Study on the Combined Toxicities and Quantitative Characterization of Toxicity Sensitivities of Three Flavor Chemicals and Their Mixtures to Caenorhabditis elegans.

It was shown that flavor chemicals with high toxicity sensitivities mean that small changes in their effective concentrations can lead to significant changes in toxicity. Flavors are widely used in personal care products. However, our study demonstrated that some flavor chemicals and their mixture rays have high toxicity sensitivities to Caenorhabditis elegans (C. elegans), which may have an impact on human health. In this paper, three flavor chemicals (benzyl alcohol, phenethyl alcohol, and cinnamaldehyde) were used as components of the mixture, and three binary mixture systems were constructed, respectively. Five mixture rays were designed for each mixture system by a direct equipartition ray design method. The lethal toxicities of the three flavor chemicals and mixture rays to C. elegans at three exposure volumes were determined. A new concept (inverse of the negative logarithmic concentration span (iSPAN)) was introduced to quantitatively evaluate the toxicity sensitivity of chemicals or mixture rays, and the combination index (CI) was employed to identify the toxicological interactions in the mixtures. It was shown that the three flavor chemicals as well as the binary mixture rays have a significant concentration-response relationship on the lethality of C. elegans. The iSPAN values of the three flavor chemicals and their mixture rays were larger than 3.000, showing very strong toxicity sensitivity to C. elegans. In mixture systems, the toxicity sensitivities of mixture rays with different mixture ratios were also different at different exposure volumes. In addition, it can be seen from the CI heat map that the toxicological interaction not only shows the mixture ratio dependence but also changes with the different exposure volumes, which implies that the mixtures consisting of flavor chemicals with high toxicity sensitivity have complex toxicological interactions. Therefore, in environmental risk assessment, special attention should be paid to chemicals with high toxicity sensitivities.

Introduction

Personal care products have been detected in various environmental matrices and identified as emerging pollutants[1] that also pose a threat to organisms.[2,3] As the components of personal care products and food additives, flavors are also widely used in food production, such as biscuits, frozen foods, condiments, canned foods, and beverages.[4] Recent studies have found that flavors were also detected in aquatic environments.[5,6] Most of the cosmetics contain flavors that often enter the environment through common ways, such as washing[7,8] and swimming,[9] and some flavors are also mixed with antibiotics, steroids, anti-inflammatory drugs, sedatives, and diet pills.[10] As a very complex mixture system, flavors contain many different components, and some of them have toxic effects on organisms.

Benzyl alcohol (BEA), phenethyl alcohol (PHA), and cinnamaldehyde (CID) are common flavor chemicals used in flavors. A survey shows that the content of BEA in 27 groundwater sites in Beijing and Tianjin in North China is 582 ng L–1.[11] BEA is very common in industrial chemistry. Its physical and chemical properties make it possible to be used as an anticorrosive agent in medical solutions, over-the-counter medications, local creams and lotions, perfumes, and cosmetics.[12] As an effective ingredient of flavors, it is commonly used in many daily necessities. Although it is found to be an antibacterial agent in many parenteral preparations, it is the cause of asthma syndrome in preterm infants.[13] It can also induce zebrafish embryo apoptosis, increase mortality, reduce the hatching rate, and reduce the number of body segments[12] and has acute lethal effects on rats.[14] PHA is a highly aromatic alcohol with a rose flavor. In the Kantar surveys, mean aggregate systemic PHA exposure was 7.18 μg/kg bw/day, and less than 5% of the population has systemic PHA exposure above 26.73 μg/kg bw/day.[15] Hydrosol volatiles from flowers of 10 Paeonia suffruticosa Andr. cultivars were analyzed, and it was found that the content of PHA is 48.0–79.5%.[16] It is often used as a flavoring agent,[17] which is detected in some commercial red wines and also used in the soy sauce that we eat.[18,19] In addition, it plays an important role in diet and is widely used in olfactory activity tests and cosmetic ingredients, but it can affect the behavior of mice and may cause antidepressant effects.[20] As alcohols are often used as reactants in industrial production, environmental and sustainable development challenges arise.[21] CID is the main component of cinnamon, which has antibacterial and bactericidal effects.[22] Studies have shown that CID was detected in three soils.[23] It is used to treat blood circulation disorders, dyspepsia, inflammation, and gastritis and also used in manufacturing as spices and condiments for beverages, sugar, ice cream, and chewing gum.[24] Because it is used as an effective component of common insecticides, its excessive use will cause some harm to the soil ecological environment.[25,26] It was found that CID had a strong inhibitory effect on Staphylococcus aureus and Escherichia coli (E. coli)[27]. In addition, studies have shown that Tribolium castaneum adults and Sitophilus zeamais adults were sensitive to contact toxicity of CID.[28] They are just toxicity studies of the single flavor chemical. However, flavors are mixtures, and it is not enough to study the toxicity of a single component. Moreover, it cannot reflect the change law of the combined toxicity of flavors. Further exploration is needed for its mixture. There is little research on flavor mixtures, but it is still a problem to be discussed.

