Quartz crystals (9 MHz) coated with a thin layer of chrome covered with polished gold on both sides (Maxtek Inc., Cypress, CA, USA) were used. The thickness of the quartz, chrome and gold layers was 0.2 mm, 20 nm and 50 nm, respectively, and the diameter of these layers was 12.5, 6 and 6 mm, respectively. The piezoelectric crystal was placed on a cell with two gold terminals 6 mm in diameter. The cell was covered on the upper and lower sides by two polystyrene rings. The quartz crystal microbalance (QCM) was closed with a plastic cap containing two metal pins to which tubes were attached to allow continuous flow cleaning of the crystal when a peristaltic pump was switched on (SJ 1211H, series 253512, Chromatograph Atta, Tokyo, Japan). QCM and an AUTOLAB potentiostat (Eco Chemie, Utrecht, The Netherlands) were used to form a high-resolution integrated oscillating circuit. The modular set of equipment was linked to a computer to record the measurements.

The characterization of the surface coated with gold nanoparticles was performed using an electrochemical technique involving cyclic voltammetry. Electric current measurements were performed at a constant electric potential and scan rate during the immobilization of anti-TnT.

1,6-hexanedithiol (HDT), colloidal gold (5 nm) in a hydrochloric medium, cystamine (CYS), glutaraldehyde (GLU) and glycine (GLY) were acquired from Sigma-Aldrich (St. Louis, MO, USA). The monoclonal biotinylated anti-TnT antibody (200 μg mL⁻¹) and human cardiac troponin T (10 ng mL⁻¹, lot H0505) were acquired from Roche Diagnostics (Mannheim, Germany). All other reagents used in the present study were of analytical grade purity. Phosphate buffer solutions (PBS) with different pH values and in different concentrations were prepared using adequate of Na₂HPO₄, KH₂PO₄, NaCl and KCl. In the anti-TnT concentration studies, 10 mmol L⁻¹ PBS was used, dissolving 0.24 g of Na₂HPO₄, 1.44 g of KH₂PO₄, 0.2 g of KCl, 8.0 g of NaCl in 1,000 mL of ultra-pure water, with the pH adjusted to 7.4. “Piranha” solution used for cleaning the electrode consisted of a 5:1 mixture of concentrated sulfuric acid to 30% hydrogen peroxide solution, which was diluted 1:5 in ultra-pure Milli-Q water. Due to the self-decomposition of hydrogen peroxide, this mixture was always used freshly prepared.

A pool of human serum samples was obtained from 10 healthy volunteers (five males and five females). For all serum sample collections, the blood was collected in 10-mL vacutainer serum tubes (BD, Franklin Lakes, NJ, USA), allowed to stand for 15 min and then centrifuged at 3,500 g for 25 min. The serum was transferred from the vacutainer tubes into polypropylene tubes and stored at −20 °C until analysis. This pooled serum was established as control. The amount of TnT was measured using ECLIA (Elecsys analyser 1010, Roche), which presented 0.0015 ng mL⁻¹ TnT. Serum samples with different concentrations of TnT for the evaluation of immunosensor response were obtained by adding a known amount of TnT to the pooled serum.

The gold surface of the piezoelectric crystal was cleaned twice with the “piranha” solution for 5 min and rinsed with ultra-pure distilled water. After the chemical cleaning, successive washes with distilled water followed by PBS were performed using a peristaltic pump for 10 min at a speed of 700 μL min⁻¹. The piezoelectric crystal was placed on the detection cell and sealed with an o-ring, leaving only one side of the gold electrode in contact with the solution for reactions. The surface of the crystal electrode was then incubated with a solution of 20 mmol L⁻¹ of HDT prepared in ethanol for 2 h with the flow stopped. After washing in PBS, the gold electrode was incubated for 14 h in a solution of AuNP (1 ng mL⁻¹) prepared in PBS (pH 9.0, 10 mmol L⁻¹). The nanostructured gold electrode surface was incubated with ethanol solution of 25 mM CYS (H₁₂N₂S₂·2HCl) for 2 h. Linkage of the anti-TnT was then performed through aldehyde groups by incubating with the modified surface with 2.5% GLU solution in PBS for 45 min. Finally, an anti-TnT solution (150 ng mL⁻¹) was incubated on the functionalized nanostructured gold electrode surface for 2 h. Non-specific bindings were blocked by the use of 50 mmol L⁻¹ of glycine solution for 2 h. Between all incubations step, the crystal was washed with a continuous flow of PBS for 10 min.

