Metabolite Bioanalysis in Drug Development: Recommendations from the IQ Consortium Metabolite Bioanalysis Working Group

The intent of this perspective is to share the recommendations of the International Consortium for Innovation and Quality in Pharmaceutical Development Metabolite Bioanalysis Working Group on the fit‐for‐purpose metabolite bioanalysis in support of drug development and registration. This report summarizes the considerations for the trigger, timing, and rigor of bioanalysis in the various assessments to address unique challenges due to metabolites, with respect to efficacy and safety, which may arise during drug development from investigational new drug (IND) enabling studies, and phase I, phase II, and phase III clinical trials to regulatory submission. The recommended approaches ensure that important drug metabolites are identified in a timely manner and properly characterized for efficient drug development.


Metabolite Bioanalysis in Drug Development: Recommendations from the IQ Consortium Metabolite Bioanalysis Working Group
Wenkui Li 1, * , Faye Vazvaei-Smith 2, * , Gordon Dear 3 , Jason Boer 4 , Filip Cuyckens 5 , Daniela Fraier 6 , Yuexia Liang 2 , Ding Lu 7 , Heidi Mangus 8,15 , Patricia Moliner 9 , Mette Lund Pedersen 10 , Andrea A. Romeo 6 , Douglas K. Spracklin 11 , David S. Wagner 12 , Serge Winter 13 and Xiaohui (Sophia) Xu 14 The intent of this perspective is to share the recommendations of the International Consortium for Innovation and Quality in Pharmaceutical Development Metabolite Bioanalysis Working Group on the fit-for-purpose metabolite bioanalysis in support of drug development and registration.This report summarizes the considerations for the trigger, timing, and rigor of bioanalysis in the various assessments to address unique challenges due to metabolites, with respect to efficacy and safety, which may arise during drug development from investigational new drug (IND) enabling studies, and phase I, phase II, and phase III clinical trials to regulatory submission.The recommended approaches ensure that important drug metabolites are identified in a timely manner and properly characterized for efficient drug development.
This paper was developed with the support of the International Consortium for Innovation and Quality in Pharmaceutical Development (IQ Consortium).The IQ Consortium is a notfor-profit organization of pharmaceutical and biotechnology companies with a mission to advance science-based and scientifically driven standards and regulations for pharmaceutical and biotechnology products worldwide.Within the IQ Consortium, the Metabolite Bioanalysis Working Group (WG) was formed under Translational and ADME Science Leadership Group to provide a scientific forum for the harmonization of best practice in metabolite bioanalysis in support of drug development.
Within the pharmaceutical industry, project teams are often challenged by questions related to drug metabolites.A drug metabolite may contribute to the pharmacological effect (efficacy) of the parent drug or pose a safety risk to humans due to off-target toxicity or reactivity, or more likely have no effects at all.With projects progressing to first-in-human (FIH) trials and beyond, there is a need to obtain information on circulating human metabolites and metabolite exposure coverage between humans and nonclinical species used in repeated-dose safety studies.2][3] However, some recurring questions remain to be satisfactorily addressed.These questions include but are not limited to: (i) what drivers dictate when to start or stop quantification of a pharmacologically "active" metabolite; (ii) how to effectively assess metabolite exposure margins between humans and nonclinical species from the perspective of "Metabolites in Safety Testing" (MIST); and (iii) how best to address metabolites with a human exposure < 10% of total drug related exposure (TDRE) or total drug-related material (TDRM) following a single dose human absorption, distribution, metabolism, and excretion (hADME) study but with potential for accumulation at steady-state (SS) after multiple dose administration.
Bioanalytical (BA) methods for metabolite bioanalysis can be categorized in three tiers, that is, exploratory, qualified, and validated, with each tier bearing different rigor from compliance and validity perspectives and being associated with distinct levels of investment of BA resources (cost and time). 4It appears there is still a lack of clarity in the community in three main areas of metabolite bioanalysis: (i) when a metabolite should be quantified; (ii) which tier of BA method should be used for metabolite bioanalysis that is fit-for-purpose for the intended study (in vitro and in vivo); and (iii) after deciding to quantify a metabolite, does this metabolite need to be quantified for all subsequent studies or just selected studies.To add additional ambiguity, the US Food and Drug Administration (FDA) guidance on "In Vitro Drug Interaction Studies-Cytochrome P450 Enzyme and Transporter Mediated Drug Interactions" recommends a "validated" BA method for measuring in vitro metabolite(s). 5Due to these reasons, there has been a growing conservative trend within the industry to quantify metabolite(s) too early or in too many studies, or to use fully validated method(s) to generate metabolite exposure data for projects that may never reach pivotal phase (i.e., phase III) or, in many cases, even clinical stage (i.e., FIH).This extra work only drives up the cost to develop new medicines and lengthens the time needed for important new therapies to reach patients.
In addition, chiral metabolites present some unique challenges for quantitative bioanalysis.Developing and qualifying/ validating a chiral assay is generally much more difficult and time-consuming than an achiral method.Therefore, it is highly desirable to have a well-considered decision-making process with appropriate analytical rigor to address specific needs related to chiral metabolites in a fit-for-purpose and streamlined manner. 6,7aking the above into consideration, there is a strong need for the industry to come together to discuss and harmonize scientific best practices for metabolite bioanalysis.In this undertaking, the WG conducted a cross-industry survey among the 29 member companies using a series of questions covering metabolite bioanalysis practices, concentrating on (i) pharmacologically active vs. inactive metabolites, (ii) metabolite characterization and bioanalysis in relation to safety testing, (iii) metabolite-mediated drug-drug interaction (DDI), (iv) fit-for-purpose metabolite bioanalytical methods, and (v) chiral metabolite bioanalysis.When the survey was completed, the results were thoroughly examined by the WG members to set up the basis for harmonization discussion on those topics.This output from the WG along with case studies is presented to summarize the high-level outcomes of the survey and the post-survey harmonization discussions to reflect the current positions of the IQ member companies on metabolite bioanalysis and to recommend best practices on fit-for-purpose metabolite bioanalysis.