At present, through some research results,[29−32] we find that different toxic chemicals or mixture rays have different toxicity sensitivities to different organisms. The toxicities of chemicals with strong toxicity sensitivities change significantly with the slight changes in the effective concentration. On the contrary, the toxicity of chemicals with weak toxicity sensitivity will not change due to the slight changes of effective concentration; the specific performance of the order of the magnitude span of a 20% lethal concentration (LC20) and an 80% lethal concentration (LC80) is different; the difference of the large span is one or more orders of magnitude, while that of the small span is less than half an order of magnitude. When toxicity assessment is carried out for chemicals with high toxicity sensitivities, if the change rate of each experimental concentration point between LC20 and LC80 is too large, then it will result in some intermediate effects being ignored, so the toxicity data obtained is not rigorous enough, which may lead to a large error in the experimental results and toxicity assessment.

Toxicological interaction assessment is an important part of the combined toxicity study of chemical mixtures. The combination index (CI) can well describe the interaction of mixture systems.[33,34] Many scholars also use the CI to study combined toxicity. Liu et al.[35] studied the nature of the binary and ternary interactions of chlorantraniliprole, λ-cyhalothrin, and imidacloprid on the mortality rate of silk worms. Zhang et al.[36] found that the CI was a better way of predicting the interaction of combined phenicol antibiotics than classical models. In addition, the CI is also used to study the toxicological interactions of other pharmaceuticals.[37,38] The genes of Caenorhabditis elegans (C. elegans) have high similarity to those of humans. As a classic model organism, it is sensitive to chemicals, and the toxicity assessment of C. elegans can reflect the potential impact of chemicals on human beings.[39] Zhang et al.[40] used 6-well microplates, Moyson et al.[41] used 24-well microplates, and Huang et al.[42] and Wang et al.[43] used 96-well microplates for toxicity tests. For the study of C. elegans, scholars often pay attention to the effect of the concentration. The concentration of a chemical is determined by its amount and volume. When conducting toxicity tests with C. elegans, different exposure volumes can be designed if different exposure carriers are used. However, for chemicals with high toxicity sensitivities, will the selection of different exposure volumes at the same concentration affect the toxicity assessment results of the CI? This is another question to be discussed in this paper.

The equivalent-effect concentration ratio (EECR),[44] specified proportion,[45−49] and fixed proportion[50−52] are commonly used by most of the scholars when designing mixtures. For example, Zhao et al.[53] mixed cytosine and thymine chlorination byproducts according to a fixed proportion to determine its acute toxicity to E. coli. The chronic toxicity of zebrafish larvae was determined by mixed sulfamonomethoxine, cefotaxime sodium, tetracycline, and enrofloxacin in the same proportion.[54] The toxicity of Bacillus subtilis was studied by mixed perfluorooctane sulfonic acid and chromium(VI) at a specified concentration.[55] However, using these methods for designing mixture systems, the experimental results cannot effectively represent the combined toxicity and toxicological interaction of the whole mixture system. Hence, efficient and simple methods are needed to design complex mixture systems and select representative mixture rays for mixture toxicity studies. The direct equipartition ray design (EquRay)[56,57] is used to design binary mixture systems, which can reasonably and effectively select some representative mixture concentration points from the binary mixture system as the basic concentration composition of the mixture, so as to comprehensively characterize the concentration distribution of the binary mixture system and then comprehensively investigate the toxicity change law of the binary mixture. This method usually takes EC50 as the concentration reference point, forms a line segment by connecting the EC50 of two components in the plane coordinate system, evenly divides the line segment to obtain k average points, and takes these average points as the basic concentration composition of the mixture to construct a representative mixture ray, so as to comprehensively investigate the toxicity change law of binary mixtures. Moreover, designing several representative mixture rays can study the toxicological interaction of mixture systems with different mixing ratios.[58−61]

In this paper, to explore the toxicity sensitivities of flavor chemicals and their mixture rays and the sensitivity of toxicological interactions caused by different exposure volumes, BEA, CID, and PHA were used to structure three binary mixture systems, using the EquRay to design five mixture rays of each mixture system. Twelve concentration gradients were set for each mixture ray. The combined toxicity of each mixture ray to C. elegans was determined by microplates at three different exposure volumes. A new concept was proposed to quantitatively evaluate the toxicity sensitivities of each chemical and mixture ray. The toxicological interactions of the three mixture systems were evaluated by the CI, which provides references and suggestions for ecological risk assessment and toxicological research of toxic chemicals with high toxicity sensitivities.

