Madeeha Chaudhryaf,
Dong-Kwon Lim
b,
Jeon Woong Kang
c,
Zahid Yaqoobc,
Peter Socde,
Muhammad Fahad Bhopala,
Minqiang Wangg,
Raheel Qamari and
Arshad Saleem Bhatti
*ah
aCentre for Micro and Nano Devices, Department of Physics, COMSATS University Islamabad, Tarlai Kalan, Islamabad 45550, Pakistan. E-mail: asbhatti@comsats.edu.pk
bKU-KIST Graduate School of Converging Science and Technology, Korea University, Seoul, Korea
cLaser Biomedical Research Center, G. R. Harrison Spectroscopy Laboratory, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
dDepartment of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
eDepartment of Biological Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
fDepartment of Biosciences, International Islamic University, H-10, Islamabad Capital Territory, 44000 Islamabad, Pakistan
gElectronic Materials Research Laboratory, Key Laboratory of the Ministry of Education & International Center for Dielectric Research, Shaanxi Engineering Research Center of Advanced Energy Materials and Devices, Faculty of Electronic and Information Engineering, Xi'an Jiaotong University, Xi'an, 710049, China
hVirtual University of Pakistan, M.A Jinnah CampusLahorePakistan
iScience &Technology Sector, ICESCO, Rabat, Morocco
First published on 20th January 2022
Cardiovascular diseases pose a serious health risk and have a high mortality rate of 31% worldwide. Digoxin is the most commonly prescribed pharmaceutical preparation to cardiovascular patients particularly in developing countries. The effectiveness of the drug critically depends on its presence in the therapeutic range (0.8–2.0 ng mL−1) in the patient's serum. We fabricated immunoassay chips based on QD photoluminescence (QDs-ELISA) and AuNP Surface Enhanced Raman Scattering (SERS-ELISA) phenomena to detect digoxin in the therapeutic range. Digoxin levels were monitored using digoxin antibodies conjugated to QDs and AuNPs employing the sandwich immunoassay format in both the chips. The limit of detection (LOD) achieved through QDs-ELISA and SERS-ELISA was 0.5 ng mL−1 and 0.4 ng mL−1, respectively. It is demonstrated that the sensitivity of QDs-ELISA was dependent on the charge transfer mechanism from the QDs to the antibody through ionic media, which was further explored using electrochemical impedance spectroscopy. We demonstrate that QDs-ELISA was relatively easy to fabricate compared to SERS-ELISA. The current study envisages replacement of conventional methodologies with small immunoassay chips using QDs and/or SERS-based tags with fast turnaround detection time as compared to conventional ELISA.
Digoxin is one of the most commonly prescribed cardiac glycosides with a positive inotropic effect to the patients of heart failure and atrial fibrillation.7 It strengthens the weakened heart muscles by helping them in contraction for restoring normal heart rhythm. It works by inhibiting the Na+/K+ pump (active ion transporter in human myocardium) through binding to its allosteric site on the pump. This results in the accumulation of intracellular Na+ ions, excess of which inside the cell binds to the second pump which is the Na+/Ca2+ exchanger. This effect induces intercellular transportation of one extracellular Ca2+ ion and extracellular transportation of three Na+ ions in the presence of adenosine triphosphate (ATP). The increasing of Ca2+ ions inside the cell induces cardiac action potential and leads to weak heart muscles contractions.8,9
Digoxin is administered orally or injection to cardiac patients. It has a narrow therapeutic range of 0.8–2 ng mL−1.10 Since a concentration above or below this window in the serum could be fatal, continuous and careful monitoring is required for effective treatment.11,12 Chromatography and immunoassay are commonly used for digoxin detection. High performance liquid chromatography (HPLC) or liquid chromatography (LC) combined with mass spectroscopy (MS) is the most sophisticated technique for digoxin monitoring in blood,8,13 serum14 and urine.15,16 The drawbacks of chromatographic techniques are large sample consumption, complex sample preparation procedures, and purification requirements; all are time-consuming processes and limit their application in diagnostics.17 In contrast, immunoassay procedure is easily available, requires less skilled personnel, and is easily implementable in emergency toxicology situations.18
The most commonly used immunoassay techniques include the radioimmunoassay (RIA), enzyme immunoassay (EIA), fluorescence polarization immunoassay (FPIA), enzyme multiplied immunoassay technique (EMIT), and chemiluminescent microparticle immunoassay (CMIA). These techniques are based on enzyme-linked immunosorbent assay (ELISA) detection where fluorescent molecules could be enzymes, microparticles, radioactive molecules, etc.19,20 In Pakistan, digoxin level measurement is commonly done through CMIA technique, which uses architect immunoassay analyzer and only available in private laboratories or major hospitals in metropolitan cities of Pakistan.
