Ultrasensitive determination of carcinoembryonic antigens using a magnetoimpedance immunosensor

Abstract

Herein, a giant magnetoimpedance (GMI) sensor combined with the sandwich immunoassay was used for ultrasensitive detection of carcinoembryonic antigens (CEA). Oblong-shaped Au films were enclosed with SU-8 photoresists to form several microcavities that were treated with different concentrations of CEA. The quantification of low-concentration CEA (1 pg mL⁻¹–10 ng mL⁻¹) was achieved by using one single GMI sensor, demonstrating a good linear relationship between the GMI response and the CEA concentration. The lower detection limit for CEA was found to be as low as 1 pg mL⁻¹, which provides extreme sensitivity. We observed that high-concentration CEA (>10 ng mL⁻¹) can result in a cluster of Dynabeads, thereby reducing the target signals since the adjacent magnetic dipole fields canceled each other out. Target specificity of the GMI-based magnetic immunoassay has also been verified by testing nonspecific targets. The results obtained here suggest that the GMI sensor could serve as an ultrasensitive immunosensor for quantitative determination of biomarkers.

---

Introduction

Carcinoembryonic antigens (CEA) were first described in 1965 by Gold and Freedman^1,2 when they identified an antigen that was present in both fetal colon and colon adenocarcinoma. CEA is a glycoprotein involved in cell adhesion. It is normally produced during fetal development, but the production of CEA stops before birth, therefore, it is not usually present in the blood of healthy adults. Forty years after the initial detection of CEA in serum, it is one of the most widely used tumor markers worldwide, and it is responsible for clinical diagnosis of over 95% of all colon tumors, 50% of breast tumors, as well as tumors of the lung, pancreas, ovaries, and others of epithelial tissue origin, especially of the gastrointestinal tract. Consequently, detection of CEA is very significant for clinical diagnosis of cancers of the digestive system.^3–10

The giant magnetoimpedance (GMI) effect refers to a large change in the electrical impedance of a soft ferromagnetic conductor carrying a driving alternating current (AC) when subjected to a static magnetic field.^11–15 The impedance change ratio per oersted in GMI sensors based on sandwiched films having a transverse anisotropy can reach a value of 10–100%/Oe even at relatively low frequencies.^16–19 This sensitivity is at least one order of magnitude higher then that of giant magnetoresistive sensors. Therefore, the GMI sensors based on the sandwiched films are well suited for detection of the weak magnetic field and biosensing applications.

Superparamagnetic beads are generally defined as solids less than 100 μm in all three dimensions, and are commonly composed of magnetic elements, such as iron, nickel, cobalt and their oxides.²⁰ Superparamagnetic beads are of great interest for researchers due to their good biocompatibility, easy of surface modification and good superparamagnetism, they are being used in an increasing number of biomedical applications.^21–24 In recent years, the GMI sensors have been used to detect magnetic beads^25–30 and magnetic nanoparticles^31–34 in order to develop a new generation of magnetic biosensor. Furthermore, the GMI sensors have been introduced into the field of biosensing for detection of magnetically labeled bioanalytes.^35–37 Unfortunately, most of early GMI-based biodetections^35–37 have put too much emphasis on the physical nature, whereas the biosensing exploration is very shallow. Generally speaking, quantitative data, detection limit, linear range and specificity test are needed in a bioassay, but most of which are absent in the early GMI-based biodetections.^35–37 The sensitivity for detecting alpha-fetoprotein (AFP) antigens is relatively low, only 1 ng mL⁻¹ AFP antigens were detected in previous work,³⁵ besides, quantitative detection cannot be performed by the in situ method in previous work.³⁵ In order to fill those gaps and explore the possibility of GMI sensor in clinical application, ultrasensitive and quantitative detection of CEA by GMI sensor was performed in this work. Experimental results showed high sensitivity, linearity and specificity of the GMI-based magnetic immunoassay for detection of CEA, demonstrating its strong superiority and practicability of using the GMI sensor in the detection of biomolecules.

