Introduction

Top Abstract Introduction Discussion References

If we take a long view of the history of the use of drug metabolism (DM)¹ studies in industrial drug discovery and development, we can see several phases. For many years, companies developing new drugs did not pay any attention to the question of the metabolic fate of the compound. Indeed, the question was not even asked. Later, as regulatory agencies began to expect a more complete accounting of the actions and disposition of new drugs submitted for marketing approval, companies introduced the isolation and identification of metabolites recovered from urine and feces. We may call this the "chemistry" phase. This phase was largely descriptive, with little understanding of the phenomena leading to certain metabolic pathways except the purely chemical properties of the drug compound itself.

Then, in the 1970s, we entered the "biochemistry" phase, which allows us to account for and even predict interspecies differences, individual variation, and drug-drug interactions occurring because of drug-metabolizing enzymes. Presently, we are watching the development of the "genetics" phase, in which various technologies are being introduced to allow industry to take into consideration the genetic determinants of drug metabolism during the preclinical selection and clinical evaluation of new drug candidates. Interestingly, this is not a process of serial replacement of one approach by the next, because nothing has become obsolete: all phases continue to coexist. For instance, the isolation and identification of metabolites is even more important today than it was in the past.

What, then, is the next phase? For this, one must look into the ultimate destination of medicine. In other words, what will therapeutics look like in the future? I believe that in the first quarter of the 21st century, we will enter a new era, which I will call the "biology" phase. This article will examine these phases and the evolving direction of DM, especially as it is applied to industrial drug discovery and development.

	Discussion

Top Abstract Introduction Discussion References

Present Practices. A good place to start is to consider the role of DM in present-day drug development. We can consider that DM is involved in three distinct areas of the process: discovery, nonclinical development, and clinical development. Although DM and pharmacokinetics (PK) are interrelated, they are nonetheless distinct, and this discussion will not consider PK except as the consequence of DM.

Historical Perspective. Another way to project trends in DM is to look at the forces that have driven DM evolution in the past. The history of DM in industrial drug discovery/development can be characterized as the pursuit of an increasingly greater scientific understanding of the DM aspects of the clinical behavior of new drug entities and correspondingly smaller fractions of purely phenomenological descriptions of the drug. I have, somewhat arbitrarily, divided this evolution into phases. Let us examine these phases and see where they direct us into the future.

"Chemistry" phase of industrial DM (1950-1980). Although DM has existed as a science since the 19^th century (Conti and Bickel, 1977), it was not incorporated into the development of new drugs until much later. During this "chemistry" phase, the main DM objective for the registration of a new drug entity was to account for the elimination of drug-related materials from the body. This was primarily accomplished by mass balance studies based on the recovery of radioactivity during prolonged collection periods after dosing with a radiolabeled drug. Isolation and identification of metabolites from urine was accomplished to the extent possible. In the earlier part of this period, this literally involved chemical isolation and crystallization of the metabolites, followed by traditional chemical identification methods (i.e. elemental analysis, solid derivatives, infrared spectroscopy, and mass spectrometry). Proof-of-structure was accomplished by unambiguous chemical synthesis of the putative metabolite structure. Obviously, only the major metabolites excreted in the urine were amenable to these classic chemical-identification methods. In the latter part of this period, the introduction of high-performance liquid chromatography made possible the detection of less prominent urinary metabolites as well as any major circulating metabolites. Overall, this phase was largely descriptive, with some understanding of the purely chemical phenomena resulting in a particular metabolic pathway, but little understanding of the biological determinants of the processes, such as the characteristics, regulation, and even location of the enzymes involved.

"Biochemistry" phase of industrial DM (1975-1995). In this discussion, the term "biochemistry" is used in a broad sense, encompassing not only the subcellular processes responsible for xenobiotic removal but also the structural chemistry of the enzymes and the molecular biology of the genes involved. This phase began with a decade of basic biochemical research in which the DM enzymes were studied at the molecular level. Proof of the "isozyme" hypothesis of P450 substrate selection was accomplished by the isolation and rigorous biochemical characterization of the hypothesized enzyme forms, thereby demonstrating their actual existence. Subsequent work showed the presence of many more forms of P450 than anyone had guessed existed, based merely on differential substrate metabolism patterns. By the early 1990s, the human P450 enzyme family was well-established, and it was possible to dissect such phenomena as species differences, interindividual differences, differential enzyme induction, and drug-drug metabolic interactions into the individual contributions made by discrete P450 enzymes, thereby allowing not only understanding of these phenomena in the clinical setting but also their preclinical prediction.

