1564
Chem. Res. Toxicol. 2006, 19, 1564-1569
Safety Assessment of Drug Metabolites: Characterization of Chemically Stable Metabolites W. Griffith Humphreys* and Steve E. Unger Department of Pharmaceutical Candidate Optimization, Bristol-Myers Squibb Pharmaceutical Research Institute, P.O. Box 4000, Princeton, New Jersey 08543 ReceiVed September 28, 2006
Drug molecules are typically subjected to a variety of biotransformation reactions, and the metabolites formed through these reactions must be considered when conducting safety testing programs for new chemical entities. Metabolites that are chemically stable sometimes have pharmacological activity profiles similar to those of the parent compound but rarely have potent activity against off-target receptors that is unique relative to the parent profile. This fact argues for the thorough testing of drug metabolites for their pharmacological activity. It also argues for a significantly lower need for the thorough characterization and quantitation of stable metabolites not thought to substantially contribute to the pharmacodynamic effect. Given the tremendous resource requirements involved in the thorough characterization of drug metabolites, a more flexible, tiered approach to stable metabolite characterization would seem to provide the best utilization of resources while still allowing a complete evaluation of the toxicological profile of a new drug. Contents Introduction Safety of Stable Metabolites Why Are Stable Metabolites Generally Safe? Differentiation between Metabolite Profiling for Contribution to Efficacy versus Safety Purposes How to Characterize Stable Metabolites What Are the Risks in the Quantitation of Multiple Metabolites? Conclusions
1564 1565 1566 1567 1567 1567 1568
Introduction Recently, there has been a flurry of interest in the area of safety testing of drug metabolites (1, 2). This has been driven by the need to fully reduce to practice the idea that to adequately test the safety of a new drug, the toxicology models used in safety evaluation should be exposed to the same mixture of metabolites that humans are exposed to after administration of the parent compound. This fairly simple concept proves to be quite involved when the complex mixtures of metabolites present in humans and multiple animal species are considered and has led to proposals that provide a standard framework for metabolite characterization (2, 3). Each of these proposals attempts to strike a reasonable balance between the collection of data that will have a meaningful impact on the safety evaluation of a new drug with the practical considerations that ensuring the safety of every metabolite present in humans at any level is an impossible task. A great deal of literature over the past decade has addressed the role of the metabolism of drugs to reactive metabolites as the causative event in drug-induced toxicities, especially those of an idiosyncratic nature (4-8). Although the molecular * Corresponding author. Tel: (609) 252-3636. Fax: (609) 252-6802. E-mail:
[email protected].
mechanisms linking the initial reactive metabolite-induced insult with toxicity have not been established, there is a strong positive correlation between the observation of adverse events and the metabolism of drugs to reactive metabolites. Because this correlation exists and, thus, places an additional risk associated with compounds that form reactive metabolites, it seems prudent that this property be carefully explored during the drug development process. Investigation of the formation of reactive metabolites is a very complicated undertaking because most often, the byproducts of these metabolites are the only species that can be observed. Furthermore, the observation of the downstream products of reactive metabolite formation in typical ADME studies with collection of plasma and excreta samples may give misleading information regarding the actual amount of reactive metabolite formed. Despite these difficulties, downstream markers of reactive species must be carefully monitored because the relatively minor flux down these pathways is thought to produce the toxicities that have led to the withdrawal of a number of clinical candidates and marketed drugs. Minor pathways leading to reactive metabolites become even more important when the parent is administered at relatively high doses, thus making even minor pathways important on an absolute scale. The bulk of drug metabolism produces stable metabolites and not metabolites that are inherently reactive or have arisen from reactive intermediates that leave an enzymatic active site (Note: for the purposes of this manuscript, a stable metabolite is defined as a nonreactive chemical entity, arising from a metabolic transformation but does not include (1) nonreactive metabolites that are the product of the capture of a reactive species subsequent to a metabolic transformation, e.g., glutathione conjugates, diols, etc. and (2) redox active metabolites). Complex mixtures of stable metabolites are commonly present in humans and animals and are normally observed as circulating and excreted metabolites. The contribution of stable metabolites to the desired pharmacological target efficacy has long been recognized as an important consideration in drug development,
10.1021/tx6002547 CCC: $33.50 © 2006 American Chemical Society Published on Web 11/08/2006
PerspectiVe
Chem. Res. Toxicol., Vol. 19, No. 12, 2006 1565