BIOANALYSIS OF PHARM ACO LOG ICA LLY ACTIVE VS. INACTIVE DRUG METABOLITE(S)
Metabolism is an important clearance pathway for many drugs, often resulting in more water soluble, more easily excreted, and usually more benign chemical entities.][10][11] If a metabolite has sufficient in vitro and/ or in vivo potency (pharmacological activity) against the same primary target as the parent drug and contributes substantially to the overall desired pharmacological effect, it is commonly referred as an "active" metabolite.An active metabolite has the potential to affect clinical efficacy and/or safety through increased and/or prolonged pharmacological effect beyond that expected for the parent drug alone (supra-pharmacology).As of 2003, ~ 22% of the top 50 prescribed drugs in the United States were reported to have active metabolites that contribute significantly to the overall pharmacological effect. 12In contrast, an "inactive" metabolite has no or insufficient pharmacological activity at the primary target.A metabolite may also elicit a toxicological response, which may or may not be the same as the parent drug.Given this complexity, to understand the potential contributions of metabolites to drug efficacy and safety, it is important to quantify metabolites that contribute "significantly" to the pharmacological effect or trigger treatmentrelated toxicity.
The continuum of traditional drug development spanning discovery through registration is shown in Figure 1.There are several key points at which metabolite information may be collected, including structure, potency, and exposure, which will inform on an integrated decision to start and/or stop quantification of an active metabolite(s).In drug discovery, metabolic profiles are typically generated using in vitro systems for (i) metabolic soft spot optimization, (ii) identification of potential reactive/genotoxic metabolite(s), (iii) early understanding of the possible major routes of metabolism and the drug metabolizing enzymes involved, and (iv) cross-species comparison to aid toxicology species selection.In vivo samples are often examined to establish an in vitro-in vivo correlation, which may include an early assessment of metabolites in nonclinical species.At this stage, some companies undertake investment to synthesize metabolites for further characterization. 13n nonclinical development, known metabolites categorized as per the International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH) S3A guideline on assessing systemic exposure (toxicokinetics) in toxicology studies are generally quantified 14 when: • The administered compound acts as a "prodrug" and the delivered metabolite is acknowledged to be the primary active entity. 14 The compound is metabolized to one or more pharmacologically or toxicologically active metabolites which could make a significant contribution to tissue/organ responses. 14 The administered compound is extensively metabolized, and quantification of plasma or tissue concentrations of a major metabolite is the only practical means of estimating exposure following administration of the compound in toxicity studies. 14 this stage, information about human metabolites for a drug candidate is just a projection. 15,16When the project reaches the FIH trial stage, exploratory metabolite profiling is typically performed to identify circulating human metabolites to help guide the clinical development, including pharmacokinetic (PK)pharmacodynamic (PD) assessment, and MIST-and DDI-related investigations, etc. 17 For the latter, the 2020 FDA guidance on "In Vitro Drug Interaction Studies -Cytochrome P450 Enzyme-and Transporter-Mediated Drug Interactions" clearly states that the risk of a DDI when the metabolite acts as a substrate should be evaluated for a pharmacologically active metabolite that contributes toward at least 50% of the overall activity of the parent drug. 5n practice, a common challenge is the determination of the threshold that directs quantification of a metabolite that "significantly" contributes to the overall pharmacological effect or safety outcomes of the parent drug, while remaining proportionate with BA resources (cost and time) spent for a metabolite that does not meaningfully contribute to the effect of the parent drug.From the WG survey, ~ 80% of respondents begin quantifying pharmacologically active metabolite(s) based on the potency (in vitro or in vivo) of the metabolite and stop quantifying such metabolite(s) based on the in vivo exposure in combination with PD effect of the metabolite relative to the parent drug.When sufficient information is available for the activities (in vitro or in vivo) of a metabolite, a relative potency of 50% (or higher; vs. the parent drug) may be considered a guiding trigger to initiate quantification of the metabolite using an exploratory or qualified BA method.At this point, a validated method is not generally required.Potency for consideration herein includes but is not limited to (i) inhibitory constant, (ii) concentration producing 50% of maximal inhibition, and (iii) concentration producing 50% of its maximum effect. 18As soon as clinical exposure data are obtained for the metabolite of interest, a total contribution or pharmacological activity index (PAI) of the metabolite (vs. the parent drug) should be evaluated to decide on either continued quantification of such metabolite using a qualified or validated method or discontinuation.In this assessment, metabolite protein binding and available information on whether the metabolite reaches the target (target engagement) should be included in the consideration, as in vitro potency may not always translate to in vivo activity for some metabolites that are substrates for transporters (e.g., P-glycoprotein substrates for central nervous system targets) thus limiting their access to target. 4,19,20 this integrated decision-making framework, general recommendations from the authors are the following: • A PAI of 0.5 or higher is recommended as a guided threshold to continue metabolite quantification.A PAI value of 0.5 or higher is generally considered relevant in the context of a metabolite being able to contribute significantly to the pharmacology of the parent drug, 5 for which ideally evidence has been collected that the metabolite reaches the target organ.• A PAI < 0.25 is suggested as a guided threshold to discontinue metabolite quantification.A PAI value < 0.25 is generally not considered relevant in the context of a metabolite being able to contribute significantly to the pharmacology of the parent drug. 4,19,20 A PAI between 0.25 and 0.5 should instigate a project team discussion on whether the metabolite quantification should be continued (scientific justification) and, if so, in which studies.In this discussion, safety margin of the parent drug and other safety factors should be considered.
It should be recognized that each drug candidate is unique and project-specific considerations may override the above general recommendations.These PAI assessments can be used to guide whether bioanalysis is required for active metabolites.However, in some cases, quantification of inactive metabolites is necessary to help address project-specific challenges in relation to MIST, 1 DDI, 5 unexpected PK of parent drug or off-target activity (e.g., nonselective binding to a nontarget receptor or change in selectivity against receptor or enzyme subtypes, undesired activity for hERG product and other ion channels). 21For example, quantitative analysis of an inactive metabolite may be justified if the metabolic clearance occurs through a single cytochrome P450 (CYP450) enzyme modulated by inhibition or induction.This was reflected by the WG survey, where more than a third of respondents quantitatively analyze pharmacologically inactive metabolites to address project-specific concerns and/or regulatory inquiries on a case-by-case basis.

