CTP (purity ≥ 99.5%), CTP disulfide (purity ≥ 99.5%), and 5,5’-dithiobis-(2-nitrobenzoic acid) (DTNB or Ellman’s reagent) (purity ≥ 98.5%), were purchased from Sigma (MO, USA). CTP tablets (25 mg per tablet) were obtained from a drugstore in Thailand. Other ingredients in the tablets included microcrystalline cellulose, corn starch, anhydrous lactose, colloidal silicon dioxide and talc. An iPhone 11 Pro and white LED lamps were used as the camera and a light source, respectively. UV-vis spectrophotometer (Cary 60 UV-Vis Spectrophotometer, Agilent Technologies, Germany) was used for the optimization study of the colorimetric reaction and the comparison of the assay results with the proposed method. The Agilent 1220 Infinity LC System (Agilent Technologies, Germany) was used to perform the standard assay of CTP tablets in comparison with the proposed SDIC method.

Standard solutions of CTP were prepared in the range of 2.5–25 µM using 3 mM (1 g/L) edetate disodium solution as a diluent. Sample solution was prepared by dissolving the equivalent to 25 mg of CTP from a portion of powdered tablets in 20 mL of 3 mM edetate disodium solution in a 25 mL volumetric flask and sonicating for 15 min. The solution was then made up to volume with 3 mM edetate disodium solution and filtered through a 0.45 µm membrane filter. The resulting solution was further diluted with 3 mM edetate disodium solution to obtain the solution with the CTP concentration of about 20 µM. DTNB solution was prepared by dissolving appropriate amounts of the chemicals in 0.1 M sodium phosphate buffer, pH 8.0.

The colorimetric reaction was conducted by adding 350 µL of 50 µM DTNB solution to 350 µL of the standard solution (0, 2.5, 5, 10, 15, 20, 25 µM CTP) or sample solution (about 20 µM CTP) in 1.5 mL microcentrifuge tubes. The mixture was vortexed for 10 s and incubated at room temperature for 10 min.

Subsequently, the intensities of the yellow color of the solutions were measured as RGB values using transmission mode, i.e., samples were place between a light source and camera. For this purpose, two 350-µL aliquots of the solution in each tube were pipetted and transferred into two polystyrene cuvettes with a 1 cm pathlength. To double the pathlength, both cuvettes were placed side by side in front of the light source, consisting of a light-emitting diode (LED) array illuminating white light and a white acrylic sheet acting as a light-diffusing background (Fig. 1a, b). Under an ambient lighting condition, a set of cuvettes containing the reaction solutions of standards and samples aligned in two rows were photographed in the same shot using an iPhone 11 Pro camera set on the autofocus mode with a 12 MP picture quality, f/1.8 aperture, and flash off. The distance between the smartphone camera and the cuvettes was 25 cm.

Figure 1. 

Alignment of cuvettes for photography from top view (a) and front view (b), and measurement of RGB colors on the digital image using the RGB Color Detector application (c).

From the digital image, the RGB values of the yellow color of the solutions were measured using a free mobile application, namely the RGB Color Detector, downloaded from the App Store (iOS) or Play Store (Android) (Fig. 1c). A standard curve between the concentration of CTP and the analytical signal, i.e., [B/(R+G+B)]_blank – [B/(R+G+B)] was plotted. The concentrations of analytes were calculated using a linear regression equation.

To assess the analytical performance, the method was validated according to the <1225> Validation of Compendial Procedures, described in the United States Pharmacopeia (USP) 43. The linearity was studied by constructing a standard curve of CTP over the concentration range of 2.5 to 25 µM and determining a linear regression equation and a coefficient of determination (r²). From the equation, the limit of detection (LOD) and limit of quantitation (LOQ), defined as 3.3 and 10 times of standard deviation of the Y-intercept divided by the slope, respectively were calculated. The accuracy was evaluated by spiking known amounts of standard CTP at three concentration levels into the tablet placebo (n = 3 for each level). The spiked drugs were then assayed, and the results were reported as the percentage of recovery. To determine the intra-day and inter-day precision, the assays of commercial tablets were performed within a day and on three consecutive days (n = 6), respectively. The results were reported as % relative standard deviation (% RSD) of the % labeled amount. The specificity was confirmed by the assay of the known amount of CTP in the presence of excipients used in tablet formulation and an impurity, namely, CTP disulfide.

The USP assay of CTP tablets based on HPLC (USP43-NF38 2020) was used to compare the analysis results with those obtained from the SDIC method. Briefly, the standard solution and sample solution were prepared at 1 mg/mL (4.6 mM) of CTP. The analysis was performed on a reversed-phase column (Cosmosil 5C18-MS 4.6 mm ID × 25 cm; 5 µm, Nacalai Tesque, Inc., Nakagyo-ku, Kyoto, Japan), and an isocratic elution using a mobile phase composed of methanol, water, and phosphoric acid (45:55:0.05 by volume) at a flow rate of 1.0 mL/min. The UV detector was set at 220 nm. The percentage of the labeled amount of CTP in the portion of tablet was calculated using the peak areas of the sample solution and the standard solution. The chromatograms are shown in Fig. 1 of the Supplementary Data.

