F. From the reported data and using eq 4, we have calculated LOD for HER2(+) breast cancer-derived exosomes to be 4.4 × 102 exosomes/mL. As shown in Table 3, our reported LOD is better than the LOD of several SERS-based methods reported in the literature.44−48 As a result, after proper engineering design, heterostructure-based SERS may have excellent prospects for highly sensitive analysis of exosomes.

Table 3. Comparison of Our SERS Data with Others Reported Methods for the Detection of Exosomes.

method LOD/mL dynamic range/mL type of cancer SERS substrate reference
SERS using AuNP and Raman reporter 2.3 × 106 106 to 108 pancreatic cancer SiO2 (35)
SERS using AuNS@4-MBA@Au 2.7 × 104 103 to 1010 liver cancer SiO2 (44)
SERS using polydopamine-encapsulated antibody-reporter-Ag(shell)–Au(core) multilayer (PEARL) 500 102 to 1010 pancreatic cancer SiO2 (43)
SERS using gold nanorod-coated Raman reporter 2.0 × 106 107 to 109 breast cancer SiO2 (47)
SERS using MB@SiO2@Au with Raman reporter 2.03 × 105 105 to 109 prostate cancer SiO2 (48)
label-free SERS using GO-GNS 4.4 × 102 102 to 105 breast cancer SiO2 this work
Open in a new tab

4. Conclusions

In conclusion, our finding reveals that mixed-dimensional heterostructure-based Raman substrate has the capability to be used for trace level tracking of cancer-derived exosomes. Reported experimental data show that mixed-dimensional heterostructure using 2D-GO and plasmonic GNS exhibits very high Raman EF, which is ∼1010 via combined EM and CM enhancement mechanism. Simulation data indicate that because of the formation of hot spots by GNS through sharp edge on the 2D-GO surface, more than 2 orders of magnitudes higher field enhancement occurs, which can enhance the Raman signal by 4 orders of magnitude. We have demonstrated that mixed-dimensional heterostructure-based SERS can be used for the identification of exosomes derived from TNBC-type MDA-MB-231 cells and HER2(+)-type SKBR3 breast cancer cells via their fingerprint Raman bands. Reported data shows that the LOD is 3.8 × 102 exosomes/mL for TNBC-derived exosomes and 4.4 × 102 exosomes/mL for HER2(+) breast cancer-derived exosomes. Our reported data show the potential of mixed-dimensional heterostructure-based SERS for cancer biomarker identification. Because of the high heterogeneity of exosomes as well as biological fluids, it is very important to understand how Raman spectra for exosome vary from different sources before it can be used for clinics.

Acknowledgments

Dr. Ray thanks NSF-PREM grant # DMR- 1826886 and NSF CREST grant # 1547754 for their generous funding. The research reported here made use of shared experimental facilities of the NSF Materials Research Science and Engineering Center (MRSEC) at UC Santa Barbara (DMR 1720256). R.S. acknowledges the support of the PREM Program as well.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.0c01441.

  • Detailed design and characterization of mixed-dimensional heterostructure; other experimental details; TEM image showing the morphology of freshly prepared bare GNS without PEG; XRD spectra from PEG coated GNS; and plot showing how the Raman EF from mixed-dimensional heterostructure surface varies with samples made in different batches (PDF)

The authors declare no competing financial interest.

Supplementary Material

ao0c01441_si_001.pdf (446.1KB, pdf)

