Structure-toxicity relationships for benzenes evaluated with Tetrahymena pyriformis

Toxicity data for 200 substituted benzenes tested in the two-day Tetrahymena pyriformis population growth impairment assay representing the neutral narcosis, polar narcosis, respiratory uncoupling, and weak and strong electrophilic mechanisms of toxic action were evaluated. A quantitative structure-toxicity model correlating toxic potency [log(IGC(50)(-)(1))] with hydrophobicity quantified by the 1-octanol/water partition coefficient (log K(ow)) and electrophilic reactivity quantified by the molecular orbital parameter, maximum superdelocalizability (S(max)), was developed. This model [log(IGC(50)(-)(1)) = 0.50(log K(ow)) + 9.85(S(max)) - 3.47; n = 197, r(2) = 0.816, s = 0.34, F = 429, Pr > F = 0.0001] allows for the prediction of acute potency without the a priori identification of the mechanism of action. The examination of residuals reveals that neutral narcotics with high volatility (e.g., methyl- and chloro-substituted benzenes) and highly reactive fluoro- and nitro-containing derivatives are fitted poorly. A comparison of observed (obs) and predicted (pred) toxicities on the additional set of derivatives [log(obs IGC(50)(-)(1)) = 1.05[log(pred IGC(50)(-)(1))] + 0.02; n = 20, r(2) = 0.979, s = 0.13, F = 825, Pr > F = 0.0001] validated the model as a good predictor of toxicity regardless of the mechanism of toxic action.     

Structure-toxicity relationships for selected halogenated aliphatic chemicals

Toxicity to the ciliate Tetrahymena pyriformis (log(IGC(50)(-1))) for 39 halogen-substituted alkanes, alkanols, and alkanitriles were obtained experimentally. Log(IGC(50)(-1)) along with the hydrophobic term, logK(ow) (1-octanol/water partition coefficient) and the electrophilic parameter, E(lumo) (the energy of the lowest unoccupied molecular orbital) were used to develop quantitative structure-activity relationships (QSARs). Two strong hydrophobic dependent relationships were obtained: one for the haloalkanes and a second for the haloalcohols. The relationship for the haloalkanes [log(IGC(50)(-1))=0.92 (logK(ow))-2.58; n=4, r(2)=0.993, s=0.063, f=276, Pr>f=0.0036] was not different from baseline toxicity. With the rejection of 1,3-dibromo-2-propanol as a statistical outlier, the relationship [log(IGC(50)(-1))=0.63(logK(ow))-1.18; n=19, r(2)=0.860, s=0.274, f=104, Pr>f=0.0001] was observed for the haloalcohols. No hydrophobicity-dependent model (r(2)=0.165) was observed for the halonitriles. However, an electrophilicity-dependent model [log(IGC(50)(-1))=-1.245(E(lumo))+0.73; n=15, r(2)=0.588, s=0.764, F=18.6, Pr>f=0.0009] was developed for the halonitriles. Additional analysis designed to examine surface-response modeling of all three chemical classes met with some success. Following rejection of statistical outliers, the plane [log(IGC(50)(-1))=0.60(logK(ow))-0.747(E(lumo))-0.37; n=34, r(2)=0.915, s=0.297, F=162, Pr>F=0.0001] was developed. The halogenated alcohols and nitriles tested all had observed toxicity in excess of non-reactive baseline toxicity (non-polar narcosis). This observation along with the complexity of the structure-toxicity relationships developed in this study suggests that the toxicity of haloalcohols and halonitriles is by multiple and/or mixed mechanisms of action which are electro(nucleo)philic in character.     

TETRATOX: TETRAHYMENA PYRIFORMIS POPULATION GROWTH IMPAIRMENT ENDPOINTA SURROGATE FOR FISH LETHALITY

A short-term, static protocol using the common freshwater ciliate Tetrahymena pyriformis (strain GL-C) is reported. The 50% impairment growth concentration (IGC 50) is the endpoint of choice. Cultures are reared in 50 mL of a semi defined medium in 250-mL Erlenmeyer flasks. Definitive test replicates consist of a minimum of 5 different concentrations of each test material. Duplicate flasks are inoculated to an initial density of 2500 cells/ mL with log-growthphase ciliates. Following 40 h of incubation at 27 1 C, population density is measured spectrophotometrically and 50% effect levels are determined. Toxicity data for 250 chemicals representing several mechanisms of toxic action are reported and compared to fish toxicity data. As demonstrated by the relationship, log (IGC50-1) =0.77 log (LC50-1) - 0.40; r 2 =.750; s =.546; F =744, there is favorable similarity in toxic potency between T. pyriformis (log \[IGC50-1]) and fish, Pimephales promelas, survivability (log \[LC50-1]). However, toxicokinetics and/or toxicodynamic differences between the two systems result in either potency or mechanism of action variances for several chemical classes.     