Optimization of DTNB concentration and reaction time

In the assay, the CTP containing a thiol group reacted with DTNB, to form a yellow-colored product, namely, 2-nitro-5-thiobenzoic acid (TNB-) (Fig. 2). To achieve the reaction condition in which the intense yellow color was rapidly produced, the optimal concentration of DTNB and the reaction time were investigated. At pH 8, which is used in a standard protocol for Ellman’s reaction, the optimal condition was investigated by fixing the concentration of CTP at 25 µM (i.e., the maximum drug concentration of the standard curve), and varying the concentration of the DTNB reagent (30, 37.5, 50 µM). For each reaction, the absorbance values at 412 nm were monitored over a period of 20 min. As shown in Fig. 3, the reaction in which 50 µM of DTNB solution was used, produced the highest absorbance when compared to the others. In addition, at this DTNB concentration, the yellow-colored product formed the fastest, giving the maximum absorbance in 10 min. Therefore, 50 µM of the DTNB with a 10-min reaction time was used in the assay procedure.

Figure 2. 

Reaction of CTP with DTNB.

Figure 3. 

Effect of the DTNB concentration on the reaction time course.

Enhancement of pathlength by using dual cuvettes

In UV-vis spectrophotometry, increasing the cuvette pathlength would increase the absorbance, which could aid in more accurate quantification, especially for dilute samples. However, long-pathlength (>1 cm) measurement is not practical in many laboratories since typical UV-vis spectrophotometers are equipped with a cell holder designed for the most commonly used 1 cm cuvette. While this issue leads to additional requirements of special cuvettes and instrumentation, it is not a problem or limitation for SDIC. This is because measuring color intensity of a solution in double pathlength can be done effortlessly by photographing a pair of cuvettes which are filled with the same sample and then placed next to each other. In this work, experiments were therefore conducted to demonstrate the feasibility and advantages of this technique by comparing the use of double cuvettes (two 1 cm pathlength cuvettes) with a single cuvette (one 1 cm pathlength cuvette).

As shown in Fig. 4, the standard curve of the double cuvette-based measurement had twice the slope and the higher r² than that of a single cuvette, reflecting the increased sensitivity and linearity. It was likely that when the pathlength increased, the color captured in the image was more intense. This resulted in a more accurate RGB measurement, and finally, the lowered scattering of data points around a regression line. In another experiment, the SD values which represented the variation of the analytical signals were examined by analyzing the replicate samples having three different CTP concentrations (n = 10). The results depicted in Fig. 5 revealed that the color measurement of dilute samples with a pale color encountered a problem related to the high variation of the signals obtained. For example, the greatest SD value was found when 2.5 µM of the CTP solutions were measured, especially in a single cuvette. When the CTP concentration increased, the effect was less pronounced. Interestingly, the color measurements done by using double cuvettes could reduce the variation in signals, due to the acquisition of more intense color in the image. This finding suggested that the use of double cuvettes offered the advantage in giving more precise analytical signals.

Figure 4. 

Standard curves obtained from signal acquisition using a single cuvette and double cuvettes.

Figure 5. 

Standard deviations (SD) of B/(R+G+B) signals measured using single cuvette versus double cuvettes at different CTP concentrations (n = 10).

In addition to being proven by the higher slope of the standard curve, the increment of sensitivity due to using double cuvettes was shown by the analysis’s LOD and LOQ. Compared to that of a single cuvette, the standard curve based on double cuvettes had a higher slope and lower SD of the Y-intercept. Consequently, their LOD and LOQ as calculated by the ratio of the SD of the Y-intercept to the slope, were lower. For double cuvettes, the LOD and LOQ were 0.70 and 2.13 µM, respectively. For the single cuvette, the LOD and LOQ were 2.74 and 8.30 µM, respectively. The lower values of LOD and LOQ confirmed that using double cuvettes enhanced sensitivity. From these results, it can be concluded that increasing the pathlength by using double cuvettes in the measurement of color intensity helped improve sensitivity, linearity, and precision for the SDIC, especially in the assay where the color intensity of the sample is low or the sample solution is dilute. To the best of our knowledge, this simple but effective technique to improve the analytical performance of the SDIC has never been reported in the literature.

While a longer pathlength can be helpful in obtaining more accurate analytical signals of dilute sample, there are also limitations. Too many stacks of cuvettes and/or exceedingly long rows might produce an image with an overly wide angle of the relevant objects. In other words, the cuvettes positioned far away from the center might be obscured by the walls of the others. This, in turn, rendered the RGB measurement difficult and unreliable. Therefore, two rows with a maximum of 10 cuvettes aligned in each row were used in this study.

