key: cord-0567621-f892y952
authors: Chugh, Vinit Kumar; Wu, Kai; Krishna, Venkatramana D.; Girolamo, Arturo di; Bloom, Robert P.; Wang, Yongqiang Andrew; Saha, Renata; Liang, Shuang; Cheeran, Maxim C-J; Wang, Jian-Ping
title: Magnetic Particle Spectroscopy (MPS) with One-stage Lock-in Implementation for Magnetic Bioassays with Improved Sensitivities
date: 2021-05-26
journal: nan
DOI: nan
sha: 6e39b4f81b09e2da75d4f615d72bf7277e731cca
doc_id: 567621
cord_uid: f892y952

In recent years, magnetic particle spectroscopy (MPS) has become a highly sensitive and versatile sensing technique for quantitative bioassays. It relies on the dynamic magnetic responses of magnetic nanoparticles (MNPs) for the detection of target analytes in liquid phase. There are many research studies reporting the application of MPS for detecting a variety of analytes including viruses, toxins, and nucleic acids, etc. Herein, we report a modified version of MPS platform with the addition of a one-stage lock-in design to remove the feedthrough signals induced by external driving magnetic fields, thus capturing only MNP responses for improved system sensitivity. This one-stage lock-in MPS system is able to detect as low as 781 ng multi-core Nanomag50 iron oxide MNPs (micromod Partikeltechnologie GmbH) and 78 ng single-core SHB30 iron oxide MNPs (Ocean NanoTech). In addition, using a streptavidin-biotin binding system as a proof-of-concept, we show that these single-core SHB30 MNPs can be used for Brownian relaxation-based bioassays while the multi-core Nanomag50 cannot be used. The effects of MNP amount on the concentration dependent response profiles for detecting streptavidin was also investigated. Results show that by using lower concentration/amount of MNPs, concentration-response curves shift to lower concentration/amount of target analytes. This lower concentrationresponse indicates the possibility of improved bioassay sensitivities by using lower amounts of MNPs.

Magnetic particle spectroscopy (MPS) for magnetic bioassays was first reported in 2006. 1, 2 It is a technology that derived from magnetic particle imaging (MPI), which relies on the nonlinear magnetization curves of magnetic nanoparticle (MNP) tracers for medical tomographic imaging. 3 While, on the other hand, MPS monitors the dynamic magnetic responses of MNPs in liquid phase and assists in the analysis of the nanoparticles' binding status. To be specific, MNPs dispersed in the liquid adds an additional degree of rotational freedom that allows for bioassays directly from liquid phase. Upon the application of external AC magnetic fields (also called driving fields or excitation fields), the magnetizations of MNPs follow the field direction through a Brownian relaxation process, which is a physical rotational motion of nanoparticles. The dynamic magnetic responses of MNPs can be transformed to real-time voltage signal and monitored by using a pair of pick-up coils. The signal spectrum contains higher harmonics that are uniquely generated by MNPs. MPS-based bioassays use these harmonics of oscillating MNPs as a measure of the rotational freedom, i.e., the bound status of MNPs to target analytes from liquid phase. With appropriate chemical modifications, MNPs can be surface functionalized with proteins (such as antibodies, antigens, streptavidin, biotin, etc.), nucleic acids (DNA and RNA), and polymers, customized according to different bioassay purposes. [4] [5] [6] These surface functionalized MNPs are nanoprobes that can bind to target analytes with high specificity and have shown great promise not only for MPS-based bioassays but also other magnetic bioassays and medical applications. Some target applications include magnetoresistive and magnetic impedance biosensors, and nuclear magnetic resonance biosensors. [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] In liquid phase MPS-based bioassays, the binding of MNPs to targe analytes will hinder or even block the Brownian relaxation of MNPs and thus, causes a phase lag between their magnetizations and the external AC magnetic fields. The binding of MNPs to analytes causes weaker dynamic magnetic responses of MNPs and, as a result, the harmonic amplitudes drop is expected. Thus, this assay mechanism allows for development of one-step, wash-free, and quantitative detection of target analytes directly in liquid phase. 18 In the past decade, various MPS platform designs have been reported, such as two AC (also called frequency mixing) and one AC driving fields methods based on how the excitation fields are applied, as well as the surfaceand liquid phase-based bioassays (based on how the MNP are bound). 11, 12, [18] [19] [20] [21] [22] [23] [24] [25] However, a common issue with all MPS systems is the presence of feedthrough signal corresponding to the driving magnetic fields which can be orders of magnitude higher than that of the MNP signal and can be a limiting factor in the MPS-based bioassays.

