Electronic Bio-Reconfigurable Impedance Platform for High Sensitivity Detection of Target Analytes

Abstract

The present research presents a portable bioelectronic platform for multiplex detection to read biosensor chips with several sensing sites for real-time analyte capture. The technique is based on Differential Impedance Sensing (DIS) of the target through functionalized nanoparticle amplification. Gold-interdigitated microelectrodes are the core of the biosensing system. They are designed in a differential configuration, reference and active sensor, to counteract all possible mismatches such as temperature fluctuations and variations in the ion content of the solution. The surface of the sensor is biochemically functionalized with a synthetic probe specifically developed for the selected target. The successful combination of all of these elements allowed the system to detect IgG antibodies spiked in buffer with a limit of detection of below 100 pg/mL. In a real case study for viral infection diagnosis, the system has been challenged with infected human serum samples for digital counts of anti-dengue virus antibodies, achieving the detection of clinically relevant target concentrations. Also, the bio-reconfigurability of the system has been successfully tested with oligonucleotide detection down to pM target concentration. To allow the portability of the entire measurement setup, the setup has been equipped with a custom electronic board based on an FPGA module allowing a multiplexing approach for the parallel reading of several electrodes. The final system provides simple and effective bio-reconfigurability, exploiting advances in bio-recognition through proper probe selection and boosting the possible use of multiplex sensing to a broad spectrum of needs.

1 Introduction

Detection of viruses is essential for the control and prevention of viral infections. For the diagnosis, it is possible to directly detect the whole virus [1] or to determine the antibodies produced against virus proteins during and/or after the virus incubation period [2, 3]. The analytes as viral nucleic acids (DNA and RNA), viral proteins, intact viral particles and antibodies can all be used for the diagnosis [4]. Conventionally, they are detected using traditional methods like polymerase chain reaction (PCR) [5], virus culture, and enzyme-linked immunosorbent assay (ELISA) [6]. Still, these tests are usually time and reagent-consuming and do not have multiplex capability, thus not allowing the detection of several targets simultaneously [7]. In recent years, there has been a focus on simpler and faster detection methods with biosensors solutions. Fundamental to these systems is the biotransducer that converts the concentration of the targeted analyte into a proportional signal as a combination of a biological recognition layer and a physical transducer. Among the transduction mechanisms, biosensors can rely on dyes or fluorescent labels [8, 9] or innovative approaches such as functionalized tapered optical fiber [10], silicon microring resonators for active opto-magnetic biosensing [11] or Surface Plasmon Resonance-based technologies (SPR) [12]. These techniques have further improved the quality of the measurement but still require complex components and can be hardly miniaturized into a multipurpose platform. To overcome these limitations, an electrical approach using impedimetric biosensors can be convenient. As a device, planar interdigitated electrodes (IDEs), which consist of a pair of comb-shaped micro-band array electrodes and a pad at the bottom, are widely applied to the field of sensors due to their simple manufacturing process. They can be miniaturized, integrated in a microfluidic system and fabricated with bio-compatible materials. The currents/voltages between the IDEs can be evaluated to directly detect binding [13] or indirectly detect the change of the electrical properties (complex impedance, resistance or capacitance) of the medium around the electrodes [14]. Between non-faradaic [15] and faradaic [16] approach, the latter application requires a wide-range frequency sweep and the use of potentially hazardous redox mechanisms for current induction resulting in complex experimental protocol and electronic reading system. Non-faradaic impedance systems can be used in the simpler fixed frequency to provide transient information [17] about rapid changes in impedance thanks to increasing sampling [18, 19]. As the main result of this thesis, it has been demonstrated for the detection of Dengue Virus infection a non-faradaic based biosensor system that combines a specific biochemistry recognition mechanism with an electronic platform based on Differential Impedance Sensing [20]. Due to the target features, the impedance signal is amplified by hybridizing the probe-target with a functionalized polystyrene nanoparticle. The increased signal-to-noise ratio helps to overcome drawbacks such as laborious surface preparation, the usually limited storage time of the sensor and surface biofouling that occurs in unpurified samples. Moreover, the single-frequency measurements and data analysis on the enhanced signal simplify the platform hardware and software, enabling portable analysis.