especially as it relates to PK/PD relationships and effects of drug-drug interactions on efficacy. However, there has not been a great deal of literature on the contribution of stable metabolites to adverse drug reactions caused through off-target toxicities. Optimally, the safety of stable metabolites is determined through the exposure in the species used for toxicological evaluation after administration of the parent. However, in the unusual example of a metabolite that is formed at appreciable levels in humans but not found in animals, direct administration of the stable metabolite to the toxicological species may be required. More typically, a common set of stable metabolites circulate in animals and humans, but at variable ratios compared to the parent or total drug-related material. The initial quantitative metabolite profiles for animals and humans are usually generated following a single dose of a radiolabeled version of the drug candidate. The information from these studies is then used to address issues of metabolite characterization. Such issues can include which metabolites require synthesis of authentic standards, which metabolites need to be measured and in what studies, whether there is a need for GLP bioanalytical assays, and whether the metabolites should be profiled for liabilities. In determining how to develop a metabolite monitoring plan, a sponsor may choose to measure all potential stable human metabolites with validated assays and test for liabilities very early in IND toxicology studies and in first-in-human trials. More assays could be added after the human study with the radiolabeled drug. This approach would yield significant information early in clinical development on the relative levels of metabolites in animals and humans after a variety of dosing situations. However, this plan would also consume tremendous resources and may be of questionable benefit in an overall safety assessment because the majority of adverse drug reactions are associated with molecules known to form reactive metabolites. The relative contribution of stable metabolites to adverse drug reactions is not well documented. Given that current proposals treat metabolites in a generic fashion, it is important to examine the utility of the extensive characterization of stable metabolites. This perspective will review the available literature on offtarget toxicities of stable metabolites and provide some discussion on the resources and risks associated with metabolite monitoring in an effort to ascertain the cost/benefit relationship of in depth characterization of these compounds during the drug development process.
Safety of Stable Metabolites Although it is generally accepted that reactive metabolites are linked to drug toxicities, there are relatively few examples of stable metabolites causing significant off-target toxicity not already elicited by the parent compound. This observation regarding stable metabolites is based purely on empirical evidence gained from an examination of the literature on druginduced toxicity and would be very difficult to demonstrate experimentally. The observation also makes some degree of intuitive sense on the basis of structure-activity arguments. Perhaps it is not surprising that many stable metabolites retain activity toward the intended pharmacological target or related targets because the majority of oxidative metabolic transformations introduce relatively minor structural alteration. The small structural modifications produced by oxidative metabolic transformation are often not enough to produce compounds that fall far enough outside of the parent chemotype structure-activity relationship to be truly inactive. Indeed, many marketed drugs produce metabolites that display pharmacological activity (traditionally referred to simply as active metabolites). Typically,
these metabolites have slightly diminished pharmacological activity relative to that of the parent. Although they can often retain target activity, there are very few examples of stable metabolites that have a potent interaction with a receptor unique from the parent and, in doing so, produce an off-target toxicity different from those found with the parent. As described below, there are examples of metabolites that interact with receptors closely related to the pharmacological target, but a wide body of empirical evidence shows that these interactions are unlikely to produce a stable metabolite that mediates a unique off-target toxicity. There are many examples of CNS drugs that target biogenic amine receptors, where metabolites are thought to play a role in the efficacy of the parent and in some cases are thought to mediate aspects of the side effect profile (9-12). An illustrative case of this relationship between parent and metabolite is seen in two of the early tricyclic antidepressants, clomipramine and amoxapine. With clomipramine, both parent and two hydroxylated metabolites inhibit serotonin reuptake, while the des-methyl metabolite blocks noradrenaline uptake (9). In the case of amoxapine, the 7- and 8-hydroxylated metabolites are noradrenaline uptake inhibitors with a potency similar to that of the parent, but the 7-hydroxylated metabolite is a more potent dopamine receptor antagonist (9). The probability of a metabolite mediating a side effect would likely be greater when the target belongs to a receptor family with multiple closely related members, for example, the biogenic amine receptors, the peroxisome proliferator-activated receptor families, and other closely related kinase receptor families. It is logical to assume that the probability of a metabolite having significant potency against a receptor related to the target would increase in cases where the parent compound already interacts with multiple members of a receptor family. Although many biotransformation reactions lead to molecules with relatively small molecular changes, there are examples where larger changes occur, with the resultant metabolites having substantially different molecular properties. These types of transformations include (1) hydrolytic or oxidative cleavage of internal bonds to give substantial fragments or (2) the formation of drug or metabolite conjugates. Two examples of an internal cleavage leading to metabolites with interesting activity are the formation of m-chlorophenylpiperazine from trazodone or nefazodone and the formation of 1-(2-pyrimidinyl)piperazine (1-PP) from buspirone. As stated, these small fragments possess activity, but as with other metabolites of the antidepressant/antianxiolytic class, the metabolites generally have overlapping activity with the parent and display similar pharmacology. 