METABOLITE CHARA CTE RIZ ATION AND BIOANALYSIS IN MIST ASSESSMENT
The current health authority (HA) guidelines emphasize the possible safety concerns when exposure of a circulating human metabolite exceeds 10% of TDRE at SS, and stress that metabolite(s) circulating at disproportionately higher levels in humans than in Parent AUC × Fu parent × pharmacological activity of parent pharmacological activity of metabolite nonclinical species should be considered for safety testing and phase II conjugates other than acyl glucuronides are not generally required for further evaluation. 1,2The guidelines also state that, where exposure of a metabolite of concern exceeds 10% of TDRE, the contribution of the metabolite to the overall toxicity of the parent drug is considered established if one species used in general toxicity, one species used in carcinogenicity testing (or in vivo micronucleus study if required), and one species used in embryo-fetal development study at no observed adverse effect level (NOAEL) or higher (e.g., maximal tolerated dose) forms the same metabolite with systemic exposure at SS at least 50% of that observed in humans at the projected therapeutic dose level.However, if the metabolite comprises the majority of the TDRE at SS in humans, it is prudent for exposure to the metabolite in nonclinical species to exceed that in humans.For drugs with a daily administered dose < 10 mg, greater fractions of the TDRE might be more appropriate triggers for testing. 2,3hese requirements prompted the pharmaceutical industry to swiftly adopt exploratory metabolite profiling (or metabolite scouting) using samples collected from FIH trials, ideally the multiple ascending dose cohort, to identify potential major circulating human metabolites of the parent drug. 22In this effort, area under the drug concentration-time curve (AUC) pooling 23,24 in combination with mixed matrix matching followed by liquid chromatography-mass spectrometry (LC-MS; commonly high resolution MS) response comparison, LC-MS semiquantification (e.g., using biological standards derived from radiolabeled ( 14 C or 3 H) studies in nonclinical species or species matrix, e.g., hepatocytes) or LC-MS quantification by an exploratory or qualified method (if reference compound is available) are commonly used.Other techniques, including liquid chromatography-ultraviolet detection and nuclear magnetic resonance spectroscopy, are applied on a case-by-case basis.This is routinely followed by exposure comparison of the major circulating metabolites in humans, ideally, at the projected therapeutic dose level, vs. nonclinical species (rodent and non-rodent) dosed with parent drug at NOAEL or higher.][27][28][29][30][31][32][33][34][35][36][37] Among the WG member companies, criteria of animal/human metabolite exposure ratio generally align with that conveyed in the HA guideline 3 with almost two-thirds of respondents quoting 0.5 or higher as appropriate to establish coverage, whereas a smaller portion (~ 20%) of the respondents utilize a ratio of 1 or higher, depending on the methodology used in generating the ratio values.If the mixed matrix matching and LC-MS response comparison method is used, a ratio of animal/human metabolite exposure of 1 or higher is considered appropriate.If a qualified or validated LC-MS BA method is used, an animal/human exposure ratio of 0.5 or higher is considered adequate.It is worth noting that for a drug candidate targeting patients with advanced cancer, an evaluation of MIST-relevant metabolite exposure coverage is generally not warranted. 2,38Furthermore, for a drug candidate under development for serious or life-threatening diseases other than cancer, metabolite exposure coverage between nonclinical species and humans can be determined on a case-by-case basis, for which the sponsors should communicate with HA to determine an approach. 1he most comprehensive assessment of the biological fate of an administered drug candidate in humans can be obtained from an hADME study using radiolabeled (commonly 14 C labeled) parent drug, from which all drug-related metabolites that retain the label can be quantitatively monitored.0][41] An hADME study can be conducted in any clinical phases but is commonly performed before the end of phase IIb prior to initiating large-scale clinical trials, as reflected by the survey results (> 90%; Table 1), which are in line with previously published data 17 and the current FDA draft guidance on "Clinical Pharmacology Considerations for Human Radiolabeled Mass Balance Studies." 42Although the hADME study is generally performed late in drug development, the timing and study design have been continually discussed as reflected by a recent cross-industry white paper. 43A key advantage of using radiolabeled material for definitive metabolite quantification and calculation of TDRE (%) in relation to ICH M3(R2) 2 is that metabolites with little or no MS response can be detected, and the TDRE (%) of metabolites in human circulation can be easily obtained by measuring the radioactivity of individual metabolites vs. total radioactivity (TR).With this regard, care should be taken on potential metabolic loss of the radiolabel.
A single dose administration of radiolabeled drug for an hADME study provides a reasonable estimate of TDRE (e.g., AUC 0-inf ) and is an adequate basis for calculating if any circulating human metabolites are ≥ 10% of TDRE. 3 Regulatory guidance emphasizes that minor metabolites with exposure < 10% of TDRE at SS generally do not require further safety testing.On the other hand, nonclinical characterization of metabolites with an identified cause for concern (e.g., unique human metabolite) should be considered on a case-by-case basis. 2 A metabolite with measured exposure of < 10% TDRE in a definitive single dose hADME study with radiolabeled drug could still exhibit an exposure exceeding this threshold at SS, for example, due to a long terminal half-life. 44,45An early indication of the possible accumulation of a metabolite of interest in humans at SS

