Mimicking the complexity of the in vivo environment with in vitro systems may be challenging, particularly for novel TP modalities with several biodistribution mechanisms | • A combination of in vitro systems, e.g., co-cultures and 3D systems, could help to understand and quantify the ‘contribution of components’ of overall biodistribution for TPs with multiple biodistribution mechanisms |
There are currently few established in vitro– in vivo correlations or in silico predictive tools for PK and/or biodistribution of TPs in the early discovery setting | • Cross-industry and academia collaborations and publications of large datasets required, accounting for different experimental settings, or providing recommended standard experimental conditions • Machine learning techniques will be key for interpreting large datasets to generate reliable QSPKR and IVIVC, with clarity on the appropriateness of extrapolating or interpolating within or across TP types, and regular retrospective analyses to ensure validity |
Using systemic PK as the input for preclinical PK/PD relationships may not be an appropriate surrogate for PK at the target site for some TPs | • Use both experimental and modeling approaches to understand major factors driving tissue distribution and PK/PD relationships prior to deciding whether systemic PK can be used in the absence of biodistribution data for PK/PD-related decision-making • Measure immunogenicity where possible, as anti-drug antibodies may affect the relationship between systemic PK, efficacy and toxicity, and should be accounted for in PK/PD modeling |
Quantitative interpretation of ‘traditional’ in vivo biodistribution study data are often dependent on the technique used to measure it | • Rational tagging/labeling location on TP is crucial to ensure that the appropriate moiety is being quantified • Improvement and innovation in bioanalytical techniques, e.g., mass spectrometry for unlabeled TPs; and knowledge of their applicability and/or limitations for different TP modalities will inform the suitability of this data for different applications, e.g., molecule ranking, translational PK/PD modeling |
Translation of preclinical PK/PD data and/or models to humans for novel TP modalities may require more complex approaches than allometry or systemic target-mediated drug disposition models | • Mechanistic models, e.g., physiologically-based pharmacokinetic/PD, quantitative systems pharmacology, will become essential when there is complex target biology, e.g., target in multiple physiologic locations, target shedding; and for TPs with multiple targeting domains or modes of action • In vitro–in vivo correlation or quantitative structure PK relationship models could be integrated into mechanistic models for use during early stages of development |
There are currently no clear rules on which biodistribution data, or in silico models based on biodistribution data, should be included in regulatory submissions | • In the absence of specific regulatory guidance on this topic, it remains at the company’s discretion. Similarly to PK data, biodistribution data which is critical to the understanding of the key PK/PD relationships for efficacy or toxicity, and/or which has been used in human dose predictions, should be included in submissions • Industry and regulatory agencies may wish to collaborate to produce future guidance on this topic, to avoid ambiguity and further expedite the submission process |