CC BY
20
UDC 613.26:632.951:616-036.3-037(047.37)
https://doi.Org/10.26641/2307-0404.2019.3.181892
A.M. Antonenko 1, O.P. Vavrinevych 1, S.T. Omelchuk 2,
B.I. Shpak 3
HYGIENIC SUBSTANTIATION OF CALCULATING MODELS FOR PREDICTING TOXICITY
OF DIFFERENT CLASSES INSECTICIDES
(first part)
Bogomolets National Medical University Hygiene and ecology department N11 Hygiene and ecology institute 2 Peremohy av., 34, Kyiv, 03057, Ukraine «Syngenta» LCC 3
Kozatska str., 120/4, Kyiv, 02000, Ukraine
Нацюнального медичного yuieepcumem M. О.О. Богомольця
кафедра гтени та екологИ № 11
1нститут гтени та екологИ2
пр. Премоги, 34, Кш'в, 03057, Украша
ООО «Сингента» 3
вул. Козацька, 120/4, Кшв, 02000,Ukraine e-mail: antonenko1985@ukr.net
Цитування: Медичт перспективы. 2019. Т. 24, № 3. С. 106-112 Cited: Medicni perspektivi. 2019;24(3):106-112
Key words: insecticide, toxicology, calculation models, regression equations Ключовi слова: шсектицид, токсикологiя, pospaxyuKoei моделi, pieuMHHM регреси Ключевые слова: инсектицид, токсикология, расчетные модели, уравнение регрессии
Abstract. Hygienic substantiation of calculating models for predicting toxicity of different classes insecticides (first part). Antonenko A.M., Vavrinevych O.P., Omelchuk S.T., Shpak B.I. This work is the first part of our study to develop alternative experimental mathematic models for predicting toxicity of insecticides. In the first stage, calculations will be carried out and the most reliable models will be proposed. In the second - a statistical analysis and comparative estimation of the toxicometric parameters obtained experimentally and calculated according to the proposed equations. The purpose of the research is the scientific substantiation of the calculation models for predicting toxicity of insecticides of different classes. Data on the physico-chemical properties and toxicometry parameters of fungicides are taken from the PPDB pesticides database. Insecticides of such chemical classes as derivatives of tetram and tetronic acids, benzoylureas, carbamates, neonicotinoids, pyrethroids, organophosphorus compounds, avermectins were selected for analysis. It has been established that there is a significant positive correlation between NO(A)EL in the chronic experiment of all insecticides, the median lethal doses at oral administration (LD50 per os) ofpyrithoids and neonicotinoids, and the molecular weight (at p<0.05). There is a significant negative correlation between the toxicometry parameters of all insecticides and their individual groups (pyrithoids, neonicotinoids, organophosphorus compounds) and melting temperature and the octanol-water partition coefficient, log Po/w (at p<0.05). It is proved that the proposed calculation models for predicting insecticide hazards are adequate according to Fisher's criterion, and the coefficients of regression equations are reliable according to Student's criterion (p<0,05).
Реферат. ГМешчне обгрунтування розрахункових моделей прогнозування токсичносп iнсектицидiв рпних клаав (перша частина). Антоненко А.М., Вавршевич О.П., Омельчук С.Т., Шпак Б.1. Ця робота е першою частиною нашого до^дження з розробки альтернативних експериментальних математичних моделей для прогнозування токсичностi iнсектицидiв. На першому етат будуть проведет розрахунки й запропонован найбшьш надiйнi моделi. На другому - буде проведено статистичний аналiз i порiвняльна оцтка токсикометричних параметрiв, отриманих експериментально i розрахованих за пропонованими рiвняннями. Метою до^дження було наукове обгрунтування розрахункових моделей прогнозування токсичностi iнсектицидiв рiзних клаав. Дам про фiзико-хiмiчнi властивостi та токсикометричм параметри фунгiцидiв взятi з бази даних пестицидiв PPDB. Для аналiзу були обран iнсектициди таких хiмiчних клаав: похiднi тетрамово'1 i тетроново'1 кислот, бензоысечовини, карбамати, неонiкотиноïди, пiретроïди, фосфорорга^чн сполуки, авермектини. Встановлено, що кнуе достовiрний прямий кореляцшний зв'язок мiж NO(A)EL у хромчному експериментi всiх iнсектицидiв, середньосмертельними дозами при пероральному надходженнi (ЛД50 per os) пiритоïдiв i неонiкотиноïдiв i молекулярною масою. 1снуе достовiрний негативний кореляцшний
МЕДИЧН1ПЕРСПЕКТИВИ / MEDICNIPERSPEKTIVI
зв'язок мiж параметрами токсикометри ecix iнсектицидiв та Их окремих груп (тритоИди, неоткотиногди, фосфороргатчт сполуки) i температурою плавлення та коефщентом розподшу октанол-вода, log Po/w. Доведено, що запропонованi розрахунковi моделi для прогнозування небезпеки iнсектицидiв е адекватними за критерieм ФШера, а коефiцiенти рiвнянь регресИ е надшними за критерiем Стьюдента.