References

  1. Zong C.; Xu M.; Xu L.-J.; Wei T.; Ma X.; Zheng X.-S.; Hu R.; Ren B. Surface-Enhanced Raman Spectroscopy for Bioanalysis: Reliability and Challenges. Chem. Rev. 2018, 118, 4946–4980. 10.1021/acs.chemrev.7b00668. [DOI] [PubMed] [Google Scholar]
  2. Ding S.-Y.; You E.-M.; Tian Z.-Q.; Moskovits M. Electromagnetic theories of surface-enhanced Raman spectroscopy. Chem. Soc. Rev. 2017, 46, 4042–4076. 10.1039/c7cs00238f. [DOI] [PubMed] [Google Scholar]
  3. Reguera J.; Langer J.; Jiménez de Aberasturi D.; Liz-Marzán L. M. Anisotropic Metal Nanoparticles for Surface Enhanced Raman Scattering. Chem. Soc. Rev. 2017, 46, 3866–3885. 10.1039/c7cs00158d. [DOI] [PubMed] [Google Scholar]
  4. Cardinal M. F.; Vander Ende E.; Hackler R. A.; McAnally M. O.; Stair P. C.; Schatz G. C.; Van Duyne R. P. Expanding Applications of SERS Through Versatile Nanomaterials Engineering. Chem. Soc. Rev. 2017, 46, 3886–3903. 10.1039/c7cs00207f. [DOI] [PubMed] [Google Scholar]
  5. Li J.-F.; Zhang Y.-J.; Ding S.-Y.; Panneerselvam R.; Tian Z.-Q. Core–Shell Nanoparticle-Enhanced Raman Spectroscopy. Chem. Rev. 2017, 117, 5002–5069. 10.1021/acs.chemrev.6b00596. [DOI] [PubMed] [Google Scholar]
  6. Sinha S. S.; Jones S.; Pramanik A.; Ray P. C. Nanoarchitecture Based SERS for Biomolecular Fingerprinting and Label-Free Disease Markers Diagnosis. Acc. Chem. Res. 2016, 49, 2725–2735. 10.1021/acs.accounts.6b00384. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cialla-May D.; Zheng X.-S.; Weber K.; Popp J. Recent Progress in Surface-Enhanced Raman Spectroscopy for Biological and Biomedical Applications: From Cells to Clinics. Chem. Soc. Rev. 2017, 46, 3945–3961. 10.1039/c7cs00172j. [DOI] [PubMed] [Google Scholar]
  8. Matricardi C.; Hanske C.; Garcia-Pomar J. L.; Langer J.; Mihi A.; Liz-Marzán L. M. Gold Nanoparticle Plasmonic Superlattices as Surface-Enhanced Raman Spectroscopy Substrates. ACS Nano 2018, 12, 8531–8539. 10.1021/acsnano.8b04073. [DOI] [PubMed] [Google Scholar]
  9. Li M.; Kang J. W.; Dasari R. R.; Barman I. Shedding Light on the Extinction-Enhancement Duality in Gold Nanostar-Enhanced Raman Spectroscopy. Angew. Chem., Int. Ed. 2014, 53, 14115–14119. 10.1002/anie.201409314. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Khoury C. G.; Vo-Dinh T. Gold Nanostars For Surface-Enhanced Raman Scattering: Synthesis, Characterization and Optimization. J. Phys. Chem. C 2008, 112, 18849–18859. 10.1021/jp8054747. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Jones S.; Sinha S. S.; Pramanik A.; Ray P. C. Three-dimensional (3D) plasmonic hot spots for label-free sensing and effective photothermal killing of multiple drug resistant superbugs. Nanoscale 2016, 8, 18301–18308. 10.1039/c6nr05888d. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fan Z.; Kanchanapally R.; Ray P. C. Hybrid Graphene Oxide Based Ultrasensitive SERS Probe for Label-Free Biosensing. J. Phys. Chem. Lett. 2013, 4, 3813–3818. 10.1021/jz4020597. [DOI] [Google Scholar]
  13. Singh A. K.; Khan S. A.; Fan Z.; Demeritte T.; Senapati D.; Kanchanapally R.; Ray P. C. Development of a Long-Range Surface-Enhanced Raman Spectroscopy Ruler. J. Am. Chem. Soc. 2012, 134, 8662–8669. 10.1021/ja301921k. [DOI] [PubMed] [Google Scholar]