1-PP may mediate a beneficial effect on bladder function because of its R2-adrenergic receptor antagonist activity (13). An extreme example of a small metabolic fragment displaying a new toxicology is the cyanogenic glycoside amygdalin (Leatrile). In this case, the compound releases cyanide subsequent to an enzymatic glycoside cleavage reaction (14-16). Although the example with amygdalin demonstrates that stable metabolites can mediate toxic side effects, it deviates significantly from what would be considered a normal druglike chemical structure. Chemically stable conjugates of parent molecules or of oxidative metabolites generally do not have activity against the intended pharmacological target (the O-6 glucuronide of morphine is a notable exception). Also, there are few, if any, documented examples where a stable conjugate produces a significant off-target mediated toxic response. Several potential exceptions to this rule are (1) the sulfate conjugate of troglita-
1566 Chem. Res. Toxicol., Vol. 19, No. 12, 2006
zone, which has been shown to inhibit the BSEP and OATP transport proteins (17, 18), thus hypothetically leading to cholestasis due to the inhibition of the biliary elimination of bile acids and/or other endogenous compounds and (2) the glucuronide of gemfibrozil, which has been shown to inhibit CYP2C8 (19, 20) and may contribute to the significant drugdrug interactions between gemfibrozil and CYP2C8 substrates such as cerivastatin. A notable finding in both of these cases is that the metabolite interaction at the off-target site was more potent than the parent interaction. Also, in the case of gemfibrozil glucuronide, the metabolite was a potent time-dependent CYP inhibitor, whereas the parent was a reversible inhibitor. Importantly, both of these examples are hypothetical, and for the troglitazone sulfate case, there is considerable data demonstrating that sulfation does not represent a bioactivation pathway (21). A potentially important observation regarding the hypothetical off-target receptor interactions of troglitazone sulfate and gemfibrozil glucuronide is that in both cases the interactions were intrahepatic, that is, directly at the site of metabolite generation, and thus, these targets would be expected to be more susceptible to metabolite effects. The importance of localized metabolite generation is illustrated by 1-methyl-4-phenyl1,2,3,6-tetrahydropyridine (MPTP). MPTP is a prototypical neurotoxin that produces an irreversible and severe parkinsonian syndrome. The molecule is thought to be transported first to the brain after which it is oxidized to a pyridinium species, MPP, predominantly in glial cells (22, 23). The pyridinium species is then transported into neuronal cells and is responsible for toxicity through a number of mechanisms, primarily through the inhibition of complex 1 of the mitochondrial respiration chain. In addition, toxicity results from the generation of local concentrations of active oxygen species through redox cycling mechanisms. (It should be noted that although MPP can redox cycle enzymatically, it likely does not redox cycle through the purely chemical mechanisms that are thought to play a major role in the toxicity of the structurally related paraquat (24).) The toxicity of MPP is thought to be exacerbated by the fact that the charged species can access neurons through the dopamine transporter, but there is no efficient mechanism for efflux. There has been speculation that some of the extrapyramidal side effects seen with haloperidol are mediated through the types of mechanisms outlined for MPTP, that is, intraneuronal generation of a pyridinium species similar to MPP, but the speculation is purely based on common metabolic pathways and structural similarities between haloperidol metabolites and MPTP (25). Another class of toxicities speculated to arise from the localized generation of a metabolite comes from examples where a metabolite had significantly different physical properties than the parent and results in toxicity through a nonreceptor mediated event. Examples where metabolites have been conjectured to play a role in this type of toxicity are hepatic phospholipidosis caused by ketoconazole and the urolithiasis seen after guaifenesin administration as well as several reports of other drugs during clinical development. It has been demonstrated that desacetyl ketoconazole orients in phospholipid membranes similarly to classical cationic amphiphiles known to produce phospholipidosis but in a different manner than that of ketoconazole itself (26). The differential membrane association that results from des-acetyl ketoconazole has been conjectured to lead to membrane disruption, accumulation of the metabolite, and eventual phospholipidosis (27). Similar toxicity has been described for ABT-770, a matrix metalloproteinase inhibitor in
Humphreys and Unger
