Efeito mutagênico da exposição aguda e crônica ao herbicida ácido 2,4-diclorofenóxiacético (2,4-D)

Abstract

Introduction: The herbicide 2,4-dichlorophenoxyacetic acid (2,4-D) was the first selective herbicide produced in 1940, belonging to the class of phenoxyacetics. Ingestion, breathing or dermal contact with this herbicide can cause loss of appetite, nausea, vomiting, muscle fasciculation and even cancer. Objective: To evaluate the mutagenic effect of acute and chronic exposure to the herbicide 2,4-D to rodent bone marrow cells and to evaluate the performance of learning algorithms in detecting differences in exposure groups, through recognition performed based on mutagenic characteristics . Material and methods: For the experiment, three concentrations of 2,4-D were used: 3.71 x 10-3 grams of active ingredient per hectare (g.i.a / ha) - considered low concentration; 6.19 x 10-3 g.i.a / ha - medium concentration; and 9.28 x 10-3 g.i.a / ha - high concentration. In the acute phase, 88 Swiss mice were used, randomly divided: C - Saline group; BG - Low Concentration Group; MC - Medium Concentration Group; AC - High Concentration Group and CPa - Positive control group for acute exposure. All mice were exposed to the recommended nebulizations for each group for 15 minutes and at different time intervals: 24, 48, 72 and 192 hours. For the experiment in the chronic phase, 88 albino Wistar rats were used, divided into: CI - Inhalation Control; CO - Oral Control; BCI: Low Inhalation Concentration; BCO - Low Oral Concentration; MCI - Average Inhalation Concentration; MCO - Average Oral Concentration; ACI - High Inhalation Concentration; ACO - High Oral Concentration and CPc - Positive control of chronic exposure. The animals of the first eight groups were euthanized 6 months after the beginning of the experiment and the animals of the CPc, 24 hours after. Femur bone marrow cells from acute and chronic rodents were collected for the micronucleus test and “comet assay”. For the statistical analysis, the Kruskal-Wallis test was applied, followed by the Dunn test. Pattern recognition was performed based on mutagenic characteristics (micronuclei and comets), using machine learning algorithms. The level of significance used for each test was 5%. The confusion matrix was created based on the data set divided into categories, the total of these results being displayed in a matrix and for the computational experiments 5 versions of the collected data set were considered. The data were evaluated by the following pattern recognition algorithms: K-Nearest Neighbors (k-NN), Support Vector Machine (SVM) and Decision Tree (DT). Results: In acute exposure, it was observed that medium and high concentrations induced DNA damage, regardless of the number of nebulizations by the comet assay, but not an increase in micronucleated cells. In chronic exposure, there was an increase in micronuclei and also damage to the comet assay DNA in all exposed groups, regardless of the exposure pathway. Machine learning, in the acute phase, was observed that the MC group presented very good accuracy (73%) using DT and CPa excellent accuracy (95%) using k-NN. In chronic exposure, we observed excellent accuracy in CO (100%) and CI (84%). The classes BCI + MCI + ACI + BCO + MCO + ACO (98%) and CPc (100%) showed excellent accuracy. Conclusion: Our data showed that both acute inhalation exposure and chronic oral and inhalation exposure to 2,4-D can cause genotoxic effect, regardless of concentration and route of exposure. In machine learning, the algorithm recognized the groups with excellent accuracy, showing that they adequately represent the groups exposed and not exposed to 2,4-D. Thus, greater care is needed in the application of 2,4-D in crops, since its use can cause health risk.