An Early Cancer Diagnosis Platform based on Micro-magnetic Sensor Array Demonstrates Ultra-high Sensitivity

In this paper a highly sensitive detection methodology for obtaining the concentration of protein tumor markers via the selective incorporation of super paramagnetic iron oxide nanoparticles (MNP) onto an antigen is reported. The tumor marker concentration is measured with a nano-oxide layer inserted giant magneto-resistive (GMR) sensor that determines the attenuation of an external magnetic field that is induced by the MNP. The 15 nm MNPs, used herein, exhibit a super paramagnetic behavior with saturation magnetization of 36 emu/g. The output voltage signal of the Specular GMR sensor is 500 μV/Oe. The patterned GMR multiple-bio-sensor array, has demonstrated real-time measurements of carcinoembryonic antigen (CEA) protein concentrations down to femtomole level (10-15 mole) in a variety of clinically relevant media with a linear dynamic range of over five orders of magnitude. The sensitivity of the GMR bio-array platform is 1000 times higher than conventional enzyme-linked immunosorbent assay (ELISA) technique. The arrays of magnetoresistive sensors offer great promise in appli cations for early cancer diagnosis. *Corresponding authors: Wei Zhang, State of Key Laboratory of Materialsoriented Chemical Engineering and School of Chemical Engineering, Nanjing Tech University, Nanjing, Jiangsu, 210009, PR. China, Tel: 86 25-83587060; E-mail: zhangw@njtech.edu.cn Xuefeng Hu, State of Key Laboratory of Materials-oriented Chemical Engineering and School of Chemical Engineering, Nanjing Tech University, Nanjing, Jiangsu, 210009, PR. China, E-mail: xuefeng.hu@njtech.edu.cn Received November 25, 2015; Accepted December 29, 2015; Published January 10, 2016 Citation: Yang N, Li T, Zhang PP, Chen X, Hu X, et al. (2016) An Early Cancer Diagnosis Platform based on Micro-magnetic Sensor Array Demonstrates Ultrahigh Sensitivity. J Nanomed Nanotechnol 7: 344. doi:10.4172/2157-7439.1000344 Copyright: © 2016 Yang N, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.


Introduction
Colorectal cancer (CRC) is one of the most commonly diagnosed types of cancer worldwide. Carcinoembryonic antigen (CEA) in human serum is reported to be an important biomarker for early CRC diagnosis [1][2][3]. Because CEA's concentration in human serum is very low (approximately 2.5 ng/mL), a highly sensitive measurement technique for the identification of the presence of CEA is required. Currently Enzyme-linked immunosorbent assay (ELISA) [4,5] is a widely used protein readout method based on a fluorescent or colorimetric transducer response. However, inherent auto fluorescence and optical absorption of either the matrix of biological samples or reagents is a major limiting factor of ELISA technique. Several other technologies, such as, nanowires [6], carbon nanotubes [7], electrochemical biosensors [8], protein microarrays [9], and quantum dots [10], has been also investigated as diagnostic tools for protein detection. The detection physics for these non-optical methods are rooted in charge-based interactions between either the protein or tag and the sensor, which limits the reliability of these sensor types due to solutions of varying pH and ionic strength.
Bio-sensing strategies based on magnetic nanoparticles (MNP) have received considerable attention [11,12]. Magnetic particles possess unique magnetic properties and easily functionalized with stable, nontoxic protective coatings for bioconjugation [13][14][15]. Magnetic background in biological samples is minimal, since these are composed predominantly of diamagnetic molecules, and even large magnetic fields are compatible with biochemical processes. In particular, the stray magnetic fields from the beads, as well as externally applied field gradients which can be used to generate forces on the beads to remotely manipulate them are not screened in an aqueous biological environment. Tiny size, which exhibits a superparamagnetic property, allows for the formation of stable dispersion in reaction solution, making them well suitable for target separation [14,16,17], targeted delivery of therapeutic agents [18], and biosensor based diagnostics [19].
Among these platforms or concepts using magnetic nano-particle labels and magnetometers, the first demonstration of detecting biological molecules by employing magnetic labels was achieved with superconducting quantum interference device (SQUID) [20]. In 1998, Baselt et al. introduced the bead array counter (BARC) for the detection of a single molecule combined with a magnetic particle [21]. This platform was then developed into a system for biological warfare agents detection [22]. In addition, the efforts paid on substituting giant magnetoresistive (GMR) sensors for traditional magnetic sensors in immunoassay or labon-a-chip-based rapid point-of-care testing (POCT) has been significantly increasing [23][24][25]. Most recent studies [26][27][28][29][30][31][32][33][34][35] have demonstrated the usage of GMR biosensors that detect either as few as 600 streptavidin coated magnetic particle (<10-21 mol, zeptomol) or CEA with at lower limit of detection around 50 fM (10)(11)(12)(13)(14)(15). For CEA sensing, it is ideal to have the particle labels either 20 nm or smaller in diameter [36]. However, the magnitude of the detectable magnetic field decrease with particle size, thus, a more sensitive sensors and measurement systems is required [37].
