Organ-on-Chip Evidence Should Constrain Natural Product Pharmacology Models Through Dynamic Exposure, Tissue Interfaces, Multicellular Interactions, and Human-Relevant Physiology

Download PDF

Natural-product pharmacology is commonly modeled from nominal concentrations, endpoint responses, simplified barriers, and isolated cell systems, even when the relevant biology depends on changing exposure, transport across tissue interfaces, metabolism, mechanical context, or communication among cell types and organs. Organ-on-chip and related microphysiological systems can reproduce selected features of these conditions, but their value for computational pharmacology is often framed too broadly as increased physiological realism. This methodological framework instead asks what chip measurements can legitimately constrain. Evidence from dynamic absorption, material-associated drug loss, barrier models, airway pharmacology, gut–liver crosstalk, microbiome-dependent metabolism, liver toxicity, and coupled-organ pharmacokinetics supports treating exposure history, permeability, organ-specific transformation, and context-dependent response as separable empirical restrictions on model behavior. We propose a chip-constrained pharmacology framework in which measured chip observables are qualified for provenance and context, transformed only through explicit scaling or computational assumptions, and used to narrow admissible model states rather than replace simpler assays or in-vivo evidence by default. Contradictions among platforms trigger localized model updating and investigation of alternative explanations instead of automatic preference for the more complex system. For natural products, this approach is especially relevant where parent compounds, metabolites, mixtures, and tissue-specific exposure complicate nominal-dose reasoning. Major limitations include sparse direct natural-product chip evidence, material sorption, biological and device variability, uncertain multi-organ scaling, incomplete standardization, and limited prospective validation. The framework is therefore a proposal for disciplined evidence integration, not a validated prediction standard.