Modalities based on both active and passive cancellation for such signal have been explored. [26] [27] [28] [29] [30] In this present work, we are reporting a modified version of two AC magnetic fields-based MPS system with the addition of one-stage lock-in scheme for passive cancellation of feedthrough signal and improved detection sensitivity. The performance and sensitivity of this one-stage lock-in MPS system is firstly evaluated by determining the lowest amount of MNPs detectable in liquid phase. Then we explored the impact of MNP amounts on the ability of MPS system to detect varied concentrations of target analytes by using a streptavidin-biotin binding system in liquid phase.

Materials. The Nanomag50 MNPs are 50 nm superparamagnetic dextran iron oxide composite nanoparticles functionalized with biotin, with weight concentration of 5 mg/mL and particle concentration of 91.36 nM, purchased from micromod Partikeltechnologie GmbH (product no. 79-26-501). The SHB30 MNPs are 30 nm iron oxide nanoparticles functionalized with biotin, with weight concentration of 1 mg/mL and particle concentration of 34 nM, provided by Ocean NanoTech. Streptavidin from Streptomyces avidinii is purchased from Sigma-Aldrich (product no. S4762). Figure 1 (a) for recording the magnetic response which in principle cancels out any EMF generated due to the applied magnetic field and permits exclusive recording of the magnetic response from MNPs. However, in practice, the feedthrough signal (i.e., EMF due to driving fields) tends to be a real problem. Figure 1 (f) shows the FFT spectrum of MNP response and it can be clearly observed that the feedthrough signal corresponding to fH is two orders of magnitude higher than the 3 rd harmonic responses of MNPs. Herein, a one-stage lock-in based approach is used to remove the feedthrough signals corresponding to driving field frequencies and to capture only the MNP responses. Figure 1 (a) & (d) depicts schematic diagrams of the signal decoding topology with and without the one-stage lock-in implementation and the corresponding captured signals in temporal and frequency domains (Figure 1(b) , (c), (e), and (f)). From Figure 1 (c), we can clearly observe that the one-stage lock-in approach significantly removes the feedthrough signals. Another advantage that lockin based approach provides is to reduce sampling frequency requirements. Experiments on a MPS system without lock-in design (Figure 1(d) ) require a sampling frequency of up to 500 KSPS to obtain optimal SNR performance.

However, the lock-in based approach shifts the MNP spectra to a lower frequency range and hence allows for better SNR performance with lower sampling rate (100 KSPS) that is easily adaptable to on a handheld system setting.

Circuit Design of One-stage Lock-in MPS System. The circuit for one-stage lock-in MPS system can be divided into 3 main parts: (1) power, (2) excitation coil driver, and (3) signal processing. Figure 2 Excitation Coil Driver. The coil excitation circuit consists of sine wave generation followed by voltage source implementation. A 2-channel DDS IC AD9958 (Analog Devices) is used for generation of sinusoids for highfrequency (fH) driving field and for phase shifted reference to the synchronous demodulator in signal processing stage. Differential output from AD9958 is passed through instrumentation amplifier INA128 (Texas Instruments) to convert into the single ended signal. DDS AD9833 (Analog Devices) is used for generation of sinusoid for low-frequency (fL) driving field. An inverting amplifier topology utilizing OPA548 (Texas Instruments) is used for the voltage source implementation for driving the primary and secondary coils.