Experimental details

Fabrication of GMI sensor and microcavity

The GMI sensor was fabricated by Micro-Electro-Mechanical-Systems (MEMS) technology, which has been reported elsewhere.²⁹ A microcavity formed by Au and SU-8 for immunoassay was also fabricated by MEMS technology, the fabrication procedure is schematically illustrated as follows: (1) a round piece of glass substrate with a diameter of 7.62 cm and a thickness of 2 mm was washed with acetone, ethanol and deionized water; (2) the glass substrate was dried out in a oven at 70 °C for 8 h; (3) 300 nm thick Au/Cr films were deposited on the glass substrate by radio frequency sputtering system (Z-550); (4) 500 μm thick negative photoresist was spun on the Au film and dried out in an oven, after that, it was exposed by ultraviolet light at a wavelength of 400 nm, followed by the development. Until now, the microcavity was obtained, as shown in Fig. 1. The microcavity has good biocompatibility and is very suitable for immunoassay.

Fig. 1 The microcavity formed by Au and SU-8 for immunoassay (a) 4 × 5 mm² (b) 3 × 5 mm².

In our previous work,³⁵ we found it impossible to perform the quantitative detection of AFP antigens by using different GMI sensors due to the performance difference between them. In this work, a separated-type method based on a single GMI sensor was proposed for quantification of CEA. This separated-type method is similar to that used for detection of magnetic beads,²⁹ namely, the CEA with different concentrations were firstly captured and labeled in different microcavities, then the microcavities were successively fixed on a position very close to the GMI sensor for detection, this method will be illustrated in more detail at a later stage.

Instruments and reagents

Hewlett–Packard (HP) 4194A, Scanning Electron Microscopy (SEM) system [ULTRA^1M 55, Energy Disperse Spectroscopy (EDS)-INCA PentaFET-x3], Atomic Force Microscope (AFM) [UHV-SPM] and X-ray Photoelectron Spectroscopy (XPS) [AXIS Ultra^DLD ] were used to characterize the samples. Human CEA antigens, mouse anti-human CEA monoclonal antibody, and biotinylated mouse anti-human CEA monoclonal antibody were purchased from Linc-Bio Science Co. Ltd (Shanghai, China). N-Hydroxysuccinimide (NHS) was purchased from Medpep (Shanghai Medpep Co. Ltd.). 11-Mercaptoundecanoic Acid (11-MUA) was purchased from J&K chemical Co. Ltd (Shanghai, China). 1-Ethyl-3-[3-dimethylaminopropyl]carbodiimide (EDC) hydrochloride was purchased from Aladdin Chemistry Co. Ltd (USA). Phosphate buffer tablets (pH 7.4) were purchased from Medicago AB (Uppsala, Sweden). Dynabeads® MyOne™ Streptavidin C1 were purchased from Invitrogen. Hydrochloric acid (HCl) was purchased from Sinpharm Chemical Reagent Co. Ltd (Shanghai, China). Sodium hydroxide (NaOH) was purchased from Pinghu chemical reagent (Pinghu, China). Bovine serum albumin (BSA) was purchased from Wegene (Shanghai, China). Ethanol (CH₃CH₂OH) was purchased from Yangyuan Chemical Co. Ltd (Changshu, China). Acetone (C₃H₆O) was purchased from Lingfeng chemical reagent Co. Ltd (Shanghai, China). In the experiments, deionized water was used.

Immunoassay procedure

A sandwich immunoassay using Dynabeads as the magnetic labels and monoclonal antibodies as the molecular probes was performed in the microcavity for selective capturing and tagging of CEA. (1) The microcavity was first soaked and rinsed in 1 M NaOH and then soaked and rinsed in 1 M HCl for washing; (2) the microcavity-encapsuled Au film was treated with 30 mmol L⁻¹ 11-MUA at room temperature for 3 h prior to the activation with NHS and EDC for 1 h; (3) 10 μL of CEA monoclonal antibody (0.5–1 mg mL⁻¹) was dropped on the microcavity-encapsuled Au film for use as specific probe, after storing the sample at 4 °C for 24 h, the nonspecific adsorption sites were blocked using 1% BSA; (4) 10 μL of desired CEA was dropped into the microcavity for immune reaction, taking approximately 1 h; (5) biotin labeling for CEA was done by adding 8 μL of biotinylated CEA antibody to the microcavity, incubating at room temperature for 1 h; (6) at last, 10 μL of streptavidin-coated Dynabeads was injected into microcavity for magnetically labeling CEA, followed by the incubation for 30 min at room temperature. During the experiment, washing steps with phosphate buffered solution (PBS, pH 7.4) were required.

