Chem. Res. Toxicol. 2009, 22, 263–266
263
Approaches to the Assessment of Stable and Chemically Reactive Drug Metabolites in Early Clinical Trials Thomas A. Baillie* School of Pharmacy, UniVersity of Washington, Box 357631, Seattle, Washington 98195-7631 ReceiVed NoVember 18, 2008
The recommendations of the FDA’s final Guidance document on Safety Testing of Drug Metabolites provide a framework for devising preclinical toxicology assessment paradigms, where necessary, for human metabolites of small molecule drug candidates. Importantly, these recommendations carry implications for the qualitative and quantitative analysis of circulating drug metabolites in early human trials, which typically are performed without the benefit of a radiolabeled tracer. In this perspective, an approach to these goals is outlined based on recent work at Merck Research Laboratories involving the use of ultraperformance liquid chromatography-mass spectrometry analysis, performed on a highresolution time-of-flight mass spectrometer, of first-in-human study plasma samples. With the aid of a fractional mass filtering algorithm, drug metabolites are distinguished from endogeneous background materials and subsequently identified on the basis of their accurate masses, product ion mass spectra, and computer-assisted structure elucidation software routines. Semiquantitative analysis then is based on calibration of the MS response to each analyte with reference to radioactivity data from in vitro metabolic profiles. In the case of chemically reactive drug metabolites, which are excluded from consideration in the Guidance, a proactive approach is advocated whereby potent (low dose) drug candidates with only a limited propensity to form electrophilic intermediates are advanced into development. Overall, a decision on the need to conduct separate evaluation of the safety profile of a human drug metabolite(s) should take into consideration all of the available information on the compound of interest and be based on a case-by-case approach employing sound scientific principles.
I. II. III. IV.
Contents Drug Metabolites in Safety Testing Chemically Stable Drug Metabolites Chemically Reactive Drug Metabolites Conclusions
263 264 265 266
I. Drug Metabolites in Safety Testing The issue of the role played by drug metabolites in the toxicity associated with their respective parent compounds has been a topic of growing interest to both the pharmaceutical industry and the regulatory agencies since publication of the “Metabolites in Safety Testing” (MIST)1 paper in 2002 (1). Particular emphasis has been placed on those cases where metabolites circulating in humans either are absent in the animal species employed for toxicology testing or are present at much lower levels than in humans, since it may be argued that the safety profile of the drug and its associated metabolites cannot be fully assessed by means of a conventional preclinical toxicology program. A number of commentaries on the subject have appeared, dealing with issues such as the safety assessment of stable (vs chemically reactive) metabolites (2, 3); the need for systemic exposure to drug metabolites to be measured on an * To whom correspondence should be addressed. Tel: 206-616-9254. Fax: 206-685-9297. E-mail:
[email protected]. 1 Abbreviations: MIST, metabolites in safety testing; ADME, absorption, distribution, metabolism, and excretion; AUC, area under the plasma concentration vs time curve; LC-MS, liquid chromatography-mass spectrometry; LC-MS/MS, liquid chromatography-tandem mass spectrometry; UPLC-MS, ultraperformance liquid chromatography-mass spectrometry; AMS, accelerator mass spectrometry; GMP, Good Manufacturing Practices; FDA, U.S. Food and Drug Administration; iTRAQ, isobaric tags for relative and absolute quantitation.
absolute, as opposed to relative, basis (4); the resource implications of separate safety testing of metabolites (3); and, in general, under what circumstances it would be appropriate to consider specific studies on one or more metabolites of a new drug candidate, given that there appear to be relatively few good examples where metabolites have been shown to contribute to drug toxicity (5). Individual pharmaceutical companies have developed somewhat different approaches to the problem, as exemplified by reports from Bristol-Myers Squibb (3), Pfizer (5), and Eli Lilly (6). In 2005, the U.S. Food and Drug Administration (FDA) issued a draft Guidance document entitled “Safety Testing of Drug Metabolites”, which was followed, in 2008, by the final version of the Guidance that outlined the Agency’s current thinking on the topic (7). The latter document recommended that while there is a need to maintain a flexible, case-by-case evaluation of the need to perform additional safety studies on human drug metabolites, those metabolites present at disproportionately higher levels in human plasma than in any of the animal test species should be considered for safety assessment.2 Importantly, it is stated that “Human metabolites that can raise a safety concern are those formed at greater than 10 percent of parent drug systemic exposure at steady state” (7). Moreover, the Guidance underscores the need for sponsors to conduct studies on the metabolic fate of drug candidates at an early stage of clinical development, such that issues of disproportionate human metabolites may be addressed prior to the initiation of 2 An exception to this position, noted in the Guidance document (7), relates to drug-related materials that represent stable conjugates (e.g., ether glucuronides and phenolic sulfates) of either the parent drug or one or more of its metabolites.