The advancements in microelectromechanical systems (MEMS) technology combined with the miniaturized ELISA coupled with a detection module on the chip provides an efficient way for conducting study of antigen–antibody assays. The detection could be electrochemical, mechanical or optical. Optical detection method is one of the preferred techniques because of its sensitivity, robustness and simple instrumentation.21 Among various optical detection methods employed are fluorescence, absorbance, chemiluminescence and surface plasmon resonance (SPR) or surface enhanced Raman spectroscopy (SERS).21,22 A number of microfluidic devices offering on-chip drug level detection have been reported, e.g., fluorescence-based detection of theophylline,23 antithrombotic drug24 using FPIA, and vancomycin.25 Similarly, reported a bioluminescent biosensor for the detection of a wide range of pharmaceutical preparations like immunosuppressant, antiepileptic, anticancer and antiarrhythmias.26
There are different types of nanomaterials such as quantum dots, carbon nanotubes (CNTs), nanowires, Au-NPs, and nanoshells used for detection and diagnosis in microfluidic platforms. Studies have reported use of MoS2/graphene heterostructures and MoS2 nanosheets for biomedical applications.27,28 Another study reported small analyte detection on quartz crystal microbalance based immunosensor using gold nanoparticles.29 Among these, the most preferred for diagnostic applications are Au-NPs and QDs due to their small size, large surface area to volume ratio, ease of synthesis, bioconjugation and tunable optical properties.30 Optical properties of QDs include size-dependent emission spectra, 20 times brighter and 100 times more stable than traditional dyes offering high sensitivity and fluorescence stability with a broad excitation spectrum.31 On the other hand Au-NPs changes the optical behavior of organic molecules on conjugation26,32 by multiplexing through SERS tag as its spectra give information of a whole range of molecules present in the sample.30,33
Recent studies have also reported digoxin detection through electrochemical method using Au coated magnetic nanoparticles,31,34–36 optical and electrochemical aptasensors37–40 immunochromatographic strip using colloidal Au-NPs, fluoroimmunoassay,41 bioluminescent,26 amperometric,42 low surface plasmon resonance of AuNPs,43 and impedance spectroscopy.44 However, no extensive or a detailed study is available for digoxin optical and electrochemical detection through QDs so far.
In the current study, we present simple and efficient digoxin detection using photoluminescence (PL) technique and compare with SERS based technique, which can easily be employed in microfluidic chip. Small immunoassay chips of quartz slide were prepared for both PL and SERS detections employing the sandwich immunoassay technique. Quartz slides were used instead of conventional polystyrene or polypropylene based microplate ELISA reader because quartz offer less background noise or absorbance especially in UV range as compared to the aforementioned.45 The chips were first coated with digoxin antibody (Ab) followed by incubation with digoxin and then QD or Au-NP tagged Ab completing the sandwich immunoassay format. The prepared chips were characterized with the photoluminescence spectroscopy and Raman spectroscopy, respectively for QDs and Au-NPs. Finally limit of detection (LOD) was determined for both techniques for comparison of accuracy.