Signal Processing. The differential signal from pick-up coils is amplified using the precision instrumentation amplifier INA828 (Texas instruments) for removing the common mode noise. Signal at this stage consists of MNP response centered around 5 KHz excitation frequency as can be seen in Figure 1 (f). The amplified signal is processed using a lock-in based implementation consisting of an AD630 synchronous demodulator (Analog Devices) with phase shifted 5 KHz (fH) reference signal followed by a bandpass filter to reject signal images at 0 Hz and around 10 KHz. The band-pass filter is implemented using the Sallen-key scheme. The filtered signal is sampled using LTC2368-24, 24-bit SAR ADC (Linear Technology, Analog Device) at 100 KSPS sampling rate.

STM32F747 microcontroller (STMicroelectronics) is used for handling and storing the sampled data which is then transmitted to a PC using UART communication protocol for further processing and analysis. The Nanomag50 MNPs were diluted up to 8192 times, from 5 mg/mL (400 μg per vial, no dilution) to 610 ng/mL (48.8 ng per vial, 8192-fold dilution). Three independent MPS readings were carried out on each sample. Figure 3 , the amplitudes of higher harmonics (i.e., from the 3 rd to the 15 th harmonics) linearly decreased as the amount of Nanomag50 MNPs decreases. The inset of Figure 3 (g) shows a zoomed in view of the 3 rd harmonic amplitude collected from lowest amount of Nanomag50 MNPs. It was concluded that the minimum detectable amount of Nanomag50 MNPs by MPS system is 781 ng (512-fold dilution). In addition, varying amount of SHB30 MNPs were prepared by two-fold dilutions in the same manner. The SHB30 MNPs are diluted up to 2048 times, from 1 mg/mL (80 μg per vial, no dilution) to 488 ng/mL (39 ng per vial, 2048-fold dilution). As shown in Figure 4 , the amplitude of higher harmonic linearly decreases as the SHB30 amount decreases in the samples. The minimum detectable amount of SHB30 MNPs by the developed one-stage lock-in topology based MPS system is 78 ng (1024-fold dilution) . The modified MPS system shows a 50-fold improvement in sensitivity for detection of iron oxide MNPs when compared to our previous work (limit of detection was 4 µg) not utilizing lock-in based approach as shown in Figure 1 1 (a.4) ), the large cluster matrices are disrupted, and MNP clusters no longer form the majority in liquid phase. Individual MNPs when saturated with streptavidin in this scenario will exhibit harmonic response that is greater than that of large clusters but smaller than that of the free MNPs due to an increment in effective hydrodynamic size. Table 1 . clustering and hinder their Brownian relaxation. We observed a weak reversal of harmonic signals from its nadir for groups I to III (in vials #1-4), however the phenomenon was not as prominent as observed for group IV, even at the highest (400 nM) concentration of streptavidin tested. It is plausible that 400 nM streptavidin is insufficient to completely saturate MNP associated biotins in groups I -III. These three different response zones are highlighted in blue (excessive amount of streptavidin), orange (linear response region), and green (inadequate streptavidin) regions in Figure 5 (e-h) as well as in Scheme 1(d). Overall, the higher harmonics show parallel concentration-response curves in the 3 rd harmonic signals. It is observed that the linear response region of concentration-response curve moves towards lower streptavidin quantities (as shown by the arrow in Figure 5 (a) and summarized in Table 2 ) with the dilution of SHB30 MNPs from group I to group IV. Figure 5(b-d) compares the 3 rd harmonic amplitudes of SHB30 MNPs from groups I -IV detecting the same concentration/amount of streptavidin. Interestingly, the best SNR performance was not observed from the 3 rd harmonic response of the MNPs but, instead the least standard deviation was observed consistently when 9 th and 15 th harmonic were used to represent MNP binding information. MNPs, whose Brownian relaxation is intrinsically blocked cannot be used for liquid phase MPS-based bioassays.