10.1021/tx800439k CCC: $40.75 2009 American Chemical Society Published on Web 02/16/2009
264
Chem. Res. Toxicol., Vol. 22, No. 2, 2009
Baillie
Figure 1. LC-MS data from the analysis of a plasma sample obtained from an early clinical trial with an investigational anti-HIV agent. The traces correspond to (A) the TIC chromatogram and (B) the “phase I” and (C) the “phase II” fractional mass-filtered chromatograms. The unchanged drug elutes at a retention time of 10.7 min, while glucuronide conjugates of an N-demethylated metabolite and parent drug elute at 7.4 and 7.9 min, respectively. In-source fragmentation of the glucuronides results in detection of the corresponding aglycones in the “phase I” chromatogram (B). Reprinted with permission from ref (11). Copyright 2008 Wiley.
large-scale clinical trials. These recommendations have a number of key practical implications for drug development, namely, that (1) comparative profiles of circulating metabolites in animals and humans be defined, both qualitatively and quantitatively (or at least semiquantitatively), as early as possible; (2) steadystate conditions apply for the human evaluation, thereby requiring a multiple dose regimen; and (3) the desired early assessment of metabolites in human plasma would occur prior to the availability of radiolabeled drug that has been manufactured and formulated in a manner suitable for the traditional absorption, distribution, metabolism, and excretion (ADME) study, which generally is accepted as the “gold standard” method for defining the fate of a drug candidate in man3 (8). With regard to the latter constraint, it is possible to obtain preliminary information on the metabolic profile of a drug candidate in human plasma by use of a 14C tracer in a “microdose” or “lightly labeled” approach (which requires much less investment of chemistry and toxicology resources prior to the human study); in this scenario, detection of metabolites is achieved by accelerator mass spectrometry (AMS) following off-line LC separation with fraction collecting (9, 10). However, identification of these metabolites requires a separate analysis, for example, by liquid chromatography-tandem mass spectrometry (LC-MS/MS) techniques, since AMS methodology involves degradation of carbonaceous samples to graphite prior to analysis. From the foregoing discussion, it will be evident that compliance with the recommendations outlined in the FDA Guidance document will require new analytical methodology that provides an unbiased, and at least semiquantitative, survey of circulating drug metabolites in the plasma of human subjects 3 Human ADME studies normally are conducted according to a single dose protocol using a 14C-labeled variant of the drug candidate that has been synthesized under Good Manufacturing Practices (GMP) conditions and appropriately formulated for administration to humans. The need for prior dosimetry evaluation in animals, together with an estimate of the likely therapeutic dose of the drug (that would be adopted for the ADME study), dictate that the ADME study usually is not performed until after proof-ofconcept has been obtained in humans (i.e., post-phase IIA); this may not occur until 1-2 years following the first-in-human trials. In contrast, metabolic profiling studies in vitro and ADME studies in animals typically are performed at the preclinical stage.
dosed to steady state with the investigational agent, in most cases without the benefit of a radioactive isotope.