Au-NPs were purchased from Ted Pella, Inc. (Redding, CA). Raman reporter 4-nitrothiophenol (NTP) (technical grade, 80%) for coating the Au-NPs, sodium silicate solution (reagent grade), bovine serum albumin, BSA (≥96%), (3-aminopropyl)trimethoxy silane (APTMS) (97%), and N-(3-dimethylaminopropyl)-N-ethylcarbodiimide (EDC) (≥97%), N-hydroxyl succinimide (NHS) (98%) for antibody conjugation and digoxin dilution (1 mL in methanol) were acquired from Sigma-Aldrich. Polyclonal antibody was purchased from Abcam and phosphate buffer saline (PBS) for antibody dilution was obtained from Thermo Fisher Scientific (Gibco™, 1× pH 7.2). For carboxyl group activation on Au-NPs and quartz slide (3-triethoxysilyl) polysuccinic anhydride (TEPSA) (95%) was purchased from Gelest (Morrisville, PA). Quartz cover slips (25.4 mm dia × 0.15–0.25 mm thick) were purchased from Alfa Aesar (Lancashire, UK). Silicone based press to seal isolator wells (2.5 mm diam. × 0.5 mm depth) on quartz slide based microchip devices were also purchased from Sigma Aldrich.
![]() | ||
Fig. 1 Biofunctionalization scheme of (a) CdSe/ZnS QDs involving coating with MAA and the antibody, (b) AuNPs coating with reporter, silica shell, TEPSA and antibody. |
The reporter was added dropwise and the solution was kept stirring for 30 minutes. After 30 minutes, silane APTMS solution in ethanol (30 μL, 0.1 M) was added and left stirring for another 30 minutes. The pH 9.0–10 was adjusted using 1 M NaOH during the stirring process. Au-NPs were coated with a protective silica layer, which prevented aggregation, process of photobleaching and improved their water solubility. Silane was introduced to make NPs surface vitreophilic to make it reactable with silica. APTMS coating makes vitreophilic; receptive to silica groups followed by binding to trimethoxy silane group which is exposed to the environment.47,48 After silane coating, sodium silicate solution was added to make a protective silica shell around the Au-NPs surface. Three different volumes of sodium silicate (300 μL, 500 μL, 1 mL; 0.60 wt%) solution were added dropwise while continuously stirring, followed by stirring overnight. 5 mL anhydrous ethanol was added to the solution 24 hours later to obtain a uniform silica layer and the solution was left standing for another day. The purification of SERS tag modified Au-NP was done using anhydrous ethanol and 1× PBS, 3–4× at 6000 rpm for 10 minutes. The SERS modified Au-NP were dispersed in 1 mL 1× PBS for future use.
Immunoassay chips or devices for drug level monitoring were prepared by fixing silicone-based isolator with holes (wells) on the cleaned quartz slide. Two sets of wells were prepared, where each well of both sets was then used for different dilutions of digoxin as shown in (Fig. 2a and b). As shown in Fig. 2a, the first set was prepared for SERS detection using Au-NPs, where each well was first coated with carboxyl group for Ab conjugation by putting TEPSA buffer solution (10 μL, 100 mM). In the second set wells were prepared for QDs, carboxyl group coating was done by the adding MAA (10 μL, 0.1 M). Both sets of wells were left overnight to complete the conjugation reaction.
The excess TEPSA and MAA were removed from both sets of wells by washing with ethanol and PBS 1×. In the next step activation of the coated carboxyl groups in each well was done using EDC/NHS (10 μL, 0.1 M). 2 hours later PBS buffer washing step was performed to remove excess EDC/NHS from both sets of wells. The antibody conjugation with carboxyl group was performed by overnight incubation of all wells (for QDs and SERS Au-NPs) with Ab (10 μL, 1:
10 dilution in PBS). After antibody conjugation, all the wells were washed with PBS 2–3×. Then 2 μL of BSA (0.1%) was added in each well to prevent non-specific binding of Ab to the wells and left for an hour. This was followed by washing with PBS solution and incubation with various dilutions of digoxin in each well and left for an hour. After an hour excess digoxin was removed by washing with PBS, and finally the addition of prepared Ab conjugated SERS tags (Au-NPs) and QDs in wells. The scheme of preparation of immunoassay sandwich ELISA chips with various dilutions of digoxin is shown in (Fig. 2a and b) as explained above.
