Ask Me Anything About In Vitro ADME, Drug-Drug Interaction Approaches and Strategies

Last month, Dr. Joanna Barbara joined AAPS for an online Q&A in their ongoing “Ask the Experts” series. As regulatory agencies continue to expect ADME data earlier and earlier in the process of developing new therapeutics, we logged in to give the community access to one of our experts and obtain insights into ADME regulatory expectations, investigatory objectives, and practical concerns for such studies in order to successfully achieve investigatory goals and regulatory submission. In vitro metabolism and drug-drug interaction (DDI) investigations are critical for framing downstream decision making, however, few companies have experts with a comprehensive understanding of this science in-house.

Q: How early should the DDI tests begin and can you please elaborate on the best strategy/approach to evaluate DDIs during the course of drug development from early discovery stage to clinical phase?

A: There is no one-size fits all answer to this great question. It depends on each program’s regulatory strategy and risk-tolerance level. For a middle of the road approach, I would suggest starting with some basic in vitro metabolism (clearance and met ID), and some CYP induction and inhibition, with a screen for inhibition of Pgp and BCRP transporters fairly early on, once you have a strong lead compound or two. The more definitive enzyme mapping work could come next to help identify victim potential. Then finally more definitive transporter inhibition and substrate. Answering the basic metabolism and CYP questions first can build some confidence before spending resources on the more expensive portions of DDI work. That said, for a very conservative company with minimal risk tolerance, it is sometimes preferable to do a full DDI package early on once the lead is identified. Conversely, a company with high risk-tolerance and a cost-focused approach may opt for CYP induction and inhibition screening early on and proceed with other work if no red flags are observed.

Q: As a follow up to the DDI questions, if you don’t identify any DDI liabilities during the lead op phase (or say you do), would you suggest testing in preclinical animal efficacy models the two drugs in combination? (for example: SOC+ drug)? What is the best approach to triage from invitro work to, animal studies and then to clinical. Are there good references that capture these strategies?

A: This is a great question. There are 2 components to this answer. Animal models are usually not very useful for evaluating DDIs. That is because most are enzyme or transporter based and human enzymes/transporters are very different from animal ones. For example, the human CYP3A equivalent in rat is 2C9 – but they don’t behave identically, so in general I would not recommend animal models to study DDI.

The second component is the combination aspect. The intent of getting data on the enzymes is so that you can avoid performing a lot of combination studies. So let’s say your drug is a CYP3A substrate but you know the patients will take an antibiotic that inhibits 3A. You would model that in silico using known data for the antibiotic and then the kinetic parameters measured for just your drug in your in vitro studies. That way you can avoid all the combination studies, which are actually very difficult to design and also difficult to deconvolute the results for.

Q: When should we think about CYP2C induction?

A: CYP2C induction is tricky. It is hard to measure, as the mRNA endpoint doesn’t work well. It becomes important if CYP3A4 induction is positive. There is cross-talk with pathways, so the FDA requires 2C data when there is observed or known CYP3A4 induction. If a team knows or suspects that their compound will likely induce CYP3A, then it is prudent, and most cost-effective overall, to include 2C induction measurement using the activity endpoint along with the common 1A2, 2B6 and 3A4 work. Then there are compound- or indication-specific reasons for consideration of 2Cs, but the most generally applicable situation is the 3A4-based reasoning. Read more about this topic.

Q: How do preclinical PK assays work to describe the ADME of a biotherapeutic?

A: They can help with clearance and dose prediction to some extent, but due to the traditionally long half-lives it can be more useful to do some in vitro ADME work with longer term human models, depending on what aspect of ADME is of interest.

Q: Which are the best ADME assay predictor tools available online?

A: I am a big believer in just generating some data rather than relying on prediction. Of course that isn’t always possible, but as a result I cannot speak to which online tools are the best.

Q: Can anybody please suggest good papers on in vitro ADME study?

A: We rely pretty heavily on chapter 6 from Casarett and Doull’s Toxicology: The Basic Science of Poisons written by my former and current colleagues. It is named Biotransformation of Xenobiotics and explains a lot about things to think about. However, I strongly recommend the xenotech website for some really good initial content.

Q: What are the 4 phases of pharmacokinetics?

A: The 4 phases are characterized by the ADME term. This acronym encompasses Absorption – drug getting into bloodstream; then Distribution to tissues, then Metabolism when the drug is converted to more water soluble (theroretically) chemicals by drug metabolizing enzymes in order to promote excretion, and finally Elimination – the exit of the drug from the body. All are important and can contribute to safety and efficacy of the drug.

Q: What are common factors that affect absorption of a drug?

A: The big players are the lipid solubility, formulation, and route of administration. Most people start by worrying about solubility, which is logical as there isn’t much you can do with a drug before solubilizing it. But the formulation part is really important, too. Sometimes in animal studies, people get vastly different PK results with different formulations.

Q: What is the use of a reaction phenotyping study?

A: Reaction phenotyping means enzyme mapping. Basically, you try to figure out which enzyme family, or specific enzyme if it is a well-characterized family like CYPs, is involved in the clearance of the drug. The use is that then you can predict what other drugs might cause an enzyme-based DDI. For example, maybe you determine that the main clearance route for your drug is 2C19 during your reaction phenotyping study. Then you can look at the other medications you would expect your patient population to be on to see whether they inhibit 2C19 and evaluate the likelihood of a serious impact on clearance of your drug.

Q: How do I pick species for tox studies?

A: People usually want their tox species to be rat and dog because they are cost-effective, well-known and easily accessible. However, the best way to figure out which rodent and non-rodent species will actually be predictive of human behavior is by combining in vitro metabolite identification data from all potential species and from human with PK data from the potential species. The metabolite ID data tells you whether the animals can produce all the human metabolites and whether they produce them at similar levels. The PK data tells you whether the exposures will be relevant and helps you with dose selection

In my experience, the answer is not often clear-cut for dog and rat, so my advice would be to start out comparing in vitro data from mouse, rat, dog, monkey and human. I almost always recommend all 5 to be run at once because it is much better to have one easily comparable dataset.