Testing method

In order to achieve quantitative determination of biomolecules, a separated-type method was used to quantify CEA in this work. Differing from the in situ method in the early study,³⁵ this separated-type method can be briefly stated as follows: the CEA with different concentrations were first labeled by Dynabeads on different microcavities, then the microcavities were successively fixed on the edge of the sensor for testing. During the test, the two electrodes of GMI sensor were connected to HP 4194A which supplied an AC of 10 mA, an external magnetic field (H_e) of 0–100 Oe was applied in longitudinal direction in order to change the transverse permeability in soft magnetic films through modifying the penetration depth of AC. Meanwhile, the CEA-conjugated Dynabeads would be magnetized by the longitudinal external magnetic field and thus produce a measurable stray magnetic field that can be detected by the sensor, thereby altering the magnetoimpedance of the sensor and providing a detection signal in the form of a voltage change, Fig. 2 shows the schematics of the experimental setup. The relative change in impedance (GMI ratio) was defined as:
where Z(H) and Z(H_max) represent the impedance in an external magnetic field H_e, and in the maximum external magnetic field H_max, respectively.

Fig. 2 The schematics of the experimental setup for quantitative detection of CEA.

Results and discussion

AFM, XPS and SEM characterizations

AFM characterizations (Fig. 3) show the significant difference in surface topography of the Au film before and after treatment of 11-MUA. As can be seen from the Fig. 4A and B, the main peak can be found at the binding energy of S element after modifying 11-MUA on Au film, which confirms the self-assembled monolayer on Au film. Dramatical increase in N element can be observed in the samples after immobilizing CEA antibody on Au film, as shown in Fig. 4C and D. As can be seen from Fig. 5, the number of Dynabeads increases with increasing CEA concentration.

Fig. 3 AFM characterizations (A) Blank sample (B) 11-MUA-self-assembled sample.

Fig. 4 XPS characterizations (A) Blank Au film (B) 11-MUA-self-assembled Au film (C) before immobilizing CEA monoclonal antibody on Au film and (D) after immobilizing CEA monoclonal antibody on Au film.

Fig. 5 SEM characterizations of Dynabeads (A) Blank (B) 1 pg mL⁻¹ CEA (C) 100 pg mL⁻¹ CEA (D) 10 ng mL⁻¹ CEA.

Magnetic field dependence and frequency dependence of GMI response

Fig. 6 shows the magnetic field dependence of the GMI ratio for detecting different concentrations of CEA, the small changes of GMI ratio at low magnetic fields (<10 Oe) is probably attributed to the low-level magnetization in the Dynabeads and the low field-sensitivity in sandwiched films at low magnetic fields. There is no obvious change in GMI ratio under overlarge magnetic fields (>40 Oe), which is related to the stray magnetic field of Dynabeads becoming strongly overwhelmed by the overlarge external magnetic field, and the soft magnetic films closing to the magnetic saturation. Large changes of GMI ratio can be observed at around the peak field, this is because the transverse permeability has risen considerably due to the strong rotational magnetization occurred around the anisotropy field according to the eddy current model.¹⁵ Moreover, the Dynabeads were fully magnetized in the large magnetic fields (10–40 Oe), producing strong stray magnetic field that can be sensitively detected by the sensor.

Fig. 6 Magnetic field dependency of the GMI ratio for detecting CEA (1 pg mL⁻¹–100 ng mL⁻¹) with the separated-type method (A) full view (B) partial enlargement.

Frequency dependence of the GMI ratio in different concentrations of CEA is shown in Fig. 7. The skin effect is very strong in the frequency range of 1–6 MHz, and the large transverse permeability can be obtained due to not only the domain wall motion but also the spin rotation,¹⁵ both of which contribute to the GMI effect. In other words, the GMI sensor possesses high field-sensitivity at high frequency (1–6 MHz) and can be highly sensitive to the stray magnetic field, thereby resulting in large changes of GMI ratio in the presence of the Dynabeads at high frequency (1–6 MHz). When the frequency is below 1 MHz, the change of GMI ratio is very small. This is because the GMI effect mainly origins from the magneto-inductive effect at low frequency, thus leading to a low detection sensitivity. The target signals start to decrease at higher frequency (above 6 MHz), which is related to the domain wall motion becoming strongly damped by the eddy currents,¹⁵ resulting in continued reductions in field-sensitivity of GMI.

