2014 Volume 62 Issue 12 Pages 1252-1258
The objective of this study was to evaluate an improved bitterness sensor which has been developed to allow the precise and sensitive prediction of the bitterness of acidic bitter pharmaceutical active ingredients, using as examples nine non-steroidal anti-inflammatory drugs (NSAIDs). The bitterness of the nine NSAIDs was measured using a multichannel taste-sensing system incorporating a bitterness sensor, C00, which has a membrane surface with high hydrophobicity, and was developed to allow an enhanced hydrophobic interaction with acidic bitter substances. The sourness intensities of the nine NSAIDs were also determined in gustatory sensation testing and by a taste sensor using a sourness-sensitive membrane, CA0. The ‘Change in membrane Potential caused by Adsorption’ (CPA) of sodium diclofenac and etodolac were also determined in the presence of increasing concentrations of tartaric acid using membrane C00. Multiple regression analysis performed on the data on bitterness intensity obtained using the taste sensor and in gustatory sensation testing showed that CPA values from C00 could be used to predict the bitterness of the NSAIDs. The derived equation was y=−0.0413×CPA+0.3164, where y represents the predicted bitterness intensity. There were concentration-dependent changes in the bitterness intensities of diclofenac sodium and etodolac without any change in their sourness intensities. For diclofenac sodium and etodolac, there was a good correlation between predicted and actual bitterness intensities in the presence of increasing concentrations of tartaric acid. The taste sensor may be useful for predicting the bitterness intensity of acidic bitter pharmaceutical active ingredients such as NSAIDs.
Taste plays an important role in determining the acceptability of a pharmaceutical formulation. Many active pharmaceutical ingredients exhibit an unpleasant taste, making taste-masking an important step in formulation development. The use of an ‘electronic tongue’ or taste sensor for pharmaceutical purposes is a useful innovation, as it reduces dependence on human gustatory sensation testing. The taste sensor is an analytical sensor array system which is able to detect substances with specific chemical properties (or tastes) using electrochemical techniques and involving different artificial membranes.
Uchida et al. have reported a quantitative analytical method for the evaluation of the bitterness of various basic pharmaceutical products using a taste sensor.1–10) This taste sensor, an ‘electric tongue’ with global selectivity, was initially developed by Toko.11) It comprises several lipid/polymer membranes capable of transforming information about the substances which produce taste into electrical signals.12,13) The sensor output exhibits different patterns for chemical substances with different taste qualities such as saltiness, sourness, bitterness and umami, while exhibiting similar patterns for chemical substances with similar tastes.
Non-steroidal anti-inflammatory drugs (NSAIDs) are clinically efficacious in a variety of disease states based on their antipyretic, anti-inflammatory, antithrombotic and analgesic properties. These effects are principally achieved through inhibition of cyclogenase (COX) which inhibits prostaglandin synthesis.14,15) Currently, there are three recognized isoforms of COX, known as COX-1, COX-2 and COX-3,16) although there is a lack of consensus among investigators on whether COX-3 is a distinct human isoform.17) NSAIDs are classified on the basis of the specificity of their mechanism of action.18–20) Non-selective NSAIDs such as diclofenac, ibuprofen and indomethacin inhibit both COX-1 and COX-2 enzymes while COX-2-selective inhibitors (COXIBs) such as rofecoxib, valdecoxib and celecoxib, block COX-2 only.21) Most NSAIDs are acidic compounds.
C00 is a sensor membrane with a positive charge which has been developed to react to the bitterness of iso-α acid, the principal ingredient in hops, which are used to make beer. C00 is capable of evaluating the bitterness intensity of acidic pharmaceutical substances, which interact hydrophobically with the positively charged lipid.
The aim of this study was to investigate the ability of the taste sensor to evaluate the bitterness intensities of NSAIDs.
Nine NSAIDs were used in the present study: aspirin, diclofenac sodium, loxoprofen sodium, indometacin, etodolac, flurbiprofen, naproxen, meloxicam and piroxicam. All NSAIDs, as well as quinine hydrochloride and tartaric acid, were purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). All other reagents were special reagent grade.