clinical development (28). A related type of toxicity is the formation of renal calculi due to the generation of a poorly soluble metabolite. The calculi isolated from patients taking the expectorant guaifenesin is primarily composed of a metabolite of the drug (29). This type of toxicity has also been reported to be the cause of nephrotoxicity associated with an experimental cognition enhancer (30). In this example, a hydroxylated metabolite was found in human urine at concentrations at the equilibrium solubility of the compound in urine, and it was speculated that the metabolite was precipitating in urine, and the resulting crystaluria produced the observed nephrotoxicity. The bile duct and the gall bladder may also be sights expected to be sensitive to these types of injury because of the exposure to high concentrations of stable metabolites during drug elimination. Another related instance of this type of toxicity is that displayed in male rats when they were administered d-limonene or 2,2,4-trimethylpentane. In this example, it is not the metabolite itself that precipitates, but the complex of the metabolites, either d-limonene oxide or 2,4,4-trimethyl-2pentanol, with R2-microglobulin that forms protein droplets and ultimately nephrotoxicity (31-33). The examples described for off-target pharmacology mediated through oxidative metabolites of pharmaceutical agents were confined to cases where, although a new activity was displayed by the metabolite, off-target activity was due to the interaction with a receptor closely related to the target receptor. This type of receptor family cross-talk has been observed with metabolites of CNS drugs and may mediate side effects of some agents, but it could be argued that it does not truly constitute new pharmacological activity, given that the receptors are quite similar. The troglitazone sulfate and gemfibrozil glucuronide conjugate examples are both cases where an interaction was hypothesized to be mediated by a stable metabolite. More research is necessary to determine if these cases are indicative of a general class of metabolite interaction or whether they are isolated examples. In any case, a review of the literature shows that it is a challenging task to prepare a list of representative drugs or natural products in which a stable metabolite elicits a toxic off-target effect that is not displayed by the parent compound. Cases such as MPTP and amygdalin, where truly new pharmacology/toxicology is displayed by a stable metabolite, seem to be far more of the exception rather than the rule.
Why Are Stable Metabolites Generally Safe? Modern drug discovery revolves around the development of detailed structure-activity relationships for a given molecular target and produce compounds with excellent potency and selectivity. The great majority of compounds entering the development phase have been optimized such that the molecules have potent binding affinities with their pharmacological target(s) that are often measured in the low nM range or even in the pM range. In general, these compounds have also been profiled extensively for off-target activity, including screening against related receptors and receptors known to mediate unwanted sideeffects, such as the HERG or Na-channel. Because metabolic transformations typically produce molecules with relatively minor molecular changes, it is, therefore, not surprising that these metabolites often have significant potency against the molecular target. It is also not surprising that compounds that are structurally similar to the parent do not generally have potent activity against receptors that the parent compound does not already interact with. In other words, both parent and metabolite follow similar structure-activity relationships for target and offtarget interactions; therefore, although it is likely that the pair
PerspectiVe
will be active against the target, it is equally unexpected that a significant new activity would be found for only the metabolite. This argument is somewhat based on the premise that the concentration-response relationship for the parent and metabolites is in a potency range that is low relative to the same relationships for off-target activities, that is, often in the nM range for the target and the µM range for off-target activities. The relationship would be much less likely to hold for compounds that lack potency and must achieve high plasma concentrations (i.e., >10 µM) and makes it much more likely that metabolites of high dose drugs could play a role in offtarget toxicities as has been pointed out previously (2). Another factor that makes off-target activity for metabolites less likely is the pharmacokinetic parameters of metabolites. In general, stable metabolites do not circulate at concentrations greater than the parent, on the basis of the plasma area under the curve. Again, this places the metabolite on a similar concentration-response curve because the parent for target and offtarget effects, together with the receptor potency arguments above, significantly lower the probability for new off-target activity for metabolites. Also, the metabolite plasma concentration versus time profiles will in general show a considerably lower Cmax value than that of the parent, even for metabolites with significant exposure. Obviously, this lowers the probability of any untoward off-target activity, where the concentration-response relationship is related to Cmax exposure. However, these effects may be partially offset by lower protein binding and higher free fraction values often displayed by more polar metabolites. Although metabolites in general circulate at levels lower than that of the parent, their concentration at the site of formation may be significantly higher. Thus, the interaction site, that is, intrahepatic versus extrahepatic, should be considered when investigating off-target toxicities mediated by metabolites.
