Diabetes Mellitus (DM) encompasses a series of chronic metabolic diseases characterized by inadequate glucose metabolism1. It is quickly becoming a worldwide epidemic, involving nearly 24 million people in the United States, and costing nearly 250 billion dollars2. According to the American Diabetes Association, by the year 2034 the number of diagnosed and undiagnosed people with diabetes will increase from 23.7 million to 44.1 million3. With such an increase in prevalence, there has also been a large need for next generation technology to help manage the disease with better portability and increased sensitivity4. Currently, diabetes management involves monitoring glucose levels daily, either discretely or continuously, and glycated hemoglobin (HbA1c) levels periodically5,6.
Embodiments disclosed herein relate to a rapid and label-free insulin biosensor with high sensitivity and accuracy. In certain embodiments, an insulin biosensor prototype capable of detecting insulin in a physiological range without complex data normalization is disclosed.
Further embodiments relate to electrochemical impedance spectroscopy use to identify an optimal frequency specific to insulin detection on a gold disk electrode with insulin antibody immobilized, which can be accomplished by conjugating the primary amines of an insulin antibody to the carboxylic bond of the self-assembling monolayer on the gold surface.
Other embodiments relate to the use of imaginary impedance to detect insulin concentration and to establishment of an optimal frequency of insulin at 810.5 Hz, which is characterized by having the highest sensitivity and sufficient specificity.
These and other aspects are further described in the following figures and detailed description of certain embodiments.
All publications, patents, and patent applications mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference.
Embodiments herein relate to apparatus and methods for detecting one or more analytes in a bodily fluid sample utilizing Electrochemical Impedance Spectroscopy (EIS). In particular embodiments, an electrochemical sensor is configured to (e.g., operably configured to) provide an electrochemical impedance measurement of an analyte and includes a diabetes-related target-capturing molecule immobilized (e.g., through a chemical linker) to a working electrode.
Presently, detection and monitoring of metabolic status in DM is achieved through detection of a single molecular marker: glucose. Other molecular markers, such as insulin, are important but currently not easily measured at the point of care. The development of multi-marker detection assays is desirable; generally speaking, many studies have shown that monitoring multiple biomarkers associated with a complex disease can enhance the accuracy of disease diagnosis, prognosis, management, and treatment7-10.
Among the many diabetes-related target molecules in
The momentum for developing electrochemical insulin sensors has been increasing in the past few years16-19. The inventors recently have showed that, using the imaginary impedance of EIS, a biomarker will have an optimal binding frequency (OBF) at which the change in imaginary impedance best correlates to the change in target concentrations20. The inventors have already characterized glucose previously using EIS and have shown its feasibility in glucose detection15. Additional biomarkers can be explored to build a multi-marker sensing platform monitoring all the major biomarkers of DM, providing the most accurate information for medical intervention and glycemic control.
In some embodiments, the devices and methods include a diabetes-related target-capturing molecule, such as an antibody, an aptamer, or other molecule recognized in view of the teachings in this application by those of ordinary skill in the art as suitable for their specific applications. Aptamers are single-stranded, synthetic oligonucleotides that fold into 3-dimensional shapes capable of binding non-covalently with high affinity and specificity to a target molecule. The diabetes-related target-capturing molecules may capture targets such as insulin, glucose, or other diabetes-related molecules. Moreover, the detection of binding of such molecules may be continuous in certain sensor embodiments.
In some embodiments, multiple targets are captured and imaginary impedance measurements are taken at distinct frequencies to then determine the binding (and related concentration) of each target.
All chemical reagents were purchased from Sigma (St Louis, Mo., USA) unless stated otherwise. The 10 mM phosphate buffer saline (PBS) tablets were purchased from Calbiochem (Gibbstown, N.J., USA), potassium hexacyanoferrate (III) from EMD Chemicals (Billerica, Mass., USA), and sulfo-derivative of N-hydroxysuccinimide sodium salt (NETS) from Toronto Research Chemicals (Toronto, Ontario, Canada). The redox probe reagent used was 100 mM potassium ferricyanide dissolved in pH 7.4 PBS.
