Detection of carcinoembryonic antigen using functional magnetic and fluorescent nanoparticles in magnetic separators
- First Online:
- Cite this article as:
- Tsai, H.Y., Chang, C.Y., Li, Y.C. et al. J Nanopart Res (2011) 13: 2461. doi:10.1007/s11051-010-0138-5
- 161 Views
We combined a sandwich immunoassay, anti-CEA/CEA/anti-CEA, with functional magnetic (~80 nm) and fluorescent (~180 nm) nanoparticles in magnetic separators to demonstrate a detection method for carcinoembryonic antigen (CEA). Determination of CEA in serum can be used in clinical diagnosis and monitoring of tumor-related diseases. The CEA concentrations in samples were deduced and determined based on the reference plot using the measured fluorescent intensity of sandwich nanoparticles from the sample. The linear range of CEA detection was from 18 ng/mL to 1.8 pg/mL. The detection limit of CEA was 1.8 pg/mL. In comparison with most other detection methods, this method had advantages of lower detection limit and wider linear range. The recovery was higher than 94%. The CEA concentrations of two serum samples were determined to be 9.0 and 55 ng/mL, which differed by 6.7% (9.6 ng/mL) and 9.1% (50 ng/mL) from the measurements of enzyme-linked immunosorbent assay (ELISA), respectively. The analysis time can be reduced to one third of ELISA. This method has good potential for other biomarker detections and biochemical applications.
KeywordsImmunoassayFunctional nanoparticlesMagnetic separatorDiagnosticsNanomedicine
Carcinoembryonic antigen (CEA) is a tumor marker associated with many cancers such as colon cancer, lung cancer, ovarian cancer, and breast cancer (Duffy et al. 2003). Determination of CEA in serum can be used in clinical diagnosis and monitoring of tumor-related diseases. Therefore, fast, sensitive, and selective detection methods of CEA are important in medical diagnosis. Enzyme-linked immunosorbent assay (ELISA) is a widely used method for detection of CEA. This method is based on the principle of immunoassay with enhanced detection of enzymes for biomolecules in the field of life science (Haugland 2002; Kemeny 1991; Price and Newman 1997). Although ELISA is a useful and sensitive assay, it is very laborious and time consuming. Therefore, improvement of ELISA analysis is important since it is used for many biochemical analyses.
Functional magnetic nanoparticles have been used for many biochemical analyses (Kouassi and Irudayaraj 2006; Hamley 2003; Trindade 2001; Tsai et al. 2006, 2007). We tried to improve the efficiency of the antigen–antibody reaction by integrating the sandwich immunoassays with functional magnetic and fluorescent nanoparticles in magnetic separators. This method has many potential advantages. The amounts of protein immobilized on the particles are consistent in the same batch, which can be used for several reactions. The attainable surface area to sample volume ratio of nanoparticles are higher than those of the plate or channel. Thus, this method can provide more available binding sites for analytes (Farrell et al. 2004). Furthermore, the antibody-labeled nanoparticles can be dispersed in the solution to form a pseudo-homogenous reaction with antigen. Thus, the reaction of antigens with antibodies labeled on nanoparticles is more efficient than those labeled on microplates (Farrell et al. 2004). In addition, the magnetic nanoparticles can be easily separated from the unreactive substances by applying magnetic force and can be redispersed again in the solution after removing the magnetic force.
We used CEA as a model analyte to demonstrate our detection method for tumor marker. In this study, the magnetic nanoparticles were labeled with anti-CEA. We also prepared fluorescent nanoparticles by doping rhodamine 6G (R6G) in silica nanoparticles to increase the stability of the fluorescent signal. The fluorescent nanoparticles were then conjugated with anti-CEA to react with CEA. A sandwich immunoassay, anti-CEA/CEA/anti-CEA, combined with functional nanoparticles for CEA detections in a magnetic separator was formed and the fluorescent intensities of sandwich nanoparticles were related to the amount of CEA present in the sample, respectively.
Chemicals and materials
Tetraethoxysilan, carcinoembryonic antigen (CEA), anti-CEA, rhodamine 6G (R6G), and phosphate-buffered saline (PBS) were purchased from Sigma-Aldrich (Saint Louis, Missouri, USA). Ferrous and ferric chlorides were purchased from J. T. Baker (Philipsburg, New Jersey, USA). Sera were from Chung-Shan Medical University Hospital (Taichung, Taiwan) and Puli Veteran Hospital (Nantou, Taiwan). Sera samples could be analyzed directly or stored at 4 °C for less than 48 h before use. Sera and used utensils were handled carefully according to safety regulations to avoid biohazard.
