What is LogP?

Lipophilicity has been called “the most important molecular property” for its ability to impact every part of a drug’s ADMET profile.  It describes a molecule’s affinity for lipids or other non-polar environments. This is important as many key processes are influenced by a molecule’s interaction with non-polar environments. As a drug passes through the body, it moves from a polar aqueous phase (blood) across cell membranes, or to protein binding sites, which are mostly non-polar/hydrophobic. Lipophilicity is also correlated with other key properties of drug-like molecules, such as solubility and metabolic stability.

LogP is the most common measurement of lipophilicity. It refers to the base 10 logarithm (Log₁₀) of P, where P is defined as the partition coefficient of the molecule. This is a ratio of the concentration of un-ionised molecules in an octanol-water biphasic system. If a molecule has a LogP of 3, this means the concentration of that molecule in the octanol layer is 1000x higher than in the water layer. Typical values for LogP are between 0 and 5. Octanol is used to represent the non-polar or lipid-like phase. It is an ideal solvent, as it is almost immiscible with water, but most drug-like molecules are still soluble in it due to its polar hydroxyl group (as opposed to octane). The equation shows that if a molecule’s concentration in octanol is high, then its LogP will also be high. This demonstrates how molecules which are more lipophilic, or have a high LogP, have a greater affinity and solubility in octanol, rather than water.

There are multiple ways to measure LogP experimentally, but the Shake Flask method and HPLC are two of the most common.  The shake flask method is considered the gold standard for measuring LogP. It involves adding the compound to a mixture of octanol and water and agitating the mixture until it reaches equilibrium. The concentration of the compound in each layer can then be determined using UV/VIS spectroscopy or LC-MS. Traditionally, this would have been done using laboratory glassware such as a separating funnel. This has now been converted into an automated 96-well plate format, allowing for high-throughput measurements. HPLC is also commonly used to determine LogP. In this case, the compound’s retention time is correlated to a LogP value based on reference compounds with known LogP’s. Although these values need to be corrected if the molecule is capable of forming hydrogen bonds of sufficient strength.

LogP measures the concentration of uncharged species in octanol vs water. This works for neutral compounds. But many drug-like molecules contain acidic or basic groups, which can be charged depending on the pH. In this case, it makes sense to use LogD, where D is the distribution coefficient. This is a measure of the concentration of all species (charged and uncharged) in both layers. LogD can be measured at any pH, but pH 7.4 is commonly used as this matches the pH of blood. To measure LogD, the same experiments can be used as for LogP, with the pH adjusted to 7.4. For neutral compounds, LogP = LogD across all pH values. If a molecule is ionised at pH 7.4, its LogD will be lower than its LogP as the charged species will have a higher affinity for the water layer.

Although they can be measured experimentally, these values are usually predicted using algorithms. It is common to assume that when a chemist refers to LogP/D, they are referring to a calculated value (cLogP), rather than a measured value. This involves using an algorithm (of which there are many) to calculate the cLogP of a molecule based on its structure. While it is standard to report calculated LogP values, these values often contain systematic errors. For LogD, it is also important to calculate the pK_a of the molecule correctly, which can introduce further error. Other errors occur when trying to calculate LogP values for compounds which are significantly different from the training sets used for these algorithms. This is especially true when looking at cyclic peptides, where cLogP is overestimated most of the time. Derek Lowe illustrates this using cyclosporine A, which has a cLogP~14, but a measured value of 3.

In general, there are clear trends of LogP’s effect on various ADMET properties. These should be treated as a “rule of thumb”, as exceptions can be found for specific cases.

Compounds with high LogP usually have poor aqueous solubility, which makes sense when relating LogP to a concentration in octanol vs water. This leads to the common strategy of reducing LogP as a way of improving compound solubility. This strategy appears to be more effective for neutral and basic compounds, rather than acidic and zwitterionic compounds. Because the solubility of a compound increases exponentially with changes in pH and pKa, introducing ionisable groups is also a common strategy for increasing solubility.

Permeability is another property that is easy to relate to the octanol-water partition, and was one of its intended uses. These two layers can mimic the polar aqueous phase (blood) and the non-polar interior of cell membranes. In general, as LogP rises, compounds become more cell-permeable. Studies have reported various relationships between permeability and LogP, but they all show poor permeability at low values of LogP. This is due to the high energy penalty needed to desolvate the molecule and decreased attraction to the lipid bilayer.

The metabolism of a drug is a complex process which involves many different processes within the body.  That being said, higher LogP values are often associated with lower metabolic stability. This can be rationalised by understanding that the enzymes which are mainly involved in this process (cytochrome P450s) evolved to convert lipophilic compounds into more polar molecules to be removed from the body. It is also accepted that the binding sites of these enzymes are generally lipophilic, meaning compounds which are more lipophilic will have a greater affinity for these binding sites. Reducing LogP is a common strategy for improving metabolic stability. It is important to note that, like with the other properties we have discussed, changes in LogP are difficult to separate from structural changes that occur to the molecule. For example, reducing the LogP of a compound usually involves either removing lipophilic groups or introducing polarity into these areas. As these groups are usually the main sites of metabolism, the improvement in metabolic stability may be due to chemical structure. Molecules with very low LogP will also suffer from high renal clearance. These highly polar compounds have low permeability and are not reabsorbed by the kidneys.

It is commonly accepted that increasing a molecule’s lipophilicity will increase its binding affinity to all targets due to increased hydrophobic interactions. This can lead to lipophilic molecules being more promiscuous. This can mean binding to anti-targets, which cause toxicity such as the hERG ion channel and also inhibiting cytochrome P450s. It has been reported that the probability of encountering issues from these targets increases with LogP. Although this can be dependent on ionisation class, with basic lipophilic molecules having a higher chance of a hERG liability.

Many of LogP’s effects on ADMET properties oppose each other. This means there is an acceptable range for LogP, rather than a single value. Compounds with low LogP are likely to suffer from poor permeability and high renal clearance. While compounds with high LogP are likely to have poor solubility and metabolic stability, and may also display toxicity through binding to anti-targets. In Lipinski’s rule of 5, he suggests a LogP < 5 from compounds that reached Phase II clinical trials. It is accepted that a LogP/D between 1 and 3 provides the highest chance of achieving an acceptable ADMET profile to not hinder drug development. This is a narrow range, and many successful current and future drugs will sit outside this. This is especially true when considering new modalities and beyond rule of 5 compounds (bRo5), such as PROTACs and cyclic peptides.

In general, it is far more common to encounter problems from high LogP as opposed to low LogP. One reason for this is that molecules usually become larger and more lipophilic during drug development to achieve higher potency.

There are two common strategies for reducing LogP. The first is removing or replacing non-polar groups, such as halogens and phenyl rings. The other approach is to introduce polar atoms, such as one or more nitrogens to phenyl rings and oxygen and nitrogen atoms into aliphatic rings and chains, to form ethers and amines.

While it is possible to reduce the LogP of a molecule, doing so will likely cause a drop in potency due to decreased hydrophobic interactions.  A better approach would be to focus on molecules with inherently lower LogP and avoid large increases in LogP during development. This can be done by using ligand efficiency (the next article on this topic), which considers LogP and potency, allowing changes in potency to be benchmarked against changes in LogP.