Hansch predictions

The attraction of lipophilicity in medicinal chemistry is mainly due to Corwin Hansch s work and thus it is traditionally related to pharmacodynamic processes. However, following the evolution of the drug discovery process, lipophilicity is today one of the most relevant properties also in absorption, distribuhon, metabolism, excretion and toxicity (ADMET) prediction, and thus in drug profiling (details are given in Chapter 2). 

Hansch and Leo [13] described the impact of Hpophihdty on pharmacodynamic events in detailed chapters on QSAR studies of proteins and enzymes, of antitumor drugs, of central nervous system agents as well as microbial and pesticide QSAR studies. Furthermore, many reviews document the prime importance of log P as descriptors of absorption, distribution, metabolism, excretion and toxicity (ADMET) properties [5-18]. Increased lipophilicity was shown to correlate with poorer aqueous solubility, increased plasma protein binding, increased storage in tissues, and more rapid metabolism and elimination. Lipophilicity is also a highly important descriptor of blood-brain barrier (BBB) permeability [19, 20]. Last, but not least, lipophilicity plays a dominant role in toxicity prediction [21]. 

Hansch analysis marked the breakthrough of QSAR. The method was soon extended with additional parameters with the aim of improving the fit between biological and physicochemical data and for the prediction of drugs with optimal... 

A difficulty with Hansch analysis is to decide which parameters and functions of parameters to include in the regression equation. This problem of selection of predictor variables has been discussed in Section 10.3.3. Another problem is due to the high correlations between groups of physicochemical parameters. This is the multicollinearity problem which leads to large variances in the coefficients of the regression equations and, hence, to unreliable predictions (see Section 10.5). It can be remedied by means of multivariate techniques such as principal components regression and partial least squares regression, applications of which are discussed below. 

The octanol-water partition coefficient Kow is widely used as a descriptor of hydrophobicity. Variation in /fow is primarily attributable to variation in activity coefficient in the aqueous phase (Miller et al. 1985) thus, the same correlations used for solubility in water are applicable to /fow. Most widely used is the Hansch-Leo compilation of data (Leo et al. 1971, Hansch and Leo 1979) and related predictive methods. Examples of Kow correlations are ... 

The measurement of the solubility of drugs in polar and non-polar media is very important in the pharmaceutical field. One method proposed to describe this solubility is the partition coefficient between octanol and water. The mathematical calculation of an octanol-water partition coefficient from values for functional groups was first proposed by Hansch et al. as Hansch s n constants,1 and was later developed by Rekker as hydrophobic fragmental constants (logP).2 This method was further improved by the use of molecular connectivities.17 The prediction of logP values can be performed by either a computer program or by manual calculation. For example, approximate partition coefficients (log P) have been calculated by Rekker s method ... 

Chemoinformatic Analysis of the Predicted Hansch Substituent Constants 415... 

Chemoinformatic Analysis of the Predicted Hansch Substituent Constants of the Diversity Reagents for Design of Vector Exploration Libraries... 

Fig. 15.22 Display of predicted Hansch parameters of the sum of inductive and resonance constants (F + R) versus lipophilicity (jr) for 302 commercially available carboxylic acids, acid chlorides, sulphonyl chlorides and isocyanates. Symbol size corresponds to larger molecular size (MR). Circles represent the selected R groups and triangles represent the unselected compounds. Colors convey the segmentation of the diversity reagents into nine sectors for selection.

Fig. 15.23 Selected examples for the selection of diversity reagents of 302 choices according to nine sectors selected according to Hansch parameter predictions.

Every component of an organic compound has a defined lipophilicity and calculation of partition coefficient can be performed from a designated structure. Likewise, the effect on log P of the introduction of a substituent group into a compound can be predicted by a number of methods as pioneered by Hansch [5-8] (k values), Rekker [9-10] (fvalues) and Leo/Hansch [5-7, 11-12] (f values). 

The construction of the Training and Test Sets can have a significant impact on the ability of the model. In the traditional QSAR portion, Bioheavy models were able to adequately predict the original bioactivities for the Training and Test Set for the Hansch ( R2 = 0.86, Q2 = 0.78, R2 = 0.58) and MOE ( R2 = 0.79, Q2 — 0.69, R2 — 0.66) descriptors. This was not the case when the Biolite models were confronted with the same task. The Biolite models were unable to predict the original bioactivities for the Test Sets even though the models were able to predict the bioactivities for the Training Set Hansch descriptors ( R2 = 0.91, Q2 = 0.86, R2 = 0.00) and MOE Descriptors ( R2 = 0.84, Q2 = 0.77, R2 = 0.09). 

There is a long history of efforts to find simple and interpretable /i and fi functions for various activities and properties (29, 30). The quest for predictive QSAR models started with Hammett s pioneer work to correlate molecular structures with chemical reactivities (30-32). However, the widespread applications of modern predictive QSAR and QSPR actually started with the seminal work of Hansch and coworkers on pesticides (29, 33, 34) and the developments of various powerful analysis tools, such as PLS (partial least squares) and neural networks, for multivariate analysis have fueled these widespread applications. Nowadays, numerous publications on guidelines, workflows, and... 

Another area where this intermediate level of moieties has been invoked is the use of molecular fragments to predict certain physical properties of molecules (Lyman, 1982). In one example, the prediction of partition coefficients as a measure of lipophilicity is steadily evolving in several laboratories (see below). The central theme of these efforts is the dissection of a molecule into fragments followed by an evaluation of their individual contribution to the physical property. From there a simple summation of contributions (i.e. increments), mitigated by a variety of factors encoding constitutive properties, is made to model the property (Rekker, 1977 Hansch and Leo, 1979). 

Since the early works of Hansch and Leo (28), and the fragmental constants of Rekker (29), many papers and a book (30) have appeared on how to predict retention from molecular structure. Retention increases with an increase in hydrophobicity. When hydrophobicity order is not straightforward, a rule of thumb is to consider that elution is correlated with the solubility in the mobile phase. 

Hammett s equation was also established for substituted phenols from the elementary hydroxyl radical rate constants. The Hammett resonance constant was used to derive a QSAR model for substituted phenols. The simple Hammett equation has been shown to fail in the presence of electron-withdrawing or electron-donating substituents, such as an -OH group (Hansch and Leo, 1995). For this reason, the derived resonance constants such as o°, cr, and o+ were tested in different cases. In the case of multiple substituents, the resonance constants were summed. Figure 5.24 demonstrates a Hammett correlation for substituted phenols. The least-substituted compound, phenol, was used as a reference compound. Figure 5.24 shows the effects of different substituents on the degradation rates of phenols. Nitrophenol reacted the fastest, while methoxyphenol and hydroxyphenol reacted at a slower rate. This Hammett correlation can be used to predict degradation rate constants for compounds similar in structure. 

---