Fig. 7 Frequency dependency of the GMI ratio for detecting CEA (1 pg mL⁻¹-100 ng mL⁻¹) with the separated-type method (A) full view (B) partial enlargement.

Detection limit and linear range

The experimental results (Fig. 6 and 7) show that the GMI ratio is increased in different degrees due to the presence of CEA with different concentrations, and the GMI ratio increases with increasing CEA concentration. This is because the CEA were labeled by Dynabeads that can generate the magnetic signals, the more CEA captured on the Au film, the more Dynabeads will be conjugated (Fig. 5), and the more stronger the target signals are. The linear fitting results of GMI ratio and CEA concentration are shown in Fig. 8, it can be seen that the GMI ratio is linearly dependent on the based-10 logarithms of the CEA concentration (lg C) in the range of 1–10 000 pg mL⁻¹, the regression equation is Y = 175.81–2.724X, R = −0.99013. Hence, the GMI response is linearly proportional to the CEA concentration in the range of 1 pgmL⁻¹–10 ng mL⁻¹, namely, the quantification of low-concentration CEA was achieved by the GMI sensor in this work.

Fig. 8 The linear fitting results of GMI ratio and CEA concentration.

As can be seen from Fig. 6 and 7, the GMI response starts to decrease when CEA concentration exceeds 10 ng mL⁻¹, namely, the GMI sensor seems incapable of quantifying the high-concentration CEA. In our previous work,³⁸ the used high-concentration Dynabeads can make Dynabeads cluster together, and enabling the adjacent exciting fields of Dynabeads to cancel each other out, resulting in the decrease of the fringe field and thus the reduced GMI signals. A similar result was reported by J. Devkota et al.³⁹ who found that the GMI signals began to go down when high-concentration magnetic nanoparticles were tested. G. V. Kurlyandskaya et al.⁴⁰ also found the similar phenomenon that high-concentration nanoparticles can cause the reduction of the target signals. Similar results can also be found in the detection of magnetic beads based on giant magnetoresistive effect,⁴¹ it is reported that the opposing dipole fields of adjacent beads would reduce the average stray magnetic field as the bead density increased, thus the resultant giant magnetoresistive signals were getting weaker in the presence of high-density magnetic beads. In this work, since the high-concentration CEA (above 10 ng mL⁻¹) can bind too many Dynabeads, causing a high-density cluster in Dynabeads, thus, the GMI signals begin to decrease exceeding 10 ng mL⁻¹. Briefly speaking, the upper detection limit should be about 10 ng mL⁻¹. We also found that no obvious target signals were detected below 1 pg mL⁻¹ CEA, thus, the lower detection limit for CEA was 1 pg mL⁻¹. Thus, the GMI sensor is very suitable for ultrasensitive detection of low-concentration biomolecules.

Specificity, reproducibility and stability of the magnetic immunoassay

To demonstrate the specificity of the magnetic immunoassay, Blank group, 1% BSA and 100 pg mL⁻¹ AFP (Control group) and 100 pg mL⁻¹ CEA (Experimental group) were tested, respectively. Fig. 9 shows the experimental results, evidently, there is a significant increase in GMI ratio due to the presence of CEA, but no obvious changes in GMI ratio have been observed in the control group. The experimental results indicated the good specificity of the GMI-based magnetic immunoassay for the detection of CEA. A relative standard deviation of 2.86% was obtained by performing 5 independent measurements on 100 pg mL⁻¹ CEA, indicating an acceptable reproducibility of the magnetic immunoassay. It was found that there was no obvious change in the GMI response after storing the GMI sensor at 4 °C for two weeks, suggesting the good stability of the GMI sensor. As is stated above, it was impossible to quantify high-concentration CEA (>10 ng mL⁻¹) since the adjacent magnetic dipole fields canceled each other out due to the high-density cluster and the large size of Dynabeads, superparamagnetic nanoparticles are probably a better choice for magnetic labeling in the GMI-based bioassay owing to their smaller size. Since blood or serum samples are still unavailable for now, it is impossible for us to do clinical trials. We have been looking for a cooperator who can provide us blood or serum samples, clinical trials may be reported in future work.