Gustatory Sensation TestsThe protocol and experimental design for all gustatory sensation tests was approved in advance by the ethical committee of Mukogawa Women’s University. Six healthy female subjects, 27±10 years old, participated in the tests in which various tastes and textures were evaluated. No subject reported having a cold or other respiratory tract infection in the week prior to testing. The subjects were asked to refrain from eating, drinking, or chewing gum for at least 1 h prior to testing. All subjects were non-smokers and signed an informed consent before the experiments. The experimental protocol was approved in advance by the ethical committee of Mukogawa Women’s University. The experimental protocol of this study (No. 12–45) was approved on May 18, 2013 by the ethical committee of Mukogawa Women’s University.
The gustatory sensation test to measure bitterness or sourness intensity was performed with six well-trained volunteers according to a previously described method,22) using quinine hydrochloride at concentrations of 0.01, 0.03, 0.1, 0.3 and 1.0 mM as the standard for bitterness. Scores of 0, 1, 2, 3 and 4 were allocated to increasing concentrations of the bitterness standard solution. Tartaric acid at concentrations of 0.17, 0.60, 1.73, 4.66 and 11.99 mM was used as the standard for sourness, with scores of 1, 2, 3, 4 and 5 allocated to increasing concentrations of the sourness standard solution. Before testing, the volunteers were asked to keep the standard quinine hydrochloride or tartaric acid solutions in their mouths for 10 s, and were told the concentration and bitterness or sourness score of each solution. They were then asked to give each of the samples bitterness or sourness scores. Each sample was kept in the mouth and evaluated after 10 s. After tasting each sample, subjects gargled well and waited for at least 20 min before tasting the next sample.
The Taste SensorThe taste sensor, SA402B (Intelligent Sensor Technology Inc., Atsugi, Japan) was used to determine the bitterness and sourness intensities of the sample solutions using sensors C00 and CA0, developed specifically to detect the bitterness and sourness, respectively, of acidic substances. In the first step of the procedure, a reference solution (corresponding to saliva) is measured and the electric potential obtained (mV) is defined as Vr. Then a sample solution is measured and the electric potential obtained defined as Vs. The relative sensor output (R), represented by the difference (Vs−Vr) between the potentials of the sample and the reference solution, corresponds to the ‘taste immediately after putting in the mouth.’ The electrodes are subsequently rinsed with a fresh reference solution for 6 s. When the electrode is dipped into the reference solution again, the new potential of the reference solution is defined as Vr0. The difference (Vr0−Vr) between the potentials of the reference solution before and after sample measurement is the ‘Change in membrane Potential caused by Adsorption’ (CPA) and corresponds to the so-called ‘aftertaste.’ The value obtained when CPA is divided by R is defined as the adsorption rate.
In this study, the relative sensor outputs of CA023) and the CPA of C00 were taken as the bitterness and sourness intensities, respectively, of the NSAIDs tested. The measurement of each sample was repeated six times and the average value of the last three measurements used in the data analysis.
pH DeterminationThe electrode of pH meter (F-52, Horiba, Ltd., Kyoto, Japan) was inserted in the sample solution, and its pH value recorded.
ExperimentsThe bitterness intensities of the nine NSAIDs were evaluated in human gustatory sensation tests and by the taste sensor using membrane C00. Multiple regression analysis was performed on the bitterness intensity by gustatory sensation test and taste sensor outputs using membrane C00 to investigate the correlation between them. The data obtained using the taste sensor (R, CPA and CPA/R) were used as explanatory variables to derive a multiple regression expression with a partial regression coefficient obtained from multiple regression analysis using the data obtained by gustatory sensation testing as an object variable. The regression equation to predict bitterness intensity score from sensor outputs by taste sensor was derived. The sourness intensities of the nine NSAIDs were also evaluated in human gustatory sensation tests and by the taste sensor using membrane CA0. The correlation between the sourness intensities and relative values of CA0 was evaluated. As the bitterness intensities of sodium diclofenac and etodolac were over the bitterness threshold, τ1 (score of 1 in bitterness intensity) at concentrations over 1 mM, in order to investigate the influence of sourness intensity on bitterness intensity, the correlation between the actual and predicted bitterness intensities of sodium diclofenac and etodolac was evaluated in the presence of tartaric acid.
Statistical AnalysisEkuseru-Toukei 2010 (Social Survey Research Information Co., Ltd., Tokyo, Japan) was used for statistical analysis. The Bonferroni test was used for multiple comparisons. Correlation was examined using Pearson’s correlation test. The 5% level of probability was considered significant.