In the present study, we have reported a method for passive cancellation of feedthrough signal for dual-frequency (2 AC driving fields) MPS methodology. The sensitivity of this one-stage lock-in MPS system is first evaluated by detecting two-fold dilutions of commercial iron oxide MNPs: SHB30 and Nanomag50. The lowest amount detectable by the system was confirmed to be 78 ng for the single-core SHB30 and 781 ng for the multi-core Nanomag50 iron oxide MNPs. In addition, using a streptavidin-biotin binding system as a model, we explored the effects of MNP amount on concentration-responses profiles for detecting target analytes. By fine tuning the MNP amount/concentration in the sample, we were able to shift the linear response region for streptavidin detection. Results confirmed that the liquid phase bioassay scheme shows improved sensitivity on our one-stage lock- our one-stage lock-in MPS system is able to detect as low as 800 fmole of streptavidin using 1.0625 nM concentration of MNPs. However, this detection limit can be further improved by using lower amounts (higher dilutions) of MNPs. In addition to improved sensitivities, the cost per assay can be further reduced by using less amount of MNPs. Low-cost options allow the point-of-care assays to be available in impoverished regions with scarce medical resources.

Our future plans include improving the sensitivity of MPS system by using active feedthrough cancellation techniques. The passive feedthrough cancellation reported in the present study help remove the feedthrough signals before ADC sampling. Therefore, with proper amplifications in place, we can take advantage of true ADC resolution for improved sensitivity. However, this method of removing feedthrough signals and amplifying remnant signals still holds disadvantage as the intrinsic noise components also get amplified through the amplification units. Active cancellation of feedthrough can help improve the signal-to-noise ratio before the instrumentation amplifier stage and hence allow for better sensitivity in MPS bioassay applications. We believe this passive feedthrough cancellation methodology in MPS system design and fine tuning MNP amount/concentration to shift linear response region will elucidate new ways of increasing detection sensitivity of MPS-based bioassays.

Supporting Information: S1. Magnetic Properties of Nanomag50 and SHB30 MNPs.; S2. Low noise powerline split implementation.; S3. Experimental Designs for Varying Amount of Nanomag50 and SHB30 MNPs.; S4.

Multi-core Nanomag50 MNPs for Streptavidin Detection. S1. Magnetic Properties of Nanomag50 and SHB30 MNPs.

The static magnetic hysteresis loops of Nanomag50 and SHB30 MNPs are measured by PPMS and shown in Figure S1 . External magnetic fields are swept from -5000 Oe to +5000 Oe and -500 Oe to +500 Oe. The saturation magnetizations (Ms, emu/g) of Nanomag50 and SHB30 MNPs under 5000 Oe field are 30.7 emu/g and 36.8 emu/g, respectively. The specific magnetic magnetizations (M, emu/g) of Nanomag50 and SHB30 MNPs under 500 Oe field are 18.2 emu/g and 29.2 emu/g, respectively. Nanomag50 MNPs show superparamagnetic properties with zero magnetic coercivity while, on the other hand, SHB30 shows a coercivity field of 36 Oe. Due to the increased inter-particle distances introduced by surface functional groups (biotin in this case for SHB30), this negligible magnetic coercivity won't cause SHB30 to cluster in the absence of magnetic fields. In addition, the magnetic moment per particle (m, emu/particle) is also calculated based on the particle concentrations, as shown in Table S1 . Figure S1 . The static magnetic hysteresis loops of Nanomag50 and SHB30 nanoparticles measured by PPMS with external magnetic field ranges of (a) & (c) 5000 Oe and (b) & (d) 500 Oe, respectively. S4 S2. Low noise powerline split implementation. Figure S2 . Schematic for low-noise voltage generation of ±24 V for the excitation coil driver setup.

A total of 15 vials containing two-fold dilutions of Nanomag50 multi-core MNPs are prepared, with concentrations from 5 mg/mL (vial #1) down to 610 ng/mL (vial #14). Vial #15 is a negative control group without loading any MNPs. Each vial contains 80 μL of MNP solution. The Nanomag50 MNP amount per vial as well as dilution folds are summarized in Table S2 . Table S3 . 