2 The Differential Impedance Sensing Approach

2.1 Platform Overview and Experimental Protocol

The sensors based on IDEs have been designed in a differential configuration in order to evaluate the total value of the impedance and the beads contribution. In this structure, the selected probe is spotted only over the active sensor, while the reference sensor operates as a negative control without surface functionalization. The resulting measurement is unaffected by macroscopic temperature fluctuations and variations in the liquid medium (i.e., ion concentration) that determine the assay’s conductance. A schematic representation of the biosensing system is shown in Fig. 1. The gold IDEs have fingers of 3-$\upmu$m width, 90-$\upmu$m length and 3-$\upmu$m spacing. The active area of each sensor (90 $\upmu$m $\times$ 90 $\upmu$m) and the comb dimensions of 3 $\upmu$m have been optimized considering the resolution of the spotting machine for the functionalization of the sensor surface with the probe, the full coverage of the active area and the beads dimension (800 nm).

Fig. 1

The Biosensing system. a The workflow for electrode realization: COMSOL simulation of the differential IDE structure, process steps for microfabrication of gold IDE and photographs of the final biosensor structure. b A schematic view of the microfluidic system of the differential electronic impedance measurement setup. c shows the configuration to perform the impedance measurement in a digital differential mode and the full spectrum of impedance versus frequency in PBS solution. The selected working frequency around 1 MHz ensures that we are working in a region sensitive to variations in the liquid conductivity

The IDEs detection properties have been optimized and validated using finite element method simulations (COMSOL Multiphysics) to reach the maximum impedance variation for one-bead detection. The resulting sensing area of the IDE leads to a dynamic range of 100 ppm. A single bead in the sensing volume over the electrodes would give a signal variation of about 15 m$\Omega$ (over 935 $\Omega$ of total impedance), allowing a direct correlation between the electrical signal, the number of beads and the number of biological targets captured by the probes. The fabrication process of the electrodes was carried out at PoliFAB (the micro and nanotechnology center of the Politecnico di Milano) on 3-inch borosilicate wafers using standard microfabrication techniques. The interdigitated gold electrodes are exposed to liquid with micro windows patterned on a SU8 layer (thickness of 5 $\upmu$m) used as a passivation layer over metal connections to limit leaky current paths.

The experimental protocol, reported in Fig. 2, consists of an incubation phase targeting specific primary antibody binding and labeling by secondary antibody, followed by a beads counting phase targeting nanobeads binding, impedance measurement and beads count. While this second phase is always performed in the microfluidic chamber, the incubation phase can also be conducted outside the microfluidic system.

Fig. 2

Scheme of the protocol of the assay. The impedance measurement to take into account for the differential detection are the ones before and after the beads injection in buffer condition

The limit of detection (LOD) of the system has been investigated and the same protocol has been replicated for different concentrations of primary antibody, from 50 $\upmu$g/ml down to 100 pg/ml. The LOD was extrapolated from the impedance value of blank samples plus 3 standard deviations (3$\upsigma$) using the linear range of the calibration curves.

$$\begin{aligned} LOD = 3.3\cdot \dfrac{\sigma }{S} \end{aligned}$$

(1)

where S is the calibration curve in Fig. 3 and $\upsigma$ is the standard deviation of impedance background in the control sample.

Fig. 3

Impedance variation as a function of serial dilutions of anti-E01 IgG primary antibody at decreasing concentrations to assess the platform sensitivity in terms of LOD. Error bars show standard error. The figures show the coverage of the active sensor by beads. All measurements have been performed with a 50 mV amplitude signal at 1 MHz

Figure 3 summarizes these results. For antibody concentrations above 50 $\upmu$g/mL, no significant change was observed due to the saturation of the binding sites with anti-IgG antibodies. On the other extreme, the system demonstrates to operate down to a concentration of 100 pg/mL with signals (700  m$\Omega$) well above the noise ground limit of 150 m$\Omega$ as present at the output of the lock-in amplifier.