Differentiation between Metabolite Profiling for Contribution to Efficacy versus Safety Purposes As mentioned, there is a logical relationship between the plasma concentration of parent (and dose) and the probability of untoward off-target effects caused by metabolites (2). However, because metabolites have a reasonable probability for contribution to the overall pharmacological activity, there remains a need to characterize the metabolite profile regardless of the dose in order to derive PK/PD relationships and to predict the effects of concomitantly administered medications on the efficacy profile of the candidate drug in question. The profiling and identification of metabolites will also aid in the complete determination of clearance pathways and the prediction of potential drug-drug interactions and/or polymorphic metabolism.
How to Characterize Stable Metabolites Various proposals have been put forward regarding the details of metabolite characterization for new drug molecules (2, 3) and the FDA recently published a draft guidance on this topic. The metabolites in safety testing (MIST) manuscript and the FDA draft guidance have proposed that a threshold limit of metabolite quantity relative to that of the parent should serve as the trigger for detailed metabolite characterization and quantitation. As has been argued above, the relative potential to cause the toxicity of a new medicine because of the the presence of a metabolite is not equivalent from metabolite to metabolite and will vary widely depending on relative reactivity, pharmacological activity, off-target activity, and representation
Chem. Res. Toxicol., Vol. 19, No. 12, 2006 1567
in the toxicological species. The unique and complex nature of the metabolic profile in humans and animals makes setting a reasonably low threshold for initial characterization appear rational but argues against a uniform strategy for the full evaluation of metabolites, including rigorous quantitation. If, as it seems likely, there is a recommended threshold for the characterization for metabolites present in humans, then there should be a significant amount of flexibility when determining where and how to quantitate metabolites. The last point is particularly important because, as will be discussed below, there are tremendous amounts of resources involved in performing a GLP analysis on multiple components over multiple studies.
What Are the Risks in the Quantitation of Multiple Metabolites? Metabolites of candidate drugs can be quantitated by various methods and with varying degrees of rigor. Often, during the discovery stage, there will be some knowledge of prominent metabolites from in Vitro or in ViVo studies that may provide information for metabolite standard synthesis and lead to metabolite quantitation via specific LC-MS assays during the early stage of development. Having a synthetic reference standard confirms the proposed structure, allows in Vitro screening, and affords a calibrator for accurate analysis. When exact structures are unknown or reference compounds cannot be synthesized, quantitation may involve the use of biological isolates as reference standards. Alternatively, radiolabeled metabolites have served as calibrators for MS response to quantify metabolites from nonradiolabeled studies. Either approach is only a preliminary estimate of exposure that may define its significance in considering future GLP analyses. This tiered approach to metabolite quantitation obviates the need to develop and validate numerous metabolite assays when exposures are insignificant (as judged by the various exposure thresholds that have been proposed). Ideally, one should know information on animal metabolism in ViVo and human metabolism in Vitro at the start of IND studies so that highly specific assays can be established for toxicology and clinical single-ascending dose (SAD) studies. The use of accelerator mass spectrometry in exploratory IND or in combination with the SAD study can afford an early evaluation of metabolite coverage and quantitative needs. Alternatively, LCMS profiling of SAD samples for target and unknown metabolites affords similar qualitative information except for metabolites that are poorly ionized or highly dissimilar to the parent drug. Additional metabolite quantitative information is gained while carrying out more detailed ADME studies with the radiolabeled drug in animals and humans. Finally, for metabolites deemed important enough for rigorous quantitation, fully characterized metabolite standards need to be synthesized followed by assays with fully validated analytical methods. The resource demands rise significantly as one progresses to the final stage of characterization with fully validated assays. It could also be argued that the first steps of this process are by far the most important in adding to the knowledge base for a new drug. The isolation, structural identification, and activity testing against the pharmacological target for human metabolites are by no means trivial exercises but are often far less labor intensive than what is required to mount a fully validated LC-MS assay. There are basically two types of risks associated with the analysis of multiple metabolite components using the fully validated assay during the toxicological and clinical study of a candidate drug: (1) the opportunity costs associated with the time and resources necessary for standard synthesis and
1568 Chem. Res. Toxicol., Vol. 19, No. 12, 2006
Humphreys and Unger
Table 1. Typical Costs Associated with the Development of a Fully Validated Bioanalytical Assay task
cost estimate
method development
$40,000 per assay
synthesis and characterization of reference standard
$25,000 per metabolite
synthesis and characterization of internal standard
$25,000 per metabolite
method validation
$25,000 per assay
partial validation
$15,000 per assay
human urine validation
$25,000 per assay
sample analysis sample analysis
