Review examines translational bottlenecks for plant-derived anti-inflammatory compounds

A review led by Professor Weidong Zhang at the School of Pharmacy, Naval Medical University, was published in Volume 2 of the journal Immunity & Inflammation on October 27, 2026. The article systematically examines the translational bottlenecks and potential solutions for plant-derived anti-inflammatory compounds. The authors establish a research framework that evaluates candidate molecules across several dimensions: chemical properties and active forms, proximal or direct molecular targets, target binding strength, administered dose, disease-relevant mechanisms, and clinical maturity. The framework is intended to prevent downstream pathway changes, transcriptional responses, or phenotypic improvements from being equated with direct molecular targeting.

Chronic inflammation is a shared driver of autoimmune diseases, metabolic disorders, neurodegenerative diseases, and cardiovascular disease. Current anti-inflammatory treatments face incomplete responses, long-term safety concerns, high costs for some targeted agents, and substantial patient heterogeneity. Medicinal plants are regarded as an important source of anti-inflammatory drugs because of their structural diversity and potential for multi-target regulation. However, the path from plant-derived molecules to therapeutic agents is more complex than is often assumed.

A notable feature of plant-derived compounds is pleiotropy. Many can influence multiple inflammatory pathways, including NF-κB, MAPK, arachidonic acid metabolism, JAK-STAT, Nrf2, and NLRP3. Yet such broad activity does not mean that these compounds directly engage key targets. Different compounds can differ substantially in their binding strength to targets within the same pathway. The review summarizes mechanistic studies on natural compounds and points out that “research on natural products should not conclude that a compound directly targets a pathway merely because it inhibits that pathway.” Direct target binding, downstream pathway changes, and computational predictions represent different levels of evidence and should be distinguished.

Dose is another major bottleneck. Many compounds show low micromolar activity in vitro, but after oral administration, the free parent compound may be present at very low or even undetectable plasma concentrations. Even if target binding evidence is established, in vitro mechanisms cannot be directly extrapolated to clinical benefit if free compound concentrations in diseased tissue are insufficient to reach pharmacologically effective thresholds. The review therefore emphasizes that “dose must be evaluated together with target engagement evidence.“

Clinical evidence also requires careful matching. In knee osteoarthritis, the curcumin formulation BCM-95 combined with diclofenac and resveratrol combined with meloxicam have shown adjunctive improvements in pain and function in randomized trials. However, these studies did not demonstrate that the clinical effects were mediated by binding to IKKβ or SIRT1. Clinical data for Tripterygium preparations in rheumatoid arthritis largely derive from extracts or glycoside mixtures and cannot be directly used as clinical validation of pure triptolide. Colchicine, by contrast, represents a relatively complete evidence chain: structural biology evidence supports its binding to tubulin, clinical trials show cardiovascular endpoint benefits, and in 2023, it received FDA approval for reducing the risk of atherosclerotic cardiovascular disease.

The review also discusses new tools that are changing research approaches. Network pharmacology and multi-omics can narrow candidate compounds and generate hypotheses, but they cannot replace target binding experiments. Activity-guided isolation and molecular networking can improve dereplication efficiency and reduce repeated isolation of known compounds. Virtual screening and artificial intelligence models can prioritize candidates, but their predictions are limited by training data bias, stereochemistry, and applicability domain, and still require prospective experimental validation. Three-dimensional co-culture systems, organoids, and organ-on-a-chip models can test compound effects on immune–stromal–vascular interactions in environments closer to diseased tissue. For delivery, extracellular vesicles, biomimetic nanocarriers, and carrier-free self-assembled nanoparticles may improve compound stability and tissue-specific targeting, but scalable manufacturing and batch consistency remain prerequisites for translation.

The review’s discussion of combination therapy is also noteworthy. Plant-derived compounds are more realistically positioned as adjuncts to existing targeted therapies rather than replacements. Combined benefit does not equal pharmacological synergy. Future studies need to distinguish simple additive effects from true synergy and to evaluate dose-sparing effects, pharmacokinetic interactions, and additive toxicity.

In summary, the central translational bottleneck for plant-derived anti-inflammatory compounds is not whether they have anti-inflammatory activity, but whether they can act on a defined molecular target at achievable pharmacological concentrations in diseased tissue and produce disease-relevant effects. “Moving mechanistic and pharmacological validation earlier, separating direct target evidence from downstream pathway changes, and separating clinical evidence for extracts from evidence for pure compounds are key to advancing the field from correlative discoveries to actionable candidates,” the authors highlighted.