Today's pesticides are a very effective in increasing yields, but at the same time a serious risk factor for human health because of contamination of environmental objects [6, 7]. Methods for determining the toxicological parameters of pesticides are long-term, labor-intensive and require significant financial and resource costs [9, 14]. In solving this problem, the important role is played by methods of mathematical modeling and predicting of toxicity of xenobiotics, the results of which can be used both for substantiation of toxicological parameters, and at the stage of experiment planning, this will reduce the probability of error and reduce the duration of the study [9].
Foreign laboratories and institutes have been using modeling dependence of toxicometric parameters on the physico-chemical properties of xenobiotics for a long time [5, 11].
In Ukraine today, there are such models for fungicides and herbicides proposed by the specialists of the Hygiene and Ecology Institute of Bogomolets National Medical University [10, 15], as for insecticides, only attempts were made to develop hazard prediction models for prognosis of the risk of neoni-cotinoid insecticides [2].
This work is the first part of our study to develop alternative experimental mathematic models for predicting toxicity of insecticides. At the first stage, calculations will be carried out and the most reliable models will be proposed. At the second - a statistical analysis and comparative estimation of the toxicometric parameters obtained experimentally and calculated according to the proposed equations.
The purpose of the research is the scientific substantiation of the calculation models for predicting toxicity of insecticides of different classes.
MATERIALS AND METHODS OF RESEARCH
For the development and substantiation of the calculation models in the hygienic assessment of the studied insecticides hazard, an array of experimentally established values of LD50 (median lethal dose) after oral and percutaneous administration, LC50 (median lethal concentration) at inhalation and NO(A)EL (no observed (adverse) effect level) [12], physical and chemical properties (molecular mass, water solubility, vapor pressure, melting point, octanol-water partition coefficient (LogP Ko/w), surface tension) was taken. Data on the physico-chemical properties of fungicides are taken from the PPDB pesticides database [12].
For analysis, we selected insecticides of chemical classes the most widely used in the world [4, 8]: 3 compounds of derivatives of tetram and tetronic acids class (spiromesifen, spirodiclofen and spiro-tetramat); 3 benzoyl-ureas (diflubenzuron, navalu-rone, teflubenzuron); 4 compounds of the carbamates class (carbosulfan, methomyl, carbaryl, pheno-xycarb); 5 substances of the neonicotinoid class (acetamiprid, thiacloprid, imidacloprid, thiame-thoxam, clothianidin); 12 active ingridients - pyre-throids (zeta-cypermethrin, lambda-cyhalothrin, beta-cyflutrin, tau-fluvalinate, gamma-cyhalothrin, tefluthrin, bifentrine, esfenvalerate, beta-cypermethrin, cypermethrin, deltamethrin, alpha-cypermethrin); 6 organophosphorus compounds (chlorpyrifos-methyl, dimethoate, malathion, fosalon, pyrimifos-methyl, phenythrothion); 2 avermectins (abamectin and ema-mectin benzoate).
Statistical processing of the results was performed using the package of licensed statistical software IBM SPSS StatisticsBase v.22 and MS Excel. Correlation and regression analyses were carried out, taking into account the determination coefficient, which most closely approximates the connection between the selected toxicological parameters and physical-chemical properties. The significance of the obtained regression equations was checked by Fisher's F-criterion, the reliability of individual coefficients in the regression equation (a, b) -according to the Student's t-criterion.