QSAR models for predicting in vivo aquatic toxicity of chlorinated alkanes to fish

Quantitative structure-activity relationship (QSAR) models are expected to play a crucial role in reducing the number of animals to be used for toxicity testing resulting from the adoption of the new European Union chemical control system called Registration, Evaluation, and Authorization of Chemicals (REACH). The objective of the present study was to generate in vitro acute toxicity data that could be used to develop a QSAR model to describe acute in vivo toxicity of chlorinated alkanes. Cytotoxicity of a series of chlorinated alkanes to Chinese hamster ovary (CHO) cells was observed at concentrations similar to those that have been shown previously to be toxic to fish. Strong correlations exist between the acute in vitro toxicity of the chlorinated alkanes and (i) hydrophobicity [modeled by the calculated log K ow (octanol-water partition coefficient); r (2) = 0.883 and r int (2) = 0.854] and (ii) in vivo acute toxicity to fish ( r (2) = 0.758). A QSAR model has been developed to predict in vivo acute toxicity to fish, based on the in vitro data and even on in silico log K ow data only. The developed QSAR model is applicable to chlorinated alkanes with up to 10 carbon atoms, up to eight chlorine atoms, and log K ow values lying within the range from 1.71 to 5.70. Out of the 100204 compounds on the European Inventory of Existing Chemicals (EINECS), our QSAR model covers 77 (0.1%) of them. Our findings demonstrate that in vitro experiments and even in silico calculations can replace animal experiments in the prediction of the acute toxicity of chlorinated alkanes.     

The toxicity of acetylenic alcohols to the fathead minnow, Pimephales promelas: narcosis and proelectrophile activation

1. The 96-h LC50 values for 16 acetylenic alcohols in the fathead minnow (Pimephales promelas) were determined using continuous-flow diluters. The measured LC50 values for seven tertiary propargylic alcohols agreed closely with the QSAR predictions based upon data for other organic non-electrolytes acting by a narcosis mechanism. 2. Four primary and four secondary propargylic alcohols were 7 to 4600 times more toxic than the respective narcotic toxicity estimated by QSAR. Metabolic activation to electrophilic alpha,beta-unsaturated propargylic aldehydes or ketones is proposed to account for the increased toxicity. 3. 3-Butyn-1-ol and 4-pentyn-2-ol, primary and secondary homopropargylic alcohols, were 320 and 160, respectively, times more toxic than predicted. In this case an activation step involving biotransformation to an allenic electrophile intermediate was proposed.     

The Toxicity of Acetylenic Alcohols to the Fathead Minnow,Pimephales Promelas: Narcosis and Proelectrophile Activation

1. The 96-h LC₅₀ values for 16 acetylenic alcohols in the fathead minnow (Pimephales promelas) were determined using continuous-flow diluters. The measured LC₅₀ values for seven tertiary propargylic alcohols agreed closely with the QSAR predictions based upon data for other organic non-electrolytes acting by a narcosis mechanism.

2. Four primary and four secondary propargylic alcohols were 7 to 4600 times more toxic than the respective narcotic toxicity estimated by QSAR. Metabolic activation to electrophilic α,β-unsaturated propargylic aldehydes or ketones is proposed to account for the increased toxicity.

3. 3-Butyn-1-o1 and 4-pentyn-2-ol, primary and secondary homopropargylic alcohols, were 320 and 160, respectively, times more toxic than predicted. In this case an activation step involving biotransformation to an allenic electrophile intermediate was proposed.     

