In the Harlow laboratory, Tsai studied cyclin-dependent kinases in order to identify their role in cell division. Tsai became interested in CDK5, which she found was not only inactive in cancer cells, but inactive in all other tissue cells except for the brain. She also found that Cdk5 requires p35 to be active. After moving to Harvard Medical School, she began to investigate the function of CDK5 and p35. Tsai found that mice lacking p35 displayed cortical lamination defects and were prone to seizures, and that CDK5-p35 activity was essential for neurite outgrowth during neuronal differentiation. Tsai also discovered that while Cdk5 activity is essential to proper brain development and function, overexpression of Cdk5 was associated with Alzheimer’s disease. Tsai observed that a truncated version of p35 called p25 accumulated in diseased or damaged brain tissue in mice and in tissue samples from deceased Alzheimer’s patients. In an experiment with genetically-engineered mice, Tsai found that increased expression of CDK5 led to the development of Alzheimer’s-like symptoms such as a decline in learning and cognition, profound neural loss in the forebrain, and that amyloid plaques developed within weeks. After moving to MIT in 2006, Tsai began to investigate how to ameliorate or reverse Alzheimer’s symptoms. In a 2007 study, Tsai trained mice to find and remember a platform submerged in a murky pool. When she induced Alzheimer’s-like symptoms, the mice could no longer find the platform; however, after spending some time in an enriched environment, those same mice could locate platform immediately, indicating their memories had returned. Tsai was able to replicate the same effects as the enriched environment by treating the mice with a drug that inhibited a chromatin-remodeling class of enzymes called histone deacetylases, or HDACs. In later studies, Tsai showed that HDAC2 creates an epigenetic blockade of genes that regulate structural and synaptic plasticity and that some cognitive function could be restored by inhibiting HDAC2 activity. Tsai has elucidated the role of structural and epigenetic mechanisms in Alzheimer's disease, showing in two 2015 studies that the DNA breakage necessary to learning was also responsible for cognitive decline, due to decline in DNA repair systems with age, and that the genetic component of Alzheimer’s primarily affects the regulatory circuitry of immune processes, rather than neuronal processes as expected. In 2016, Tsai demonstrated that visual stimulation of mice with an LED flashing at 40 hertz substantially reduces the beta amyloid plaques associated with Alzheimer’s disease, likely by inducing gamma oscillations. In more recent work, Tsai has created a lab-engineered model of the Blood-Brain Barrier to investigate how Alzheimer disease risk genes, namely APOE, contribute to breakdown of the brain's vasculature.