Number of responses
Before the end of phase IIb study 93. 1  27   During phase III study 24. 1  7   Issue driven a 24. 1  7   Abbreviations: HA, health authority; hADME, human absorption, distribution, metabolism, and excretion.a If efficacy of the parent drug is well-demonstrated and the project is highly accelerated due to its importance, the radiolabeled hADME (mass balance) study can be expedited significantly.On the other hand, when the pharmacology (efficacy) and safety of the parent drug is well-understood and the projected therapeutic dose is low and almost the entire administered dose can be accounted for as unchanged form of the parent drug molecule in urine (with minimum metabolism), the need for the radiolabeled hADME study can be obviated upon consultation with HA.
(vs. after a single dose) can be assessed using samples collected from conventional clinical studies (i.e., non-radiolabeled) following a single and multiple dose administration (ideally at the projected therapeutic dose level) via a fit-for-purpose method (e.g., AUC pooling followed by LC-MS response comparison or LC-MS quantification to evaluate possible accumulation vs. parent).This assessment can be performed immediately upon completion of the FIH multiple ascending dose trial and provides information for the design of the hADME study.Once single dose exposure data of the metabolites are available, modeling can also be used, if feasible, to simulate/estimate metabolite exposure at SS. 46 The accumulation factors obtained can facilitate project team decision making for additional assessments if needed.
If hADME data indicate accumulation of a metabolite may lead to SS exposure > 10% of TDRE, analysis of SS samples collected from clinical trial(s) (ideally at projected therapeutic dose level) and nonclinical studies (at NOAEL or higher) should follow to establish unequivocal exposure coverage of the metabolite, for which additional nonclinical studies (e.g., repeat dose PK study) might be necessary.If exposure coverage cannot be established for an accumulating metabolite in this scenario, this then becomes MIST relevant and requires safety characterization in nonclinical species. 47In this effort, metabolite properties (active/inactive/reactive, exposure of parent drug, etc.) and anticipated therapeutic index and safety profiles of the parent drug should be included in the consideration.
Last, according to the HA guidelines, conjugated metabolites (e.g., Oglucuronides, phenol-sulfates, and quaternary N +glucuronides) are generally pharmacologically inactive, more water soluble for elimination, and commonly not considered to be a safety risk requiring further evaluation even with exposure > 10% of TDRE and no coverage from nonclinical species.However, in cases where glucuronide and sulfate conjugates are unique to humans, circulate at relatively high levels, and/or possess long halflives, justification should be in place to explain why such human metabolites may not need further safety evaluation.[50][51]

METABOLITE BIOANALYSIS IN SUPPORT OF IN VITRO DDI STUDIES
As described previously, both the 2020 FDA "In Vitro Drug Interaction Studies -Cytochrome P450 Enzyme-and Transporter-Mediated Drug Interactions" draft guidance and the 2013 European Medicines Agency (EMA) "Guideline on the Investigation of Drug Interactions" 5,52 explicitly state the need to assess potential DDI risk for metabolites with significant plasma exposure or pharmacological activity in relation to the overall safety and efficacy of the parent drug.A fundamental principle of these guidelines is the appreciation of major circulating human metabolites as potential inhibitors or inducers (perpetrators) of a CYP450 enzyme(s) or transporter protein(s) in relation to the parent drug.Related thinking is also being applied to the ICH M12 draft guideline on "Drug Interaction Studies." 53][56][57][58] Interestingly, for BA support of in vitro drug metabolism and DDI studies, the 2020 FDA guidance states that "The sponsor should develop validated and reproducible analytical methods to measure levels of the parent drug and each metabolite." 5 As surveyed, > 90% of the WG member companies do not use methods that were fully validated, as detailed in the HA guidelines on BA method validation [59][60][61] (Table 2).It was agreed within the WG during harmonization discussion that the term "validated" should not be interpreted as a validation as referenced in the FDA, 59 EMA, 60 or ICH M10 61 guideline on BA method validation in support of regulated studies.Instead, the use of a well-characterized fit-for-purpose assay is encouraged as referred herein and in other publications. 62,63Typically, wider accuracy and precision acceptance criteria than in validated assays are used for the BA method in support of metabolite quantification in in vitro drug metabolism (including probe substrates) and DDI studies.5][66] In practice, it is recommended to predefine requirements and associated acceptance criteria through internal procedure documents in support of these studies.