Results and Discussion

The CRCs and Toxicity Sensitivities of the Three Flavor Chemicals

The concentration–response (mortality) profiles of three flavor chemicals BEA, CID, and PHA to C. elegans are shown in Figure . All fitted CRCs at different exposure volumes of 100, 200, and 400 μL can be effectively characterized by the nonlinear Weibull function, and the regression parameters (location α and shape β) as well as the goodness-of-fit (determination coefficient R2 and root-mean-square error RMSE) are listed in Table . The negative logarithms of the median lethal concentration (pLC50), 20% lethal concentration (pLC20), and 80% lethal concentration (pLC80) as well as the toxicity sensitivity defined as an inverse of the negative logarithmic concentration span (iSPAN) values obtained from the fitting functions of the three flavor chemicals are also listed in Table .

Figure 1

Concentration–responses of the three flavor chemicals where triangles, squares, and circles refer to exposure volumes of 400, 200, and 100 μL, solid lines to the fitting curves, and dashed lines to the 95% observation-based confidence intervals (OCIs).

Table 1

The Physical Properties and Weibull Fitting Parameters (α and β), Fitting Statistics (R2 and RMSE), pLC20, pLC50, pLC80, and Toxicity Sensitivity (iSPAN) at Three Exposure Volumes (EV) for the Three Flavor Chemicals

M.W.: molecular weight.

From the values of pLC50 of the three flavor chemicals in Table , for BEA and PHA, the means ± 2 standard deviation of pLC50 at the three exposure volumes were 1.613 ± 0.171 and 1.743 ± 0.046, respectively, implying that their toxicities were weak and had no significant difference. However, the toxicity of CID (3.130 ± 0.032) was significantly stronger than those of BEA and PHA, which is stronger than those of some pesticides,[62] equal to those of common heavy-metal pollutants such as zinc chloride (3.102) and cadmium chloride (3.191), but less than that of copper chloride (4.014).[41] Compared with organic pollutants, CID has a lower toxicity than 2,4-dichlorophenol (3.408) but a higher toxicity than 4-nitrophenol (2.609), glyphosate (2.469), dichlorvos (2.448), 1-butyl pyridine chloride (2.151), 1-butyl pyridine bromide (2.062), and 2, 4-dichlorophenoxyacetic acid (2.425). There is no significant difference compared with 4-chlorophenol (3.186).[62,63]

As shown in Figure , for a substance at three exposure volumes, except for the CRCs and confidence intervals of BEA that did not overlap completely, those of CID and PHA were basically overlapped. It was particularly pointed out that the lethal toxicities of the three flavor chemicals to C. elegans were very sensitive to the slight changes in concentration. In other words, the subtle change in concentration may lead to a significant change in toxicity. According to the iSPAN (Table ), the smallest iSPAN was 3.460 (BEA at 400 μL), and the largest iSPAN was 25.641 (PHA at 400 μL). The change of the exposure volume does not produce a significant effect on the iSPAN of any of the three flavor chemicals, and the means of iSPANs (± 2 times standard deviation) of BEA, CID, and PHA were 4.222 (± 1.430), 8.594 (± 1.934), and 20.055 (± 9.748), respectively. The iSPANs of the flavor chemicals under study were significantly larger than those of the other pollutants reported (Table ). The values of the iSPAN in Table were roughly estimated from the CRCs in the literature reported. From Table , all the iSPAN values are not greater than 2.0. For example, the iSPAN values of 24 h lethal toxicity of dichlorvos (no. 8), copper sulfate (no. 6), and 1-hexyl-3-methylimidazole ammonium bromide (no. 11) to C. elegans are 1.250–2.000, and that of 3-methyl-1-octylimidazole chloride (no. 12) is 1.000–1.250 and so on.[31,41,62−65] The least one is only 0.40, the iSPAN of cadmium chloride (no. 4). It had been shown that CID has a high iSPAN value or high toxicity sensitivity so that small changes in concentration will produce significant changes in toxicity. Because CID is often used as the main component of insecticides,[26] special attention should be paid to the use of CID; otherwise, it will cause harm to crops, animals, and the soil ecological environment.[25]

Table 2

The iSPAN Values of Some Chemicals on the Lethality Endpoint of C. elegans at 24 h, Estimated from the Literature

no.	chemical	iSPAN	references
1	1-butylpyridinium bromide	0.833–1.000	Feng et al., 2017
2	1-butylpyridinium chloride	0.833–1.000	Feng et al., 2017
3	4-chlorophenol	1.250–2.000	Feng et al., 2017
4	CdCl2	0.400–0.500	Moyson et al., 2018
5	CuCl2	0.455–0.556	Moyson et al., 2018
6	CuSO4	1.250–2.000	Tang et al., 2016
7	2,4-dichlorophenol	1.250–2.000	Ju et al., 2019
8	dichlorvos	1.250–2.000	Tang et al., 2016
9	gallic acid	1.250–2.000	Verdu et al., 2020
10	glyphosate	1.000–1.250	Feng et al., 2017
11	1-hexyl-3-methylimidazolium bromide	1.250–2.000	Tang et al., 2016
12	3-methyl-1-octylimidazolium chloride	1.000–1.250	Tang et al., 2016
13	4-nitrophenol	1.250–2.000	Feng et al., 2017
14	nonylphenol ethoxylate	0.500–0.667	De la Parra-Guerra et al., 2020
15	ZnCl2	0.500–0.667	Moyson et al., 2018

