What Is The Role of GLPs in ADME?

Good Laboratory Practice (GLP) regulations have long been employed in non-clinical safety evaluations to help ensure a measure of confidence in the accuracy and quality of these data, which are used to determine potential risks to humans given a new molecular entity (NME) in clinical development.  Strict adherence to GLPs has not been historically expected during the conduct of non-clinical Absorption-Distribution-Metabolism-Excretion (ADME) studies, in vivo or in vitro, except for bioanalytical methods supporting toxicokinetic (TK) evaluation in GLP animal toxicity studies.  However, I have personally observed a recent trend in this realm, which has witnessed some sponsors performing non-clinical ADME studies according to strict GLP regulations, expecting that this extra effort may be needed in order to ensure acceptability in a regulatory filing.  This is in spite the fact that such studies do not directly provide safety data, which is the type of work for which GLPs were specifically developed.  Regulatory agency guidance on this issue has been sporadic or lacking.

I recently presented a talk with the title “What is the role of GLPs in ADME?” at the ADME and Predictive Toxicology Congress in Barcelona, Spain (12 April 2013), in which I attempted to address a number of the issues surrounding the question of whether GLPs have a larger place in non-clinical ADME research.  The outline of the talk is provided in the accompanying slide deck.  I will let the slides speak for themselves for the most part, and will embellish a few points here.

By way of background, GLP regulations were promulgated by the Food and Drug Administration (FDA) in the late 1970’s in response to suspicions about the quality of a number of submissions from a major Pharma company, which revealed evidence of inconsistencies in the data as well as unacceptable laboratory practices.  Through the course of intensive investigation and “for cause” audits, the agency confirmed, to their profound chagrin, that this was a systemic problem within the industry.  Even more troubling, their investigations unearthed egregious breaches in ethical behavior at a few key contract research laboratories.

With congressional support and funding, FDA established the Bioresearch Monitoring Program in 1976 which, pursuant to extensive investigation and evaluation of numerous alternatives, elected to institute GLPs to provide strict guidelines that would help ensure adequate quality of the data underpinning future regulatory submissions.   The first GLPs were put in force in 1978 (21 CFR, Part 58).  Other agencies – notably the Environmental Protection Agency (EPA) and the Organization for Economic Cooperation and Development (OECD) –followed suit.  In all cases, the guidelines remain “living” documents, undergoing periodic updates to keep abreast of advances in science and technology.

The breadth and depth of topics that these regulations encompass reflect the devolution of a once comfortable relationship (Pharma and Regulatory Agencies in this case) into one sadly bereft of mutual trust.  The areas covered by GLPs are listed in the slide deck (Slides 10-15) and are nicely expounded upon by Peterson in a chapter titled “FDA/GLP Regulations” in “Good Laboratory Practice Regulations” (4th ed. Edited by S. Weinberg. Informa Healthcare USA, 2007. pp 25-110).

The focal area of GLP regulations is safety.  As such, the inclusion of TK support (most importantly, bioanalytical methods) under the GLP umbrella is clear and logical, with the regulations insinuating themselves into ADME development at well defined junctures (Slide 17).  Very clear, so end of story, right?  Not necessarily.  If we dig into it a bit deeper, we see shades of ambiguity.

It is uniformly acknowledged that the focus of GLPs is on safety.  But do ADME studies support safety?  Consider that Peterson, in the section of his book chapter summarizing the scope of GLPs, states, “The following are examples of studies to which the GLPs can apply: …(xxi) target animal absorption, distribution, metabolism, and excretion (ADME)…”  (FDA/GLP Regulations.  In “Good Laboratory Practice Regulations,” 4th ed. Edited by S. Weinberg. Informa Healthcare USA, 2007, pp 25-110).  Perhaps there is a nuanced interpretation of the word “target” as a qualifier that renders this pronouncement somehow logical.  And he is not alone.  In my own experience, I was recently taken aback when a former FDA Pharm/Tox reviewer who was evaluating a proposed IND package that I had designed questioned why the ADME studies weren’t being done under GLP (all of them!).  In a separate instance, another former FDA Pharm/Tox reviewer was of the firm belief that the broad CYP inhibition profile displayed by a new chemical entity in non-clinical development was a safety issue that could prove to be a challenge in IND review.  Although there is certainly no overwhelming trend in this regard, the existence of these examples fosters a mild undercurrent of uncertainty.  Realistically, it is, in fact, fairly easy to connect the dots from an in vitro ADME study to safety assessment.  Say, for example, we find in an in vitro protein binding study that an NME is highly bound in plasma in a non-linear fashion with respect to concentration, and that the effect of concentration on binding is species dependent (Slide 20).  Let’s also consider that the target organ for toxicity is the CNS, which is highly dependent on free concentrations in the plasma.  To adequately understand the safety implications involved in extrapolating animal exposure data to humans in this case, it is crucial to know the protein binding behavior of this molecule and how it relates to concentration and species.  This makes protein binding data a critical safety indicator for this compound, yet protein binding studies would not typically be evaluated under GLP.