TIERED APPROACH OF METABOLITE BIOANALYSIS IN SUPPORT OF IN VIVO STUDIES
The tiered approach of bioanalysis, specifically with respect to metabolite bioanalysis, has been widely adopted in the pharmaceutical industry. 4,29,64,65,67,68For the purposes of this discussion, the three tiers applied to metabolite bioanalysis include (i) exploratory, (ii) qualified, and (iii) validated methods.All recommendations below are related to LC-MS quantification.It is worth noting that, regardless of which tier of method is to be used, stability of metabolite(s) of interest in the intended study samples, particularly in the banked/stored samples, is essential to ensure reliable quantification.In this regard, available stability knowledge on the metabolite(s) of interest and common instability indicators should help guide effective stability measure in sample collection, processing, extraction, and LC-MS analysis. 4,29,31,68,69Exploratory BA method: An exploratory BA method is intended for quantification, semiquantification, or estimation of metabolite(s) of interest with limited method development.Metabolite reference standard is often not available and biological standards, if available, derived from radiolabeled (e.g., 14 C or 3 H) studies in nonclinical species may be used for semiquantification.The assay is not bound by regulatory requirements and acceptance criteria.If a reference standard is available, each assay batch should include a standard calibration curve and quality controls (QCs) at a minimum for absolute quantification.The resulting data are generally used for internal decision making but not for regulatory submission.• Qualified BA method: A qualified method provides absolute quantification using a reference standard of known purity.Often, a stable-labeled internal standard is not available.The resulting data are used for internal decision making and regulatory consideration, if included in the submission.A qualified method could include the following tests: (i) at least one accuracy and precision run using native matrix QCs across the assay range; (ii) selectivity in at least one lot of native matrix; (iii) carryover assessment, and (iv) stability in the intended matrix (e.g., stability in blood, benchtop stability, freeze/thaw stability, and/or short-term stability).Slightly expanded acceptance criteria (e.g., 25% or 30% at the lower limit of quantification) is acceptable.Method qualification can be standalone or included in study runs.Additional tests may be performed at the discretion of the bioanalytical scientist to address analyte-specific concerns.It is strongly recommended that a priori requirements and acceptance criteria are put in place for method qualification.In many cases, due to lack of certificate of analysis (CoA), a well-scrutinized BA method can only be claimed as qualified rather than validated from compliance point of view. 4,68 Validated BA method: A validated method is considered the most robust and suitable for absolute quantitation of metabolite(s) of interest using well-characterized reference standard (generally with CoA) and usually stable-labeled internal standard.The use of data generated using a validated method is generally intended for regulatory submission.[61] The three tiers of BA methods discussed herein cover all needs of metabolite bioanalysis in drug development, from non-Good Laboratory Practice (GLP) dose range-finding toxicity studies through IND-enabling GLP toxicity studies, phase I, phase II, and phase III clinical trials, along with long-term and other toxicity studies and clinical DDI studies.As discussed above, when deciding on initiating metabolite(s) quantification, known information on the metabolite(s), such as pharmacologic activity (potency), in relation to the parent drug and possible structural safety alert should be taken into consideration.
In nonclinical development, a general recommendation from the authors is to not quantify metabolites until there is a good understanding of their relevance to humans unless such measurement falls in the requirements of the ICH S3A. 14 Any "nice to have" (simply because we can) request for metabolite quantification should be scrutinized to ensure such effort has the added value to the respective project.If justified, a qualified BA method is sufficient to help understand the exposure of the metabolite(s) of interest in support of non-GLP toxicity or nonclinical ADME studies or the non-GLP portion of GLP toxicity studies as reflected by the survey where > 70% of respondents employ qualified BA method for non-GLP quantification of metabolite(s) (Table 3).In the same survey, > 80% of the respondents agreed that a fully validated BA method should be implemented for quantification of active metabolite(s) for compliance when supporting GLP toxicity studies (Table 4).
With project progression from nonclinical development to the FIH trial in clinical development, additional information on metabolites, particularly metabolite exposure coverage-related information, is obtained.If metabolite exposure coverage complies with ICH M3(R2), from a MIST perspective, there is no further need for quantitative analysis of metabolites in nonclinical species after this milestone unless justified as per the ICH S3A requirements.
Through clinical development progression, knowledge and understanding of the metabolite(s) of interest and associated practice of bioanalysis is most likely to evolve.From the survey, at phase I clinical development stage, more than 70% of the respondents validate the respective BA method for the metabolite(s) of interest whereas ~ 30% of the respondents use exploratory or qualified methods to help understand metabolite exposure or demonstrate metabolite exposure coverage.With the FIH and other assessment readouts (e.g., PAI, MIST, DDI, and PK-PD) being available, a metabolite that was previously considered trivial might become important.Consequently, a new BA method may need to be developed and validated/qualified for the metabolite (not quantitively analyzed before), or a previously used exploratory or qualified BA method may need

WHITE PAPER
to be revamped, followed by full validation for the quantification of the metabolite to support the progression of the project until regulatory submission.This was reflected by > 50% and 30%, respectively, of the respondents validating the respective BA method for metabolite concentration data generation in support of phase II and phase III trials (Table 5).In this effort, the project team should also decide whether the metabolite(s) of interest should be quantified for all subsequent clinical trials or selected clinical trials to ensure (i) appropriate use of BA resources and (ii) no unplanned studies later due to missing piece of important data in relation to metabolite(s) prior to regulatory submission.On the other hand, a metabolite that was previously considered important might become trivial.This should trigger an integrated decision by the project team to stop bioanalysis of such metabolite(s), regardless of whether a validated or qualified BA method is in place (because we can), as continued quantification of such metabolite(s) would have no added value to the project.
It is worth noting that, although a qualified BA method is considered sufficient for metabolite quantification in support of clinical trials, if metabolite data are important for regulatory decision making, it is recommended to validate the respective BA method if resources (time and cost) permit and other factors (e.g., CoA) are supportive.This was reflected by the survey where > 80% the respondents use a validated BA method for metabolite quantification in support of clinical trials in patient populations and DDI studies (Table 6).