Aquatic toxicity and abiotic thiol reactivity of aliphatic isothiocyanates: Effects of alkyl-size and -shape

Aquatic toxicity data in the TETRATOX assay and reactivity data in an abiotic thiol assay were collected for a series of aliphatic isothiocyanates. These compounds can act as Michael-type acceptors with N-hydro-C-mercapto-addition to cellular thiols as a molecular mechanism of action. Comparison of both toxicity and reactivity among the analogues revealed that derivatives with a branch hydrocarbon moiety, especially branched in the β-position were less toxic and less reactive. In contrast, the di-isothiocyanate and the allyl and propargyl derivatives are more toxic than their 1-alkyl homologues. The toxicity and reactivity differences are consistent but except for the tert-butyl-derivative not remarkable. The differences are due to variations in steric hindrance at the reaction center. For the mono-isothiocyanates compounds toxicity (IGC(50)) is linearly related to thiol reactivity (EC(50)): log(1/IGC(50))=1.33(log(1/EC(50)))-0.41; n=23, s=0.24, r(2)=0.911, q(2)=0.907, F=215.     

A quantitative structure--activity relationships model for the acute toxicity of substituted benzenes to Tetrahymena pyriformis using Bayesian-regularized neural networks

We have used a new, robust structure-activity mapping technique, a Bayesian-regularized neural network, to develop a quantitative structure-activity relationships (QSAR) model for the toxicity of 278 substituted benzenes toward Tetrahymena pyriformis. The independent variables used in the modeling were derived solely from the molecular structure, and the model was tested on 20% of the data set selected from the whole set by cluster analysis and which had not been used in training the network. The results show that the method is robust and reliable and give results for mixed class compounds which are comparable to earlier QSAR work on single-chemical class subsets of the 278 compounds and which employed measured physicochemical parameters as independent variables. Comparisons of Bayesian neural net models with those derived by classical PLS analysis showed the superiority of our method. The method appears to be able to model more diverse chemical classes and more than one mechanism of toxicity.     

A Qsar Study Of Anti-Inflammatory N-Arylanthranilic Acids

A detailed quantitative structure-activity relationship (QSAR) analysis of a series of 112 anti-inflammatory N-arylanthranilic acids has been performed to determine which physicochemical properties of these compounds are responsible for their anti-inflammatory activity. The results indicate that activity is modelled best by molecular shape parameters. The angle between the planes of the two benzene rings, dictated by the substitution pattern of the compounds, also appears relevant to activity. Dipole moments show some significance, but log P and other physicochemical parameters correlate poorly with activity. The best QSAR obtained was: log(1/MED) = 1.601B₁₍₂₎ + 0.576B₁₍₃₎ ? 1.187B₃₍₄₎ + 0.522B₁₍₆₎ ? 1.681 μ(bond)₍₂₎ ? 0.208 μ(bond)₍₅₎ ? 0.265 μ (bond)₍₆₎ ? 0.226 n = 112, r = 0.855, r² (adj) = 0.716, F = 40.33, s = 0.511, r(CV)² = 0.665 where B₁ and B₃ are Verloop substituent width parameters and μ(bond) is bond dipole (position in parentheses).     

Toxicity assessment of organic pollutants: Reliability of bioluminescence inhibition assay and univariate QSAR models using freshly prepared Vibrio fischeri

The toxicity of 14 industrially relevant organic chemicals was determined using freshly grown Vibrio fischeri bioluminescence inhibition assay. The results were compared to lyophilized V. fischeri, 96h fish, 48h Daphnia magna and 95h green algae bioassays. Reliability of octanol-water partition coefficient (K(ow)), and first order simple and valence molecular connectivity index ((1)chi, (1)chi(v)) based regression models for predicting toxicity to V. fischeri was studied. Correlations were obtained between freshly grown V. fischeri data (Log(EC(50))) and Log(K(ow)), molecular connectivity indices ((1)chi, (1)chi(v)), energy of the highest occupied (E(HOMO)) and lowest unoccupied (E(LUMO)) molecular orbitals, and their difference (E(LUMO)-E(HOMO)). A good match was observed between V. fischeri assay conducted with freshly grown and lyophilized culture (r(2)=0.90). Good correlations (r(2)>0.95) were obtained with all the other bioassays after excluding compounds with Log(K(ow)) less than 2.0. Available regression models based on Log(K(ow)) and (1)chi(v) yielded lower toxicity values. V. fischeri bioassay showed fairly good correlation with Log(K(ow)), (1)chi and (1)chi(v) (r(2)>0.75) but poor correlation with E(HOMO), E(LUMO) and (E(LUMO)-E(HOMO)) in presence of polar compounds. E(HOMO) and E(LUMO) values are affected by polarity and can be used along with Log(K(ow)) and (1)chi(v) for generating better predictive models.