The toxicity indexes of most of the chemicals to organisms reported in the literature are often EC50 or LC50,[29,31,66] which cannot reflect the toxicity sensitivities of chemicals or the speed of toxicity changing with the concentration. For the chemicals with high iSPANs (such as greater than 3), slight changes in concentration will lead to significant changes in toxicity, so it is necessary to obtain other effective concentrations, such as EC20 and EC80 to obtain the iSPANs, which are particularly important for the toxicity assessments of chemicals with high toxicity sensitivities. Otherwise, it may lead to inaccurate results of toxicity assessments.

The Combined Toxicities and Toxicity Sensitivities of Binary Mixture Rays

Three flavor chemicals form three binary mixture systems, BEA-CID, BEA-PHA, and PHA-CID, where five representative mixture rays are selected for each of mixture systems by EquRay.[56] The concentration–response profiles of the 15 mixture rays in three binary mixture systems are shown in Figure . It can be seen that the combined toxicities of all mixture rays increased with the increase in concentration. Their concentration–responses can be effectively fitted by the two-parameter Weibull function and the fitted RMSEs for most of the CRCs were less than 5%, while the determination coefficients (R2) were larger than 0.9500 (seeing Table ).

Figure 2

(a–c) Concentration–responses of 15 binary mixture rays in three mixture systems (BEA-CID, BEA-PHA, and PHA-CID) where triangles, squares, and circles refer to exposure volumes of 400, 200, and 100 μL, solid lines to the fitting curves, and dashed lines to the 95% observation-based confidence intervals (OCIs).

Table 3

The Mixture Ratios, Weibull Fitting Parameters (α and β), Fitting Statistics (R2 and RMSE), pLC20, pLC50, pLC80, and Toxicity Sensitivity (iSPAN) of Each Mixture Ray at Three Exposure Volumes (EV) for the Three Mixture Systems

mixture ray (A-B-Rk)	mixture ratio (LC50,A:LC50,B)	EV (μL)	α	β	RMSE	R2	pLC20	pLC50	pLC80	iSPAN
BEA-CID-R1	5:1	100	14.16	8.92	0.0466	0.9739	1.756	1.629	1.534	4.505
		200	18.05	10.64	0.0433	0.9881	1.837	1.731	1.652	5.376
		400	17.69	9.72	0.0294	0.9916	1.974	1.858	1.771	4.926
BEA-CID-R2	4:2	100	18.51	11.35	0.0220	0.9871	1.763	1.663	1.589	5.747
		200	29.61	17.04	0.0158	0.9982	1.826	1.759	1.710	8.621
		400	18.70	10.16	0.0463	0.9857	1.988	1.877	1.794	5.155
BEA-CID-R3	3:3	100	23.78	12.83	0.0245	0.9972	1.970	1.882	1.816	6.494
		200	16.07	8.50	0.0535	0.9901	2.067	1.934	1.835	4.310
		400	19.95	10.64	0.0107	0.9994	2.016	1.909	1.830	5.376
BEA-CID-R4	2:4	100	29.66	15.12	0.0231	0.9975	2.061	1.986	1.930	7.634
		200	28.37	14.30	0.0245	0.9968	2.089	2.010	1.951	7.246
		400	26.51	13.27	0.0335	0.9943	2.111	2.025	1.962	6.711
BEA-CID-R5	1:5	100	37.82	16.34	0.0358	0.9927	2.406	2.337	2.285	8.264
		200	30.89	13.32	0.0230	0.9966	2.432	2.347	2.283	6.757
		400	26.16	11.05	0.0342	0.9905	2.503	2.401	2.324	5.587
BEA-PHA-R1	5:1	100	23.56	14.96	0.0526	0.9802	1.675	1.599	1.543	7.576
		200	24.03	14.98	0.0345	0.9924	1.704	1.629	1.572	7.576
		400	25.32	15.39	0.0286	0.9944	1.743	1.669	1.614	7.813
BEA-PHA-R2	4:2	100	36.72	22.10	0.0311	0.9947	1.729	1.678	1.640	11.236
		200	28.69	17.42	0.0368	0.9913	1.733	1.668	1.620	8.850
		400	35.96	21.66	0.0388	0.9908	1.729	1.677	1.638	10.989
BEA-PHA-R3	3:3	100	28.81	17.81	0.0420	0.9816	1.702	1.638	1.591	9.009
		200	28.92	17.38	0.0385	0.9896	1.750	1.685	1.637	8.772
		400	41.78	24.72	0.0264	0.9959	1.751	1.705	1.671	12.500
BEA-PHA-R4	2:4	100	28.27	16.85	0.0591	0.9686	1.767	1.700	1.650	8.547
		200	41.38	23.85	0.0371	0.9926	1.798	1.750	1.715	12.048
		400	50.20	28.97	0.0196	0.9982	1.785	1.745	1.716	14.706
BEA-PHA-R5	1:5	100	69.67	40.97	0.0555	0.9764	1.737	1.709	1.689	20.833
		200	45.95	26.29	0.0324	0.9932	1.805	1.762	1.730	13.333
		400	84.16	48.21	0.0240	0.9969	1.777	1.753	1.736	24.390
PHA-CID-R1	5:1	100	30.85	18.17	0.0394	0.9894	1.780	1.718	1.672	9.174
		200	37.41	21.28	0.0229	0.9976	1.828	1.775	1.736	10.753
		400	58.48	32.81	0.0571	0.9863	1.828	1.794	1.768	16.667
PHA-CID-R2	4:2	100	19.94	10.71	0.0481	0.9773	2.002	1.896	1.817	5.435
		200	41.07	22.71	0.0323	0.9943	1.875	1.825	1.788	11.494
		400	33.51	17.89	0.0267	0.9948	1.957	1.894	1.847	9.091
PHA-CID-R3	3:3	100	35.80	18.43	0.0608	0.9767	2.024	1.962	1.917	9.346
		200	33.29	17.39	0.0337	0.9932	2.001	1.935	1.887	8.772
		400	39.01	19.78	0.0331	0.9937	2.048	1.991	1.948	10.000
PHA-CID-R4	2:4	100	36.14	16.97	0.0330	0.9919	2.218	2.151	2.102	8.621
		200	30.20	14.05	0.0270	0.9930	2.256	2.176	2.116	7.092
		400	38.82	18.13	0.0514	0.9819	2.224	2.161	2.115	9.174
PHA-CID-R5	1:5	100	25.26	10.19	0.0564	0.9658	2.626	2.515	2.432	5.155
		200	27.34	11.18	0.0620	0.9584	2.580	2.478	2.403	5.650
		400	29.90	11.79	0.0600	0.9669	2.663	2.567	2.496	5.952

