Mauderly J L, Kracko D, Brower J, Doyle-Eisele M, McDonald J D, Lund A K, Seilkop S K
Lovelace Respiratory Research Institute , Albuquerque, NM , USA .
Inhal Toxicol. 2014 Sep;26(11):691-6. doi: 10.3109/08958378.2014.947448.
An experiment was conducted to test the hypothesis that a mixture of five inorganic gases could reproduce certain central vascular effects of repeated inhalation exposure of apolipoprotein E-deficient mice to diesel or gasoline engine exhaust. The hypothesis resulted from preceding multiple additive regression tree (MART) analysis of a composition-concentration-response database of mice exposed by inhalation to the exhausts and other complex mixtures. The five gases were the predictors most important to MART models best fitting the vascular responses. Mice on high-fat diet were exposed 6 h/d, 7 d/week for 50 d to clean air or a mixture containing 30.6 ppm CO, 20.5 ppm NO, 1.4 ppm NO₂, 0.5 ppm SO₂, and 2.0 ppm NH₃ in air. The gas concentrations were below the maxima in the preceding studies but in the range of those in exhaust exposure levels that caused significant effects. Five indicators of stress and pro-atherosclerotic responses were measured in aortic tissue. The exposure increased all five response indicators, with the magnitude of effect and statistical significance varying among the indicators and depending on inclusion or exclusion of an apparent outlying control. With the outlier excluded, three responses approximated predicted values and two fell below predictions. The results generally supported evidence that the five gases drove the effects of exhaust, and thus supported the potential of the MART approach for identifying putative causal components of complex mixtures.
进行了一项实验,以检验以下假设:五种无机气体的混合物能够重现载脂蛋白E缺陷小鼠反复吸入柴油或汽油发动机尾气所产生的某些中心血管效应。该假设源于之前对吸入尾气和其他复杂混合物的小鼠组成-浓度-反应数据库进行的多重加法回归树(MART)分析。这五种气体是对最能拟合血管反应的MART模型最重要的预测因子。高脂饮食的小鼠每天暴露6小时,每周7天,持续50天,暴露于清洁空气或含有30.6 ppm一氧化碳、20.5 ppm一氧化氮、1.4 ppm二氧化氮、0.5 ppm二氧化硫和2.0 ppm氨的空气混合物中。气体浓度低于之前研究中的最大值,但处于导致显著影响的尾气暴露水平范围内。在主动脉组织中测量了五个应激和促动脉粥样硬化反应指标。暴露增加了所有五个反应指标,各指标的效应大小和统计学显著性有所不同,并取决于是否纳入一个明显的异常对照。排除异常值后,三个反应接近预测值,两个低于预测值。结果总体上支持了这五种气体导致尾气效应的证据,因此支持了MART方法在识别复杂混合物假定因果成分方面的潜力。