Cyclic voltammetry studies were carried out using an electrode system composed of an anti-TnT/AuNP/HDT/Au electrode as the working electrode, Ag/AgCl as the reference electrode and Pt wire as the auxiliary electrode. To discern the role of the individual components, cyclic voltammograms of the bare Au electrode, HDT/Au electrode, AuNP/HDT/Au and Gly/Anti-TnT/AuNP/HDT/Au electrode were recorded in 25 mmol L⁻¹ PBS (pH 7.2) containing 0.1 M K₃Fe(CN)₆^3/4− in aqueous KCl as the redox probe at a scan rate of 0.0 to +1 mV s⁻¹ at an interval of 50 mV s⁻¹.

Cyclic voltammetry is commonly used to monitor the process of biomolecule immobilization. Cyclic voltammograms were recorded using 10 mmol L⁻¹ K₃Fe(CN)₆^3/4− in 0.1 M aqueous KCl for the redox probe. To ensure a good Au-S bond, the electrode surface should be properly cleaned [26]. The gold electrode was pretreated with a “piranha” solution in order to clean the surface and form gold oxide, which facilitates the linkage with -SH groups from HDT [Equation (2)]: (2) Au + HS − ( CH 2 ) 6 − SH   →   Au − S − ( CH 2 ) 6 − SH + ½ H 2

To obtain greater effectiveness in the linkage of the thiol groups on the gold electrode, the surface was pre-polarized to 0.7 V potential vs. Ag/AgCl during 4 min. The proximity of the peak potential of oxidation and reduction exhibited by Fe(CN)₆^3−/4− redox probe of the bare gold electrode with a difference less than 0.1 V [Figure 1(A)] is an indicative of the fact that the cleaning procedure was effective. Considerable attention has been focused on dithiol films on gold electrodes [27]. Compared to the response of the bare gold electrode, a substantial reduction of the redox peaks was observed. This behavior is attributed to film deposition on the electrode surface through HDT. An increased oxidation current compared to the reduction current reinforced by a reductive desorption of these monolayers in thiol according to Equation (3) was also observed [28]: (3) RS − Au + e − → Au + RS −

Figure 1(B) displays cyclic voltammograms for a gold electrode in HDT and AuNPs solutions. The HDT-modified electrode exhibited significant oxidation and reduction currents starting at around 0.7 and −0.2 V, respectively, indicating a fast follow-up chemical reaction [27]. The AuNP-modified gold electrode exhibited a very weak electrochemical response over this potential range as well a decrease in the area of the voltammogram, suggesting the formation of a blocking film on the electrode surface, likely due to the addition of AuNPs negative charges which are linked to the QCM electrode surface by the HDT that increased conductivity on the electrostatic electrode surface [28,29].

Although a remarkable reduction wave is observed for the AuNP electrode in comparison to HDT, the AuNP electrode exhibited (a) a larger anodic current than the (b) cystamine, (c) anti-TnT and (d) glycine electrodes, indicating that the nano-gold-HDT connection was successful. The overlapping areas of the voltammograms for anti-TnT and glycine reveal that nanogold-cystamine immobilization allowed the adequate coating of the electrode surface [Figure 1(C)]. The formation of SAMs building on gold surfaces allows control of the amount and orientation of the antibodies. However, the use of planar gold as the transducing surface may lead to a limited number of accessible antibodies [30]. Gold nanoparticles have been successfully used to enhance the sensitivity of biomolecular detection assays by improving antibody immobilization, as AuNPs increase the surface contact of the electrode, which leads to a corresponding increase in the number of immobilized antibodies per unit of area on the electrode.

The effects of the anti-TnT concentration immobilized on the quartz electrode were investigated in relation to the maximum sensitivity of the immunosensor. Solutions of different concentrations of anti-TnT prepared in PBS (50, 100, 150 and 200 ng mL⁻¹) were immobilized on the nanostructured electrode and the optimal anti-TnT concentration was established through the maximal response to conjugation with the TnT antigen suspended in PBS (pH 7.4, 10 mmol L⁻¹). Figure 2 displays the regression correlation “r” values obtained from calibration curves with different concentrations of immobilized anti-TnT. Maximal equilibrium between antibody and antigen was achieved at 150 ng mL⁻¹ anti-TnT, as evident by the plateau reached, which may be attributed to an equivalence zone. Thus, the TnT concentration of 150 ng mL⁻¹ was used in all remaining experiments.