RESULTS AND DISCUSSION
Initially, by the Pirson method we analyzed the correlation relations between the physical and chemical properties of the derivatives of tetram and tetronic acids, benzoylureas, carbamates, neonico-tinoids, pyrethroids, organophosphorus compounds, avermectins and their parameters of toxicometry. Statistically significant results are given in Table 1.
The results of the correlation-regression analysis revealed a reverse (negative) correlation between the "octanol-water" partition coefficient and LD50 percut. and NO(A)EL of all insecticides and LD50 per os of neonicotinoids (r= -0.47; -0.39 and -0.83, respectively, at p<0.05) on the array of 35, 28 and 5 active ingredients of the corresponding classes. Using the determination coefficient (R2), the proportion of the effect of the investigated factor on the parameters of toxicometry was determined and it was established that the share of influence of the
19/ Том XXIV/ 3
107
«octanol-water» distribution coefficients was 22.1; 15.0 and 69.0%, respectively.
The obtained results can be explained by the fact that fat-soluble compounds are slowly metabolized and eliminated from the body, tend to accumulate [13], this increases their toxicity and reduces NO(A)EL and LD50. This dependence is typical, since log Po/w is an indicator of the bioavailability of toxic substances [1, 3]. At the same time, for herbicides, the opposite dependence was found: their toxic manifestations decreased with increasing log Po/w, which can be explained by the presence of
alternative routes of herbicide penetration into the organism due to slightly lower molecular weight and higher solubility in water [10].
There is an inverse (negative) correlation between the melting point and LC50 in inhalation and NO(A)EL of organophosphorus compounds, as well as LD50 per os of pyrethroids (r= -0.88; -0.92 and -0.62, respectively at p<0,05) on an array of 5 and 12 active substances of the corresponding classes. Percentage of impact of the melting temperature on the parameters of toxicometry is quite high and makes 78.0; 84.0 and 38.8%, respectively.
Table 1
Correlation between the parameters of toxicometry and physical and chemical properties of studied insecticides
Chemical class Resulting variable Factor variable Statistical parameters*
correlation coefficient determination coefficient, % n
Insecticides LD50 percut., mg/kg log Po/w -0.47 22.1 3Й
NO(A)EL, mg/kg -0.39 15.0 28
NO(A)EL, mg/kg molecular weight 0.51 25.8 27
Pyrethro-ids LD 50 per os, mg/kg melting temperature, °C -0.62 38.8 12
molecular weight 0.72 51.3 11
Oganophosphor NO(A)EL, mg/kg melting temperature, °C -0.92 84.0 s
us compounds
LC 50 inhal., mg/m3 -0.88 78.0 s
Neoni-cotinoids LD 50 per os, mg/kg log Po/w -0.83 69.0 s
molecular weight 0.85 71.6 s
Notes: «*» - significant results are given (at p <0,05); n - number of observations; log Po/w - the logarithm of octanol-water partition coefficient; LD50 per os - median lethal dose after oral consumption; LD50 per cut - median-lethal dose when applied to the skin; LC50 inhal. - median lethal concentration at inhalation; NO(A)EL - no observed (adverse) effect level.
A similar correlation of the melting point with the parameters of toxicometry (with an increase in the melting point LD50 per os and inhal., NO(A)EL, values decrease, it means that toxicity of these compounds increases) was also found in the chemical class of herbicides pyrazoles-triketones [10]. One of the reasons for this effect may be the need for a higher temperature for the degradation of the mentioned compounds. At body temperature they retain their initial physical and chemical
properties longer and, accordingly, have more time to implement toxic manifestations.
Positive (direct) correlation was found between molecular weight and LD50 per os of pyrethroids (n=11) and neonicotinoids (n=5); NO(A)EL of all insecticides without phosphorus compounds (n=27) (r=0.72; 0.85 and 0.51, respectively, at p<0.05). The proportion of the influence of the molecular weight, as well as the melting point, is quite high and amounts to 51.3; 71.6 and 25.8%, respectively.
MEffHHHI nEPCnEKTHBH / MEDICNIPERSPEKTIVI
Detected dependence is due to the fact that compounds with very high molecular weight poorly penetrate into the body, in contrast to low molecular weight compounds that easily pass through histohemic barriers [1, 3, 13]. It should be noted that for fungicides, the relationship between molecular weight and toxicometry parameters is reversible, since most of them, in contrast to insecticides, form
isomers that greatly increase the specificity of their action and toxicity [15].