II. Chemically Stable Drug Metabolites Recently, Tiller and co-workers at Merck Research Laboratories (11) reported on the application of liquid chromatography-mass spectrometry (LC-MS), employing high-resolution time-of-flight mass analysis, to the detection, identification, and semiquantitative assessment of drug metabolites in human plasma following administration of “cold” (nonradioactive) doses of a number of representative drug candidates. Through acquisition of high-resolution mass spectral data on all components eluting from the LC column, accurate m/z ratios for MH+ ions are obtained to a high degree of precision (typically 5 parts per million), thereby allowing determination of the elemental composition of all analytes of interest. The fractional mass (i.e., the integers following the decimal point in an accurate mass) reflects the elemental composition of the molecule (or ion) in question and affords a mechanism through which drugrelated components of the sample may be recognized against a background of endogenous materials. Recognition of xenobiotic as opposed to endogenous constituents by this approach relies on the fact that the great majority of drug molecules exhibit a so-called “negative mass defect” as compared to products of intermediary metabolism as a consequence of their relative deficiency in hydrogen atoms (i.e., high degree of unsaturation due mainly to multiple ring systems), a property that may be enhanced significantly by the presence of one or more halogen atoms (F, Cl, and Br). Computer-based algorithms have been devised to search LC-MS data sets specifically for those components of a biological sample that possess fractional masses that are consistent with either unconjugated or conjugated metabolites of the drug of interest, in a process termed “fractional mass filtering” or “mass defect filtering” (12). By means of this technique, drug metabolites may be distinguished from background and then subjected to structural elucidation by collision-induced dissociation MS/MS analysis. An example of the power of this approach is shown in Figure 1, which compares the total ion current (TIC) chromatogram (panel A)
PerspectiVe
from a plasma sample taken from a volunteer given a dose of an exploratory anti-HIV drug with the corresponding “phase I” (panel B) and “phase II” (panel C) fractional mass-filtered chromatograms (the terminology “phase I” and “phase II” referring here to unconjugated and conjugated metabolites, respectively). The component eluting at 10.7 min exhibited the LC retention time and accurate mass of the parent drug, while those with retention times of 7.4 and 7.9 min had accurate masses consistent with glucuronide conjugates of a demethylated metabolite and the parent drug, respectively. The reason that signals were detected for the two conjugates in the “phase I” chromatogram likely was due to in-source fragmentation leading to loss of the conjugating moiety, a well-known characteristic of this conjugate class. Confirmation of these structural assignments was based upon product ion spectra and computer-assisted analysis of fragmentation patterns (13). This example serves to demonstrate that it is now possible to obtain a rapid, comprehensive, and unbiased profile of drug metabolites in human plasma following administration of the unlabeled parent compound in an early clinical trial. Indeed, the speed and versatility of this ultraperformance liquid chromatography (UPLC)-highresolution MS approach has led to the adoption of a new paradigm for drug metabolite profiling in Merck‘s drug discovery and lead optimization programs (13). Having detected and identified metabolites circulating in humans, the remaining step is to assess their quantitative importance to establish whether exposure to the same metabolites in animals exceeds that in humans and, if not, whether such “disproportionate” human metabolites exceed 10% of the area under the plasma concentration vs time curve (AUC) of unchanged parent. Because individual analytes may differ appreciably in their ionization efficiency, it is necessary to first calibrate the mass spectrometric response of metabolites relative to parent drug. This is accomplished by comparing the metabolic profiles obtained from in vitro and animal in vivo studies using radiolabeled drug with the corresponding LC-MS data from the same samples; a direct comparison of responses from flowthrough liquid scintillation counting LC and LC-MS analyses provides the necessary calibration between MS response and mass for each metabolite and parent and thereby allows for the semiquantitative assessment of all drug-related species in human plasma. When this approach is combined with proportional pooling of human plasma specimens (14), estimates may be made of the relative AUCs of metabolites and parent drug, thereby providing the necessary information as to whether any particular metabolite should be subject to further examination based on the 10% threshold alluded to in the FDA Guidance. It should be noted that this methodology is applicable to the situation where reference standards of metabolites are not available, provided that there is reasonable assurance that the metabolites observed in the in vitro and animal in vivo studies are the same as those found in the human plasma samples. In the event that this is not the case and that human plasma contains a metabolite(s) that has not been detected in any nonclinical studies, it would be necessary to isolate, identify, and synthesize the compound in question such that its exposure in humans relative to the parent can be assessed. While relatively straightforward, the potential shortcomings of this new methodology are that some metabolites may ionize so poorly under the conditions of the LC-MS experiment that they go undetected. Also, metabolic reactions, such as heteroatom dealkylations or hydrolyses that occur at an internal site and result in cleavage of the drug molecule into fragments that differ appreciably in size (and thus elemental composition) from
Chem. Res. Toxicol., Vol. 22, No. 2, 2009 265
the parent, may not be recognized by the fractional mass filtering technique. However, such events usually would be anticipated on the basis of nonclinical metabolism studies, and adjustments to the analytical protocol would be made accordingly. In the future, it may be of interest to apply stable isotope labeling techniques, such as the use of isobaric tags for relative and absolute quantitation (iTRAQ) reagents currently employed in proteomics research, for the detection of putative amines resulting from amide hydrolysis or N-dealkylation reactions. Overall, the application of fractional mass filtering, combined with classical metabolic profiling, represents a practical approach to address the need for semiquantitative data on circulating drug metabolites in first-in-human studies conducted with nonradiolabeled drug candidates. This paradigm now has been implemented at Merck for all multiple ascending dose safety/ tolerability studies with new drug candidates, while samples from the corresponding single dose first-in-human trials are employed for exploratory human metabolite profiling work.