+ $10 per sample + $70 per sample
characterization, assay method development, and assay conduct and (2) the risk of overall assay failure due to the analysis of multiple components. Typical costs estimates for each process involved in the development, validation, and implementation of a GLP bioanalysis are provided in Table 1. Obviously, more analyses translates into greater drug development costs. For instance, each time a new metabolite needs to be measured in animal and human plasma, an additional $100,000 in method development and validation costs would be incurred. Continued measurements can incur added expenses, as much as $700,000 if a separate assay is used to quantify 10,000 samples. Although cost alone should not be a barrier to the conduct of these studies, the resource requirements to complete these studies early in clinical development when the attrition rate for candidate drugs is still relatively high become significant and often will lead to resource trade-offs with other early development activities. Whenever possible, a single assay is developed and used to measure the parent and metabolites. This greatly reduces added sample analysis costs to those associated with preparing more complex standards and QC samples, detecting, integrating and quantifying additional peaks, and reporting the resulting data. When several metabolites are measured, an added cost of $100,000 for 10,000 samples would be typical. What is often undervalued is that an increased failure rate should be expected when attempting to measure multiple analytes. Each assay has an associated failure rate that is increased by adding additional analytes. Failures rates of >20% are coming under increased scrutiny from the FDA. Investigation into the causes of these failures and prompt remediation is expected. What is less well appreciated is that an assay containing five analytes will have a 23% increased risk of failure over what was determined for the single analyte assay. Although the original result for the successful analytes need not be repeated, the failed analyte is re-assayed. For many combination assays, their use in studies is the only manner in which a valid statistical assessment of the individual failures is exposed. Combination assays often accept more risk because it is unlikely that all metabolites will have comparable properties that allow a similar extraction, separation, and detection. Metabolites often have exposures significantly lower than that of the parent drug, placing further demands on detection sensitivity and dynamic range. Detection sensitivity is related inversely to the square root of the number of analytes. Therefore, when measuring the parent and three metabolites, one should expect a 2-fold decrease in sensitivity from an assay that measures only
comment applicable to all species plasma and urine (including initial stability assessments) assumes that the bioreactor is incapable of generating quantity of pure material assumes that the stable label is needed for accurate quantification full validation in one species/matrix (including stability) partial validation in other species (including stability) human urine validation (including stability) assumes combination assay assumes separate assay (dissimilar metabolite)
the parent drug. Dynamic range must also be considered because response can vary considerably for structurally dissimilar metabolites. Although easily optimized for the efficient measurement of the parent, an assessment of the appropriate range for metabolites is complicated by intraspecies and intersubject differences. Often, samples need to be re-assayed both with dilution to measure the parent and without dilution to measure the low level metabolites. One must be careful to ensure that dilutions are properly executed and that if two values are obtained, then the appropriate result is reported. All these factors can impact the integrity of exposure data. By moving forward to simultaneously measure too many metabolites, one can inadvertently compromise the measurement of the parent drug.
Conclusions Although there is no way to systematically determine the toxicological liabilities of stable metabolites of marketed drugs, a survey of the literature demonstrates few rigorous examples where a stable metabolite displayed significant off-target activity that was not present in the parent compound. There are examples where additional efficacy or side effects have been ascribed to metabolite interactions with receptors related to the target receptor. As with the examples cited for CNS drugs, this type of interactions would be expected to be much more likely when the pharmacological target has many closely related family members. The obvious case where there is a potential for crosstalk due to metabolite interactions with related receptors is with kinase inhibitors. Because of the great many kinase inhibitors that have been recently introduced and are in clinical development, further data with this drug class will likely clarify whether the off-target toxicity of both the parent drug and metabolites will remain a theoretical area for concern or will represent a legitimate toxicological issue. This theoretical concern aside, the bulk of the evidence points toward a very low impact of stable drug metabolites toward the overall toxicological/side effect profile of drugs and provides little justification for the tremendous amount of resources required to fully characterize these metabolites in humans and toxicological species. A flexible, tiered approach to the characterization and quantitation of stable drug metabolites would seem to provide the best utilization of resources and still provide the necessary information to fully evaluate the safety profile of a new drug molecule. Acknowledgment. We acknowledge Scott Grossman, Lois Lehman-McKeeman, and Richard Robertson for their critical comments during the preparation of this manuscript.