  14. Demeritte T.; Viraka Nellore B. P.; Kanchanapally R.; Sinha S. S.; Pramanik A.; Chavva S. R.; Ray P. C. Hybrid Graphene Oxide Based Plasmonic-Magnetic Multifunctional Nanoplatform for Selective Separation and Label-Free Identification of Alzheimer’s Disease Biomarkers. ACS Appl. Mater. Interfaces 2015, 7, 13693–13700. 10.1021/acsami.5b03619. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Pramanik A.; Gao Y.; Gates K.; Begum S.; Ray P. C. Giant Chemical and Excellent Synergistic Raman Enhancement from a 3D MoS2–xOx–Gold Nanoparticle Hybrid. ACS Omega 2019, 4, 11112–11118. 10.1021/acsomega.9b00866. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Le Ru E. C.; Meyer M.; Etchegoin P. G. Proof of single-molecule sensitivity in surface-enhanced Raman scattering (SERS) by means of a two-analyte technique. J. Phys. Chem. B 2006, 110, 1944–1948. 10.1021/jp054732v. [DOI] [PubMed] [Google Scholar]
  17. Begum S.; Pramanik A.; Davis D.; Patibandla S.; Gates K.; Gao Y.; Ray P. C. 2D and Heterostructure Nanomaterial Based Strategies for Combating Drug-Resistant Bacteria. ACS Omega 2020, 5, 3116–3130. 10.1021/acsomega.9b03919. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Ling X.; Xie L.; Fang Y.; Xu H.; Zhang H.; Kong J.; Dresselhaus M. S.; Zhang J.; Liu Z. Can Graphene Be Used as a Substrate for Raman Enhancement?. Nano Lett. 2010, 10, 553–561. 10.1021/nl903414x. [DOI] [PubMed] [Google Scholar]
  19. Wang Y.; Polavarapu L.; Liz-Marzán L. M. Reduced Graphene Oxide-Supported Gold Nanostars for Improved SERS Sensing and Drug Delivery. ACS Appl. Mater. Interfaces 2014, 6, 21798–21805. 10.1021/am501382y. [DOI] [PubMed] [Google Scholar]
  20. Hu Y.; López-Lorente Á. I.; Mizaikoff B. Graphene-Based Surface Enhanced Vibrational Spectroscopy: Recent Developments, Challenges, and Applications. ACS Photonics 2019, 6, 2182–2197. 10.1021/acsphotonics.9b00645. [DOI] [Google Scholar]
  21. Ling X.; Huang S.; Deng S.; Mao N.; Kong J.; Dresselhaus M. S.; Zhang J. Lighting Up the Raman Signal of Molecules in the Vicinity of Graphene Related Materials. Acc. Chem. Res. 2015, 48, 1862–1870. 10.1021/ar500466u. [DOI] [PubMed] [Google Scholar]
  22. Tan Y.; Ma L.; Gao Z.; Chen M.; Chen F. Two-Dimensional Heterostructure as a Platform for Surface-Enhanced Raman Scattering. Nano Lett. 2017, 17, 2621–2626. 10.1021/acs.nanolett.7b00412. [DOI] [PubMed] [Google Scholar]
  23. Pan X.; Li L.; Lin H.; Tan J.; Wang H.; Liao M.; Chen C.; Shan B.; Chen Y.; Li M. A graphene oxide-gold nanostar hybrid based-paper biosensor for label-free SERS detection of serum bilirbin for diagnosis of jaundice. Biosens. Bioelectron. 2019, 145, 111713. 10.1016/j.bios.2019.111713. [DOI] [PubMed] [Google Scholar]
  24. Krishnan S. K.; Godoy Y. C. Deep Eutectic Solvent-Assisted Synthesis of Au Nanostars Supported on Graphene Oxide as an Efficient Substrate for SERS-Based Molecular Sensing. ACS Omega 2020, 5, 1384–1393. 10.1021/acsomega.9b02759. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Jalani G.; Cerruti M. Nano graphene oxide-wrapped gold nanostars as ultrasensitive and stable SERS nanoprobes. Nanoscale 2015, 7, 9990–9997. 10.1039/c4nr07473d. [DOI] [PubMed] [Google Scholar]
  26. Ling X.; Fang W.; Lee Y.-H.; Araujo P. T.; Zhang X.; Rodriguez-Nieva J. F.; Lin Y.; Zhang J.; Kong J.; Dresselhaus M. S. Raman Enhancement Effect on Two-Dimensional Layered Materials: Graphene, h-BN and MoS2. Nano Lett. 2014, 14, 3033–3040. 10.1021/nl404610c. [DOI] [PubMed] [Google Scholar]