For the BEA-CID mixture system (Figure a), the CRCs of five mixture rays (R1, R2, R3, R4, and R5) at three exposure volumes were shifted from right to left, indicating that their toxicities increased gradually from R1 to R5. If pLC50 was taken as the toxicity index, then at 100 μL, the toxicities of five mixture rays from R1 to R5 were 1.629, 1.663, 1.882, 1.986, and 2.337, respectively. At 200 μL, the toxicities of five mixture rays from R1 to R5 were 1.731, 1.759, 1.934, 2.010, and 2.347, and at 400 μL, the toxicities of five mixture rays from R1 to R5 were 1.858, 1.877, 1.909, 2.025, and 2.401 (Table ), which indicate that the toxicities of the mixture rays of BEA-CID increased with the increase in the mixture ratio of CID, and R5 has the highest toxicity. As shown in the results, the toxicity of a mixture ray is related to its mixture ratio. In this study, the mixture ratios of three rays for R3 were the equivalent-effect/toxicity concentration ratio (EECR),[67−69] so the mixture ray with the EECR is not necessarily the most toxic ray in a mixture system with a certain chemical composition. Therefore, it is necessary to systematically investigate the toxicity changes of many mixture rays with multiple mixture ratios to reveal the toxicity change law in a mixture system. It should be noted that the CRCs of the six mixture rays for R1 and R2 at three exposure volumes of 100, 200, and 400 μL were different from each other, which are not parallel, and the toxicities increased with the increase in the exposure volume. However, the CRCs of the other nine mixture rays for R3, R4, and R5 were almost overlapped at three exposure volumes. That is, for R3, R4, and R5, the change of the exposure volume had no significant effect on the mixture rays’ combined toxicities.

The concentration–response profiles of five mixture rays in the BEA-PHA system shown in Figure b were different at three exposure volumes, but the differences were less distinct than those of the mixture rays of R1 and R2 in the BEA-CID system. The change of the exposure volume had little effect on the toxicity of the mixture ray BEA-PHA-R2. According to pLC50, the toxicity gradually increased from R1 to R5, i.e., BEA-PHA-R5 had the highest toxicity in the BEA-PHA system. Figure c shows the results of the experiments of the PHA-CID system. The shapes of CRCs in Figure a,c show similar toxicity change trends. Similarly, the toxicities of the five mixture rays in the PHA-CID system were gradually increased from PHA-CID-R1 to PHA-CID-R5 (Table ). For the BEA-CID system, the maximum and minimum values of pLC50 of mixture rays were 2.401 and 1.629, and those for the other two systems, BEA-PHA and PHA-CID, were 1.762 and 1.599 and 2.567 and 1.718. Obviously, there was little difference between the most toxic mixture ray and the least toxic mixture ray in the BEA-PHA system, but toxicities of the mixture rays in the BEA-CID or the PHA-CID system were different from each other, which may be due to the toxicity of CID being significantly greater than those of BEA and PHA.

