Lipophilicity determination of active pharmaceutical ingredients is an important step in drug discovery since lipophilicity affects the absorption, distribution, metabolism, excretion, and toxicity of drugs (ADMET properties). When we mention lipophilicity, quantitative determination with the logarithm of partition coefficient – logP comes to mind. Even though we now know many computer-based models for calculating logP based on the chemical structure, these results come with a calculation error in many cases. On top of that, we know many different models with different logP results for the same compound. The most traditional method for experimental determination of logP is the shake flask method which is still the gold standard for logP determination even though it is not time or cost-effective. In this case, it involves a direct determination of lipophilicity. However, we also know indirect determination. In this master’s thesis, we focused on the indirect determination of lipophilicity with the gradient reversed-phase high-pressure liquid chromatography (HPLC) method where we determined the chromatographic hydrophobic index (CHI). CHI value is usually between 0 and 100. The higher the CHI value, the more hydrophobic the analyte is. In comparison to the shake flask method, the HPLC method is faster, more sensitive, and more selective, and it also requires less sample volume.
In some articles, we also came across a UHPLC method which was presented as a faster and more optimized method in comparison to HPLC. However, the data for UHPLC determining lipophilicity was scarce. Therefore, we decided to optimize the method ourselves. We determined the logD – logarithm of the distribution coefficient which takes into account the ionization of the compound at the specific pH. We decided to determine logD at physiological pH – logD7,4 experimentally to get a realistic insight into the lipophilic properties of the drug. We studied the effect of ionization on retention and logD7,4/logP values and how they differ based on ionization.
Following a successfully established method, we determined the CHI value of more than two hundred unique compounds experimentally and compared the obtained results with computationally obtained data. We observed a strong correlation between the calculated and experimental values of logD7.4. In comparison to compounds in development, our optimized method achieved comparable results in terms of correlation and the average difference between the values. Our method has proven to be more accurate for neutral/basic compounds and less accurate for acids.
|