Permanent magnets with 6 mm in diameter and 13 mm in length were put firmly in the well of a microplate and the assembled magnetic microplate was then placed under another microplate to form a magnetic separator. The magnetic field strengths of the magnets were 2.5 ± 0.1 kG at the bottom well of separator.
The crystal structure of nanoparticles was measured by an X-ray diffractometer (Shimazu XRD-7000, Kyoto, Japan). The magnetization curves of particle were studied by a superconducting quantum interference device (SQUID) magnetometer. Transmission electron microscopy (TEM) and atomic force microscopy (AFM) were used to characterize the size and morphology of nanoparticles. Samples were prepared by placing few drops of dispersed nanoparticles on copper grids and allowing to be dry at room temperature. A Fourier transform infrared spectrometer (Perkin Elmer RX-1, Norwalk, Connecticut, USA) was used to measure the functional groups of particles. A spectrometer (Varioskan, Thermo Electron, Waltham, Massachusetts, USA) was used to measure the fluorescent intensity. The excitation wavelength was 545 nm and the emission wavelength was 575 nm for fluorescent detection. All fluorescent intensities were measured in triplicate.
Magnetic nanoparticles were prepared by chemical precipitation as described in our previous study (Tsai et al. 2007). Briefly, ferric and ferrous chlorides with a mole ratio of 2 to 1 were dissolved in distilled water to form 0.06 M of iron solutions. Five milliliters of ammonium hydroxide solutions were added and mixed well with ferric and ferrous solutions to form iron oxide nanoparticles. Then, the obtained iron oxide nanoparticles were magnetically separated and washed three times with distilled water. The core/shell nanoparticles were prepared by mixing 0.15 g of iron oxide nanoparticles with 0.5 mL of ammonium hydroxide [28% (v/v)] and 20 μL of tetraethoxysilane for 1 h. The resulting core/shell nanoparticles were magnetically separated and washed three times with ethanol. Then, the core/shell nanoparticles were mixed with 1 mL of 3-aminopropyl-triethoxysilane for 1 h to form an amine-modified surface.
Fluorescent nanoparticles were prepared by mixing 5 mg of R6G with 1.5 mL of tetraethoxysilane, 0.4 mL of ammonium hydroxide [28% (v/v)], and 15 mL of ethanol with stirring for 2 h. The resulting fluorescent nanoparticles were washed five times with distilled water. The surface of fluorescent nanoparticles was also modified to form an amine group using 1 mL of 3-aminopropyl-triethoxysilane.
The coupling of particles with proteins used the same method as our previous work (Tsai et al. 2007). Briefly, the coupling reagents of magnetic and fluorescent nanoparticles with various proteins used 0.153 g of 1-ethyl-(dimethylaminopropyl) carbodiimide and 0.023 g of N-hydroxysuccinimide in 10 mL of phosphate-buffered saline (PBS) solution and incubated with proteins for 2 h. Functional magnetic and fluorescent nanoparticles were prepared by mixing 2 mL of 0.35 mg/mL anti-CEA with coupling reagents, and 2 mL of 15 mg/mL magnetic and fluorescent nanoparticles for 2 h, respectively. The labeled particles were centrifuged and washed three times with PBS solution to remove unreacted labels.
Magnetic sandwich immunoassay
One hundred microliters of CEA with concentration ranging from 10−8 to 10−18 M were added into each well of a microplate which contained 100 μL of 0.15 mg/mL anti-CEA-labeled magnetic nanoparticles. The solutions of CEA and anti-CEA were mixed completely and put on the magnetic microplate for 5 min to settle down the complex from unreacted species. These procedures were repeated during each washing step. The magnetic nanoparticles were attracted to the bottom of the plate and the supernatant was removed. The resulting complexes on the magnetic nanoparticles were washed two times with 100 μL of PBS buffer solutions to remove unreacted CEA. Then, 100 μL of 0.75 mg/mL anti-CEA labeled fluorescent nanoparticles were added into each well and mixed completely with the anti-CEA/CEA complex to form the anti-CEA/CEA/anti-CEA sandwich nanoparticles in buffer solutions. The unreacted fluorescent nanoparticles were removed by washing two times with buffer. The sandwich nanoparticles were resuspended in 100 μL of buffer solutions and the fluorescent intensity was measured.
A reference plot of fluorescent intensity was established by plotting the measured fluorescent intensity of sandwich nanoparticles versus the various known concentrations of CEA added in the diluted serum.