Fig. 9 GMI ratio for specificity test on CEA with the separated-type method (A) full view (B) partial enlargement.

Conclusion

In this study, an ultrasensitive, quantitative and specific magnetic immunoassay combined with GMI sensor has been developed for determination of CEA. A sandwich immunoassay using Dynabeads as the magnetic labels of CEA was performed in SU-8-enclosed microcavities, and a separated-type method was used to quantify CEA. The results show that the GMI response is linearly proportional to CEA concentration in the range of 1 pg mL⁻¹–10 ng mL⁻¹, and the minimum detectable concentration for CEA is approximately 1 pg mL⁻¹. The target signal became weaker upon exceeding 10 ng mL⁻¹, which is due to the high-density cluster of Dynabeads. The GMI-based magnetic immunoassay was highly specific for CEA since no positive signals were detected by using nonspecific targets. In a word, the GMI sensor is capable of quantifying low-concentration CEA and is ideally suited for ultrasensitive detection of biomolecules.

Acknowledgements

This work is supported by The National Natural Science Foundation of China (no. 61074168 and no. 61273065), National Science and Technology Support Program (2012BAK08B05), Natural Science Foundation of Shanghai (13ZR1420800), the Analytical and Testing Center in Shanghai Jiao Tong University, the Center for Advanced Electronic Materials and Devices in Shanghai Jiao Tong University.

