关键词: 菜豆, 抗营养因子, 细胞毒性, 毒性评价, 预测模型, 小鼠

Abstract: 【Objective】 The objective of this study is to analyze the toxicity of varieties of kidney beans in mice cells, and construct a toxicological prediction model of kidney beans, thus providing a basis for breeding low poisonous bean variety. 【Method】The maceration extract of 27 varieties of kidney beans whose content of four anti-nutritional factors had been measured were prepared by methods of homogenization, centrifugation and ultrafiltration. Mouse lymphocyte was isolated by density gradient centrifugation and treated with kidney bean extract. MTS assay was used for detecting the influence ratio of lymphocyte proliferation. A multiple regression equation was generated by SPSS statistical software, combined the content of four anti-nutritional factors with the influence ratio of lymphocyte proliferation, taken influence ratio of lymphocyte proliferation (Y) as dependent variable and the content of four anti-nutritional factors as independent variable. Then, the toxicological prediction model of kidney beans was established by stepwise regression analysis. An acute toxicity experiment was performed on healthy mice in order to verify the feasibility of toxicological prediction model of kidney beans. The mice were separated randomly into three groups. The mice in experimental group were given kidney bean extract by intragastric administration. At the same time, the mice in positive control group were received standard PHA, and the mice in negative control group were treated with only physiological saline. The survival time was checked every day. After death of the subjects, the biopsies of liver, spleen, kidney, stomach and intestine were taken for histological examination under the microscope using HE staining. 【Result】The results of MTS assay showed that the kidney bean extract expressed toxicity to mice lymphocyte, and the toxicity of varieties kidney beans was positively correlated with lectin content. There was a significant difference in the influence ratio of lymphocyte proliferation after treated by kidney bean extract, which was ranging from 13.77% to 85.23% (P＜0.01). A multiple regression equation (Y=-20.88+4.902 X₁+0.258 X₂-3.506 X₃+18.298 X₄) was generated by SPSS statistical software, combined the content of four anti-nutritional factors with the influence ratio of lymphocyte proliferation, taken influence ratio of lymphocyte proliferation (Y) as dependent variable and the content of four anti-nutritional factors (X) as independent variable. Then, the toxicological prediction model (Y=30.837+4.5 X₁) of kidney bean was constructed by stepwise regression analysis. The results of acute toxicity experiments showed that the fatality rate in a week was high in group with high influence ratio of lymphocyte proliferation. The mice that treated with inactivated kidney bean extract by heating could survive more than a week. It indicated that the main toxic substance of kidney bean was protein. The further study of anatomy and pathology analysis displayed that congestive, edema and the inflammatory reaction appeared in the liver, spleen, kidney, stomach, and intestine in experimental group mice, similar to the mice in positive control group. It illustrated that the major endogenous toxin of kidney bean was lectin. 【Conclusion】This study will provide a theoretical basis and technical methods for the genetic improvement of kidney bean quality and utilization.

Key words: kidney bean, anti-nutritional factors, cytotoxicity, evaluation of toxicity, predictive model, Mus musculus