As summarized in Table 2, the proposed SDIC assay performed by the sample measurement in double cuvettes showed good linearity and sensitivity over a low concentrations range of CTP. Compared with the USP assay in which the sample solution for HPLC analysis was prepared at 4.6 mM, in the SDIC method the target concentration for sample preparation in the SDIC method was as low as 20 µM. The proposed method was accurate (% recoveries within 98–102% at all three concentrations tested) and precise (% RSD of less than 2% for both repeatability and intermediate precision). In terms of specificity, the SDIC method was free from the interference by excipients in tablets as evidenced by the nearly 100% of recoveries in the accuracy study. In addition, the method was unaffected by an impurity, namely, CTP disulfide which is possibly found in CTP. Without a free thiol group, the CTP disulfide did not produce the yellow color with the DTNB, thereby giving an analytical signal which did not differ from that of the blank (Table 3). In addition, when this impurity co-existed with the drug, CTP’s analytical signal did not change.

For the evaluation of robustness, the influence of two parameters, i.e., room lighting levels and LED brightness, which might affect the image capture were studied by the assay of the CTP solutions (≈20 µM) under different conditions. The results revealed that while the image acquisition in this work was normally carried out in the area in which the room ceiling lights were on, turning off the lights, resulting in a dimmer room, did not affect the assay results (Table 4). It was probable that the LED lights placed behind the cuvettes dominated the ambient light in the photography. In addition, lowering the brightness of the LED from a full to a half level did not cause a statistically significant difference between the obtained assay results. This can be explained by the fact that different LED brightness might produce images with different RGB intensities. However, if all the standard solutions and sample solutions were photographed in the same shot and the calculations were made using their own standard curve, there was no difference between the different conditions in the results of the final analysis. From these findings, the proposed assay was considered robust, and additional equipment such as a dark box used to control ambient light was not required.

The economical and green features of the proposed method were demonstrated by comparison with previously reported methods. Compared with the smartphone-based assay relying on the formation of complex with expensive palladium (II) chloride (Ravazzi et al. 2018), the present method is cheaper, safer, and eco-friendlier since palladium ion can be harmful by interfering with enzymatic metabolisms and organ functions (Tang et al. 2020). In addition, palladium is very toxic to aquatic organisms, thereby causing environmental hazards and pollution (Zimmermann et al. 2017; Merck 2023). Compared with the spectrophotometric assay of CTP using DTNB and 1-cm cuvette for the absorbance measurement (Hosseinimehr et al. 2004), the SCIC method proposed in this work was capable of analyzing CTP in the linear range of about 8 times lower concentrations than the spectrophotometric method, and significantly saved >10 times the amount of DTNB used (Table 5). In addition, the SDIC assay in the present work was compared with the SDIC assay of another thiol drug, namely D-penicillamine, which was also based on Ellman’s reaction (Phadungcharoen et al. 2019). In our experiment, it showed that D-penicillamine reacted with DTNB to form a yellow product in the same way as CTP, as evidenced by their standard curves with negligible difference (Fig. 2 in Supplementary data). This was because both drugs have one thiol group in a molecule, and the colorimetric reaction produces the same TNB- product. As shown in Table 5, the CTP assay in which the color measurement was performed by using double cuvettes (total pathlength of 2 cm) was capable of determining the analyte in a linear range with lower concentrations, and having a lower LOD and LOQ than that of D-penicillamine in which the color analysis was done on the samples contained in a 96-well microplate (pathlength of about 0.88 cm). Consequently, despite more reaction volume, the CTP assay required 4.3 times less DTNB reagent per reaction (on a molar basis) than the D-penicillamine assay. This feature helped reduce the cost of chemicals. In addition, minimization of the colorimetric reagent consumption aligned with the 12 principles of Green Analytical Chemistry (Gałuszka et al. 2013) and White Analytical Chemistry (Nowak et al. 2021) in which the environmental impacts and human safety are carefully concerned during the development and implementation of analytical procedures. In addition to reagent saving benefits, the equipment that was used in this assay was less expensive. For example, instead of a tablet gadget, LED lamps and a white acrylic sheet were used as the light source for photography.

Compared with the USP method, the proposed assay was greener since it did not involve the use of an organic solvent and did not require expensive HPLC equipment and column. Since the SDIC method was facile, rapid, and capable of analyzing multiple samples simultaneously, the assay could be achieved in a significantly shorter time. Excluding the time required for column equilibration and washing, the USP method analyzed samples individually with a run time of 10 min per sample. In contrast, the SDIC method could analyze a set of 6 standards and 6 samples within 15–20 min, indicating a higher sample throughput. Using the Analytical Greenness Metric (AGREE), greenness had a high score of 0.77 with 0.7 being considered green. (Fig. 6a). In addition, the Green Analytical Procedure Index (GAPI) illustrated that the SDIC method possessed 8, 2, and 4 green, yellow and red shaded sections, respectively (Fig. 6b). These findings confirmed that the method proposed in the present work is greener, faster, and more affordable.

Figure 6. 

Greenness of the proposed SDIC assay assessed by (a) Analytical GREEnness (AGREE) and (b) Green Analytical Procedure Index (GAPI).