Are mouse models of Alzheimer's disease truly flawed, or are we misusing them?

The first transgenic mouse models of Alzheimer's disease (AD) developed in the 1990 s were hailed as a major breakthrough. As newer generations of models have emerged over the decades after, they still face growing criticism for translating poorly into clinical outcomes, even if the effects of newer therapeutic compounds, including monoclonal antibodies, align well with those seen in mice. Like any disease model, AD mice have substantial limitations. The real issue is that they are often treated as miniature humans with AD, which they are not. Overreliance on specific pathogen-free (SPF) housing may also blunt immune-related endpoints and compromise translational relevance. The impact of biological sex and ambient temperature may also differ between animal models and humans. In this review, we argue that despite these limitations, they remain highly useful for specific scientific questions and preclinical development. They are far better suited for pharmacokinetic and biodistribution studies, and for testing target engagement in vivo on well-defined neuropathological processes. Thus, while criticisms of these models are justified, researchers have often used them with misaligned objectives. As discussed herein, these models should not be discarded, but used more thoughtfully, in ways that align with their true strengths.