Metabolite profiling in early clinical drug development: current status and future prospects

KEYWORDS:

• Metabolite profiling
• drug development
• metabolomics
• human ADME
• micro-dose
• micro-tracer

1. Introduction

Metabolite profiling in context of drug development aims at identifying and quantifying metabolites including their intermediates of an investigational product (IP). Such data are considered of critical value for understanding the routes of elimination, predicting drug–drug interactions, and anticipating safety concerns in humans.

Traditionally, human absorption, distribution, metabolism, and elimination (hADME) studies have been conducted at a relatively late stage of drug development to generate metabolite data, typically from plasma, feces, and urine samples. Mostly, liquid chromatography-mass spectrometry has been applied to quantify metabolites that commonly had previously been identified in vitro or in animals [1].

Most recently, it has been encouraged by the Food and Drug Administration (FDA) agency to initiate metabolite profiling in humans as early as possible during drug development [2]. This is primarily driven by the interest to prevent unexpected safety issues that may arise from disproportionate (i.e. human-specific) toxic metabolites potentially leading to program termination eventually at a late stage of drug development [2]. Hence, an early understanding of the metabolite profile in humans may increase drug safety which is of greatest regulatory and public health interest. In addition, such knowledge may contribute to reduced attrition rates and overall development costs.

In terms of efficacy, metabolite profiling could also facilitate describing the time course of drug effects in case of active metabolites. This is particularly interesting in case one of the active metabolites may have favorable properties over the parent compound. Such positively differentiating metabolite may then be subjected to further clinical development as stand-alone product which can be a valuable ‘by-product’ arising from metabolite profiling. A number of marketed drugs such as cetirizine, fexofenadine, or desloratadine actually represent active metabolites of a parent compound with less favorable properties [3].

Finally, metabolite profiling allows estimating the relative importance of drug-metabolizing enzymes/pathways for the elimination of the IP which is of relevance for concomitant drug use requirements, need for drug–drug interaction studies, and pharmacokinetic (PK)-based differentiation aspects.

Two technologies are being applied for metabolite profiling purposes, namely nuclear magnetic resonance spectroscopy (NMR) and mass spectroscopy (MS) [4]. NMR-based metabolite profiling is best suited for comprehensive, structural identification of metabolites, while MS-based approaches allow for metabolite quantification with high sensitivity provided availability of synthesized metabolite standards. Recent advancements in bioanalytical technology allow integration of metabolite profiling in first-in-human (FIH) or phase-0 studies using micro-tracers.

2. Metabolite profiling in ‘conventional’ hADME studies

Most pharmaceutical companies still perform dedicated hADME studies in healthy subjects to derive metabolite profiling data. It is a highly reliable approach to identify and quantify human metabolites that has gained wide regulatory acceptance and typically includes single-dose administration of the mostly ¹⁴C radiolabeled compound.

The radioactive dose must be sufficiently high to allow adequate quantification by liquid scintillation counting (Figure 1). Therefore, one disadvantage of such ‘conventional’ hADME study is the requirement to determine the individual radiation burden based on quantitative whole-body autoradiography data from animal studies. According to the International Commission on Radiological Protection (ICRP), radiation doses ≤1 mSv are acceptable in the context of biomedical research (ICRP Category IIa) [5]. Moreover, use of good manufacturing practice (GMP) material may eventually be required leading to substantially increased study costs (Table 1).

Table 1. Comparison of conventional vs. novel approach to metabolite profiling.

	‘Convential approach’ (=hADME study)	‘Novel approach’(=micro-tracer or micro-dose study)
Ethics/safety aspects	Radiation burden needs to be calculated based on QWBA data	No radiation burden determination required.
		Radiation burden is decreased if using micro-tracers (e.g. useful for drugs with long t1/2 or administration to children)
		Micro-dosing of study drug (e.g. useful in case of narrow safety margin)
CMC aspects	Use of non-GMP material may not be acceptable	Non-GMP material can be used
Technology	Liquid scintillation counting	Accelerated mass spectrometry
Regulatory aspects	Well-established approach proven to be adequate for regulatory submissions	Limited toxicology data sufficient (e.g. no genotoxicity data required for micro-dose studies)
Costs [6]	≃1.5–3.0 Mio. USD	≃0.3–0.5 Mio. USD
Data availability	Phase II	Phase I
Pharmacological aspects	Pharmacologically active dose	Non-pharmacologically active dose, i.e. lack of PD and safety data

CMC: chemical manufacturing control; GMP: good manufacturing practice; hADME: human absorption, distribution, metabolism, and elimination; PD: pharmacodynamics; QWBA: quantitative whole-body autoradiography.

Figure 1. Different approaches of metabolite profiling in clinical development.

In view of the high-attrition rates during clinical development, hADME studies are still often conducted at a relatively late stage to avoid up-front investments at risk. However, these studies may provide unexpected findings impacting safety monitoring, restrictions on concomitant drug use, or need for additional nonclinical or clinical studies.

3. Metabolite profiling in FIH studies using a micro-tracer approach

Given the advancement of bioanalytical technologies, it is now possible to markedly reduce the radioactive dose to 1–2 μCi or even below (so-called micro-tracer) and integrate metabolite profiling in FIH studies [7]. This requires quantification by highly sensitive analytical platforms (e.g. accelerator MS). Administration of such low radioactive doses corresponds to a radiation burden <0.1 mSv (ICRP Category I) and does not require any radiation burden determination or use of GMP-compliant ¹⁴C-radiolabeled material [8] (Table 1).