With that as a backdrop, where do things stand in the “real world?”   A cursory web search revealed a mix of positions that left me vaguely unsatisfied.  So, to help get a better sense of the landscape, I conducted a survey in which I solicited opinions from a number of large contract research organizations (CROs).  All were well-known labs that I selected because I have had contact with them in recent years, and because they all offered the full range of non-clinical ADME services.  Nine CROs were solicited and eight responded (as is obvious, this was not intended to be a statistically sound assessment).  A number of questions were asked; a few key queries are listed in Slide 25.

The extent and nature of GLP ADME study offerings are captured in a series of pie charts (Slides 26-34).  The results are quite interesting.  Of note, in vivo animal studies distinguished themselves as the models that were most uniformly available as GLP compliant.  In a sense, this is not surprising insofar as they most closely mirror the pivotal toxicity studies (i.e. in vivo) for which GLPs were originally developed.   Protein binding (offered GLP by 62% of the labs) and RBC distribution studies (50%) would fairly easily lend themselves to GLP conduct, having aspects in common with bioanalytical methods, which as noted above are the crux of GLP TK support.  GLP compliance is offered in high proportion for other studies that have become more codified by FDA Guidances, such as CYP inhibition in microsomes (62%) and hepatocytes (50%), CYP induction in hepatocytes (62%), and transporter assessments (soon to be 63%).  On the other hand, few labs offer GLP-compliant metabolic stability (37%), metabolite profiling (12%), and reaction phenotyping (37%), as these studies require more flexibility in approach and are more challenging to adapt to standard operating procedures (SOPs).

The responses to the open ended questions were suffused with a strong bias that GLP really has no place in ADME whatsoever, and that this is an idea that should be squashed immediately, lest it have a chance to take root.  In fact there was in these comments an undercurrent of abject fear that even mentioning the two acronyms in the same sentence could be accompanied by the specter of elevated FDA scrutiny… “hmmm, GLPs in ADME… now that is worth thinking about…”  In that respect, it is perhaps comforting that the overall extent of GLP requests is not high and does not appear to be increasing (Slide 38).

By way of full disclosure, my own position is directly in line with the prevailing CRO mindset.  In my consulting work I have encountered clients who request (or insist!) that their non-clinical ADME studies be conducted according to GLPs.  I have consistently counseled that this is unnecessary.  And I will continue to do so for the foreseeable future. The most reasonable approach, in my opinion and the opinion of most of the CRO respondents, is to conduct ADME studies in the “Spirit of GLP,” which can be characterized as “GLP in quality, but without the onerous paperwork and strict QA oversight and audits.”  Certainly one must do whatever one can to ensure quality of the work, but the additional requirements imposed by GLP regulations are, in this context, daunting, and add little value.  This is captured conceptually in the graphic I present in Slide 46.

So, whither GLPs in ADME?  It is interesting to note that, of the individuals to whom I posed this idea off the cuff, several of those who felt that ADME will ultimately fall under the aegis of GLP were former FDA reviewers.  Considering that GLPs were spawned by FDA, this is a worrisome trend, and one to which we should pay continued attention.  The former reviewers and others provided some salient examples of studies that were not originally covered by GLPs, but gradually succumbed to the regulations, such as Safety Pharmacology and TK analysis (i.e. the mathematical analysis of concentration-time data).

In the end, there is no reason to suspect that ADME research performed in the “Spirit of GLP” or “consistent with the GLPs” is not of sufficient quality to be submission ready.  We can ill afford to constrain the drug discovery and development process any further.  The added burden on time and resources that would be incurred by regulating ADME to the level of full GLP compliance could be significant.  Unfortunately, the impact would be felt most acutely in smaller Pharma and Biotech, which have been the source of significant pharmaceutical innovation over the past decade.  Maintaining our current practices should continue to yield robust data packages that will withstand the highest level of regulatory scrutiny.  That said, it is impossible to predict which direction the winds of change will blow (hopefully at our backs).  So stay tuned.

Do We Need to Overhaul Human ADME Studies?