The next step was to carry out a regression analysis with the verification of the obtained equations (tabl. 2, 3) according to Fisher's F statistics, and separate coefficients in the regression equations (a, b) - according to Student's t-criterion.
Table 2
Models of predicting insecticides hazard (linear equations)
Chemical class n N Regression equations Model adequacy indices Coefficients reliability indices
Fisher's criterion Approxim ation accuracy (R2) a b 4 ** iKp.
F Fkp. ** t t
Insecticides 35 1 LD50 percut. = 3.1136Xi + 3371 9.34* 4.15 0.003 3.51* 3.06* 2.04
27 2 NO(A)EL = -0.0038X2 + 5.0773 8.68* 4.24 0.113 4.26* 2.95* 2.06
28 3 NO(A)EL = -0.6959X1 + 6.4644 4.58* 4.22 0.150 4.10* 2.14* 1.71
Pyrethroids 11 4 LD 50 per os = 4.1975X2 - 1690.1 9.48* 5.12 0.513 2.86* 3.08* 2.26
12 5 LD 50 per os = -2.6492X3 + 268.28 6.33* 4.96 0.388 3.99* 2.52* 2.23
Oganophos- 5 6 NO(A)EL = -0.405X3 + 18.196 15.69* 10.13 0.840 4.53* 3.96* 3.18
phorus
compounds
5 7 LC50 inhal. = -0.0541X3 + 3.7237 10.64* 10.13 0.913 6.15* 3.26* 3.18
Neonico- 5 8 LD 50 per os = 21.016X2 - 4844.9 7.57 10.13 0.716 2.48 2.75 3.18
tinoids
5 9 LD50 per os = -981.92X1 + 1172.1 6.66 10.13 0.690 3.74* 2.58 3.18
Notes: «*» - statistical significant results; «**» - at p=0.05 and k1=1, k2=n-2); n - number of observations; X1 - octanol-water partition coefficient, log Po/w; X2 - molecular weight; X3 - melting temperature, °C.
The results of the regression analysis given in Table 2 indicate that the 4 derived linear regression equations N 1-14 are adequate according to Fisher's criterion (p<0.05).
The assessment of coefficients «a» and «b» reliability showed that in all regression equations, they are significant according to Student's t-criterion (p<0,05), except for the equations N 8 and 9. In this equations, the coefficients «a» and / or «b» were unreliable, since the absolute value of the criterion t
is less than tcr, which may be due to a small number of observations (n=5).
We also used the exponential, logarithmic, polynomial and step functions in addition to the linear one. In Table 3 the obtained mathematical models are given. Except for the formulas 10, 11, 17, 18, 19, the rest showed rather high indicators of the adequacy of the model by the coefficients of approximation.
19/ TOM XXIV/ 3
109
Table 3
Models of predicting insecticides hazard (nonlinear equations)
Insecticides
Pyrethroids
Oganophos-phorus compounds
Neonicotinoids
35 35 27 27 27
27
28 28 28 28
11 11 11 11 12 12
5 5 5 5
5 5 5 5 5 5
10 11 12
13
14
15
16
17
18
19
20 21 22
23
24
25
26
27
28
29
30
31
32
33
34
35
LD50 percut. = 2247.1e-0037X1 LD50 percut. = 59.94Х12 - 418.17Х1 + 2731 NO(A)EL = 3.8947e-0002X2 NO(A)EL = -3.782ln(X2) + 25.928 NO(A)EL = 9572.6 X2-1447 NO(A)EL = 9х10-6х22 - 0.0223Х2 + 10.655 NO(A)EL = -О.1125Х12 + 0.1261Х1 + 5.5048 NO(A)EL = 3.7507e-0182X1 NO(A)EL = -0.404ln(X1) + 3.8851 NO(A)EL = 1.7218X1-"."14
LD50 per os = 0.1234X22 - 108.4X2 + 23844 LD50 per os = 0.029e00184X2 LD50 per os = 1864.9ln(X2) - 11191 LD50 per os = 3x10-20X281429 LD50 per os = 160.23e-0012X3 LD50 per os = 0.0418X32 - 6.2181X3 + 291.86
NO(A)EL = 5.7067e-0059X3 NO(A)EL = 0.0104X32 - 0.7223X3 + 11.339 LC50 inhal. = 0.0006X32 - 0.073X3 + 3.3139 LC50 inhal. = 3.2317e-0022X3
LD50 per os = 0.4334X22 - 203.01X2 + 23888 LD50 per os = 0.0567e00338x LD50 per os = 5239.7ln(x) - 28501 LD50 per os = 1E-18x84697 LD50 per os = 1170.4X12 - 2217.4X1 + 1220.3 LD50 per os = 804.89e-1 424x
0.015 0.044 0.222 0.216 0.265 0.273 0.165 0.086 0.017 0.0002
0.818 0.300 0.490 0.285 0.248 0.520
0.696 0.996 0.947 0.754
0.919 0.619 0.676 0.592 0.890 0.486
Notes: n - number of observations; Xi - log Po/w; X2 - molecular weight; X3 - melting temperature, °C.