It was shown that CID was often used as a flavoring ingredient in food production, such as beverage, ice cream, and so on.[24] It has been found that adding CID to chicken feed can promote the proliferation of probiotics in the digestive tract,[70] and Lactobacillus in the posterior segment of chicken intestines inhibits the growth of Escherichia coli(71) and produces a vasodilative effect on the aorta of rats.[72] This study demonstrates that the toxicities of the mixture rays including CID are closely correlated to the mixture ratios of CID. Therefore, it is necessary to rationally control the dosage and proportion of CID in food production, medicines, and so on.

The values of iSPANs of various mixture rays at three exposure volumes can be obtained by substituting the LC20 and LC80 data in Table into eq , and the results are included in Table . It was shown that all the iSPANs but three were greater than 5.00, implying very high toxicity sensitivities of all binary mixture rays of flavors. For the BEA-CID system, the iSPANs of 15 mixture rays were between 4.310 and 8.621, and the lowest iSPAN value of 4.310 occurs in the mixture ray of R3 at an exposure volume of 200 μL, which was the mixture ray with the lowest toxicity sensitivity. The mixture ray with the highest toxicity sensitivity (the iSPAN is 8.621) was R2 at 200 μL. For the BEA-PHA system, the iSPANs of 15 mixture rays were between 7.576 and 24.390 where the iSPAN increased gradually from R1 to R5. The maximum iSPAN of the mixture ray in this system was about three times larger than that of the BEA-CID system, indicating that the mixture rays of the BEA-PHA system were more sensitive than those of the BEA-CID system. For all 15 mixture rays of the PHA-CID system, the iSPANs were between 5.155 and 16.667 and less than that of PHA, which illustrates that the toxicity sensitivities of the mixture rays in the mixture system were weaker than that of the single chemical PHA. The above results show that the combined toxicity of the mixture ray composed of chemicals with high iSPANs is also sensitive to slight changes in concentration. For the mixture rays with the same components, their toxicity sensitivities are different across different mixture ratios. Moreover, even for the mixture ray with the same mixture ratio, its toxicity sensitivity is also different at different exposure volumes, which may be the characteristic of chemicals with high toxicity sensitivity.

Toxicological Interactions in the Binary Mixtures

Figure depicts the CI heat maps of three mixture systems, BEA-CID, BEA-PHA, and PHA-CID, i.e., various CI values with different effects of 15 mixture rays at three exposure volumes. The abscissa represents the effect and is positively correlated with the concentration, and the vertical number in the heat maps represents the CI value. Blue, white, and red blocks represent the toxicological interactions of SYN, ADD, and ANT in mixtures, respectively. The depth of the color reflects the strength of interaction. Using the heat map, we can more intuitively see the change trend of toxicological interaction in the mixture systems. For the BEA-CID mixture system, the toxicological interactions of mixture rays of R1 and R2 changed from ADD/ANT to SYN when the concentrations or effects increased, and the concentration range inducing SYN became wider with the increase in the exposure volume. The concentration range of ADD/ANT was narrowing gradually. The SYN range of the large exposure volume was wider than that of the small exposure volume. It can be seen that the different exposure volumes will affect the toxicological interaction. However, at the three exposure volumes, the toxicological interactions of mixture rays of R4 and R5 had almost no change. The mixture rays of R4 were ANT at medium and high concentration levels, and the mixture rays of R5 were SYN at medium and high concentration levels.

Figure 3

CI heat map of 15 rays in the three mixture systems at three exposure volumes of 100, 200, and 400 μL, where blue, white, and red refer to synergistic interaction (SYN), additive action (ADD), and antagonistic interaction (ANT), respectively. Here, the deeper the color is, the stronger the interaction will be.

With the increase in the exposure volume, the toxicological interactions of most of the mixture rays were ADD/ANT at a low concentration level, and those were SYN at a high concentration level in the BEA-PHA system; for R1 and R2, the ranges of SYN at medium and high concentration levels were gradually narrowing. It can be seen from the depth of the color that the SYN strength of the ray of R2 at 100 μL was the strongest in this mixture system. At 400 μL, the concentration range of SYN gradually widened from rays of R1 to R5, indicating that the change of the mixture ratio and the expansion of the exposure volume resulted in the reduction of the minimum SYN concentration of mixture rays in the system. SYN is a very noteworthy issue, which also confirms the importance of mixture toxicological interaction assessment in risk assessment.[67,73] It was worth noting that for rays of R1 and R2 at a low concentration level, the toxicological interactions changed from ADD to ANT from 100 to 400 μL, and the depth of red deepened, which means that the change of the exposure volume may increase the intensity of ANT with some mixture ratios.