The photoluminescence (PL) spectroscopy of the bare, antibody and drug conjugated QDs was performed by Dongwoo Optoron LabRAM spectrometer. The excitation was done with a He–Cd laser emitting at 325 nm with a power of 50 mW. All spectra were recorded at room temperature and corrected for the detector response. The PL spectra were fitted with a Gaussian function to determine the peak positions and integrated intensities at each step of conjugation and for each dilution.
SERS measurements were performed through a custom-built inverted confocal Raman microscope. The excitation source was a Ti:Sapphire (3900S, Spectra-Physics, Santa Clara, CA) laser with 785 nm wavelength. A 60× water immersion objective lens with 1.2 NA (UPLSAPO60XWIR 60×/1.20, Olympus, Shinjuku-ku, Tokyo) was used for laser light focusing and collection of back scattered light from the sample. The back scattered light was collected by a multi-mode fiber (M14L01, Thorlabs) and then directed to the spectrograph (Holospec f/1.8i, Kaiser Optical Systems, Ann Arbor, MI) and finally to thermoelectrically cooled CCD detector (PIXIS: 100BR_eXcelon, Princeton Instruments, USA).
Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) measurements were performed in a three-electrode system in the presence of 0.1 M potassium ferricyanide electrolyte solution. Measurements were performed using Electrochemical Analyzer (Autolab, Metrohm 85695, Netherlands) and analyzed with Nova 1.1 software. In a typical CV graph, during the positive bias, i.e., cathodic cycle, analyte deposited on the electrode loses electrons to the ionic solution and Fe3+ in the electrolyte solution reduces to Fe+2 and during the reverse voltage scan, i.e., the anodic cycle, Fe+2 loses electrons to the analyte deposited on the electrode and gains back its oxidized state of Fe+3. In case of bare QDs, the electron transfer would be difficult, however in case of QDs modified with MAA and Ab efficient electron transfer mechanism would be observed due to the presence of electron donating (COO−) species. Anodic peak represents the oxidation potential (Epa) and current (Ipa) and cathodic peak, reduction potential (Epc) and current (Ipc), respectively. Similarly, EIS measurements were performed for the same aforementioned setup at 0.1 V applied across the working electrode. The Nyquist plot between the imaginary and real impedance was obtained in the frequency range of 100 kHz to 100 mHz. The Randle circuit was drawn after fitting the Nyquist plots using Nova software.
![]() | ||
Fig. 3 TEM images of (a) bare and (b) MAA coated QDs; (c) bare and (d) reporter and sodium silicate coated Au-NPs. |
Fig. S2† shows the UV-Vis spectra of bare (red color) and 4-NTP reporter coated Au-NPs (black color) with three different reporter concentrations (a: 100 μL, b: 250 μL and c: 500 μL; 10 μM). It was observed that within 30 minutes, the color of solution B and C turned purple as shown in the inset images of Fig. S2b and c,† which was attributed to the aggregation of Au-NPs. It was noted on the other hand that the color of solution A did not change during the same interval of time. It was thus considered that sample a with 100 μL of reporter was the optimal for coating of Au-NPs and subsequent experiments.
In the second step, three different volumes of sodium silicate d (300 μL), e (500 μL) and f (1000 μL) were coated on reporter (100 μl) coated Au-NPs. Fig. S2d–f† shows the UV-Vis spectra of three quantities of sodium silicate coated Au-NPs (solution d: Au-NPs coated with 300 μL; solution e: Au-NPs coated with 500 μL; solution f: Au-NPs coated with 1000 μL). A blue shift of 8 nm in the absorption peak of silica coated Au-NPs from 538 nm to 530 nm was observed. It was again observed that the sample solutions e and f turned purple after 30 minutes (as shown in insets of (e) and (f)) due to aggregation of silica coated Au-NPs. The absorption spectrum shown in (Fig. S2d†) demonstrated that the sample d (silica-reporter coated Au-NPs with 300 μL silica) was the best concentration of sodium silicate where no particle aggregation was observed.
