Abstract / Summary
Alzheimer's disease (AD) remains a major global health challenge, with a high rate of clinical trial failures often attributed to the limited translational predictive value of traditional experimental models. In vitro systems, ranging from induced pluripotent stem cell (iPSC)-derived neurons to complex three-dimensional (3D) organoids, have emerged as promising alternatives to investigate disease mechanisms in a controlled experimental setting. However, their rapid expansion has generated fragmented nomenclature and heterogeneous classification schemes. In this systematic review, we propose a hierarchical framework that categorizes in vitro AD models according to their architectural complexity and the biological questions they are designed to address. We organize these platforms into four progressive tiers: genetically defined iPSC-derived cell-autonomous models, stimulus-based reductionist models, interaction-dependent neuron-glia co-cultures, and advanced multicellular 3D or neurovascular platforms. Our findings indicate that low complexity models (Tier 1 and Tier 2) are particularly suited to dissect cell-intrinsic genetic drivers and acute toxic signaling, but they often fail to capture the chronic multicellular feedback loops and vascular dysfunction characteristic of AD. In contrast, high-complexity systems (Tier 3 and Tier 4), including organ-on-chip platforms and vascularized organoids, provide the spatial and physiological context necessary to model neuroinflammation and network-level degeneration, although they present greater technical variability. Finally, we propose an integrative hierarchical strategy that aligns model selection with specific pathological questions. By linking cellular complexity with mechanistic objectives, this framework provides a structured roadmap to improve experimental rigor, facilitate cross-study comparability, and support the development of more predictive experimental models for AD.