The last five years have seen remarkable advances in in vitro and ex vivo pain research. Human DRG organoids, iPSC-derived sensory neurons, organ-on-chip platforms, and high-throughput electrophysiology assays have opened new windows into human nociceptive biology. These tools are valuable, and the field is right to invest in them.
But there is a growing tendency to frame these technologies as replacements for in vivo animal models rather than complements to them. That framing deserves a closer look, because the data tell a more nuanced story.
Cell-based and organoid systems excel at molecular-level questions. They allow researchers to study ion channel pharmacology, receptor signaling, and gene expression in human-derived tissue under controlled conditions. For target identification, mechanism-of-action characterization, and early compound screening, they are powerful and efficient tools.
iPSC-derived sensory neurons can be generated from patient populations, enabling disease-specific phenotyping. Human DRG organoids can recapitulate aspects of peripheral neurogenesis. Organ-on-chip platforms can model tissue interfaces. These are real capabilities that complement animal studies at the early stages of a program.
The FDA Modernization Act 2.0 (2022) has rightly encouraged the development and adoption of these new approach methodologies (NAMs). Ethical and scientific progress often align.
No systemic biology. Pain is not a single-cell event. It involves peripheral sensitization, spinal cord processing, descending modulation, immune cell recruitment, and behavioral output. An organoid or cell culture cannot replicate the interplay between the peripheral nervous system, central nervous system, immune system, and musculoskeletal system that defines a clinical pain state. There is no vascular supply, no immune infiltration, no endocrine signaling.
No behavioral endpoints. The clinical experience of pain is defined by what patients do and report: guarding, sleep disruption, functional impairment, analgesic consumption. In vitro systems measure calcium flux, action potential firing, and gene expression. These are useful surrogate markers, but they are not the endpoints that regulators and clinicians use to evaluate analgesic efficacy.
Maturation and variability challenges. Human DRG organoids still face significant hurdles in replicating the full complexity of in vivo neurogenesis. Variability across differentiation batches, between cell lines, and across research groups confounds data interpretation and reproducibility. Organoids lack the neural network complexity needed for later waves of neuronal subtype specialization.
No pharmacokinetics or duration-of-action data. Clinical pain management is often determined by how long a drug works, not just whether it works. Extended-release formulations, depot injections, and combination products require in vivo PK data in a relevant species to characterize absorption, distribution, and duration. An organoid cannot provide this.
No spontaneous pain readout. The most clinically relevant pain behaviors are spontaneous: unprompted guarding, altered gait, weight redistribution, social withdrawal. These behaviors integrate peripheral and central processing in ways that no in vitro system can replicate. They are also among the strongest predictors of clinical analgesic efficacy.
The FDA Modernization Act 2.0 removed the strict requirement for animal testing before clinical trials, opening the door for NAMs including organoids and in silico approaches. This was a meaningful policy shift. But it did not eliminate the need for in vivo data. It expanded the menu of acceptable evidence.
In practice, the FDA continues to expect robust preclinical efficacy and safety data for analgesic candidates. For pain programs, this means demonstrating not just target engagement but clinically relevant analgesia with appropriate duration, dose-response, and safety margins. The FDA's Animal Model Qualification Program (AMQP) specifically incentivizes the use of validated animal models that predict clinical outcomes.
The regulatory trend is not away from animal models, but toward better animal models that demonstrate clinical predictive validity, multimodal endpoints, and translational fidelity.
The case for translational large animal models in pain research rests on a specific set of capabilities:
Spontaneous pain behaviors that parallel clinical presentations and can be quantified using validated tools. In porcine models, these include guarding, weight shifting, altered gait, and avoidance, scored via distress behavior scales and human approach testing.
Multimodal endpoints in a single subject. Behavioral, electrophysiological (SNCV, cMAP, SNAP), histological (IENF density, CGRP, NaV1.7), pharmacokinetic, and biomarker data can all be collected from the same animal, providing an integrated dataset that mirrors the multidimensional clinical assessment of pain.
Porcine pain models have correctly predicted both clinical success (ZYNRELEF) and compounds that did not translate to the expected clinical outcome (aprepitant for neuropathic pain), and have reproduced human drug rankings for approved local anesthetics (Castel et al., 2017, J Pain Res)."
Pharmacokinetic and duration data in a species with human-like skin and tissue architecture, critical for topical, injectable, and extended-release formulations.
The most effective preclinical pain programs are not choosing between in vitro and in vivo but integrating both.
In vitro systems for target validation, mechanism-of-action, and early compound triage. Rodent models for proof-of-concept and dose-ranging. Translational large animal models for clinically predictive efficacy, safety, and PK data that support IND filings and de-risk Phase II.
The field is shifting toward smarter integration of tools, not a replacement of one with another. For pain drug development, large animal models remain the bridge between molecular insight and clinical confidence.
MD Biosciences provides integrated preclinical pain programs spanning in vitro, rodent, and translational pig platforms. For guidance on building a preclinical strategy that incorporates both emerging and established methodologies, contact neuro@mdbiosciences.com.