Probabilistic neural networks modeling of the 48-h LC50 acute toxicity endpoint to Daphnia magna.

Two modeling experiments based on the maximum likelihood estimation paradigm and targeting prediction of the Daphnia magna 48-h LC50 acute toxicity endpoint for both organic and inorganic compounds are reported. The resulting models computational algorithms are implemented as basic probabilistic neural networks with Gaussian kernel (statistical corrections included). The first experiment uses strictly D. magna information for 971 structures as training/learning data and the resulting model targets practical applications. The second experiment uses the same training/learning information plus additional data on another 29 compounds whose endpoint information is originating from D. pulex and Ceriodaphnia dubia. It only targets investigation of the effect of mixing strictly D. magna 48-h LC50 modeling information with small amounts of similar information estimated from related species, and this is done as part of the validation process. A complementary 81 compounds dataset (involving only strictly D. magna information) is used to perform external testing. On this external test set, the Gaussian character of the distribution of the residuals is confirmed for both models. This allows the use of traditional statistical methodology to implement computation of confidence intervals for the unknown measured values based on the models predictions. Examples are provided for the model targeting practical applications. For the same model, a comparison with other existing models targeting the same endpoint is performed.

An oral carcinogenicity and toxicity study of senna (Tinnevelly senna fruits) in the rat.

Senna (Tinnevelly senna fruits), a known laxative derived from plants, was administered by gavage to Sprague-Dawley (Crl:CD (SD) BR) rats once daily at dose levels of 0, 25, 100 and 300 mg/kg/day for up to 104 consecutive weeks. Based upon clinical signs related to the laxation effect of senna, the highest dose (300 mg/kg/day) was considered to be a maximum tolerated dose. Sixty animals per sex were assigned to the control and dose groups. Assessments included clinical chemistry, hematology, full histology (control and high-dose groups; in addition, low and mid dose: intestinal tract, adrenals, liver, kidneys, brain and gross lesions) and toxicokinetics. The primary treatment-related clinical observation was mucoid feces seen at 300 mg/kg/day. When compared to controls, animals administered 300 mg/kg/day had slightly reduced body weights, increased water consumption and notable changes in electrolytes in serum (increases in potassium and chloride) and urine (decreases in sodium, potassium and chloride). The changes in electrolytes are most likely physiologic adaptations to the laxative effect of senna. At necropsy, dark discoloration of the kidneys was observed in animals in all treated groups. Histological changes were seen in the kidneys of animals from all treated groups and included slight to moderate tubular basophilia and tubular pigment deposits. In addition, for all treated groups, minimal to slight hyperplasia was evident in the colon and cecum. These histological changes, together with the changes seen in the evaluation of clinical chemistry and urine parameters, have been shown to be reversible in a previous 13-week rat study of senna. No treatment-related neoplastic changes were observed in any of the examined organs. Based upon these data, it is concluded that senna is not carcinogenic even after daily administration for 2 years at dosages of up to 300 mg/kg/day in Sprague-Dawley rats.

A 13-week oral toxicity study of senna in the rat with an 8-week recovery period.

Senna was administered by gavage to Sprague Dawley rats once daily at dose levels of 0, 100, 300, 750 or 1500 mg/kg for up to 13 consecutive weeks followed by an 8-week recovery period for selected animals. Dose- and treatment-related clinical signs included abnormal feces, which were seen to varying degrees from animals at 300 mg/kg per day and more. Animals receiving 750 or 1500 mg/kg per day had significantly reduced body weight gain (males only) and, related to the laxative properties of senna, increased water consumption and notable changes in electrolytes in both serum and urine. At both the terminal and recovery phase necropsy, an increase in absolute and relative kidney weights was seen for male and female animals receiving 750 and/or 1500 mg/kg per day. A dark discoloration of the kidneys was observed at necropsy along with histopathological changes in the kidneys (slight to moderate tubular basophilia and pigment deposits) at 300 mg/kg and above. However, there were no indications in laboratory parameters of any renal dysfunction. In addition, for all treated groups, minimal to slight hyperplasia was recorded in the forestomach and large intestine. Following 8 weeks of recovery, with the exception of the brown pigment in the kidneys, there were no histopathological abnormalities. Thus, the biochemical and morphological changes seen following 13 weeks of treatment of senna significantly reversed following 8 weeks of recovery.