The influence of ionic strength on the immobilization of anti-TnT was investigated (Figure 3). There was an increase in frequency variation when the concentration of PBS increased from 5 mmol L⁻¹ to 25 mmol L⁻¹, whereas a concentration above 25 mmol L⁻¹ led to a reduction in frequency variation. It may be that the concentration of 25 mmol L⁻¹ during the immunoreaction promoted greater recognition of the TnT antigen by the immobilized antibody.

The piezoelectric quartz crystal was used to study the influence of pH on the response of the immunosensor. The TnT antigen was incubated at a concentration of 1 ng mL⁻¹ and suspended in 10 mmols L⁻¹ of PBS at different pH values (Figure 4).

The choice of the 6.8 to 7.6 range was made taking into account the possible changes in blood pH that could occur in patients whose condition may result in acidosis or alkalosis (reduction or increase in plasma pH). This variation in blood pH is seen in conditions in which the organism loses its capacity to balance the input of inspired oxygen with the carbon dioxide gas to be eliminated. Moreover, as the pH of the medium decreases, it approaches the isoelectric point of TnT and this fails to interact with the anti-TnT [31]. Figure 4 displays a rise in the frequency variation when the pH of the medium went from 6.8 to 7.2 and a subsequent decline in the curve until a pH value close to 7.5. The greatest frequency variation measured on the QCM was obtained in PBS solution with pH near 7.2, suggesting that this is the optimal pH value, likely due to the fact that this value is less favorable to electrostatic repulsion during immunoreactivity [24]. Thus, pH 7.2 for the PBS solution was established as optimal for the formation of the anti-TnT and TnT antigen complex.

The performance of the immunosensor was studied using human serum samples spiked with troponin. The sera were 1:10 diluted in PBS in order to obtain a final concentration of 1 ng/mL TnT. Figure 5 displays a higher slope for TnT serum samples spiked with TnT than control samples indicating a high discrimination of the cardiac TnT in the real samples.

The sensitivity of the immunosensor was assessed by a test involving the control serum pooled from the volunteers, with the addition of 0.1 ng L⁻¹ of TnT solution prepared in PBS (pH 7.2, 25 mmol L⁻¹). The measuring procedure for TnT was in flow. During the injection and incubation, the flow was stopped. The measurement of the frequency was performed in duplicate (Figure 6). Serum is a complex medium that contains innumerable biochemical substances such as proteins, enzymes, amino acids, carbohydrates, lipids and ions. Non specific binding by adsorption of these substances may lead to false positive responses by frequency changes. Then, different serum samples were used to evaluate the selectivity of the immunosensor. The immunosensor displayed a better performance in the analysis of the undiluted serum than the diluted serum (data not shown). This demonstrates the QCM nanosensor is less sensitive to matrix effects than other analytical techniques [8]. The specificity of the analysis is ensured by monoclonal anti-TnT. Table 1 displays the statistical parameters of the experiments for the detection of TnT in human serum using the nanostructured immunosensor.

The immunosensor exhibited a range from 0.1 ng mL⁻¹ to 0.5 ng mL⁻¹ (Y = 3.85518 – 63.81882X; SD = 1.019), according to the calibration curve in Figure 5. These data were processed using the ORIGIN program version 8.0 (Microsoft, Redmond, WA, USA) and produced high linearity (R = 0.989; p < 0.001; n = 5), combined with a relative error of 6.4%. The coefficient of variation was 7.0%, obtained with human serum with a detection limit of 0.001 ng mL⁻¹. The detection limit was calculated as the concentration of analyte giving a frequency shift equivalent to three standard deviations of the blank. The immunosensor exhibited detection limit values comparable with those of ECLIA methods and is also expected to have no limitations concerning sample color or matrix, which is often a problem in ELISA [32]. The nanosensor demonstrated the ability to measure the concentration of TnT in serum, which makes it a useful method for diagnosing myocardial damage in clinical practice.

The World Health Organization establishes the desirable blood level of TnT to be equal to or less than 0.01 ng mL⁻¹ for healthy individuals and beginning at 0.3 ng mL⁻¹ as the criterion for the diagnosis of myocardial infarction based on this biomarker [1,2]. The immunosensor proposed in the present study proved to be a sensitive method for measuring the concentration of TnT below the cutoff point for the diagnosis of myocardial infarction and is faster than conventional methods for the diagnosis of acute myocardial infarction.