For the negative controls, the same protocol has been applied with zero concentration of primary antibody, resulting in a mean value of 372 $\pm$ 90 m$\Omega$, well below the signal measured at the minimum concentration of 100 pg/mL. Figure 3 also shows the photographs of the active sensor surface taken at the end of the measurement after drying the chip. They clearly exhibit the corresponding decrease in beads density on the active area of the sensors as the primary antibody concentration decreases. The platform tracks the number of beads that can be counted by impedance measurement down to a few tens of beads, certifying the immunosensor concept and operation.

2.2 Detection of IgG Antibodies in Human Serum for Dengue Virus Diagnosis

To validate the clinical effectiveness and robustness of the adopted technique, the biosensing system has been tested for Dengue Virus detection. The target biomarkers for this experiment are the IgG antibodies anti-DNGV in human sterilized serum samples of infected patients. As negative control in order to avoid false-positive results, human serum samples from healthy patients negative to IgG antibodies anti-DNGV have been tested as well. The protocol used is the standard one described in the previous section and provides in addition all the stages of incubation of the serum for probe-target hybridization inside the microfluidic chamber. Figure 4 summarizes the measurements at different dilutions of the serum sample positive to anti-Dengue antibody (from 1:25 to 1:100000). Healthy serum samples were included in the analysis as negative controls. Negative samples show a ground noise that we consider as a cut-off signal, with a differential impedance value below 1.23 $\Omega$. As a consequence, the 1:100000 serum dilution provides the minimum detectable signal. The photographs of the active sensor surface visually certify the effectiveness of the detection mechanism for few exemplary dilutions and the number of beads over the active sensor decreases with sample dilution down to about one hundred. As can be seen in the magnification from Fig. 4 also clusters of beads are formed during washing and/or PBS evaporation contributing to the single concentrations error bar.

Fig. 4

Differential Impedance sensing variation versus dilution in a buffer of human serum positive to anti-Dengue Virus antibody. On the right, photographs of the active sensors after the final wash for four exemplary cases. Error bars show standard error. All measurements have been performed with 50 mV amplitude signal at 1 MHz

As reported in [20], the detection with the DIS platform has achieved results that outperform the one obtained over the same positive and negative serum samples with a standard ELISA test.

3 Multiplex Differential Impedance Sensing

3.1 Electronic Multichannel Reading Board

A compact and portable electronic board has been designed to operate a lock-in processing on up to 7 sensors in parallel in a differential architecture (1 input of the 8 available is used for the reference sensor). As shown in Fig. 5, the analog sections of the board have been developed around an XEM7310 FPGA module by Opal Kelly. The FPGA controls in real-time the generation of the signals to stimulate the sensors, the acquisition of the responses from the sensors and their processing in parallel by implementing dual-phase digital lock-in and transmission to an external PC with a USB port. The Sensor forcing section comprises two counter-phase sinusoidal voltage generators (frequency at 1 MHz and tunable amplitude up to 1V), generated with a fast DAC (AD9706, 12-bit, 175 Mbps). They can be operated independently or can be simultaneously applied to the active electrode and to the control electrode of a physical differential sensor pair when local on-chip differential sensing is desired. The 8 transimpedance amplifiers (TIA) read the currents from 8 sensors. The TIA uses capacitive feedback to combine wide bandwidth and low noise. The output is further processed by a gain amplifier digitally selectable from 0.1 to 1000 and converted in the digital domain by 12-bit ADC (LTC2250) at a sampling rate of 100 Msps. The board also houses an additional lock-in channel for the readout of the on-chip temperature sensor. This Temperature control module on the board allows a precise measurement of the local temperature of the biosensor liquid by tracking the resistance (around 1 k$\Omega$) of a serpentine made of the same gold thin-film of the IDEs.

