Working Paradigm for Lead Optimization Bioactivation Covalent Binding and Cytotoxicity

Many questions remain unanswered regarding idiosyncratic reactions and we do not yet have a clear and consistent marker for compounds that cause idiosyncratic reactions. We can not identify one property of compounds that we can use to eliminate or optimize lead compounds without risking losing valuable compounds. The complete elimination of all idiosyncratic reactions may not be a necessary endpoint. Our goal, however, is to reduce the frequency of idiosyncratic reactions so that the risk associated with a compound is acceptable. Therefore, the current goal for lead optimization is to develop a set of characteristics that identify high risk compounds. The preceding discussion raises several key points that can be used in a step-wise approach in an attempt to minimize the likelihood of compounds causing idiosyncratic reactions.

Step 1 (Figure 3): The available evidence strongly supports the contention that protein reactivity is a necessary characteristic of idiosyncratic toxins. Bioactivation may be required for most compounds, although some compounds (e.g. penicillamine, penicillins) may be inherently reactive. Reactive intermediate

Figure 3. Step 1 in the assessment of risk of a compound being associated with idiosyncratic reactions.

formation increases the risk of idiosyncratic reactions and minimization of bioacti-vation potential through elimination of toxicophores is a logical first step. Identification of reactive metabolite formation is a useful approach. The methods for identifying reactive intermediate formation have been reviewed elsewhere and they include demonstration of glutathione conjugates, covalent binding, trapping experiments, and suicide inhibition of metabolizing enzymes. None of these methods are fool-proof. Covalent binding studies in vitro and/or in vivo are a logical part of any pre-screening program and may be included in either Step 1 or Step 2. If bioactivation with or without covalent binding is identified, the

Figure 4. Step 2 in the assessment of risk of a compound being associated with idiosyncratic reactions.

compound may be either modified to remove this property or further characterization to better assess risk undertaken.

Step 2 (Figure 4): Once reactive intermediate formation is identified or predicted, covalent binding in vitro (if not previously performed; rat and human systems are both potentially useful) and in vivo should be evaluated. In vivo studies are generally carried out in rodents. If a series of related compounds is going to be evaluated or re-evaluated, it may be worthwhile to generate antibodies for use in immunochemical detection, otherwise such studies are typically performed with radiolabelled material. There are currently no standardized testing procedures. It has recently been suggested that standard doses of 20 mg/kg per os be used for in vivo studies and 10 |M for 1 hour be used for in vitro studies with human or rat hepatic microsomes (referred to here as "The Merck Paradigm"; (Evans et al. 2004)). While these are somewhat arbitrary doses or concentrations, if a bank of data in a standardized format is generated by consistent application of these doses/concentrations, sufficient data may be generated to allow for retrospective analysis. However, it is also recommended that studies be carried out at doses or concentrations that are 5 - 10 times the expected peak plasma concentrations to ensure that nothing is missed. In vivo studies should also include standard assessment for cytotoxicity, such as serum liver-related enzyme activities and histopathological analysis. If significant or extensive covalent binding is observed in vivo in the absence of evidence of toxicity, then this compound should be considered at an increased risk of being associated with idiosyncratic drug reactions. Not all compounds that are associated with idiosyncratic reactions lead to covalent binding in rodent models when given as the parent compound (e.g. sulfamethoxazole covalent binding is not observed in rodents after administration of the parent, but only after administration of the metabolites (Cribb et al. 1996)). In some cases, studies in isolated cells or with synthetic reactive intermediates are required.

The major outstanding question is what constitutes significant covalent binding? The Merck Paradigm suggests 50 pmol/mg protein in vivo or in vitro as a level of concern. This has not been proposed as a go/no go decision point, but as a level of heightened concern when additional assessment or consideration is required. At the moment, there are insufficient data to support or refute this as a specific level of covalent binding but a review of the literature suggests that is not an inappropriate concentration, provided that doses/concentrations that meet or exceed known or expected human exposures are used.

Step 3 (Figure 5): If covalent binding is observed, particularly in vivo, assessment of cytotoxicity can be undertaken. When covalent binding is linked to cytotoxicity in Step 2 or in further investigations in Step 3, then the compound should be treated as an intrinsic toxin and standard assessment of risks and benefits of intrinsic toxins undertaken. If the covalent binding is observed at doses or concentrations below those associated with toxicity in vivo, then additional studies in vitro may be indicated to determine the cytotoxic potential and its association with covalent binding. While this strays closer to lead selection, this information can be part of the feedback loop for lead optimization.


Figure 5. Step 3 in the assessment of risk of a compound being associated with idiosyncratic reactions.

Covalent binding in vivo in the absence of/prior to significant toxicity (eg liver - histo; elevated enzymes) Or in cellular systems not


Covalent binding in vivo in the absence of/prior to significant toxicity (eg liver - histo; elevated enzymes) Or in cellular systems not

Figure 5. Step 3 in the assessment of risk of a compound being associated with idiosyncratic reactions.

In summary, compounds that undergo bioactivation leading to covalent binding at doses or concentrations below those required to cause cytotoxicity or toxicity in vivo are likely at increased risk of causing immune-associated idiosyncratic toxicity in the clinic. Available evidence suggests that within a group of structurally related compounds, increased covalent binding and/or increased cytotoxicity of reactive metabolite increases the risk of idiosyncratic toxicity at some point. Therefore, lead compound structure should be optimized to minimize bioactivation, covalent binding, and cytotoxic potential. Minimizing the total daily dose or duration of therapy further reduces the risk of idiosyncratic reactions. However, dose reduction by increasing affinity for targets must not be undertaken at the expense of drug-like properties.

No decision on lead optimization or selection should be taken without giving consideration to the wealth of caveats or modifying risk factors that surround these decisions:

• What is the potential for chemical modification? If easy without losing other properties, it should be undertaken.

• If bioactivation is associated with additional risks such as cytochrome P450 inhibition/induction, chemical modification is indicated.

• The severity of condition treated may warrant advancement of compounds that have characteristics associated with higher risk of idiosyncratic reactions.

• If alternative therapies with good safety records exist, the risk of idiosyncratic reactions needs to be low.

• If the expected duration of therapy or expected total daily dose is high, than the risk of idiosyncratic reactions is likely higher for compounds with an "at risk" profile.

• The intended target population should also be considered.

Our ability to identify compounds that will be associated with a high incidence of idiosyncratic reactions remains unacceptably limited. Continuing investigations are required to develop suitable animal models to evaluate and modify screening and lead optimization paradigms.

Was this article helpful?

0 0

Post a comment