For the PHA-CID system, it is obvious from Figure that except for the rays of R5, the other rays show high-dose ANT. What needs special attention is that the three rays for R5 showed SYN at all concentration levels, and the intensity of SYN was the strongest in the system, even stronger than those of most of the rays in the other two systems, indicating that under this mixture ratio, the mixture system of PHA-CID may had a higher risk especially at a low concentration level. Studies have shown that the effect of chemicals on locomotion behavioral endpoints has been found to appear at concentrations below the lethal endpoint.[74] Therefore, considering the high iSPANs of the flavors and the low-dose SYN of some rays, it is necessary to conduct a neurobehavioral toxicity study of the flavor mixtures.

From the above results, the three flavor chemicals and their mixture rays had high iSPANs, and their toxicological interactions will change with different exposure volumes. In a word, high toxicity sensitivity is an important characteristic of flavors.

Suggestions on Toxicity Tests of Chemicals with High iSPANs

This study showed that different flavor chemicals can produce different combined toxicities, toxicity sensitivities, and toxicological interactions when they formed mixtures. The new concept iSPAN established in this paper is not only suitable for the toxicity sensitivity test of a single chemical but also suitable for the toxicity sensitivity test of different types of mixture rays. It can be calculated according to the concentration–response relationship, but it does not depend on the concentration–response curve. The value of the iSPAN can be calculated by knowing only two effective concentration points. Through it, we can also roughly determine the shape of the concentration–response curve. At the same time, due to the high toxicity sensitivity and the high iSPAN, the slight changes in concentration will also change the toxicity significantly. Therefore, in production and daily life, we need to pay special attention to the consumption and collocation of flavors. More comprehensive tests are needed for more species of tested organisms. In this paper, toxicity evaluations of the three flavor chemicals and their mixture rays were carried out. However, there are thousands of flavor chemicals and their combinations in actual production. More complex mixture systems and mixture rays are needed for research.

For the design of mixture rays, as mentioned above, many scholars have been inclined to design in a single way.[46−48] In this study, we used EquRay[57] to design three binary mixture systems. The results show that the 15 mixture rays of the three systems show different toxicological interactions. It is proven again that the mixture rays with the same components and different mixture ratios may show different toxicological interactions.[75] Secondly, in the toxicity test, setting more experimental concentration points between LC20 and LC80 and reducing the change rate of concentration values of two adjacent concentration points are conducive to the result evaluation, especially for chemicals with high toxicity sensitivity and high iSPANs, because if the change rate of concentration values between the two concentration points is too large, it may lead to no effect at a low concentration level and a too significant effect at a high concentration level; however, in fact, other effects between the two concentrations will be ignored, which will greatly increase the uncertainty of the evaluation results.

Many studies have shown that the combined toxicities of mixture rays depend not only on the concentration of the mixture rays but also on the mixture ratio.[75−78] Time is also an important indicator of toxicity; if we ignore the effect of time, then the toxicity results will be biased.[79,80] Now, we found that different exposure volumes will also affect the toxicological interaction results obtained by the CI, which can be understood as that the toxicological interactions of flavors are sensitive to the slight changes in the exposure volume. The exact reason for this phenomenon is not yet clear, and the study only designed three smaller exposure volumes for testing; so, considering the effect of the exposure volume on evaluation results from the CI, in future research and exploration, when the CI is used to evaluate the toxicological interaction of toxic chemicals or mixture rays with high toxicity sensitivity, more different and larger exposure volumes need to be tested.

Conclusions

In this paper, three flavor chemicals BEA, CID, and PHA were selected, and three binary mixture systems were constructed. Five mixture rays were designed for each mixture system by using the EquRay. A new concept iSPAN was used to determine the toxicity sensitivities of the flavor chemicals and their mixture rays to the C. elegans at three exposure volumes. The toxicological interaction of C. elegans was evaluated by the CI. The results showed that the single chemicals and mixture rays had obvious toxic effects on C. elegans, the combined toxicities and toxicological interactions showed sensitivity caused by slight changes in concentration and the exposure volume, and C. elegans was more sensitive to flavors than some other toxic chemicals. When the mixture ratio of the mixture ray that was composed of the same components is different, the toxicity and toxicity sensitivity will change. Different exposure volumes of mixtures with the same components and mixture ratios will also lead to changes in toxicity sensitivity, especially toxicological interaction. Considering the impact of such chemicals on the ecological environment, when evaluating the toxicities of chemicals with high toxicity sensitivities, the change rate of concentration values at multiple concentration points between LC20 and LC80 should not be too large in order to obtain more accurate results. Moreover, when the CI is used to evaluate the toxicological interaction of mixture rays with high toxicity sensitivities, the small change of the experimental exposure volume will be an important factor affecting the results. In order to better evaluate the toxicities of these chemicals, it is necessary to design experiments under multiple exposure volumes. The toxicity sensitivities of chemicals should be a meaningful part of future toxicity research, and special attention should be paid to chemicals with high toxicity sensitivities.

