Mouse anti-cortisol monoclonal antibody (association constant K_A = 1.7 × 10⁹ M⁻¹, cortisol antibody) was purchased from Abcam PLC (UK). As the cortisol analog, hydrocortisone 3-(O-carboxymethyl) oxime (hydrocortisone 3-CMO) was purchased from Sigma Aldrich (USA). PEG6-COOH aromatic dialkanethiol (PEG6-COOH) was purchased from SensoPath Technologies, Inc. (USA), and was used as a reagent to form SAMs on the chip surface. The Au sensor chip included in the SIA Kit, the N-ethyl-N-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC) included in an amine coupling kit, the N-hydroxysuccinimide (NHS) for sensor surface fabrication, a borate buffer (pH 8.5, 10 mM disodium tetraborate, 1 M NaCl), and 50 mM NaOH were purchased from GE Healthcare Bio-Science AB (Sweden). Ethylenediamine was purchased from Wako Pure Chemical Industries, Ltd. (Japan). An ELISA kit was used as a conventional salivary cortisol detection method, and Salimetrics oral swabs, used as a salivary collection material were purchased from Salimetrics LLC (USA). Water purified using a Milli-Q integral water purification system was used as the solvent. All aqueous solutions were prepared with Milli-Q water obtained from a Milli-Q system (Millipore Corporation, USA).

As a preliminary cleaning step, an Au sensor chip covered with a 50-nm-thick layer of unmodified gold was first ultrasonically cleaned in acetone for 10 min, ethanol for 2 min, and 2-propanol for 2 min. Subsequently, the sensor surface was cleaned in standard clean solution (a mixture of ammonia solution, hydrogen peroxide solution, and pure water with a ratio of 1:1:5) heated to 90°C for 20 min. Modification of the sensor chip surface was performed as follows (Figure 1): the sensing surface of the Au sensor chip was immersed in 1 mM PEG6-COOH (Lahiri et al., 1999) in ethanol for 24 h to form SAMs on the sensor surface. After cleaning with ethanol by ultrasonication, a mixture solution consisting of 0.2 M EDC in water and 50 mM NHS in water was added dropwise onto the sensor surface and incubated for 30 min to activate the terminal carboxyl group. Next, ethylenediamine in the pH 8.5 borate buffer was added dropwise onto the sensor surface and incubated for 30 min to induce amine coupling by covering the terminal carboxyl group of the SAM into an amino group. Simultaneously, 10 mM hydrocortisone 3-CMO in DMF with a carboxyl group as a cortisol analog, 0.4 M EDC in water, and 0.1 M NHS in DMF were mixed with a volume ratio of 1:1:1 and incubated for 60 min to activate the carboxyl group of the cortisol analog. The mixed solution was added dropwise onto the sensor surface and incubated for 60 min. Finally, the cortisol analog was bound to ethylenediamine immobilized on the film surface as a result of amine coupling.

Cortisol detection was performed by SPR using a Biacore J system (GE Healthcare, Japan). An indirect competitive assay was applied for cortisol measurements. 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffered saline (HBS, 150 mM NaCl, 0.05% Tween 20, pH 7.4) was used as the running buffer.

When the antibody solution is injected onto the chip surface, the antibodies bind to the cortisol analog immobilized onto the chip surface, resulting in an increased sensor response. The time from injection to reach the sensor surface is within a few seconds. The increase in the sensor response (Δθ₀) represents the number of antibodies bound to the chip surface. When a mixture of antibodies and cortisol is injected onto the chip surface, the antibodies that have already bound to the cortisol do not bind to the cortisol analog on the chip surface. Namely, the cortisol in the solution inhibits the antibodies from binding to the chip surface. Therefore, the concentration of antibodies bound to the chip surface decreases from that when only the antibodies are injected (Δθ₁). It is considered that Δθ₁ decreases as the percentage of cortisol in the mixture increases. Here, the relative change in the concentration of antibodies bound to the chip surface, i.e., the bound percentage, is calculated by (1)Δθ1∕Δθ0×100. The smaller the bound percentage, the larger the concentration of antibodies bound to cortisol. Antibodies and cortisol, each with a concentration of twice the final concentration, were prepared for the measurement.

A mixture of the antibody solution and HBS buffer was injected onto the chip surface for 5 min, followed by the injection of a regeneration solution for 3 min at a flow rate of 15 μl/min to regenerate the chip surface. Subsequently, the antibody and cortisol solutions were mixed with a ratio of 1:1 and injected onto the chip surface for 5 min at a flow rate of 15 μl/min. Next, to regenerate the sensor surface, 50 mM NaOH was allowed to flow over the sensor surface for 3 min at a flow rate of 15 μl/min.

Antibody solutions with different concentrations (0.1–2.0 ppm) were injected onto the three sensor chips with the cortisol analog immobilized on the surface for 5 min at 30 μL/min. After the injection, the sensor chip was incubated for 5 min to allow spontaneous dissociation. The obtained sensorgram was used to calculate the reaction rate constants. BIAevaluation version 4.1 software (GE Healthcare Bio-Science) was used for analysis as well as the “bivalent with mass transfer” fitting model, which assumes that antibodies undergo bivalent binding and that the association of antibodies occurs under mass transport limitation. Five sensorgrams with six concentrations of the antibody solution were simultaneously analyzed by the fitting program to calculate the association rate constant K_a and the dissociation rate constant K_d. Using these constants, the association constant K_A (=K_a/K_d) was calculated.