BIOANALYSIS OF CHIRAL METABOLITES
It is well known that health authorities have laid out the requirements on what should be done in developing chiral drugs. 70,71n enantiomeric drug may undergo in vivo chiral inversion (e.g., R to S, or S to R) to form a chiral metabolite (i.e., the antipode), which may possess a different pharmacological and/or PK profile vs. parent drug, and this difference could be species-dependent.Like many other small molecule drug candidates, absolute quantification of a chiral metabolite in study samples is required if the metabolite contributes significantly (e.g., 50% or more) to the total effect of the parent drug, if the metabolite is expected to have different pharmacological and/or PK properties (vs. the parent drug), or if the metabolite poses concerns in relation to MIST or DDI. 7,72,73It is worth noting that, from a BA perspective, development and qualification/validation of a robust chiral LC-MS method for the quantification of chiral metabolites is much more resource-intensive and time-consuming than an achiral equivalent.Therefore, sponsors are discouraged from unnecessarily developing a chiral LC-MS BA method too early without exercising due diligence.
Related to the discussion above on the three-tiered BA methods, it is generally recommended to not quantify chiral metabolite(s) in nonclinical studies unless such quantification is needed to fulfill the requirements of the ICH S3A. 14 It is also generally recommended to not start investigation into the likelihood of chiral inversion unless the parent molecule contains functional groups that are known to likely undergo chiral inversion until the respective project transitions to the clinical development stage.Then, a chiral LC-MS BA method may be developed using reference standards of the chiral metabolite(s) for exploratory analysis of selected FIH trial samples (ideally at SS at the projected therapeutic dose level or higher) to detect whether an inversion takes place.Depending on the outcome of this exploratory analysis, animal samples (ideally collected at SS at NOAEL or higher) may be examined to assess the exposure coverage of the chiral metabolite(s) of concern.[76]

CASE STUDIES
In the following section, selected case studies are presented to highlight the implementation of fit-for-purpose and stageappropriate LC-MS BA methods to address challenges associated with pharmacologically active metabolites, MIST-related metabolites (including metabolites with potential accumulation vs. parent at SS), DDI aspects of a metabolite(s), and/or chiral metabolites to support the respective development projects moving forward.
Case 1 (validated method for active metabolite in support of DDI studies) Entrectinib (Rozlytrek) is a potent and selective inhibitor of pan-TRK, ROS1, and ALK receptor tyrosine kinases for the treatment  a Metabolite quantification may be exploratory by nature as part of troubleshooting/investigation prior to decision making for the follow-up studies, for which the intended method is generally validated.

WHITE PAPER
of ROS1-positive non-small cell lung cancer and NTRK fusionpositive solid tumors.The compound is predominantly cleared by CYP3A4-mediated metabolism to metabolite M5 (Figure 2), which possesses approximately the same pharmacological effect as the parent drug.A fully validated LC-MS BA method was used for the simultaneous quantification of entrectinib and metabolite M5 to support clinical DDI studies with a CYP3A4 inhibitor, itraconazole, and CYP3A inducer, rifampicin.The results showed that the PKs of both entrectinib and M5 were altered due to DDI and the effects on entrectinib and M5 were quantitatively different.The quantified exposures of M5 were critical to the physiologically-based PK modeling in defining dosing strategies for entrectinib when co-administered with CYP3A4 inhibitors and/or inducers. 77,78se 2 (validated method for metabolite in support of DDI studies) Doravirine (Pifeltro) is a non-nucleoside reverse transcriptase inhibitor for the treatment of HIV-I.The compound is predominantly cleared by CYP3A4 to an oxidative metabolite M9 (Figure 3) which was ~ 13% of the total plasma radioactivity as demonstrated in the hADME study. 79When co-administered with a CYP3A4 inducer (e.g., rifampin), plasma doravirine concentrations decreased as anticipated, leading to contraindication with strong inducers and a dose adjustment proposal with moderate inducers.However, the effect due to CYP3A4 inducer on M9 exposure was unknown.A physiologically-based PK model suggested that whereas doravirine clearance increased up to 4.4-fold, M9 exposure increased only 1.2-fold.Therefore, a change in doravirine dosing from 100 mg once daily to twice daily would result in a 2.4-fold increase in M9 exposure.This was confirmed clinically where the exposure of M9 was explicitly quantified using a fully validated LC-MS BA method. 80se 3 (validated method for active metabolite with potential of accumulation) Ozanimod (Zeposia) is a sphingosine 1-phosphate (S1P) receptor modulator for the treatment of relapsing forms of multiple sclerosis, including clinically isolated syndrome, relapsing-remitting disease, and active secondary progressive disease, in adults.
In the hADME study using 14 C labeled ozanimod, one active metabolite, CC112273 (Figure 4), and one inactive metabolite, RP101124 were characterized at > 10% of the TR.Interestingly, CC112273 forms a downstream metabolite CC1084037 (Figure 4), which possesses a similar activity profile as ozanimod and CC112273.Although CC1084037 was only present at ~ 5% of the TR in the single dose hADME study, its parent CC112273 was the predominant active metabolite with a long half-life and accumulated upon multiple dosing.Further analysis of SS samples using a fully validated LC-MS BA method demonstrated that CC1084037 exceeded the 10% threshold and was present at 15% of the TDRE upon multiple administration of ozanimod. 44Both Figure 3 Structures of Doravirine (Pifeltro) and its major metabolite M9.

