EVENTS | VIEW CALENDAR
Multi-parameter optimization: The delicate balancing act of drug discovery
A safe and efficacious drug has a balance of many properties. Potency against the therapeutic target is essential and appropriate physicochemical and absorption, distribution, metabolism and elimination (ADME) properties are also required in order to achieve suitable in-vivo disposition. Furthermore, selectivity against off targets and an absence of non-specific and idiosyncratic toxicities are necessary to achieve acceptable side effect and safety profiles. Unfortunately, these requirements are often conflicting, with improvements in one property leading to detrimental changes in another.
The challenge of successfully achieving this delicate balancing act is illustrated by a historical view of pharmaceutical R&D, a picture dominated by increasing costs and low success rates. The causes of the high attrition rates for clinical candidates have changed over the years. In particular, the previously high failure rate due to poor pharmacokinetics (PK) has been reduced, while the failures due to toxicity have increased. The reduction in PK failures has been achieved through the introduction of early in-vitro screens to filter out compounds with potential ADME issues, and similar efforts are underway to develop early tests for toxicities. However, the overall success rate has not improved, and the hidden cost of missed opportunities due to good compounds incorrectly eliminated is also likely to be high. This suggests that an alternative approach is required, taking a holistic approach to designing compounds with a good balance of properties as early as possible in the process.
This realization has led to a recent surge in interest in methods for simultaneously optimizing multiple factors, described as multi-parameter optimization (MPO). Many MPO methods have been developed in fields such as engineering, economics and quality control that may be readily adapted to the drug discovery environment and in the context of drug discovery a good MPO method needs to satisfy the following requirements:
A number of MPO methods are being applied in drug discovery. Here, we will briefly discuss some key, illustrative examples.
The most famous example of a rule-of-thumb is Lipinski's Rule of Five (RoF), which relates the molecular weight, lipophilicity and hydrogen bonding characteristics of a compound to its likelihood of achieving good oral absorption. The RoF has since been joined in the medicinal chemists' armory by many other rules-of-thumb that relate the biological properties, development potential or safety of compounds to simple characteristics, including polar surface area, flexibility, number of sp3 carbons and number of aromatic rings.
The enormous popularity of rules-of-thumb derives from the ease with which they may be applied and interpreted. They provide clear guidelines that help chemists to focus on factors that will increase the potential to identify high quality compounds while optimizing potency and eliminate chemistries with a low chance of success early in the process.
Rules-of-thumb have been derived from extensive statistical analysis of historical data and most relate to a specific objective. Therefore, it is important to apply them in an appropriate context; for example, the RoF was developed as a guide to improving the chance of achieving oral absorption, but it is frequently applied as a definition of "drug-likeness," despite the fact that the requirements for a compound intended for other routes of administration such as intravenous or inhalation are quite different, potentially leading to inappropriate decisions.
It is also important not to be too rigid when applying rules-of-thumb, particularly when options are limited. The correlations between these simple characteristics and the biological properties of a compound are not strong. Is there a significant difference in the chance of oral absorption for a compound with a molecular weight of 501 Da versus one with 499 Da?
Probably the most common approach used in an attempt to identify compounds with an appropriate property profile is to filter out those compounds that fail to meet each criterion in turn, with the hope that one or more "ideal" compounds will emerge at the end, having satisfied all of the criteria. Early in a project, the criteria corresponding to one of the rules-of-thumb may be applied as filters, but predicted and experimentally measured properties are often compared against a target product profile in this way.
The apparent simplicity of a filtering approach hides a number of dangers. Achieving a balance of properties often requires compromise, as the requirements often conflict and it is very common for no compounds to emerge from the sequence of filters. Hard cut-offs introduce artificially harsh distinctions between options, and this is exacerbated by the effect of the uncertainty in the underlying data; combining multiple uncertain filters accumulates error and dramatically increases the chance of incorrectly discarding a good compound. As a simple illustration, if we apply 10 filters that are each 90 percent accurate in passing/failing a compound, the probability of an ideal compound emerging, even if it was present in the set being filtered, is only 35 percent; we are more likely to throw away an ideal compound than accept it.
An alternative approach that avoids hard cut-offs and allows acceptable trade-offs to be defined is provided by "desirability functions." A desirability function maps a property value onto a scale between zero and one that represents the desirability of a compound with that property value; an ideal value will achieve a desirability score of one, while a completely unacceptable value will receive a desirability score of zero. Desirability functions give excellent flexibility, as a desirability function can take any shape. For example, a slope indicating increasing desirability as the property value approaches the ideal range or even non-linear relationships, such as a sigmoid or bell curve. The desirability scores of individual properties can then be easily combined into a "desirability index" to reflect the overall quality of a compound by adding them together or taking the average or geometric mean.
Furthermore, the results may also be easily interpreted, as the impact of each individual property to the overall desirability index can be calculated to guide strategies to improve the overall quality.
A group at Pfizer described an application of desirability functions to prioritize compounds with a greater chance of success against a central nervous system (CNS) target. Its "CNS MPO" approach employs six calculated physicochemical parameters to calculate a desirability index in the range 0 to 6. They found that 74 percent of a set of marketed drugs for CNS targets achieved a CNS MPO index of ≥ 4, compared with only 60 percent of the Pfizer candidates, a statistically significant difference. They also found that high CNS MPO index correlated with positive outcomes for several key in-vitro ADME and toxicity endpoints including permeability, metabolic stability, active transport by P-glycoprotein, cytotoxicity and hERG inhibition.
The probabilistic scoring method builds on the flexibility and interpretability of desirability functions by explicitly taking into account the uncertainty in the underlying data. Thus, not only is a score calculated for each compound, representing the likelihood of achieving the required profile of properties, but an uncertainty in each overall score is also estimated. This provides much more information, as it becomes possible to see when the data allows compounds to be confidently distinguished or, alternatively, when higher resolution data or another criterion is necessary to make a clear choice. The objectivity this introduces allows projects to focus their efforts on chemistries with the highest chance of success while not missing potential opportunities.
Dealing with complex, multi-parameter data with high uncertainty is an enormous challenge. The temptation is to focus on a single property, often targeting potency, in the hope that any issues that arise can be dealt later in the process. This is a risky strategy, as once locked into a tight chemical series, it becomes difficult to break out if other necessary properties are not achievable. This leads to long optimization cycles and late-stage failures, increasing the time and cost of drug discovery and reducing productivity. It is much better to consider the overall balance of properties as early as possible in the process to focus on chemistries with the best chance of quick progress and downstream success.
Dr. Matthew Segall is director and CEO of Optibrium Ltd. Segall has a M.Sc. degree in computation from the University of Oxford and a Ph.D. in theoretical physics from the University of Cambridge. As associate director at Camitro, ArQule Inc. and then Inpharmatica, he led a team developing predictive ADME models and state-of-the-art intuitive decision-support and visualization tools for drug discovery. In January 2006, he became responsible for management of Inpharmatica's ADME business, including experimental ADME services and the StarDrop software platform. Following acquisition of Inpharmatica, Segall became senior director, where he was responsible for BioFocus DPI's ADMET division. In 2009, he led a management buyout of the StarDrop business to found Optibrium.