To construct a calibration curve, the sensor response was measured three times for each cortisol solution with concentrations of 0.01–100 ppb (1 ppb = 1 μg/l = 2.8 nM) following the above procedure. A bound percentage equal to a value three times the standard deviation was adopted as the detection limit. BIAevaluation version 4.1 software (GE Healthcare Bio-Science) was used for calculating the calibration curves.

To collect human saliva, four healthy young Japanese adults (25 ± 1.2 years) were enrolled in this study. The study protocol was approved by the Ethical Committee of Kyushu University. The subjects were not allowed to take any food or drinks, except for water, for 1 h prior to the collection. To clean the oral cavity, each subject rinsed their mouth. To collect whole saliva, the subjects inserted a salimetrics oral swab into the oral cavity under the tongue for about 3 min. Saliva was collected at 10:00 a.m., 2:00, and 6:00 p.m. The saliva was separated from the oral swab by centrifugalization at 3,000 rpm for 15 min. Then, the saliva samples were stored at −20°C until analysis.

In order to confirm that the hydrocortisone 3-(O-carboxymethyl) oxime (hydrocortisone 3-CMO) can be applied to this indirect competitive assay as a cortisol analog, the strength of association constants (K_A) of the modified sensor chip was analyzed. From the calculation result, K_A was 3.07 × 10⁷ M⁻¹ and lower than K_A between cortisol and the cortisol antibody (K_A = 1.7 × 10⁹ M⁻¹). It indicated that the cortisol analog-modified sensor chip can be applied to the indirect competitive assay.

A calibration curve for the cortisol measurement was obtained using a standard cortisol solution consisting of cortisol with various concentrations, HBS, and 2.5% ethanol (n = 3). Figure 2 shows an SPR sensorgram obtained by this measurement. When a mixture of 600 ppb antibody and HBS (final concentration, 300 ppb) was injected onto the chip, the change in the sensor response (Δθ₁) was 207.4 ± 4.7 RU (n = 18). Here, 1000 RU corresponds to approximately 1 ng/mm², and the amount of immobilized antibody was 207.4 pg/mm². Dissociation of the bound antibody from the sensing surface is required for the next immunocycle. These results indicate that the regeneration conditions are suitable, because the sensor responses (Δθ₁) were nearly unchanged.

Before injection onto the sensor surface, each mixture of the antibody and cortisol solutions was incubated for 2 h. A mixture solution containing 600 ppb antibody and the each standard cortisol solution was applied for 5 min after incubation for 2 h or without incubation before injection. Figure 3 shows the obtained calibration curves. Even without incubation, a calibration curve similar to that for the case of 2 h incubation was obtained. The detection limits of the calibration curves in the case of 2 h incubation and no incubation were 72.8 and 38.0 ppt, respectively. Alderling et al. reported that the cortisol level in human saliva ranges from approximately 2–20 nM (0.7–7.1 ppb) with circadian variation. Thus, this cortisol immuno-assay method does not need sample incubation, antibody, and cortisol solution, before sample injection to monitor human cortisol levels. Here, Δθ₁ for cortisol quantification was determined approximately 8 min after sample injection and rapid compared with other reports. Moreover, it can be used to analyze the cortisol concentrations in blood plasma and urine, which are approximately 10 times higher than that in saliva, upon controlling the sensitivity.

To evaluate the accuracy of the cortisol biosensor for detecting cortisol in human saliva, we carried out a comparison with a conventional cortisol ELISA kit. The calibration curve for the cortisol measurement was applied using standard cortisol solution included in a commercialized ELISA kit (n = 2). Figure 4 shows the calculated cortisol values using the sensor and ELISA applied to cortisol standard solution (0.12–3 ppb). The values of determination (R²) of the sensor and ELISA were 0.98 and 0.99, respectively.

In the ELISA assay, after the human saliva samples were thawed at room temperature, they were centrifuged at 3,000 rpm for 15 min and the supernatant sample was used in the ELISA assay. The measurement was performed following the manufacturer’s manual. The cortisol concentration ranged between 1.3 and 4.1 ppb (3.7 and 11.2 nM). These results were shown to be in reasonable agreement with those previously reported (Kirschbaum and Hellhammer, 2007; Hellhammer et al., 2009). In the SPR assay, after the human saliva samples were thawed at room temperature, they were centrifuged at 13,000 rpm using a Microcon centrifugal filter unit membrane (molecular weight cutoff 30 kDa, Millipore Corporation, USA) and mixed with HBS with a volume ratio of 1:3 to remove proteins. The SPR assay was carried out following the above procedure. Here, the cross reactivity of the anti-cortisol antibody to steroid hormones in saliva such as progesterone, testosterone, estradiol, and estriol was lower than 0.1% according to the data sheet of the manufacturers.

Figure 5 shows the results of the comparative experiment using human saliva samples. Linear regression analysis was used to determine the relationship between the measurement values obtained by the biosensor and the ELISA. The correlation coefficient was 0.96 and the regression equation relative the values of cortisol obtained using the SPR (Cortisol_SPR) and ELISA (Cortisol_ELISA) was Cortisol_ELISA = 0.96 × Cortisol_SPR + 0.339. The values of slope ±95% confidence interval (CI) and intercept ±95% were 0.96 ± 0.239 and 0.339 ± 0.572, respectively. Thus, it is indicated that the cortisol immuno-assay has good correlation and is capable of performing with similar accuracy to the conventional ELISA. It is necessary to develop the technology for improving sample preparation method in a short time.