References

1. P. Gold and S. O. Freedman, J. Exp. Med., 1965, 121, 439.
2. P. Gold and S. O. Freedman, J. Exp. Med., 1965, 122, 467.
3. Z. Y. Lin, G. Y. Zhang, W. Q. Yang, B. Qiuand and G. N. Chen, Chem. Commun., 2012, 48, 9918.
4. M. Liu, C. Jia, Y. Huang, X. Lou, S. Yao, Q. Jin, J. Zhao and J. Xiang, Analyst, 2010, 135, 327.
5. J. V. Jokerst, A. Raamanathan, N. Christodoulides, P. N. Floriano, A. A. Pollard, G. W. Simmons, J. Wong, C. Gage, W. B. Furmaga, S. W. Redding and J. T. McDevitt, Biosens. Bioelectron., 2009, 24, 3622.
6. Z. M. Zhou, J. Zhou, J. Chen, R. N. Yu, M. Z. Zhang, J. T. Song and Y. D. Zhao, Biosens. Bioelectron., 2014, 59, 397.
7. X. Cao, N. Wang, S. Jia, L. Guo and K. Li, Biosens. Bioelectron., 2013, 39, 226.
8. S. Hideshima, R. Sato, S. Inoue, S. Kuroiwa and T. Osaka, Sens. Actuators, B, 2012, 161, 146.
9. R. M. Wang, X. Chen, J. Ma and Z. F. Ma, Sens. Actuators, B, 2013, 176, 1044.
10. R. S. Gaster, L. Xu, S. j. Han, R. J. Wilson, D. A. Hall, S. J. Osterfeld, H. Yu and S. X. Wang, Nat. Nanotechnol., 2011, 6, 314.
11. E. P. Harrison, G. L. Turney and H. Rowe, Nature, 1935, 135, 961.
12. R. S. Beach and A. E. Berkowitz, Appl. Phys. Lett., 1994, 64, 3652.
13. V. E. Makhotkin, B. P. Shurukhin, V. A. Lopatin, P. Y. Marchakov and Y. K. Levin, Sens. Actuators, A, 1991, 27, 759.
14. L. V. Panina and K. Mohri, Appl. Phys. Lett., 1994, 65, 1189.
15. L. V. Panina, K. Mohri, T. Uchiyama and M. Noda, IEEE Trans. Magn., 1995, 31, 1249.
16. M. Senda, O. Ishii, Y. Koshimoto and T. Toshima, IEEE Trans. Magn., 1994, 30, 4611.
17. K. Hika, L. V. Panina and K. Mohri, IEEE Trans. Magn., 1996, 32, 4594.
18. T. Morikawa, Y. Nishibe, H. Yamadera, Y. Nonomura, M. Takeuchi and Y. Taga, IEEE Trans. Magn., 1997, 33, 4367.
19. S. Q. Xiao, Y. H. Liu, Y. Y. Dai, L. Zhang, S. X. Zhou and G. D. Liu, J. Appl. Phys., 1999, 85, 4127.
20. B. I. Haukanes and C. Kvam, Nat. Biotechnol., 1993, 11, 60.
21. L. Maurizi, U. Sakulkhu, L. A. Crowe, V. M. Dao, N. Leclaire, J. Vallee and H. Hofmann, RSC Adv., 2014, 4, 11142.
22. M. C. Lim, G. H. Lee, D. T. N. Huynh, C. A. M. Letona, D. H. Seo, C. S. Park and Y. R. Kim, RSC Adv., 2015, 5, 36088.
23. X. Yu, D. Kang, Y. Hu, S. Tong, M. Ge, C. Cao and W. Song, RSC Adv., 2014, 4, 31362.
24. T. Bronshtein, N. Toledano, D. Danino, S. Pollack and M. Machluf, J. Controlled Release, 2011, 151, 139.
25. G. V. Kurlyandskaya and V. Levit, Biosens. Bioelectron., 2005, 20, 1611.
26. G. V. Kurlyandskaya, J. Magn. Magn. Mater., 2009, 321, 659.
27. H. Chiriac, M. Tibu, A. Moga and D. Herea, J. Magn. Magn. Mater., 2005, 293, 671.
28. H. Chiriac, D. D. Herea and S. Corodeanu, J. Magn. Magn. Mater., 2007, 311, 425.
29. T. Wang, Y. Zhou, C. Lei, J. Lei and Z. Yang, Sens. Actuators, B, 2013, 186, 727.
30. T. Wang, Z. Yang, C. Lei, J. Lei and Y. Zhou, J. Appl. Phys., 2014, 115, 223901.
31. J. Devkota, C. Wang, A. Ruiz, S. Mohapatra, P. Mukherjee, H. Srikanth and M. H. Phan, J. Appl. Phys., 2013, 113, 104701.
32. G. V. Kurlyandskaya, M. L. Sanchez, B. Hernando, V. M. Prida, P. Gorria and M. Tejedor, Appl. Phys. Lett., 2003, 82, 3053.
33. A. Kumar, S. Mohapatra, V. F. Miyar, A. Cerdeira, J. A. Garcia, H. Srikanth, J. Gass and G. V. Kurlyandskaya, Appl. Phys. Lett., 2007, 91, 143902.
34. F. Blanc-Béguin, S. Nabily, J. Gieraltowski, A. Turzo, S. Querellou and P. Y. Salaun, Infect. Agents Cancer, 2010, 5, 2.
35. T. Wang, Z. Yang, C. Lei, J. Lei and Y. Zhou, Biosens. Bioelectron., 2014, 58, 338.
36. H. Yang, L. Chen, C. Lei, J. Zhang, D. Li, Z. M. Zhou, C. C. Bao, H. Y. Hu, X. Chen, F. Cui, S. X. Zhang, Y. Zhou and D. X. Cui, Appl. Phys. Lett., 2010, 97, 043702.
37. L. Chen, C. C. Bao, H. Yang, D. Li, C. Lei, T. Wang, H. Y. Hu, M. He, Y. Zhou and D. X. Cui, Biosens. Bioelectron., 2011, 26, 3246.
38. T. Wang, Z. Yang, C. Lei, J. Lei and Y. Zhou, Phys. Status Solidi A, 2014, 211, 1389.
39. J. Devkota, T. T. T. Mai, K. Stojak, P. T. Ha, H. N. Pham, X. P. Nguyen, P. Mukherjee, H. Srikanth and M. H. Phan, Sens. Actuators, B, 2014, 190, 715.
40. K. J. Lodewijk, E. Fernandez, A. Garcia-Arribas, G. V. Kurlyandskaya, V. N. Lepalovskij, A. P. Safronov and B. J. Kooi, J. Appl. Phys., 2014, 115, 17A323.
41. J. C. Rife, M. M. Miller, P. E. Sheehan, C. R. Tamanaha, M. Tondra and L. J. Whitman, Sens. Actuators, A, 2003, 107, 209.

---