Materials and Methods

Test Chemicals, Nematode Culture, and Mortality Tests

Three flavor chemicals, BEA, PHA, and CID, which were also components in the binary flavor mixtures, were purchased from Macklin (China). Their physical properties, the concentrations of stock, and the molecular structures are listed in Table . The maximum solubility of each flavor chemical was consistent with the concentration of the experimental stock solution. All solutions were prepared with Milli-Q water and stored at 4 °C before testing and prepared when using. All of them were soluble in water without adding a solvent. The solutions were all colorless and transparent that can be observed under a microscope.

The wild-type strains (N2) of C. elegans used in the experiments and the E. coli OP50 used as its food were all from the Institute of Medicine, Tongji University, and the culture method was the same as that in the literature.[81] The procedures of E. coli OP50 culture, nematode culture, subculture, age synchronization, and blank and treatment group design were the same as the literature.[43,62]

We set three exposure volumes of 100, 200, and 400 μL and kept the concentration of each concentration gradient consistent at the three exposure volumes. The experiments of exposure volumes of 100 and 200 μL were carried out in 96-well microplates, and those of the exposure volume of 400 μL were carried out in 48-well microplates. Each experiment included 12 concentration gradients with four parallel of each concentration. Every experiment included six blank control groups. A certain amount of a phase L4 C. elegans fluid was added to all wells to ensure that there were at least 20 nematodes in each well, and then, we counted the mortality after 24 h. No food was provided during the experiment.

Design of Various Binary Mixtures

Considering that there were numerous rays in a mixture system, so, it was unrealistic to test all the mixture rays. It was necessary to choose some representative mixture rays. In this study, three flavor chemicals, BEA, CID, and PHA, formed three binary mixture systems, BEA-CID, BEA-PHA, and PHA-CID. For every binary mixture system, five representative mixture rays (R1, R2, R3, R4, and R5) were selected by the direct equipartition ray (EquRay) method.[56] The mixture ratios of two components in 15 mixture rays of three binary mixture systems are shown in Table . For each mixture ray, 12 concentration gradients were designated by the dilution factor determined in the preliminary toxicity test.

Concentration–Response Fitting and Toxicological Interaction Evaluation

The concentration–mortality data obtained by the toxicity test can be fitted to the nonlinear Weibull function with two parameters (location α and shape β)where f(x) is the lethality to C. elegans and x is the concentration of a single chemical or a mixture ray. The determination coefficient (R2) and root-mean-square error (RMSE) were used to describe the goodness of fitting and 95% observation-based confidence intervals (OCIs) to express the uncertainty of the experimental observation and curve fitting.[43]

The combination index (CI)[42,43] was used to quantitatively evaluate the toxicological interactions in various binary mixtures. The CI equation is as followswhere m is the number of components in mixtures, EC is the concentration of the ith component that induces the x% effect when applied individually, and c is the concentration of the ith component in the mixture when inducing the x% effect. When the value of the CI is less than, equal to, and more than 1, it indicates that the mixture produces synergism (SYN), additive action (ADD), and antagonism (ANT), respectively.[42,43,82]

Toxicity Sensitivity of a Chemical or a Mixture Ray

To quantitatively and rationally describe the toxicity sensitivity of a chemical or a mixture ray with its concentration varying, a new concept, iSPAN, is proposed. The iSPAN is defined as the inverse of the negative logarithmic concentration span (iSPAN) of a chemical or a mixture ray inducing the lethality between 20 and 80%, that is, the iSPAN is based on the CRC of a chemical or a mixture ray.where p is an operator representing a negative logarithmic operation, i.e., p = −log10. From the definition of the iSPAN in eq , the larger the iSPAN of a chemical or a mixture ray is, the greater the toxicity sensitivity of the chemical or the mixture ray to the target organism will be, and the change of its toxicity is more significant with the slight change of effective concentration. Clearly, this new definition of the iSPAN is the toxicity sensitivity that the speed of toxicity changes of a single chemical or a mixture ray when the effective concentration is 20 and 80%.

All the above calculations, including the concentrations of each well, autoscaling treatment, design of mixture rays, concentration–response (lethality) curve (CRC) fitting, CI, pLC20, and pLC80, were finished by means of the APTox (assessment and prediction for the toxicity of chemical mixtures) program developed in our laboratory.[60]