In this paper, a low-amount biomarker detectable platform based on a 15 nm diameter MNPs and high sensitive specular giant magnetoresistance (GMR) sensors was investigated. The GMR ratio can be enhanced by nano-oxide layer (NOL) insertion inside the pinned layer. The magnetic particle agglomeration can be avoided due to the superparamagnetism.. The ultra-small size of MNPs can be more comparable to that of the conjugating biomolecules so that they would not block biomolecular interactions. We also demonstrate the feasibility of this detecting system for real biological applications,
As shown in Figure 1, similar to ELISA, the magnetic arrays require bio-function steps and links prior to real-time measurement. To form the base layer of the bio-functionalization, a 2% solution of polyethyleneimine (PEI) (Shanghai Qunfang Bio) in deionized water was applied to the chip surface. CEA analyte (Shanghai, NC-Bio) was diluted to the desired concentrations in the phosphate buffer saline (PBS). Twenty µL of this solution were selectively pipetted into sensors which were printed by captured antibodies onto sensors blocked by bovine serum albumin (BSA). Biotinylated antibodies to CEA (Ocean Nanotech) were diluted to a concentration of 2 µg/ml in PBS. 20 µL of this linker antibody solution were pipetted onto the surface of each sensor. Finally, 40 µL of streptavidin labeled MNPs (Ocean Nanotech) was added and incubated without stirring for 30 min at room temperature.
The size and morphology of the MNPs were characterized by Transmission Electron Microscopy (TEM). Electron Energy Loss Spectra (EELS) was used to characterize the composition of the streptavidin labeled MNP. The magnetic property measurement of the MNPs was carried out at room temperature by using a Superconducting Quantum Interference Device (SQUID). Analyte quantification was performed on a probe station (KLA-350 A), which used a source meter in a two-point configuration. Local external magnetic fields up to 3000 Oe in either the transverse (hard axis) or longitudinal (easy axis) direction relative to the sensor was generated by a quadarupole magnet [38].

Result
A typical TEM image of the MNPs is presented in Figure 2a. The MNPs exhibit a spherical shape with an average size of around 15 nm. As shown in Figure 2b, the MNPs, characterized by SQUID, display a saturation magnetization of 36 emu/g under a field of more than 40 KOe and 1.01 emu/g at a field of 100 Oe. It is estimated that 1 gram of these 15 nm MNPs will correspond to 2.63 x 1017 particles. The net magnetic moment of single MNP is 6.66 x 10-17 emu at a field of 40 kOe and 3.8 x 10-18 emu at a field of 100 Oe [39]. TEM images of streptavidin-modified nanoparticles, as shown in Figure 2c, demonstrated uniform streptavidin modification with no significant particle aggregation. The iron compositions of streptavidin-modified MNP, as characterized by EELS, as shown in Figure 2d, exhibited a streptavidin uniformly distributed across the MNP particle with no perceivable grain boundary. The aforementioned findings suggest not much significant nonspecific binding of streptavidin to the magnetic nanoparticles.
Each layer of secular GMR stack, as shown at Figure 3, has an nmlevel thickness with sharp interface between layers. Magnetoresistance ratio (MR) ratio for the full-film specular GMR sensor, measured by the field strength along the hard axis (transverse), is about 16% with a maximum sensitivity of 1.6% Oe-1 at 10 Oe field. Very slight hysteresis is shown in the loop for full film specular GMR [37,38,40,41]. The optical image of patterned specular GMR bioarrary is shown at Figure  4a. Each sensor was patterned into a square shape with 200 µm × 200 µm dimensions. After electrode deposition (250 µm × 25 µm), each sensor occupies an active area of 200 µm by 150 µm while the edge to edge distance of two sensors' separation is 100 µm. The resistance of each sensor is about 1.1 K Ω under a zero field.
To achieve statistically reliable results, for the CEA detection experiments, five sensors were selected. The concentration of CEA was about 500 femtomole (1 ng/mL, 1 µL droplet). The aforementioned second group of 5 sensors acted as negative controls. The MR loops of specular GMR sensors with CEA analyte and sensor blocked by BSA were measured before and after the spiking of the streptavidin labeled MNPs. As shown on Figure 4b, all curves exhibited giant magnetoresistance (GMR). The dR/R was 10.03% and 8.99% for patterned sensors before and after the MNPs was spiked, respectively. In contrast, a 10.01% dR/R was measured for sensors with BSA blocking under the hard-axis bias condition. Shape anisotropy due to the patterned rectangle GMR sensor caused hysteresis to appear in the loop in the transverse curve. The dR/R ratio changes before and after magnetic particle incorporation is used to identify low amounts of biomolecules. The sensitivity of the sensor before MNPs spiking is around 0.5%/Oe (and 500 µV/Oe). Each dipole moment generated from magnetic particle creates a small added field component Hdipole at the sensor. The sensor will experience an effective field strength which is slightly smaller than the external applied field (Heffect = Happlied − Hdipole). This effect generates a detectable signal on the GMR sensor arrays, normally a decrease in dR/R. The signal levels of 500 femtomole CEA is 5 × 104 µV, which is significantly larger and distinct from the nonspecific signal of 15 µV on the sensor blocked  by BSA. The longitude MR transfer curve is shown in Figures 4c and  4d as well. The longitudinal curve exhibited the same trend as did the transverse field, and a 0.5% MR drop was observed.