Entrectinib M5
active metabolites CC112273 and warranted further studies to characterize the exposure coverage in toxicity and clinical studies, for which fully validated LC-MS BA methods were used (details not shown).
Case 4 (stage-appropriate exploratory, qualified, and validated methods for MIST-relevant metabolite) KAE609 (Cipargamin) is an experimental synthetic antimalarial drug candidate in the spiroindolone class.In the hADME study with oral administration (300 mg) of 14 C-KAE609, the unchanged parent drug was confirmed as the major radioactive component (~ 76% of the TR in plasma).Metabolite M23 (Figure 5) was identified as the major circulating metabolite in humans (~ 12% of the TR in human plasma) based on both LCradioactivity detection and implementation of a qualified LC-MS BA method.In studies with 14 C-KAE609, M23 was detected via LC-radioactivity detection in feces but not in plasma of rats (10 mg/kg/day) or dogs (3 mg/kg/day), the 2 nonclinical species tested in the IND-enabling toxicity studies.M23 was present at only trace levels in rabbit (50 mg/kg/day) plasma using an exploratory LC-MS BA method.Taking the above into consideration, M23 was defined as human disproportionate metabolite and MIST relevant. 47,81Accordingly, toxicity assessment studies with M23 were conducted in rats (a 2-week GLP study with 4-week recovery) and dogs (a 2-week GLP study with 4-week recovery) along with other safety evaluation.In support of the metabolitefocused GLP toxicity studies in rats and dogs, the reference compound of M23 was synthesized and certified and the respective LC-MS BA method was fully validated according to the effective HA guidelines and internal standard operating procedures (details not shown).
Case 5 (exploratory and qualified methods to facilitate decision making on possible MIST-relevant metabolite) SAR442168 (Tolebrutinib) is being developed for the treatment of multiple sclerosis.The compound is extensively metabolized in vivo.In support of clinical studies, a fully validated LC-MS BA method was utilized for the quantification of the parent drug and an active metabolite (M2; Figure 6) in human plasma to assess sources of PK variability, and relationship between PK and efficacy/safety.From the hADME study with 14 C labeled materials, the unchanged parent molecule only represented 1.3% of the TR in plasma.Among the 19 metabolites detected in human plasma, 7 metabolites (M2, M5/M5a, M8, M10, M18, M19, and M32) had exposure greater than that of the parent drug (> 1.3% of TDRE) with exposure of metabolite M8 greater than 10% of TDRE.Accordingly, an in vitro DDI risk assessment was performed for the five characterized and synthesized metabolites (M2, M5/M5a, M8, M10, and M18; Figure 6) out of the above seven metabolites.An exploratory LC-MS BA method was developed and implemented in support of the in vitro DDI study.Because exposure of metabolite M8 exceeded 10% TDRE, selected samples from several nonclinical studies in dogs, rats, mice, and rabbits were analyzed to better understand the potential MIST relevance of this metabolite.The studies included a chronic toxicity study in dogs, a dedicated 7-day repeated toxicity study in rats, and toxicity studies in mice and rabbits.In these investigations, first the mixed matrix matching combination with LC-MS analysis was used to compare exposure coverage between nonclinical species (rats and dogs) and humans.Then, qualified LC-MS BA methods (mouse and rabbit) were used for metabolite quantification in support of decision making (details not shown).
Case 6 (validated method to confirm chiral inversion of parent drug for decision making) MK-0767 is a thiazolidinedione-containing compound that was studied for the treatment of type 2 diabetes.The drug is racemic with both enantiomers having the same in vitro potency.Rapid interconversion between (+)-(R) and (−)-(S) enantiomers (Figure 7) and stable equilibrium R/S concentration ratios were observed both in vitro and in vivo (in nonclinical species), suggesting that regardless of whether the racemate or the individual purified enantiomers are administered clinically, similar outcomes in terms of PK, safety, and efficacy could reasonably be expected.However, an additional study in healthy volunteers was required by the HA to ensure the chiral inversion in humans in vivo was rapid relative to other elements of enantiomer disposition.Therefore, a chiral LC-MS BA method was developed and fully validated in support of quantitative analysis of samples (intensive PK samples) collected from in vitro studies and in vivo human and nonclinical studies.The obtained results indicated that interconversion between (+)-(R) and (−)-(S) enantiomers of MK-0767 is rapid relative to elimination and in vivo chiral exposure ratio is constant (R/S ~ 2-2.5) regardless of whether single enantiomers or racemate were administered.The results concluded that similar safety and efficacy outcomes would be anticipated when the racemate or the individual purified enantiomers are administered to humans.Consequently, additional LC-MS chiral analysis in clinical and nonclinical studies was not required. 72,73se 7 (exploratory method to investigate the possible chiral inversion of the parent drug for decision making) Saxagliptin (Onglyza) is a marketed drug for the treatment of type 2 diabetes.An LC-MS BA method was developed and fully validated to simultaneously determine saxagliptin and its major active metabolite, 5-hydroxy saxagliptin (Figure 8) in human plasma to support clinical studies.Saxagliptin has four chiral centers (S,S,S,S configuration) and, therefore, 8 possible stereoisomers, including saxagliptin itself.The formation of any of these stereoisomers is not anticipated to occur through metabolic chiral inversion (i.e., oxidation of a secondary alcohol, conjugation of a carboxylic acid with acetyl coenzyme A).However, chiral inversion may occur through chemical reactions, either in vivo or ex vivo (during sample storage or processing).To investigate whether the diastereomers were present in humans, the authentic standards of the diastereomers were tested using the same non-chiral LC-MS method used for metabolic profiling of the hADME study with [ 14 C]saxagliptin.Via this exploratory approach, each of the available diastereomers was baseline resolved from each other and from saxagliptin.The experiment concluded that these components were not present in the tested human plasma.Nonetheless, the formation of the epimers A (S,R,S,S) and B (S,S,S,R) (Figure 8) of saxagliptin is still possible as they each are the product of a single inversion of saxagliptin at one of the two chiral centers, particularly for epimer B (S,S,S,R), which is likely to be formed under plasma sample preparation conditions.In a follow-up investigation, a mixture of saxagliptin and epimer B was examined using the same condition as the validated assay for saxagliptin alone.Epimer B was baseline separated from saxagliptin.Study samples collected from several clinical studies were further examined.No signal corresponding to epimer B was detected.Furthermore, when saxagliptin plasma QC samples after going through benchtop and frozen storage were examined for the potential presence