In addition to the tracer, the IP needs to be administered at a pharmacologically active dose in order to generate sufficient amounts of metabolites for structural identification. Two different options exist to administer the tracer: (1) to include the radiolabeled material in the IP formulation and (2) the micro-tracer is given in addition to a pharmacologically active dose of the IP (Figure 1). The latter approach is less costly, but may be complicated by differences in oral bioavailability of the IP versus micro-tracer. A higher bioavailability at a therapeutic dose may be due to saturation of P-gP-mediated efflux transport and gut wall metabolism as reported for clarithromycin [9]. Differences in the estimates of therapeutic exposure based on predictions from micro-tracer data generally become a concern in case of greater than or equal to two-fold deviations [9].

Occasionally, metabolite profiling is integrated in FIH studies by analyzing label-free compound mostly in plasma and urine. This less costly approach, however, only allows for semiquantitative assessments (i.e. relative abundance with reference to parent compound) unless synthesized metabolite standards are already available. This so-called ‘cold’ metabolite identification requires putting the human data into context with in vitro and/or animal data (e.g. retention times) and comes with the risk of missing certain metabolites due to poor ionization during MS [10].

4. Metabolite profiling in phase-0 studies using a micro-dose approach

Starting early 2000s, the concept of phase-0 studies evolved including administration of ¹⁴C radiolabeled IP at ‘micro-doses’, i.e. IP doses <100 μg or <1% of the amount predicted to yield a pharmacological effect, whichever is lower [7]. The radioactive tracer is integrated in the IP and given at low radioactive doses corresponding to a radiation burden <0.1 mSv (ICRP Category I) (Figure 1) [8].

Consequently, the preclinical toxicology package required for the conduct of micro-dose studies is leaner, e.g. genotoxicity data are not required in contrast to FIH studies [11]. The main advantage of this approach is minimal safety risk for study subjects and early data availability. However, lack of pharmacodynamic data essentially precludes predictions of the tentatively pharmacologically active dose unless exposure–response data can be derived from competitor compounds having the same mechanism of action. Also, for compounds without dose-proportional PK, a micro-dose approach may not adequately predict PK parameters at pharmacologically active doses.

5. Other applications of metabolite profiling in drug development

Over the past decade, metabolomics has evolved as a novel research field addressing the characterization of any endogenous metabolite in a given biological system [12]. These broader metabolomic approaches may provide increasing value for drug development.

This includes metabolite fingerprinting of metabolite patterns in response to pathophysiological stimuli. This untargeted, hypothesis-generating approach may be used for target identification [13].

In context of personalized medicine, pretreatment profiles of endogenous metabolites may allow for identification of target engagement or disease progression markers and thereby guide patient selection or individualized dosing to improve the benefit/risk ratio. This so-called pharmaco-metabonomic approach takes into account both genetic and environmental aspects which is perceived advantageous over pharmacogenetic/-genomic approaches [14]. Metabolomics may also contribute to the identification of diagnostic biomarkers, since alterations of the metabolome are expected to precede histopathological or pathophysiological changes associated with chronic diseases, e.g. in case of kidney dysfunction [15,16]. Early diagnosis and monitoring of disease progression as well as assessment of drug response or toxicity by means of metabolomic markers in urine appear particularly attractive in neonatology or pediatrics given its noninvasiveness [17].

In summary, scientific knowledge and technology regarding metabolomics are rapidly evolving (e.g. Human Metabolome Project), but the concrete value for drug development and ultimately for the patient is still far from being established.

6. Expert opinion

Metabolite profiling of IPs has been part of drug development for many years. Traditionally, this aspect has been addressed by means of in vitro and animal studies as well as a hADME study typically conducted rather late in the drug development process.

The feasibility of metabolite profiling applying micro-dose or micro-tracer approaches has increased in recent years. A comparison of 40 hADME studies revealed similar cumulative recovery using regular (n = 28) or low radioactive doses (n = 12) and metabolite profiling was considered successful in each study irrespective of the radioactive dose used [18]. Accordingly, the PK profile of clarithromycin, sumatriptan, propafenone, paracetamol (acetaminophen), and phenobarbital determined on the basis of micro-doses or therapeutic doses was essentially well matched [19].

Integration of metabolite profiling at the earliest possible stage of clinical drug development has most recently also been encouraged by the FDA [2]. It has several advantages:

• Early identification of reactive and/or disproportionate metabolites allows for timely conduct of targeted safety testing, thereby reducing development costs

• Apart from safety-related aspects, early metabolite profiling may allow identification of active metabolites and in turn a better understanding of the time course of drug action

• Additional up-front costs are associated with a micro-dose/-tracer approach, but these investments are considered to pay off down the road provided the IP is pushed further through the development life cycle (Table 1). In addition, micro-dose trials have been reported to be cost-effective by allowing for early go/no-go decisions of drug candidates [6].

• Integration of metabolite profiling in early clinical development is well in line with efforts to maximize the informative value of FIH trials by other complementary objectives, e.g. related to food effect, proof-of-mechanism, etc.

Despite multiple advantages, most companies have even in recent years still run dedicated hADME studies at a relatively late stage of development. However, given the most recent release of the FDA guidance regarding safety testing of drug metabolites, a paradigm shift is expected such that human metabolite profiling may be integrated more often in FIH studies.

Declaration of interest

The authors are all employees of Idorsia Pharmaceuticals Ltd or hold stock or share options. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Additional information

Funding

This paper is not funded