Determination of CEA in serum samples
Diluted serum samples were first used to react with anti-CEA on the magnetic nanoparticles to form affinity complex. Then, anti-CEA-labeled fluorescent nanoparticles were added to react with CEA of affinity complex to form sandwich nanoparticles. The fluorescent intensity of sandwich nanoparticles is proportional to the concentration of CEA in serum samples. The amounts of CEA in the samples were deduced and determined based on the reference plot using the measured fluorescent intensity of sandwich nanoparticles from the sample.
Results and discussions
Characterization of functional nanoparticles
Magnetic susceptibilities of magnetic nanoparticles decreased for each step of surface modification, as shown in Fig. 1B. The saturated magnetization decreased about 40% from about 60 emu/g of iron oxide nanoparticles to about 36 emu/g of anti-CEA-labeled magnetic nanoparticles. The high susceptibility of functional magnetic nanoparticles allows to hold more nanoparticles for reactions at fixed magnetic fields during washing steps. By using more magnetic nanoparticles to react with CEA and fluorescent nanoparticles, the sensitivity of detection can be increased. The size of fluorescent nanoparticles was found to be about 180 nm from TEM, as shown in Fig. 1D. The photostability of fluorescent nanoparticles and free dye after irradiation with a xenon lamp is shown in Fig. 1E. The rapid decrease of fluorescent intensity of the free dye can be observed clearly after photo-irradiation. The intensities of fluorescent nanoparticles with concentrations from 4.81 to 481 μg/mL stored in a dark container at 4 °C can be maintained about 90 and 80% of the original intensity even after 4 and 7 weeks, respectively. These results show that fluorescent nanoparticles are relatively stable against free fluorescent dye.
The maximum number of magnetic nanoparticles that can be used for a sandwich assay with the home-made magnetic separator was determined to be 1.5 × 1010 (~11.1 μg). The optimal numbers of magnetic and fluorescent nanoparticles were determined and fixed for the rest of the experiments at 1.1 × 1010 (~8.15 μg) and 1.5 × 1010 (~57.2 μg), respectively.
Magnetic sandwich immunoassay
Determination of CEA in serum samples
Detection of CEA biomarker in serum samples was further used to test this method. Figure 2B shows a reference plot of fluorescent intensity for fixed number of anti-CEA-labeled magnetic and fluorescent nanoparticles at various concentrations of CEA in diluted serum solutions. The averaged fluorescent intensity was 5.51 ± 0.24 for controlled selective study by replacing CEA with IgG at three different concentrations (10−10, 10−13, and 10−16 M). The linear range of CEA detection was from 1.0 × 10−10 M (18 ng/mL) to 1.0 × 10−14 M (1.8 pg/mL) with a correlation coefficient equal to 0.994. The detection limit was 1.8 pg/mL deduced from the signal to noise ratio of three. The detection limit of this method is lower and the linear range was wider than those of ELISA and other methods (He et al. 2008; Pan and Yang 2007; An et al. 2007; Wu et al. 2006). The recoveries of spiked CEA concentrations (5.0 × 10−9, 5.0 × 10−10, and 5.0 × 10−11 M) in diluted serum solutions were found to be 96, 98, and 103%, respectively. The fluorescent intensities of two serum samples were found to be 13.66 ± 0.39 and 15.30 ± 0.45 after dilution and reaction with anti-CEA-labeled magnetic and fluorescent nanoparticles. This corresponded to 5.0 × 10−13 M (0.090 ng/mL) and 3.0 × 10−12 M (0.55 ng/mL) of CEA based on the reference plot using the fluorescent intensities. The CEA concentrations in the serum were 9.0 and 55 ng/mL after 100 times dilution correction. These measurements differed by 6.7% (9.6 ng/mL) and 9.1% (50 ng/mL) from the ELISA measurements, respectively. The total analysis time can be reduced to one third of ELISA measurement. This method can be easily automated since the magnetic forces and nanoparticles can be easily adapted on the separator system.
The preliminary results of this study show that combing the sandwich immunoassay with magnetic and fluorescent nanoparticles in magnetic separators is a very efficient way to detect protein. This method has advantages of lower detection limit and wider linear range than ELISA and other methods for protein detection. The determination of CEA in serum samples from this method differed by less than 9% from those of ELISA. The analysis time can be reduced to one third of ELISA. This method has good potential for other biomarker detections (e.g. α-fetoprotein and hepatitis antigen) and biochemical applications.
This work was supported by grant NSC-96-2811-260-003 from the National Science Council of Taiwan. The authors would like to thank Ms. S. M. Tsai and Dr. F. C. Cheng for their technical assistances.