In this non-limiting example, in accordance with certain embodiments, the sensor includes 3 electrodes: working gold disk electrodes (GDEs), reference silver/silver chloride electrodes, and counter platinum electrodes acquired from CH Instruments (Austin, Tex., USA). All EIS measurements were performed at room temperature using a CHI660C Electrochemical Analyzer from CH Instrument at the electrode's formal potential from 1 Hz to 100 kHz. A Buehler felt pad with 0.05 μg grit aluminum oxide particles was used to polish the GDEs with 10 figure-eight motions, followed by a 20-minute sonication in deionized water. After electrode polishing, cyclic voltammetry (CV) from −1.0 V to 1.0 V was used to obtain the formal potential and bare electrode EIS was performed to evaluate sensor cleanliness.
Once the sensors were cleaned, the SAM (self-assembled monolayer) was created by incubating 1 mM of 16-mercaptohexadecanoic acid (MHDA) for one hour at room temperature. The sensors were then rinsed and stored dry overnight to ensure proper deposition of SAM, as SAMs takes hours to reach their final thickness and contact angles21,22. The carboxylate groups of the 16-MHDA were activated by incubating the sensor in 10 mM 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) and 80 mM sulfo-NHS for one hour at room temperature. After rinsing with DI, 100 μL of 156 μM of insulin antibody prepared in pH 7.4 PBS was incubated onto the electrode surface for one hour. After rinsing with PBS the sensors were blocked with 1% ethanolamine for 30 minutes to block any remaining active sites, completing the sensor fabrication process. The schematic of sensor preparation can be found in
Electrodes were prepared in batches of eighteen and all electrodes were analyzed using Electrochemical Impedance Spectroscopy. After measuring the post-MHDA impedance, the quality control (QC) was executed by selecting only the electrodes with similar peak frequencies and impedance magnitudes that are within 6% to 10% relative standard deviation (% RSD). Only the QC-passing sensors would then proceed with immobilization.
Once EIS was performed, the imaginary impedance values were correlated to target concentrations to calculate slope and R-square values (RSQ) across the frequency sweep. The OBF is the frequency at which the slope peaks with RSQ values above 0.85. All circuit modeling was performed using ZsimpWin software.
Using the illustrative methods described above, the impedance responses from 7 electrodes were used to determine the OBF of insulin, which was found to be 810.5 Hz (
Using ZsimpWin, one example of a benefit circuit model that well-describes the electrochemical system of insulin sensor can be obtained (
Comparing the results between
In other words, after obtaining the impedance reading from an unknown sample analyte, the number can be plugged back into the calibration curve to obtain the concentration of the unknown sample analyte. For an example involving insulin, a −7500 ohm reading at 810.5 Hz would result in 1000 pM of insulin according to
The lower limit of detection (LLD) and dynamic range are important parameters in determining the efficiency of the system. The LLD and dynamic range were calculated based off the standard deviation and slope of the system. The LLD was found to be 2.64 ρM and dynamic range from 50 pM to 1500 pM, which meets clinical needs. From a clinical standard detection of insulin, ELISA can accurately detect labeled insulin at 1.39 ρM25. This is slightly lower then what the inventors have demonstrated with this sensor prototype, but with optimization of the electrode design, the LLD may be lowered to that of ELISA. Even more so, techniques such as ELISA or high-performance liquid chromatography have labeling steps and many associated techniques that can be performed only in laboratories. EIS on the other hand, is a label free technique, and the sensor prototype can be translated into screen printed sensors, allowing the possibility of point-of-care detection with a portable device and disposable test strips similar to the setup of self-monitoring of blood glucose devices14,26.
The Food and Drug Administration requires all glucose meters to be within 20% variance from standards27. Currently, the replicated results show that across all sample concentrations the % RSDs ranges from 5% to 26%, suggesting there are still room for improvements. Although batch analysis has helped eliminate some of the variance between GDEs, polishing and reusing GDEs is a significant source of variance as surface roughness of gold can affect SAM formation28, affecting the capacitance of imperfect parallel plate capacitor (IPPC) explained in later section. Transition to screen printed sensors will reduce the variance of surface roughness under consistent manufacturing procedures and rigorous quality control.