III. Chemically Reactive Drug Metabolites It should be noted that the FDA Guidance document does not cover biologics or certain anticancer agents and is restricted to small molecule drug products. This is reasonable given the limited role that metabolism plays in the pharmacological and toxicological effects of therapeutic peptides, proteins, and nucleic acids. Also excluded from consideration (appropriately so) are chemically reactive drug metabolites, on the basis that these electrophilic species normally are short-lived and are not detectable per se in the systemic circulation of animals or humans. Although downstream products of reactive metabolites (e.g., glutathione conjugates) can serve as indices of exposure to electrophilic intermediates, such adducts usually do not circulate in plasma but are efficiently eliminated via the biliary route into the intestine and/or undergo conversion to the corresponding S-linked cysteinylglycine-, cysteine-, or Nacetylcysteine conjugates that are excreted via the kidneys into the urine. Other reactive metabolites (notably “hard” electrophiles such as iminium ions) do not give rise to stable glutathione conjugates and thus are not readily detected in vivo. In both cases, reactive metabolites bind irreversibly to cellular macromolecules, and while covalent binding to tissue proteins and DNA can be measured in animal studies, typically following administration of a radiolabeled tracer, this would be difficult to replicate in humans for ethical reasons. Aside from these practical complexities, it remains unclear how one would interpret information on exposure to reactive metabolites in humans, given our general lack of understanding of the molecular mechanisms by which electrophilic intermediates cause cellular injury in animals or humans. Thus, while alkylation of DNA has clear implications in terms of the potential for mutagenesis, teratogenesis, and carcinogenesis, the consequences of covalent binding of reactive drug metabolites to proteins remain poorly understood, even after some 40 years of research (15). What is known is that some reactive metabolites bind covalently, for example, to liver proteins and mediate the hepatotoxicity associated with the parent drug, while others bind to a comparable extent yet seemingly are benign. The reason(s) for these discrepancies remain to be established but undoubtedly relate, inter alia, to differences in the protein targets of the reactive electrophiles. In view of these observations, it is hardly surprising that two recent studies on the relationship between covalent binding and the hepatotoxic potential of a number of approved drugs led to different conclusions (16, 17). However, the topic of metabolic activation is an important one
266
Chem. Res. Toxicol., Vol. 22, No. 2, 2009
for the pharmaceutical industry, as highlighted by an estimate from one company that reactive metabolites did appear to be an issue in at least one-quarter of all toxicology-related development failures over the period 1993-2006 (18). Even if this figure is an overestimate, the extremely high cost of failure in drug development suggests that efforts to minimize the potential of drug candidates to undergo metabolic activation through appropriate predevelopment evaluation represent a prudent and practical approach to the problem (19). Promising research in the area of reactive metabolite toxicology includes studies on the identities of the protein targets of toxic vs nontoxic electrophiles (20) and on the combined application of covalent binding measurements with transcriptomic, metabonomic, and proteomic technologies in an effort to discern (and thereby predict) the characteristics of a toxic response (18). Given the state of the art of this field, it is understandable that regulatory agencies have not adopted a position on the role of chemically reactive drug MIST, as to do so would be scientifically premature at this juncture. Rather, the advancement of potent (low dose) drug candidates with only a limited propensity to form reactive intermediates would appear to be the favored strategy in today’s environment (19).
IV. Conclusions In conclusion, the topic of drug MIST is a complex one that requires that all aspects of the drug candidate of interest be taken into consideration in developing an appropriate strategy for development. While the recent FDA Guidance and both the publications cited in this perspective and those that appear in this special section of Chemical Research in Toxicology provide a helpful framework upon which to build such a strategy, flexibility is key, and a case-by-case, science-based approach to the issue will best serve the interests of the pharmaceutical industry, regulatory agencies, and consumers (21). Acknowledgment. I acknowledge the contributions of many colleagues at Merck Research Laboratories to the work described in this perspective, notably Drs. Phil Tiller, Sean Yu, Kevin Bateman, Nancy Agrawal, and Clay Frederick.