CONCLUSIONS
It has been established that there is a significant positive correlation between NO(A)EL in the chronic experiment of all insecticides, the median
lethal doses in oral admission (LD50 per os) of pyrithoids and neonicotinoids, and the molecular weight (at p<0.05).
MЕДИЧHIПЕРСПЕКТИВИ / MEDICNIPERSPEKTIVI
1. There is a significant negative correlation between the toxicometry parameters of all insecticides and their individual groups (pyrithoids, neonico-tinoids, organophosphorus compounds) and the melting temperature and the octanol-water partition coefficient, log Po/w (at p<0.05).
2. It is proved that the proposed calculation models for predicting insecticide hazards are adequate according to Fisher's criterion, and the coefficients of regression equations are reliable according to Student's criterion (p<0,05).
REFERENCES
1. Batyan AN, Frumin GT, Bazylev VN. [Basics of General and Environmental Toxicology]. S-Pb. SpetsLit. 2009;352. Russian.
2. Yermolova LV, Prodanchuk MG, Leposhkin IV. [Development of calculation models for forecasting the risk of neonicotinoid insecticides]. Sovremennye prob-lemy toksikologii. 2007;1:27-29. Ukrainian.
3. [General toxicology]. Edited BA Kurlyandskiy, VA Filova. Moskva. Meditsina. 2003;608. Russian.
4. [List of pesticides and agrochemicals authorized for use in Ukraine. Official edition]. Kyiv, Yunivest Media. 2018;1036. Ukrainian.
5. Anton C. Modeling and simulation for toxicity assessment. Math Biosci Eng. 2017;14(3):581-606. doi: https://doi.org/10.3934/mbe.2017034
6. Stavnichenko PV, Novohatska LO, Antonen-ko AM, et al. Assessment of ecotoxicological hazard and risk of groundwater contamination with different groups of pesticides. Medicni perspektivi. 2017;XXII(2):119-25. doi: https://doi.org/10.26641/2307-0404.2017.2.109845
7. Novohatska OO, Stavnichenko PV, Kondra-tiuk MV, et al. Comparative hygienic evaluation of behavior of different pesticides groups in soil, prediction of risk of ground water contamination and its danger for human health in areas with irrigation farming. Rawal Medical Journal. 2018;43(1):129-36.
8. EU - Pesticides database: Maximum Residue Levels. [cited 2019.06.20] Available from: http://ec.europa.eu/food/plant/pesticides/max_residue_l evels/index en.htm
9. Knudsen TB, Keller DA, Sander M, et al. Futu-reTox II: in vitro data and in silico models for predictive toxicology. Toxicological Sciences. 2015;143(2):256-67. doi: https://doi.org/10.1093/toxsci/kfu234
10. Antonenko AM, Vavrinevych OP, Korshun MM, et al. Hygienic substantiation of calculation models for toxicity prognosis of different herbicides classes. Sbornik nauchnykh trudov «Zdorove i okruzhayushchaya sreda». 2018;28:168-75.
11. Norto Â, Sathis1 J, Webb S, et al. Mathematical Modelling of Chronic Drug Infusion for Toxicity Assessment. UK Mathematics-in-Medicine NC3Rs Study Group. 2013;25.
12. PPDB: Pesticides property database. [cited 2019.06.20]. Available from: http://www.rupest.ru/ppdb
13. Pradhan A, Markande SK, Kurre RK. Evaluation of impact of pesticides on the basis of their physico-chemical properties. Journal of Industrial Pollution Control. 2014;30(2):223-6.