PerspectiVe
Chem. Res. Toxicol., Vol. 19, No. 12, 2006 1569
References (1) U. S. Food and Drug Administration (2005) Draft Guidance for Industry on Safety Testing of Drug Metabolites, www.fda.gov/cder/ guidance. (2) Smith, D. A., and Obach, R. S. (2005) Seeing through the mist: abundance versus percentage. Commentary on metabolites in safety testing. Drug Metab. Dispos. 33, 1409-1417. (3) Baillie, T. A., Cayen, M. N.; Fouda, H., Gerson, R. J., Green, J. D., Grossman, S. J., Klunk, L. J., LeBlanc, B., Perkins, D. G., and Shipley, L. A. (2002) Drug metabolites in safety testing. Toxicol. Appl. Pharmacol. 182, 188-196. (4) Walgren, J. L., Mitchell, M. D., and Thompson, D. C. (2005) Role of metabolism in drug-induced idiosyncratic hepatotoxicity. Crit. ReV. Toxicol. 35, 325-361. (5) Liebler, D. C., and Guengerich, F. P. (2005) Elucidating mechanisms of drug-induced toxicity. Nat. ReV. Drug DiscoV. 4, 410-20. (6) Amacher, D. E. (2006) Reactive intermediates and the pathogenesis of adverse drug reactions: the toxicology perspective. Curr. Drug Metab. 7, 219-229. (7) Guengerich, F. P. (2006) Cytochrome P450S and other enzymes in drug metabolism and toxicity. AAPS J. 8, E101-E111. (8) Baillie, T. A. (2006) Future of toxicology-metabolic activation and drug design: challenges and opportunities in chemical toxicology. Chem. Res. Toxicol. 19, 889-893. (9) Rudorfer, M. V. (1997) The role of metabolites of antidepressants in the treatment of depression. CNS Drugs 7, 273-312. (10) Rudorfer, M. V., and Potter, W. Z. (1999) Metabolism of tricyclic antidepressants. Cell Mol. Neurobiol. 19, 373-409. (11) Sanchez, C., and Hyttel, J. (1999) Comparison of the effects of antidepressants and their metabolites on reuptake of biogenic amines and on receptor binding. Cell Mol. Neurobiol. 19, 467-489. (12) Rotzinger, S., Bourin, M., Akimoto, Y., Coutts, R. T., and Baker, G. B. (1999) Metabolism of some “second”- and “fourth”-generation antidepressants: iprindole, viloxazine, bupropion, mianserin, maprotiline, trazodone, nefazodone, and venlafaxine. Cell Mol. Neurobiol. 19, 427-442. (13) Myers, R. A., Plym, M. J., Signor, L. J., and Lodge, N. J. (2004) 1-(2-pyrimidinyl)-piperazine, a buspirone metabolite, modulates bladder function in the anesthetized rat. Neurourol. Urodyn. 23, 709715. (14) Khandekar, J. D., and Edelman, H. (1979) Studies of amygdalin (laetrile) toxicity in rodents. JAMA 242, 169-171. (15) Khandekar, J. D. (1980) Amygdalin (laetrile) toxicity in rodents. JAMA 243, 2396. (16) O’Brien, B., Quigg, C., and Leong, T. (2005) Severe cyanide toxicity from ‘vitamin supplements’. Eur. J. Emerg. Med. 12, 257-258. (17) Funk, C., Pantze, M., Jehle, L., Ponelle, C., Scheuermann, G., Lazendic, M., and Gasser, R. (2001) Troglitazone-induced intrahepatic cholestasis by an interference with the hepatobiliary export of bile acids in male and female rats. Correlation with the gender difference in troglitazone sulfate formation and the inhibition of the canalicular bile salt export pump (Bsep) by troglitazone and troglitazone sulfate. Toxicology 167, 83-98. (18) Nozawa, T., Sugiura, S., Nakajima, M., Goto, A., Yokoi, T., Nezu, J., Tsuji, A., and Tamai, I. (2004) Involvement of organic anion transporting polypeptides in the transport of troglitazone sulfate: implications for understanding troglitazone hepatotoxicity. Drug Metab. Dispos. 32, 291-294. (19) Shitara, Y., Hirano, M., Sato, H., and Sugiyama, Y. (2004) Gemfibrozil and its glucuronide inhibit the organic anion transporting polypeptide 2 (OATP2/OATP1B1:SLC21A6)-mediated hepatic uptake and CYP2C8-
(20)
(21)
(22) (23) (24) (25)
(26)
(27) (28)
(29)
(30)
(31)