Figure 1 shows the bitterness intensities of 0.1, 1 and 10 mM samples of nine NSAIDs. The bitterness intensities of etodolac and diclofenac sodium increased in a concentration-dependent manner. There were no changes in the bitterness intensities of the other NSAIDs at increasing concentrations.
Data are presented as mean values±S.D. (n=6). * and **: Significant difference from 0.1 mM of drug solution at p<0.05 and at p<0.01, respectively.
Multiple regression analysis was performed on the data on bitterness intensity obtained using the taste sensor and in gustatory sensation testing. The data obtained using the taste sensor (R, CPA and CPA/R) were used as explanatory variables to derive a multiple regression expression with a partial regression coefficient obtained from multiple regression analysis using the data obtained by gustatory sensation testing as an object variable. The regression equation to predict bitterness intensity score from sensor outputs by taste sensor was derived by multiple regression analysis in this study. The derived regression equation was: y=−0.0413×CPAC00+0.3164 (r2=0.75, p<0.01) where y is the predicted bitterness intensity score, and CPAC00 is the CPA value observed using taste sensor membrane C00. There was a significant correlation between the bitterness intensity obtained in gustatory sensation testing and CPAC00 using Pearson’s correlation test (r=−0.83, p<0.01; Fig. 2). Squared cross-validated correlation coefficient, q2 for the correlation between bitterness intensity score by gustatory sensation test and predicted bitterness intensity score calculated by the derived regression equation was 0.60. CPAC00 values could therefore be used to predict the bitterness of the NSAIDs.
Data are presented as mean values±S.D. (n=6).
Figure 3(a) shows the sourness intensities of 0.1, 1 and 10 mM solutions of the nine NSAIDs obtained in gustatory sensation testing. The sourness intensity of aspirin increased in a concentration-dependent manner. None of the other NSAIDs showed any sourness at any concentration. Figure 3(b) shows the RCA0 of 0.1, 1 and 10 mM samples of the nine NSAIDs obtained using the taste sensor, where RCA0 is the R value observed using membrane CA0. The RCA0 of aspirin increased in a concentration-dependent manner, but none of the other NSAIDs showed a positive value of RCA0 at any concentration.
Data are presented as mean values±S.D. (n=6). **: Significant difference from 0.1 mM of drug solution at p<0.01.
Figure 4(a) shows the correlation between the bitterness and sourness intensities of NSAIDs as measured by gustatory sensation testing. The sourness intensity of aspirin increased in a concentration-dependent manner without any change in bitterness intensity, while the bitterness intensities of sodium diclofenac and etodolac increased in a concentration-dependent manner without any change in their sourness intensities. There were no concentration-dependent changes in the bitterness and sourness intensities of any other NSAIDs. Figure 4(b) shows the correlation between the CPAC00 and RCA0 of the NSAIDs tested in the taste sensor. The RCA0 of aspirin increased in a concentration-dependent manner without any change in CPAC00 while the CPAC00 of sodium diclofenac and etodolac increased in a concentration-dependent manner without any change in RCA0. There were no concentration-dependent changes in either the CPAC00 or the RCA0 of the other NSAIDs.
Data are presented as mean values±S.D. (n=6).
The bitterness intensities of sodium diclofenac and etodolac increased in a concentration-dependent manner without any change in their sourness intensities while the CPAC00 of sodium diclofenac and etodolac increased in a concentration-dependent manner without any change in RCA0. As the bitterness intensities of sodium diclofenac and etodolac were over the bitterness threshold, τ1, at concentrations over 1 mM, the influence of sourness intensity on bitterness intensity was examined by evaluating the correlation between their actual and predicted bitterness intensities in the presence of increasing concentrations of tartaric acid. Figure 5 shows the influence of increasing concentrations of tartaric acid on (a) the CPAC00 and (b) the pH of diclofenac sodium suspension (10 mM). Both the CPAC00 and the pH values of sodium diclofenac were decreased by the addition of tartaric acid in a concentration-dependent manner. Figure 6 shows the effect of increasing concentrations of tartaric acid on (a) the CPAC00 and (b) the pH of etodolac suspension (10 mM). The CPAC00 of etodolac remained nearly constant at increasing concentrations of tartaric acid while the pH decreased in a concentration-dependent manner. Figure 7 shows the correlation between the actual and predicted bitterness intensities of sodium diclofenac (10 mM) and etodolac (10 mM) in the presence of increasing concentrations of tartaric acid. The predicted bitterness intensities were calculated using the derived regression equation: y=−0.0413×CPAC00+0.3164 where y is the predicted bitterness intensity score. There was a high correlation between the actual and predicted bitterness intensities of sodium diclofenac and etodolac in the presence of increasing concentrations of tartaric acid (r=0.86, Pearson’s test, p<0.01).