Fig. 5

Photograph of the electronic board for multiplex experiments. All the different sections highlighted are: generation and reading of the sensor, temperature control, power supplies, references and the FPGA module. On the right, a comparison between the commercial system and the electronic reading board in detecting a signal variation of 100 ppm and the transfer function for all the 8-channels for the sensor reading, with a zoom in the region around the working frequency of 1 MHz

To characterize the control platform realized, the performances in reading the impedance are compared with the commercial lock-in system (HF2LI, Zurich Instruments). An extensive characterization is reported in [20]. Figure 5 reports the results for the 100 ppm variation comparison and the evaluation of the eight transfer functions in order to certify the correct operation and the interchangeability of the eight reading channels.

3.2 Selective DNA Detection

The system with the electronic reading board has been challenged with a multiplexing acquisition of impedance measurement of a chip realized with multiple IDEs. For this experiment, a new chip with a multi-sensor IDE configuration has been design. The layout allows to have 7 sensors to be dedicated between reference, target DNA and negative control on the same area inside the microfluidic chamber. In this case, 2 electrodes will be used as a reference electrode, 3 for target detection (specific probe) and 2 as active control (not specific probe). In the following experiment, the COCU8 is the actual probe for the target DNA COCU10-BIO, while the OLIGOX has the function of negative control and REF as reference sensor for the differential measurement configuration approach. The protocol adopted for the experiment consists as usual in an incubation phase for target hybridization and the steps inside the microfluidic system (Beads incubation and final wash). The results are summarized in Fig. 6.

Fig. 6

Signals from the multiplexing experiment. On the top, the value of electrodes in PBS solution (Rsol $=$ 1 k$\Omega$) and on the right the device under test. On the bottom, the signals from the final counting phase. The final washing step reveals the beads bounded only over the target sensors. Picture of the electrodes after the final wash. The beads are specifically bounded only over the active sensor with the probe for the COCU8. Reference and the OLIGOX electrodes are perfectly clean

Figure 6 shows the picture of the electrodes after the final wash. It is possible to notice the optical confirmation of the beads bounded only over the active sensors functionalized with the probe for the target. These results confirm the possibility to operate in a multiplexing acquisition configuration with the electronic board as an acquisition system.

4 Conclusions and Future Perspective

The developed biosensor system is based on the impedance variation between microelectrodes upon the capture of the target analyte, grafted over the biosensor surface through specific probe immobilization. A properly functionalized nanobead has been used to enhance the electronic signal since the dimensions and the structure of the target moiety would not allow direct label-free detection. The biosensor core, made of a borosilicate chip with microelectrodes, is integrated into a microfluidic path and electronically accessed to perform impedance detection by custom electronic circuits featuring high portability and multichannel operation. In this way, multiple sensing sites in parallel can be addressed, extremely important from a diagnostic point of view since they will allow performing multiplex analysis starting from a single clinical sample. The preparation of the biosensor chip by updating the antibody and the oligonucleotide linker makes this platform concept of Impedance Sensing of durable application and very practical industrial interest, yet reaching clinically breakthrough results. The experiment results show that beads count by a truly differential sensors architecture operated in a lock-in scheme is very effective in monitoring specific IgG antibodies in human serum and buffer down to a few single counts resolution, i.e. a LOD of 88 pg/mL. The sensitivity obtained by the system reaches and possibly outperforms other methods yet operating in a simple and clear protocol as demonstrated in the comparison of the proposed platform response to human serum positive to DENV with a commercial Dengue virus IgG kit. For future work, the ongoing SARS-CoV-2 pandemic represents a starting bench for the development and implementation of such a new biosensor. SARS-CoV-2 will be targeted by direct capture of the entire virus in solution, using DNA-labelled antibodies directed against the SARS-CoV-2 spike protein. This strategy will bring advantages in terms of reduced sample handling and processing (meaning less contamination and no loss of viral components) and no need for harsh chemicals nor for sample purification or amplification, resulting in a reduction of time and cost of the analysis.