论文部分内容阅读
目的:用极性表面积和分子体积预测药物对血脑屏障的穿透性。方法:用Monte Carlo法从半经验自洽场分子轨道AM1法得到最低能构型计算极性表面积和分子体积,用逐步多元回归法导出药物分子分别在脑组织和血液中的稳态浓度之比(logBB)和分子结构参数之间的回归方程式。结果:对于56个化合物组成的训练样本,logBB与氧原子和氮原子(不包括氮分子和硝基中的氮原子)的表面积之和(S_(O,N),A~2)以及分子体积(V,A~3)具有较好的相关性,回归方程式为:logBB=-1.331×10~(-5)V~2+9.228×10~(-3)V-0.02439S_(O,N)-0.4318(n=56,r=0.9043)。对于10个化合物组成的预测样本,预测值与实验值相当符合。结论:本模型简单有效,在药物设计中可以用来预测候选药物的logBB值。
OBJECTIVE: To predict the drug’s permeability to the blood-brain barrier by polar surface area and molecular volume. METHODS: The polar surface area and molecular volume were calculated from the semi-empirical self-consistent field orbital AM1 method using Monte Carlo method. The steady-state concentration ratio of drug molecules in brain tissue and blood was derived by stepwise multiple regression (logBB) and the molecular structure of the regression equation between the parameters. Results: For the training samples composed of 56 compounds, the sum of logBB (S_ (O, N), A ~ 2) and the surface area of oxygen and nitrogen (excluding the nitrogen in nitrogen and nitro) (V, A ~ 3), the regression equation is logBB = -1.331 × 10 -5 V -2 + 9.228 × 10 -3 V-0.02439S_ (O, N) -0.4318 (n = 56, r = 0.9043). Predictors of 10 compounds are in good agreement with the experimental data. Conclusion: This model is simple and effective and can be used to predict the logBB value of candidate drugs in drug design.