Various studies of the structure–activity relationships associated with bitterness have reported that ionic and hydrophobic bonds are involved in the binding of bitter substances with the bitterness receptor.24,25) Taste sensor membrane C00 is a lipid film with a positive charge. It was therefore suggested that, as the binding of bitter substances with C00 involves ionic and hydrophobic bonds, CPA values obtained using C00 could be used to predict the bitterness of acidic bitter substances such as NSAIDs.
When multiple regression analysis was performed on the bitterness intensity data obtained from the taste sensor using membrane C00 (R, CPA and CPA/R) and in gustatory sensation testing, CPA values obtained from the taste sensor were found to be capable of predicting the bitterness of NSAIDs. These results demonstrate the possibility of evaluating the bitterness intensities of acidic bitter pharmaceutical active ingredients using taste sensor membrane C00.
In general, sourness decreases bitterness intensity.26) However, it has not previously been demonstrated that sourness is capable of decreasing the bitterness intensity of acid pharmaceutical substances. In the present study, while the sourness intensity of aspirin increased in a concentration-dependent manner, there were no changes in its bitterness intensity at any concentrations. Aspirin may have a decreased bitterness intensity because it is also sour. None of the other NSAIDs demonstrated any sourness at any concentration. The bitterness intensities of etodolac and diclofenac sodium both increased in a concentration-dependent manner although there were no changes in their sourness intensities at any concentration.
Data are presented as mean values±S.D. (n=6). **: Significant difference from tartaric acid τ1 at p<0.01.
Data are presented as mean values±S.D. (n=6). **: Significant difference from tartaric acid τ1 at p<0.01.
Data are presented as mean values±S.D. (n=6).
The variation in the bitterness intensity scores of the gustatory sensation tests is large (Figs. 1, 2, 7). The gustatory sensation test to measure bitterness intensity was performed with six well-trained volunteers and the variation was almost same as the previous paper.26) They are becoming small or general variation, if the number of test persons is increased.
Both the CPAC00 and pH values of sodium diclofenac suspension (10 mM) were decreased in a concentration-dependent manner by the addition of tartaric acid. The CPAC00 values of etodolac suspension (10 mM) were slightly decreased with the addition of tartaric acid, in spite of the fact that this addition was accompanied by a change of pH. The mechanism underlying the difference between the response of the C00 sensor membrane to the addition of tartaric acid to the two solutions (sodium diclofenac and etodolac) is unclear, but it may be something to do with the different responses of ionic or hydrophobic bonds to decreased pH. It has been reported that inhibition of the human bitter taste receptor is pH-dependent,27) and the pH-dependent inhibition of response to the C00 sensor membrane is certainly similar. This may be further evidence in support of the theory that the C00 sensor membrane imitates the human bitter taste receptor.
There was a high correlation between the actual and predicted bitterness intensities of sodium diclofenac and etodolac in the presence of tartaric acid, with the predicted bitterness intensities calculated according to the derived equation using CPAC00. It is therefore suggested that the bitterness intensity of bitter acidic substances in mixed solutions with differently tasting (e.g. sour) substances, can be determined accurately by the taste sensor using sensor membrane C00.
In multiple regression analysis performed on the bitterness intensity data obtained from the taste sensor (R, CPA and CPA/R) and gustatory sensation testing, CPA values from C00 could be used to predict the bitterness of NSAIDs. The bitterness intensities of these acidic bitter substances in mixed solution with differently tasting (e.g. sour) substances, could be determined accurately by the taste sensor using sensor membrane C00. This suggests that the taste sensor, using sensor membrane C00, may be useful for predicting the bitterness intensities of acidic pharmaceutical active ingredients such as NSAIDs.
This work was supported by Grant-in-Aid for Scientific Research (C) from the Japan Society for Promotion of Science 24590226 and 25460234 (to Takahiro Uchida and Miyako Yoshida, respectively).