KAE609 (Cipargamin)
of diastereomers, no signal was detected either.The combined evidence above led to conclusion that saxagliptin does not undergo any significant in vivo or ex vivo chiral inversion.Therefore, no additional chiral analysis was warranted throughout the remaining part of the development program. 75,76

SUMMARY
In the pharmaceutical industry, there is currently little consensus on when bioanalysis of a metabolite should start, with what analytical rigor and in which studies.However, there has been a tendency of self-inflicted bar raising in metabolite quantification as often evidenced by quantification of metabolite(s) too early, in too many studies, and/or using a fully validated method unnecessarily when an exploratory or qualified method would suffice.Many new molecular entities are terminated prior to phase I or in phase I due to a lack of safety or in phase II to phase III due to failing efficacy. 82Given this high attrition, it is therefore advised that projects are managed in a cost-effective manner during the drug development process.Careful consideration is

WHITE PAPER
required before undertaking metabolite quantification within nonclinical safety assessment or FIH trial of any given project unless the metabolite(s) of interest falls the ICH S3A requirements or is deemed important from a safety or efficacy aspect.Otherwise, metabolite quantification can wait until after the FIH trial is completed; then, a comprehensive, albeit preliminary, human metabolite profiling is carried out, and metabolite exposure coverage is compared between nonclinical species (ideally at NOAEL level) and humans (ideally at SS at therapeutic dose level) to confirm the metabolite(s) of interest.The latter is often followed by necessary metabolite synthesis and characterization (e.g., activity and in vitro safety assessment) before any further evaluation.
For an active metabolite, a relative potency (pharmacological activity) of 50% or higher (vs.parent) may elicit absolute quantification using a qualified BA method.As soon as metabolite exposure becomes available, the PAI of the metabolite should be assessed to inform on continuation of metabolite quantification if PAI is 0.5 or higher or discontinuation of the quantification if PAI is < 0.25 unless the metabolite is relevant to DDI or safety (MIST or off-target activity).PAI between 0.25 and 0.5 should initiate a project team discussion regarding whether the quantification should be continued.This guided approach not only helps manage limited research capacity but also ensures that metabolites contributing significantly (e.g., 50% or higher) to the total effect of the parent drug or possessing safety or DDI liability are monitored as early as possible to prevent delay in drug development.
Every effort should be made after the FIH trial to assess whether any metabolites not adequately covered by the already tested nonclinical species are truly major circulating metabolites in humans (i.e., 10% TDRM or higher).With this regard, an integrated decision should be made on the timing of a definitive hADME study.When conducting the hADME study, diligence should be exercised to examine the terminal half-life values (both parent and metabolites) and total radioactivity (e.g., 14 C) to help assess the possible accumulation of the metabolite of interest after multiple repeated dose (SS).This exercise is particularly important for the project team to tailor a practical strategy in dealing with the potential challenge of a metabolite with exposure < 10% of TDRE after a single dose but showing a trend of accumulation at SS.
Depending on the stage of the drug development and the intent of the study, metabolite bioanalysis should be customized to address unique challenges that may arise, beginning early in nonclinical development and proceeding through FIH trial, proofof-concept study, pivotal phase III trial, and beyond.The current perspective provides recommendations on guided trigger, timing, and analytical rigor for metabolite quantification that is fit-forpurpose (e.g., internal decision making and regulatory submission) for the respective study.The case studies included herein provide It is worth noting that each drug candidate is unique from pharmacology, safety, MIST, DDI aspects and requires tailored and cost-effective bioanalytical strategies to meet the project needs.The effective HA guidance documents and the publications from leading scientists in the field, together with the continued discussion within the industry, should lead to increasingly consistent thinking and practical harmonization of metabolite characterization and bioanalysis in support of drug development.

Table 1
Timing of conducting radiolabeled hADME study in drug development (Note: One company could have multiple answers)

Table 2 A
survey summary of bioanalytical method used for metabolite quantification in support of in vitro drug metabolism or drug-drug interaction study

Table 3 A
survey summary of bioanalytical method used for metabolite quantification in support of non-GLP toxicity studies

Table 4 A
survey summary of bioanalytical method used for metabolite quantification in support of GLP toxicity studies Abbreviation: GLP, Good Laboratory Practice.a Qualified method can be employed to generate exposure data for active metabolites(s) in support of a GLP toxicity study and the nature of the method is clearly listed as an exception from the full compliance of the study.

Table 5 A
survey summary of the clinical development stage where the BA methods were validated for quantification of metabolite(s) (Note: one company could have multiple answers) Abbreviation: BA, bioanalytical.a Post-approval commitment and other studies (e.g., investigator-initiated studies) not typically classified as phase I, II, or III studies.