  27. Tan Y.; Ma L.; Gao Z.; Chen M.; Chen F. Two-Dimensional Heterostructure as a Platform for Surface-Enhanced Raman Scattering. Nano Lett. 2017, 17, 2621–2626. 10.1021/acs.nanolett.7b00412. [DOI] [PubMed] [Google Scholar]
  28. Chandra K.; Culver K. S. B.; Werner S. E.; Lee R. C.; Odom T. W. Manipulating the Anisotropic Structure of Gold Nanostars using Good’s Buffers. Chem. Mater. 2016, 28, 6763–6769. 10.1021/acs.chemmater.6b03242. [DOI] [Google Scholar]
  29. Zhao J.; Pinchuk A. O.; McMahon J. M.; Li S.; Ausman L. K.; Atkinson A. L.; Schatz G. C. Methods for Describing the Electromagnetic Properties of Silver and Gold Nanoparticles. Acc. Chem. Res. 2008, 41, 1710–1720. 10.1021/ar800028j. [DOI] [PubMed] [Google Scholar]
  30. Knight M. W.; King N. S.; Liu L.; Everitt H. O.; Nordlander P.; Halas N. J. Aluminum for Plasmonics. ACS Nano 2014, 8, 834–840. 10.1021/nn405495q. [DOI] [PubMed] [Google Scholar]
  31. Shen L.-M.; Quan L.; Liu J. Tracking Exosomes in Vitro and in Vivo To Elucidate Their Physiological Functions: Implications for Diagnostic and Therapeutic Nanocarriers. ACS Appl. Nano Mater. 2018, 1, 2438–2448. 10.1021/acsanm.8b00601. [DOI] [Google Scholar]
  32. Pramanik A.; Gates K.; Patibandla S.; Davis D.; Begum S.; Iftekhar R.; Alamgir S.; Paige S.; Porter M. M.; Ray P. C. Water-Soluble and Bright Luminescent Cesium–Lead–Bromide Perovskite Quantum Dot–Polymer Composites for Tumor-Derived Exosome Imaging. ACS Appl. Bio Mater. 2019, 2, 5872–5879. 10.1021/acsabm.9b00837. [DOI] [PubMed] [Google Scholar]
  33. Kamerkar S.; LeBleu V. S.; Sugimoto H.; Yang S.; Ruivo C. F.; Melo S. A.; Lee J. J.; Kalluri R. Exosomes Facilitate Therapeutic Targeting of Oncogenic Kras in Pancreatic Cancer. Nature 2017, 546, 498–503. 10.1038/nature22341. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Mathew D. G.; Beekman P.; Lemay S. G.; Zuilhof H.; Le Gac S.; van der Wiel W. G. Electrochemical Detection of Tumor-Derived Extracellular Vesicles on Nanointerdigitated Electrodes. Nano Lett. 2020, 20, 820–828. 10.1021/acs.nanolett.9b02741. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Zhang W.; Jiang L.; Diefenbach R. J.; Campbell D. H.; Walsh B. J.; Packer N. H.; Wang Y. Enabling Sensitive Phenotypic Profiling of Cancer-Derived Small Extracellular Vesicles Using Surface-Enhanced Raman Spectroscopy Nanotags. ACS Sens. 2020, 5, 764–771. 10.1021/acssensors.9b02377. [DOI] [PubMed] [Google Scholar]
  36. Lee K.; Fraser K.; Ghaddar B.; Yang K.; Kim E.; Balaj L.; Chiocca E. A.; Breakefield X. O.; Lee H.; Weissleder R. Multiplexed profiling of single extracellular vesicles. ACS Nano 2018, 12, 494–503. 10.1021/acsnano.7b07060. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Liu C.; Xu X.; Li B.; Situ B.; Pan W.; Hu Y.; An T.; Yao S.; Zheng L. Single-Exosome-Counting Immunoassays for Cancer Diagnostics. Nano Lett. 2018, 18, 4226–4232. 10.1021/acs.nanolett.8b01184. [DOI] [PubMed] [Google Scholar]