References (1) 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. (2) Guengerich, F. P. (2006) Safety assessment of stable drug metabolites. Chem. Res. Toxicol. 19, 1559–1560. (3) Humphreys, W. G., and Unger, S. E. (2006) Safety assessment of drug metabolites: Characterization of chemically stable metabolites. Chem. Res. Toxicol. 19, 1564–1569.
Baillie (4) 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. (5) Smith, D. A., and Obach, R. S. (2006) Metabolites and safety: What are the concerns, and how should we address them? Chem. Res. Toxicol. 19, 1570–1579. (6) Luffer-Atlas, D. (2008) Unique/major human metabolites: Why, how, and when to test for safety in animals. Drug Metab. ReV. 40, 447– 463. (7) U.S. Food and Drug Administration/CDER (2008) Guidance to Industry. Safety Testing of Drug Metabolites. www.fda.gov/cder/ guidance/6897fnl.htm. (8) Roffey, S. J., Obach, R. S., Gedge, J. I., and Smith, D. A. (2007) What is the objective of the mass balance study? A retrospective analysis of data in animal and human excretion studies employing radiolabeled drugs. Drug Metab. ReV. 39, 17–43. (9) Garner, R. C. (2005) Less is more: The human microdosing concept. Drug DiscoVery Today 10, 449–451. (10) Brown, K., Dingely, K. H., and Turtletaub, K. W. (2005) Accelerator mass spectrometry for biomedical research. Methods Enzymol. 402, 423–443. (11) Tiller, P. R., Yu, S., Bateman, K. P., Castro-Perez, J., McIntosh, I. S., Kuo, Y., and Baillie, T. A. (2008) Fractional mass filtering as a means to assess circulating metabolites in early human clinical studies. Rapid Commun. Mass Spectrom. 22, 3510–3516. (12) Zhu, M., Ma, L., Zhang, D., Ray, K., Zhao, W., Humphreys, W. G., Skiles, G., Sanders, M., and Zhang, H. (2006) Detection and characterization of metabolites in biological matrices using mass defect filtering of liquid chromatography/high resolution mass spectrometry data. Drug Metab. Dispos. 34, 1722–1733. (13) Tiller, P. R., Yu, S., Castro-Perez, J., Fillgrove, K. L., and Baillie, T. A. (2008) High-throughput, accurate mass liquid chromatography/ tandem mass spectrometry on a quadrupole time-of-flight system as a ’first-line’ approach for metabolite identification studies. Rapid Commun. Mass Spectrom. 22, 1053–1061. (14) Hop, C. E. C. A., Wang, Z., Chen, Q., and Kwei, G. (1998) Plasmapooling methods to increase throughput for in vivo pharmacokinetic screening. J. Pharm. Sci. 87, 901–903. (15) Baillie, T. A. (2006) Future of toxicologysMetabolic activation and drug design: challenges and opportunities in chemical toxicology. Chem. Res. Toxicol. 19, 889–893. (16) Obach, R. S., Kalgutkar, A. S., Soglia, J. R., and Zhao, S. X. (2008) Can in vitro metabolism-dependent covalent binding data in liver microsomes distinguish hepatotoxic from non-hepatotoxic drugs? An analysis of 18 drugs with consideration of intrinsic clearance and daily dose. Chem. Res. Toxicol. 21, 1814–1822. (17) Takakusa, H., Masumoto, H., Yukinaga, H., Makino, C., Nakayama, S., Okazaki, O., and Sudo, K. (2008) Covalent binding and tissue distribution/retention assessment of drugs associated with idiosyncratic drug toxicity. Drug Metab. Dispos. 36, 1770–1779. (18) Guengerich, F. P., and MacDonald, J. S. (2007) Applying mechanisms of chemical toxicity to predict drug safety. Chem. Res. Toxicol. 20, 344–369. (19) Evans, D. C., Watt, A. P., Nicoll-Griffith, D. A., and Baillie, T. A. (2004) Drug-protein adducts: an industry perspective on minimizing the potential for drug bioactivation in drug discovery and development. Chem. Res. Toxicol. 17, 3–16. (20) Liebler, D. C. (2008) Protein damage by reactive electrophiles: Targets and consequences. Chem. Res. Toxicol. 21, 117–128. (21) Davis-Bruno, K. L., and Atrakchi, A. (2006) A regulatory perspective on issues and approaches in characterizing human metabolites. Chem. Res. Toxicol. 19, 1561–1563.
TX800439K