Magnetic Particle Detection by Frequency Mixing for Immunoassay Applications

New Type of Biosensor Based on Magnetic Nanoparticle Detection

Tomographic Imaging Using the Nonlinear Response of Magnetic Particles

Surface Functionalization of Magnetic Nanoparticles via Palladium-catalyzed Diels-Alder Approach

New Advances Strategies for Surface Functionalization of Iron Oxide Magnetic Nano Particles (IONPs)

Magnetic Particle Spectroscopy-Based Bioassays: Methods, Applications, Advances, and Future Opportunities

Denaturation Strategies for Detection of Double Stranded PCR Products on GMR Magnetic Biosensor Array

Magnetic Particles-Enabled Biosensors for Point-of-Care Testing

Magnetic Impedance Biosensor: A Review

Magnetic Immunoassay for Detection of Staphylococcal Toxins in Complex Media

Magnetic Particle Spectroscopy for Detection of Influenza A Virus Subtype H1N1

Miniaturized Nuclear Magnetic Resonance Platform for Rare Cell Detection and Profiling

Point of Care Assessment of Melanoma Tumor Signaling and Metastatic Burden from ΜNMR Analysis of Tumor Fine Needle Aspirates and Peripheral Blood

Micro-and Nanofabrication NMR Technologies for Point-of-Care Medical Applications-A Review

Portable GMR Handheld Platform for the Detection of Influenza A Virus

Introduction to Compact NMR: A Review of Methods

Magnetic Particle Spectroscopy: A Short Review of Applications Using Magnetic Nanoparticles

A Portable Magnetic Particle Spectrometer for Future Rapid and Wash-Free Bioassays

Multiplex Biosensing Based on Highly Sensitive Magnetic Nanolabel Quantification: Rapid Detection of Botulinum Neurotoxins A, B, and E in Liquids

Multiplex Biosensing with Highly Sensitive Magnetic Nanoparticle Quantification Method

Ultrasensitive Detection Enabled by Nonlinear Magnetization of Nanomagnetic Labels

Ultrasensitive Quantitative Detection of Small Molecules with Rapid Lateral-Flow Assay Based on High-Affinity Bifunctional Ligand and Magnetic Nanolabels

Evaluating Blood Clot Progression Using Magnetic Particle Spectroscopy

Molecular Sensing with Magnetic Nanoparticles Using Magnetic Spectroscopy of Nanoparticle Brownian Motion

Analog Receive Signal Processing for Magnetic Particle Imaging

Multifrequency Magnetic Particle Imaging Enabled by a Combined Passive and Active Drive Field Feed-through Compensation Approach

Arbitrary-Waveform, Mpi Spectrometer and Relaxometer for Comprehensive Magnetic Particle Optimization and Characterization

Benchtop Magnetic Particle Relaxometer for Detection, Characterization and Analysis of Magnetic Nanoparticles

AC Magnetometry with Active Stabilization and Harmonic Suppression for Magnetic Nanoparticle Spectroscopy and Thermometry

United States c Ocean Nano Tech LLC

Multi-core Nanomag50 MNPs for Streptavidin Detection

Four experimental groups each consists of 10 samples/vials are designed and each vial contains 80 μL Nanomag50 MNP of 10.75-, 21.5-, 43-, and 86-fold dilutions are designed as shown in Table S4. The concentration-response profiles for streptavidin from groups A -D are plotted in Figure S3

680 fmole 400 nM, 32 pmole (#1) 200 nM, 16 pmole (#2) 100 nM, 8 pmole(#3) 50 nM

Group B #1-10 4.25 nM (21.5-fold dilution), 340 fmole Group C #1-10 2.125 nM (43-fold dilution), 170 fmole Group D #1-10 1.0625 nM

 Figure S3 . Concentration-response profiles for streptavidin from groups A -D, based on the 3 rd to the 15 th harmonics. Error bars represent standard errors.