  38. Shen W.; Guo K.; Adkins G. B.; Jiang Q.; Liu Y.; Sedano S.; Duan Y.; Yan W.; Wang S. E.; Bergersen K.; Worth D.; Wilson E. H.; Zhong W. A Single Extracellular Vesicle (EV) Flow Cytometry Approach to Reveal EV Heterogeneity. Angew. Chem., Int. Ed. 2018, 57, 15675–15680. 10.1002/anie.201806901. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Kruglik S. G.; Royo F.; Guigner J.-M.; Palomo L.; Seksek O.; Turpin P.-Y.; Tatischeff I.; Falcón-Pérez J. M. Raman tweezers microspectroscopy of circa 100 nm extracellular vesicles. Nanoscale 2019, 11, 1661–1679. 10.1039/c8nr04677h. [DOI] [PubMed] [Google Scholar]
  40. Yan Z.; Dutta S.; Liu Z.; Yu X.; Mesgarzadeh N.; Ji F.; Bitan G.; Xie Y.-H. A Label-Free Platform for Identification of Exosomes from Different Sources. ACS Sens. 2019, 4, 488–497. 10.1021/acssensors.8b01564. [DOI] [PubMed] [Google Scholar]
  41. Lee W.; Lenferink A. T. M.; Otto C.; Offerhaus H. L. Classifying Raman spectra of extracellular vesicles based on convolutional neural networks for prostate cancer detection. J. Raman Spectrosc. 2020, 51, 293–300. 10.1002/jrs.5770. [DOI] [Google Scholar]
  42. Dai Y.; Bai S.; Hu C.; Chu K.; Shen B.; Smith Z. J. Combined Morpho-Chemical Profiling of Individual Extracellular Vesicles and Functional Nanoparticles without Labels. Anal. Chem. 2020, 92, 5585. 10.1021/acs.analchem.0c00607. [DOI] [PubMed] [Google Scholar]
  43. Li T.-D.; Zhang R.; Chen H.; Huang Z.-P.; Ye X.; Wang H.; Deng A.-M.; Kong J.-L. An ultrasensitive polydopamine bi-functionalized SERS immunoassay for exosome-based diagnosis and classification of pancreatic cancer. Chem. Sci. 2018, 9, 5372–5382. 10.1039/c8sc01611a. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Tian Y.-F.; Ning C.-F.; He F.; Yin B.-C.; Ye B.-C. Highly sensitive detection of exosomes by SERS using gold nanostar@Raman reporter@nanoshell structures modified with a bivalent cholesterol-labeled DNA anchor. Analyst 2018, 143, 4915–4922. 10.1039/c8an01041b. [DOI] [PubMed] [Google Scholar]
  45. Ibn Sina A. A.; Vaidyanathan R.; Dey S.; Carrascosa L. G.; Shiddiky M. J. A.; Trau M. Real time and label free profiling of clinically relevant exosomes. Sci. Rep. 2016, 6, 30460. 10.1038/srep30460. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Kilic T.; Valinhas A. T. D. S.; Wall I.; Renaud P.; Carrara S. Label-free detection of hypoxia-induced extracellular vesicle secretion from MCF-7 cells. Sci. Rep. 2018, 8, 9402. 10.1038/s41598-018-27203-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Kwizera E. A.; O’Connor R.; Vinduska V.; Williams M.; Butch E. R.; Snyder S. E.; Chen X.; Huang X. Molecular detection and analysis of exosomes using surface-enhanced Raman scattering gold nanorods and a miniaturized device. Theranostics 2018, 8, 2722–2738. 10.7150/thno.21358. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Wang Z.; Zong S.; Wang Y.; Li N.; Li L.; Lu J.; Wang Z.; Chen B.; Cui Y. Screening and multiple detection of cancerous exosomes using a SERS-based method. Nanoscale 2018, 10, 9053–9062. 10.1039/c7nr09162a. [DOI] [PubMed] [Google Scholar]
  49. Avella-Oliver M.; Puchades R.; Wachsmann-Hogiu S.; Maquieira A. Label-free SERS analysis of proteins and exosomes with large-scale substrates from recordable compact disks. Sens. Actuators, B 2017, 252, 657–662. 10.1016/j.snb.2017.06.058. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ao0c01441_si_001.pdf (446.1KB, pdf)

Articles from ACS Omega are provided here courtesy of American Chemical Society

RESOURCES