To deduce the dynamic range of signal versus concentration scaling relationship, a series of quantitative calibration experiments employing varying carcinoembryonic antigen (CEA) concentration were conducted. CEA analyte was diluted in a PBS buffer to produce analyte concentrations ranging from 1 ng/mL (molar concentration 5 pM) to 1 pg/mL (molar concentration 5 fM). Eight solutions containing between 5 fM to 500 pM CEA analyte were using for testing of the sensor system. Each of the 8 solutions was pipetted onto five active sensors on the array. From the CEA concentration and sensor signal, the linear response curve (on a log-log plot) shown in Figure 5 was created. Linearity of response was obtained even at the fM concentration. These results indicate that detection of CEA is possible down to the fM-level without any biological amplification. This dynamic range of linearity is comparable to the best GMR-based detecting [30,42] and is three orders of magnitude higher than conventional ELISA (4 pM) [43][44][45].

Discussion
The enhancement of NOL (nano-oxide-layer) insertion and specular effects on the giant magnetoresistance (GMR) can be explained by the semiclassical Boltzmann theory. Spin valves (SVs) consist of two ferromagnetic (FM) nanolayers separated by a non-magnetic spacer. One of the FM layers is pinned by an adjacent AFM layer, while the other (free layer) rotates its magnetization (M) under a small magnetic field. This allows both magnetizations either parallel (low resistance) or antiparallel (high resistance) to be obtained with an externally applied magnetic field. The nano-oxide layer NOL inserted either adjacent to the pinned layer or free layers [46][47][48] can highly enhance the majority spin orientated electrons specularly scatter at NOL [49,50] and thus a high dR/R.
In order to support the observations in this study, the theoretical magnetic stray field of a fully magnetized 15 nm diameter Fe 3 O 4 nanoparticle was calculated at different distances from the free layer of GMR sensor. The stray field, H, of a magnetic dipole of length l placed perpendicularly on top of the sensing area of the GMR sensor at a distance d is approximated by the following formula (the case that d With M being the magnetic moment (5 × 10 -18 emu, of a single Fe 3 O 4 nanoparticle from SQUID measurement at 100 Oe), and l is the characteristic particle dimension (in this case the 15 nano meter diameter). The calculation shows the theoretical active distance of the sensor (at 3 × 10 -4 Oe) lies between 200 and 250 nm above the sensor. The stray field of the 2 Oe observed in Figure 4b, in principle, can be produced by 6 × 10 4 magnetic nanoparticles of size 15 nm passing the sensor at a height of approximately 200 nm. To further enhance the detecting sensitivity of GMR bioassay, decreasing the distance between MNP and the free layer of the sensor is critical too.
The 15 nm MNP-based specular GMR assay presents a very sensitive magnetic detection method for protein arrays. With carefully screened antibodies, the sensitivity and selectivity of the MNP-based specular GMR assay is already sufficient for clinically relevant protein detection in real world serum samples [52,53]. The technique can be readily performed with the multiplex protein detection. Individual sensor microspotting will allow the accommodation and measurement of up to 22000 different probes on a current generation chip (probing throughput not withstanding). It is expected that significant sensitivity improvements can also be made on the sensor; for example, using either a biased magnetic tunnel junction sensors based on MgO barrier or by using sensors with smaller geometry [54]. In the near future, capture agents with higher affinity, and similarly small but higher magnetic moment MNPs are expected to further enhance the analytic sensitivity of MNP-based assays.
In conclusion, the highly sensitive detection of protein tumor markers incorporating a MNP based specular GMR bio-assay was demonstrated. Also, it was demonstrated the detecting system can adopt the principle of ELISA assay with improved sensitivity. The GMR bioarray, permits real-time measurements of CEA protein concentrations down to the femtomolar level with a linear dynamic range of over five orders of magnitude. This GMR-and magneticnanoparticle based detecting system, therefore, is expected to be applicable to many other biological systems for detection and quantification of various biomolecules. The high sensitivity of this detecting system opens new paths for the detection of biomolecules involved in the etiology of various diseases. Most importantly, the magnetic/electric nature of this detecting system and the associated cost benefit, scalability and versatility will help the realization of personalized medicine.
Author contributions statements: NY carried out the sample preparation and experimental measurements. TL and PPZ help to analyze the experimental result. HYF drafted the manuscript. XQC and WZ conceived and designed the experiments, revised the manuscript, and provided financial support. All authors read, review and approved the final manuscript.