The inventors have shown that the EIS method of using imaginary impedance can very well detect insulin in the physiological range. Within certain embodiments, even smaller concentration interval sizes may be employed (such as about 1 pM), which is equivalent to a gold standard ELISA to distinguish between even the smallest changes in concentration. In certain embodiments, the technologies described herein may be embodied within a point-of-care (POC) device. Notably, unlike other publications on insulin detection there was no modification to the insulin solution via pH17,18.
Generally, EIS is analyzed with equivalent circuit modeling. Typically, the best-fit circuit for a semi-circle looking Nyquist plot is the Randles circuit, which models the electrochemical interactions as a resistance-capacitor circuit in parallel. The electron transfer resistance can be used to derive a calibration curve linking back to input concentration25,29. However, recently some researchers have demonstrated the use of a modified Randles circuit that implements a constant phase element (CPE) to model the capacitance20,23,30. CPE is commonly referred to as either a leaky or imperfect parallel plate capacitor (IPPC).
The bottom plate is the surface of electrode and the top plate is the top of the SAM with Molecular Recognition Elements (MREs) immobilized owing to SAM's insulating property31. The MREs different shape, orientation and size alter the smoothness of SAM in various ways, constituting the IPPC. As binding occurs, the target-MRE complex further alters the capacitance of the IPPC, affecting the electron transferring properties and impedance signals, which is evident in
Since imaginary impedance correlates to capacitance24, the inventors used imaginary impedance to correlate target concentration to reflect the impedance signal generated from changes in CPE, which the inventors believe to have less noise than using the complex impedance approach and omits the trouble of circuit modeling. Owing to this nature, it's no surprise that the LLD in imaginary impedance (2.64 ρM) is lower than that of the complex impedance approach (14.46 ρM).
An insulin biosensor prototype POC device has been developed. Detection of insulin and other molecules affecting individuals with diabetes will greatly enhance the ability of individuals with diabetes to better control their own blood glucose levels. With a reproducible LLD of 2.26 ρM the example embodiment herein suggests that imaginary impedance based techniques are not only sensitive enough to detect physiological concentrations in purified solution of small proteins such as insulin but can also compete with current SOTA devices as well.
In certain embodiments, the biosensor described herein is embodied in a disposable strip that is capable of insulin detection in clinical samples. In certain embodiments, screen printed electrodes (SPEs) may be created using a MPM Accuflex Speedline screen printer. Depending on the embodiment and specific dimension of the sensor, machine overhead, and the amount of sensors fabricated, the current cost of a sensor can be as low as 1$ per sensor with order size of 45,000 sensors. In certain embodiments, the insulin sensor may be translated onto such SPEs.
Additionally, examples of a dual-marker detection sensor using the imaginary impedance of EIS would detect glucose and insulin simultaneously at their respective optimal binding frequencies (OBFs). For example, if the OBF of glucose is 31.5 Hz and insulin's is 810.5 Hz, the impedance reading at 31.5 Hz can be correlated to the glucose's concentration and the impedance reading at 810.5 Hz will be the concentration of insulin. In certain embodiments, the impedance reading correlated to glucose concentration may be the impedance reading at a frequency within a range of 25 Hz to 35 Hz, 28 Hz to 32 Hz, or 30 Hz to 32 Hz, or at frequency of about 31.5 Hz. In certain embodiments, the impedance reading correlated to insulin concentration may be the impedance reading at a frequency within a range of 775 Hz to 825 Hz, 790 Hz to 820 Hz, 800 Hz to 820 Hz, 808 Hz to 812 Hz, or 810 Hz to 811 Hz, or at a frequency of about 810.5 Hz. As those of ordinary skill in the art will appreciate in view of the teachings in this application, the impedance reading may be correlated to a concentration of a different diabetes-related analyte at a frequency other than those set forth in the nonlimiting examples herein (e.g., an analyte other than insulin or glucose at a frequency at or approximating the OBF for that particular diabetes-related analyte).
While certain embodiments have been shown and described herein, it will be obvious to those skilled in the art that such embodiments are provided by way of example only. Numerous variations, changes, and substitutions will now occur to those skilled in the art without departing from the disclosure herein. It should be understood that various alternatives to the embodiments described herein may be employed. It is intended that the following claims define the scope of the methods and structures described and their equivalents be covered thereby.
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/US19/22703 | 3/18/2019 | WO | 00 |
Number | Date | Country | |
---|---|---|---|
62644167 | Mar 2018 | US |