The past two years have seen a vigorous discussion in the print and online media and at various scientific meetings on the general topic of what is the proper methodology, objective, timing and sequencing of the time-honored “ADME” study in animals and humans (Penner et al., 2011).  It is not hard to see why this is happening.

The basic human ADME experiment has changed little in decades, yet many clinically relevant questions could be addressed these days if there was sufficient impetus from scientists and expectation from regulators for improvements in the way ADME studies are designed and carried out.  Regulatory guidance’s requiring early identification of major human metabolites before the traditional conduct time of radiolabeled human ADME studies have stimulated analytical chemists to invent advances in analytical technology, especially high-resolution mass spectrometric instrumentation and data-mining algorithms (Zhu et al., 2011).  This new technology can be used creatively in many additional ways, not merely to satisfy MIST requirements. And, of course, the constant pressure to make drug development more rapid and cost-effective rightfully encourages sponsors to question everything.

An example of this debate from the recent literature illustrates the interest in this topic among industry experts.  In early 2012, Obach et al. raised the question whether radiolabeled ADME studies in animals provide value in clinical drug development justifying their expense and resource commitment.  Their main point was that metabolism information derived from preclinical species such as rats was not reliably predictive of human metabolism.  Thus, it was better to leave investigation of animal metabolism until after the human metabolite profile had been established, at which point one could focus on only the human-relevant aspects of animal metabolism, such as establishing coverage of all major human metabolites in preclinical safety studies.  These authors also suggested that the use of “lightly labeled” radioactive drug combined with detection by accelerator mass spectrometry (AMS) as a way to be able to conduct the human ADME study very early in clinical development, as well as to be able to achieve true pharmacokinetic steady state by multiple administration of extremely low doses of 14C (< 50 nCi).

However, in a subsequent response to that article, another group (White et al.) indicated several benefits that, in their opinion, justified continuation of preclinical radiolabeled studies to learn as much as possible about the behavior of drug candidates in a living organism before administration of radiolabeled drug to humans.  Suggested benefits included discovery of unexpected routes of metabolism, elucidating the major clearance mechanisms, and understanding the relevance of certain metabolism-based animal toxicities to humans.  Most recently, Obach et al. defended their original proposal while acknowledging that the issue deserves a wide discussion (2013).

In the inquisitive and innovative spirit of Obach, Nedderman and Smith, I propose some additional questions that ADME scientists ought to discuss.

  1. Does the newest MS methodology offer the opportunity to obtain data from human ADME experiments that is more useful for clinical development than is currently obtained?
  2. Can we extend the gathering of extensive ADME data from a handful of young, healthy male volunteers to large numbers of actual patients and investigate the effects of the variables of age, gender, ethnicity and health on metabolic profile?
  3. Can present QWBA studies in animals be enhanced to provide chemically specific assessment of tissue distribution of parent drug and metabolites in addition to the core objective of radiodosimetry?
  4. Does a human 14C-ADME study as presently conducted by most sponsors provide the information that regulatory agencies ought to have to properly assess the clinical efficacy and safety profile of a new drug candidate?
  5. Can “lightly labeled” radioactive drug with AMS detection provide ADME data of the same quality as traditional 14C-ADME studies?
  6. Can contemporary technology eliminate the need to administer radiolabeled drugs to human beings?

It is easy to think of additional questions along these lines.  The purpose of my short blog here is not to answer all these questions, but to stimulate the scientific community in the academic, industrial and regulatory sectors to share their thoughts and ideas with all of the rest of us.


Obach RS, Nedderman AN, and Smith DA (2012) Radiolabelled mass-balance excretion and metabolism studies in laboratory animals: are they still necessary? Xenobiotica 42, 46–56.

Obach RS, Nedderman AN, and Smith DA (2013) A response to “radiolabelled mass-balance excretion and metabolism studies in laboratory animals: a commentary on why they are still necessary”. Xenobiotica 43, 226–227.

Penner N, Xu L, and Prakash C. (2012) Radiolabeled absorption, distribution, metabolism and excretion studies in drug development: why, when and how? Chem Res Toxicol 25, 513-531.

White RE, Evans DC, Hop CECA, Moore DJ, Prakash C, Surapaneni S, and Tse FLS (2013) Radiolabeled mass-balance excretion and metabolism studies in laboratory animals: a commentary on why they are still necessary. Xenobiotica 43, 219–225.

Zhu M, Zhang H, and Humphreys WG (2011) Drug Metabolite Profiling and Identification by High-resolution Mass Spectrometry. J Biol Chem 286, 25419-25425.