14. Raies AB, Bajic VB. In silico toxicology: computational methods for the prediction of chemical toxicity. Wiley Interdiscip Rev. Comput. Mol. Sci. 2016;6(2):147-72. doi: https://doi.org/10.1002/wcms.1240
15. Vavrinevych OP, Antonenko AM, Korshun MM, Omelchuk ST. Hygienic substantiation of calculating models for fungicides of different classes toxicity depend on their physical and chemical properties prognosis. Environment and health. 2017;4(84):52-57. doi: https://doi.org/10.21303/2585-663.2017.00441
СПИСОК Л1ТЕРАТУРИ
1. Батян А. Н., Фрумин Г. Т., Базылев В. Н. Основы общей и экологической токсикологии. Санкт-Петербург: СпецЛит, 2009. 352 с.
2. Ермолова Л. В., Проданчук М. Г., Лепьош-кш I. В. Розробка розрахункових моделей прогнозу небезпечностi неошкотинощних iнсектицидiв. Современные проблемы токсикологии. 2007. № 1. С. 27-29.
3. Общая токсикология / ред. Б. А. Курлянд-ского, В. А.Филова. Москва: Медицина, 2003. 608 с.
4. Перелж пестицидiв i агрохiмiкатiв, дозволе-них до використання в Украiнi: офiцiйне видання. Кшв: Юнiвест Медiа, 2018. 1036 с.
5. Anton C. Modeling and simulation for toxicity assessment. Math Biosci Eng. 2017. Vol. 14, No. 3. Р. 581-606. doi: https://doi.org/10.3934/mbe.2017034
6. Assessment of ecotoxicological hazard and risk of groundwater contamination with different groups of pesticides / Р. V. Stavnichenko et al. Медичт пер-спективи. 2017. Том XXII, № 2. C. 119-125. DOI: https://doi.org/10.26641/2307-0404.2017.2.109845
7. Comparative hygienic evaluation of behavior of different pesticides groups in soil, prediction of risk of ground water contamination and its danger for human health in areas with irrigation farming / O. O. Novohatska et al. Rawal Med. J. 2018. Vol. 43, No. 1. Р. 129-136.
8. EU - Pesticides database: Maximum Residue Levels.
URL:http://ec.europa.eu/food/plant/pesticides/max_residu e_levels/index_en.htm. (date of access 20.06.2019).
19/ Том XXV/ 3
111
9. FutureTox II: in vitro data and in silico models for predictive toxicology / T. B. Knudsen et al. Toxicological Sciences. 2015. Vol. 143, No. 2. Р. 256267. DOI: https://doi.org/10.1093/toxsci/kfu234
10. Hygienic substantiation of calculation models for toxicity prognosis of different herbicides classes / A. M. Antonenko et al. Сборник научных трудов «Здоровье и окружающая среда». 2018. Вып. 28. С. 168-175.
11. Mathematical Modelling of Chronic Drug Infusion for Toxicity Assessment / A. Norto et al. UK Mathe-matics-in-Medicine NC3Rs Study Group. 2013. 25 р.
12. PPDB: Pesticides property database. (date of access 20.06.2019). URL: http://www.rupest.ru/ppdb..
13. Pradhan A., Markande S. K., Kurre R. K. Evaluation of impact of pesticides on the basis of their physico-
chemical properties. J. Industrial Pollution Control. 2014. Vool. 30, No. 2. Р. 223-226.
14. Raies A. B., Bajic V. B. In silico toxicology: computational methods for the prediction of chemical toxicity. Wiley Interdiscip Rev. Comput. Mol. Sci. 2016. Vol. 6, No. 2. Р. 147-172.
DOI: https://doi.org/10.1002/wcms.1240
15. Vavrinevych O. P., Antonenko A. M., Kor-shun M. M., Omelchuk S. T. Hygienic substantiation of calculating models for fungicides of different classes toxicity depend on their physical and chemical properties prognosis. Довтлля i здоров'я. 2017. Т. 84, № 4. С. 5257. DOI: https://doi.org/10.21303/2585-663.2017.00441
Стаття надшшла до редакцп 24.06.2019
♦