The three spectra in both cases were from (1) Ab (primary Ab) immobilized on slides, (2) digoxin attached to primary Ab, and (3) fully fabricated (QDs and SERS based) ELISA. Fig. 5a and b shows spectra with no significant signal for both QDs (PL) and SERS NPs (Raman) after the addition of digoxin Ab in the wells (a and b (1) on quartz slides). The second part in both figures shows that again no significant signal was observed after addition of digoxin drug on the digoxin Ab attached wells. However, a significant PL and Raman response (a and b (3)) was observed on addition of Ab-conjugated QDs and SERS NPs in each well employing complete sandwich format.
The blue shift was attributed to the passivation of surface defects present in bare CdSe/ZnS QDs. The loss in the PL intensity was attributed to the transfer of photoexcited carriers, i.e., electrons and holes to the HOMO and LUMO levels of conjugated MAA. The lowest unoccupied molecular orbitals (LUMO) of MAA, which acts as an electron trapping center, lie below the conduction band edge (CBE) and the highest occupied molecular orbitals (HOMO) of MAA, which acts as a hole acceptor lie above the valence band edge (VBE) of CdSe/ZnS QDs. The excited electrons from the conduction band of CdSe/ZnS were transferred to the LUMO levels of MAA and resulted in the quenching of PL intensity of QDs.49
Fig. 4b shows PL spectra of conjugation of MAA–Ab–digoxin with Ab coated QDs (Ab–QDs) detecting different dilutions of digoxin in sandwich assay format. We could clearly see a decrease in the PL intensity as the concentration of digoxin dug was decreased. In case of 2 ng dilution there is 30–40% decline in intensity as compared to 0.1 and 0.01 μg dilution. While, 80–90% decrease in PL intensity was observed for 1.0, 0.8 and 0.1 ng dilution as compared to 0.1 and 0.01 μg dilutions.
![]() | ||
Fig. 5 Plots of anodic and cathodic currents as a function of scan rate for (a) bare GC electrode (b) GCE modified with QDs–MAA–Ab. |
The rate of electron transfer and number of electrons transferred during oxidation/reduction reactions were determined: fast or slow/sluggish. In the current scenario there was a linear relationship between the reduction potential (Ep) and the scan rate, which was attribution of irreversible behavior50–52 and was described using the eqn (1):
![]() | (1) |
![]() | (2) |
The electrochemical impedance spectroscopy (EIS) was performed to determine the solution resistance and charge transfer resistance of the device and (Fig. 6) shows the Nyquist plots of the bare QDs, QDs–MAA and QDs–MAA–Ab modified GC electrodes. The plots consist of a semi-circle and a straight line. The semi-circle represents the charge transfer resistance (Rct), which varied with the configuration of electrode. The slope of the linear part represented the solution resistance (Rs).
![]() | ||
Fig. 6 EIS spectra of the bare electrode (black squares), modified with QDs (red circles), QD–MAA (blue triangles) and QD–MAA–Ab (green inverted triangles). |
The Nyquist's plots were used to draw equivalent circuit of the device as shown in the inset of Fig. 6. The equivalent circuit shows the formation of a double layer capacitance (Cdl) layer, which was indication of solution adsorption on the electrode surface, Warburg impedance representing the diffusion resistance or mass transfer. There was an increase in the diameter of the semicircle for electrodes modified with QD with higher values of Rct. Also, double layer capacitance between the electrode and electrolyte solution was smaller for QDs and similarly smaller value for mass transfer. This signified an adequate transfer of electron/charge between dots on electrode and electrolyte ionic species, which varied with electrode modification and affected by the modification processes.
As shown in Fig. 7, there is decrease in PL of dots after conjugation with MAA and then Ab, due to the interaction of excited electron with the HOMO LUMO levels of MAA as explained in SEection 3.2.1. Comparing the CV spectra of the same materials, there is increase in the flow of current on conjugation with MAA and then with Ab. This is due to the availability of electrons on the surface of MAA and Ab, followed by the interaction of these electrons with the electrolytic solution as explained in the explanation of CV in Section 2.6 (Characterization).