(32) (33)
mediated metabolism of cerivastatin: analysis of the mechanism of the clinically relevant drug-drug interaction between cerivastatin and gemfibrozil. J. Pharmacol. Exp. Ther. 311, 228-236. Ogilvie, B. W., Zhang, D., Li, W., Rodrigues, A. D., Gipson, A. E., Holsapple, J., Toren, P., and Parkinson, A. (2006) Glucuronidation converts gemfibrozil to a potent, metabolism-dependent inhibitor of CYP2C8: implications for drug-drug interactions. Drug Metab. Dispos. 34, 191-197. Kostrubsky, V. E., Sinclair, J. F., Ramachandran, V., Venkataramanan, R., Wen, Y. H., Kindt, E., Galchev, V., Rose, K., Sinz, M., and Strom, S. C. (2000) The role of conjugation in hepatotoxicity of troglitazone in human and porcine hepatocyte cultures. Drug Metab. Dispos. 28, 1192-1197. Smeyne, R. J., and Jackson-Lewis, V. (2005) The MPTP model of Parkinson’s disease. Brain Res Mol. Brain Res. 134, 57-66. Bove, J., Prou, D., Perier, C., and Przedborski, S. (2005) Toxin-induced models of Parkinson’s disease. NeuroRx. 2, 484-494. Przedborski, S., and Ischiropoulos, H. (2005) Reactive oxygen and nitrogen species: weapons of neuronal destruction in models of Parkinson’s disease. Antioxid. Redox Signaling 7, 685-693. Kalgutkar, A. S., Taylor, T. J., Venkatakrishnan, K., and Isin, E. M. (2003) Assessment of the contributions of CYP3A4 and CYP3A5 in the metabolism of the antipsychotic agent haloperidol to its potentially neurotoxic pyridinium metabolite and effect of antidepressants on the bioactivation pathway. Drug Metab. Dispos. 31, 243-249. Brasseur, R., Vandenbosch, C., Van den Bossche, H., and Ruysschaert, J. M. (1983) Mode of insertion of miconazole, ketoconazole and deacylated ketoconazole in lipid layers. A conformational analysis. Biochem. Pharmacol. 32, 2175-2180. Whitehouse, L. W., Menzies, A., Mueller, R., and Pontefract, R. (1994) Ketoconazole-induced hepatic phospholipidosis in the mouse and its association with de-N-acetyl ketoconazole. Toxicology 94, 81-95. Gum, R. J., Hickman, D., Fagerland, J. A., Heindel, M. A., Gagne, G. D., Schmidt, J. M., Michaelides, M. R., Davidsen, S. K., and Ulrich, R. G. (2001) Analysis of two matrix metalloproteinase inhibitors and their metabolites for induction of phospholipidosis in rat and human hepatocytes(1). Biochem. Pharmacol. 62, 1661-1673. Pickens, C. L., Milliron, A. R., Fussner, A. L., Dversdall, B. C., Langenstroer, P., Ferguson, S., Fu, X., Schmitz, F. J., and Poole, E. C. (1999) Abuse of guaifenesin-containing medications generates an excess of a carboxylate salt of beta-(2-methoxyphenoxy)-lactic acid, a guaifenesin metabolite, and results in urolithiasis. Urology 54, 2327. Merschman, S. A., Rose, M. J., Pearce, G. E., Woolf, E. J., Schaefer, B. H., Huber, A. C., Musson, D. G., Perry, K. J., Rush, D. J., Varsolona, R. J., and Matuszewski, B. K. (2005) Characterization of the solubility of a poorly soluble hydroxylated metabolite in human urine and its implications for potential renal toxicity. Pharmazie 60, 359-363. Lehman-McKeeman, L. D., Rodriguez, P. A., Takigiku, R., Caudill, D., and Fey, M. L. (1989) d-Limonene-induced male rat-specific nephrotoxicity: evaluation of the association between d-limonene and alpha 2u-globulin. Toxicol. Appl. Pharmacol. 99, 250-259. Lehman-McKeeman, L. D., and Caudill, D. (1994) d-Limonene induced hyaline droplet nephropathy in alpha 2u-globulin transgenic mice. Fundam. Appl. Toxicol. 23, 562-8. Borghoff, S. J., Miller, A. B., Bowen, J. P., and Swenberg, J. A. (1991) Characteristics of chemical binding to alpha 2u-globulin in vitroevaluating structure-activity relationships. Toxicol. Appl. Pharmacol. 107, 228-238.
TX6002547