"Genetics" phase of industrial DM (1990-present and into the future). Today we are seeing the application of the previous decade's basic research into industrial and clinical practice. We now routinely consider the genetic determinants of DM at the clinical stage and even the discovery stage. This consideration is still limited to phenotyping subjects in clinical trials as "poor" or "extensive" metabolizers of standard substrates for the polymorphic P450 enzymes (i.e. CYP 2C9, 2C19, 2D6) when drug candidates have been found to be predominantly metabolized by these enzymes. We will soon see the genotyping of subjects in clinical trials and, eventually, of all patients. As more information comes in from basic research, we will undoubtedly find polymorphisms in other drug-metabolizing enzymes and drug transporters and will wish to add these forms to the list of genes that are typed. "Gene-chip" technology will soon make large-scale genotyping of many enzymes and proteins a practical reality. This ability to determine the genotype must be accompanied by an understanding of the clinical consequence of the particular polymorph for a particular drug, in the same way that we presently associate CYP 2D6-deficient patients with adverse events with certain drugs. Our ability to manipulate the genes for drug-metabolizing enzymes will provide us with transgenic "humanized" animal models that might more perfectly reflect certain clinical aspects of new drugs than do our present in vitro methods, allowing more reliable prospective investigation of compounds prior to their selection as clinical candidates. On a purely utilitarian level, we have already seen the large-scale expression of single enzymes in bacterial, yeast, or insect cells, making practical their extensive use as research tools for drug development at an economically feasible cost.

"Biology" phase of DM (2010?). To extrapolate beyond a few years into the future requires an understanding of the direction of DM research and subsequent application. Toward that end, three premises are introduced here.

I.	The driving force for the evolution of DM research has been, and will remain, the need to assure safety and efficacy in the clinical application of new drug entities.
II.	Several "black box" areas exist in our current understanding of DM.
III.	Therapeutics will see a gradual decline in the use of small organic molecules in favor of peptides, proteins, nucleic acids and viral vectors to manipulate genes to restore normal physiology in diseased tissue.

1.	What is the molecular basis for the seemingly mutually exclusive phenomena of high discrimination for oxidation at a particular site on a given substrate and low discrimination among many distinctly different substrate chemotypes by P450 enzymes? This is especially true of CYP 3A4, which also exhibits some poorly understood interactions between substrates and inhibitors (Wang et al., 1997).
2.	What are the three-dimensional structures of the DM enzymes, including the P450s, the flavin-containing monooxygenases, the many types of conjugating transferases, the keto-reductases, the carboxylesterases, etc.? Although homology modeling methods are revealing much useful information (Graham-Lorence et al., 1995; De Groot and Vermeulen, 1997; Lewis and Lake, 1997; Szklarz and Halpert, 1997; Tan et al., 1997), we realize that direct structural determination would be much better.
3.	How is the reactive oxygen intermediate of P450 produced, what is its structure, and how does it oxidize substrates? Despite over 20 years of effort, the reactive oxygen intermediate has not been isolated and structurally characterized. Recent new proposals for the structure show that although there is a widespread perception that this issue is settled, it remains a subject of debate and surprises (Newcomb et al., 1995; Benson et al., 1997; Vaz et al., 1998).
4.	Can we reliably predict the human P450 induction pattern of a new drug candidate on the basis of preclinical data? There is general acknowledgment that induction in animals can be very misleading, and much better methods, perhaps based on cultured hepatocytes, are needed.
5.	How can one predict reactive metabolite-mediated toxicities? This remains possibly the greatest challenge in contemporary DM because it involves the prediction of not only the formation of reactive metabolites but also which cellular nucleophiles are susceptible to attack and what the consequence will be of covalent modification of the nucleophile. In addition, the possibility of free-radical chain reaction initiation by the reactive metabolites is only beginning to be appreciated (Nelson, 1995).
6.	What other forms of mammalian P450 have yet to be discovered, where are they localized, and what are their functions? The vast majority of P450 research has focused on the hepatic enzymes. Not until the early 1990s was the importance of intestinal CYP 3A4 realized, and we know now that P450 enzymes can be found in such diverse tissues as skin, nasal mucosa, and brain. What are all these P450 enzymes really for?