However, the Raman spectra (2–4 in Fig. 8a) collected from the different concentrations of reporter tagged Au-NPs showed a slight shift in Raman peak positions with an increase in the intensity and width of all peaks, which was independent of the reporter concentration. The above-mentioned Raman modes appeared at 1566 ± 1, 1323 ± 1, 1106 ± 1 and 1075 ± 1 cm−1 (2–4 in Fig. 8a). The increase in width was due to increase in size/diameter of Au-NP after coating with the reporter molecule.51 Comparison of Raman spectra of pure reporter with the reporter coated Au-NPs in different reporter concentrations (Fig. 8a(2–4)) showed variations in the intensities of C–H bending modes at 1100–1110 cm−1 and C–S stretching modes at 1075–1080 cm−1. In case of Raman reporter, intensity of C–H peak was more prominent as compared to C–S peak showing a decreased intensity. After coating of Au-NPs with the reporter (Fig. 8a(1–3)), signal from C–S vibrations of reporter was 3–4× enhanced and the intensity of the C–H bending modes (1100–10 cm−1) was decreased. Similar findings were observed for other reporter concentrations coated over Au NP (Fig. 8a(2–4)).
Comparison of Raman signals from all the three samples after reporter coating showed that the signal from sample B (250 μL) was enhanced as compared to sample A (100 μL) and C (500 μL). But considering the shelf life of samples, nanoparticle aggregation was observed in case of samples B and C. The images in the inset of (Fig. S2a–c†) clearly showed particle aggregation in solutions B and C, where their color turned purple while the color of solution A did not change. Thus, it was inferred that there was no or very little particle aggregation in the solution A. The enhanced signal intensity from sample B as compared to A and C could be from the aggregated Au-NPs.
Fig. 8b shows Raman spectra of reporter (1), reporter coated Au-NPs (2) and finally Ab conjugation of reporter coated Au-NPs (3), which showed more or less similar trend as observed in Fig. 8a. There was a shift in Raman reporter peak positions along with variation in Raman peak intensity and width. Fig. 8c shows the SERS plots of Au-NPs conjugated with different concentrations of digoxin Ab (1–6) and were used to determine the limit of detection of the device. All Raman modes were distinctively observed. The sensitivity was clearly observed as Raman signal was still quite appreciable even for 0.1 ng of digoxin.
![]() | ||
Fig. 9 Linear decreasing trend in PL and SERS signal with decreasing concentration of drug with LOD 0.5 ng mL−1 for QDs and 0.4 ng mL−1 for SERS. |
Previous studies have reported varied LODs achieved using optical, electrochemical and aptasensor based digoxin detection biosensors. Using AuNPs based surface plasmon resonance. Nikfarjam et al., 2017, reported LOD of 2 ng mL−1, while Ahmadi et al. reported LOD of 0.05 ng mL−1 using electrochemical biosensor. While LOD achieved using various optical and electrochemical aptasensors were 0.01 mg L−1,37 0.05 pM,38 1.85 ng mL−1 (ref. 39) and 1 pM.40
Comparing both the detection strategies, QDs based detection would be preferable as compared to SERS based. There are a lot of limitations associated with SERS based technique as the Raman intensity depends on the distance between the SERS substrate and detection molecule. Lengthy procedures for SERS tag synthesis and biofunctionalization. There is also an issue of background fluorescence while detecting any biomolecule using SERS spectroscopy. Most importantly, the equipment related to SERS based detection is costly involving Raman spectrometer, laser, optics, specific SERS detector.55 While the detection methodology associated with QDs is comparatively cost effective. The detection equipment involves laser, spectrometer that can be installed easily. QDs based detection requires easy biofunctionalization steps and small amount of excitation energy irrespective of the size of the quantum dots.56,57
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1ra07680a |
This journal is © The Royal Society of Chemistry 2022 |