While everyone in her lab was working on cancer cell lines, Tsai decided to systematically dissect out every mouse tissue and organ and perform an in vitro protein kinase assay to test for Cdk5 activity. If Cdk5 was active as a kinase, it would attach a radioactive phosphate to a test substrate, a histone H1 protein that’s a component of chromatin in eukaryotes.

“There are few times in my life when there has been a defining moment, and this was one. It was late at night, and I was exhausted, waiting for the autoradiograph to come out of the developer. I didn’t expect much because all of my other experiments had been negative. Instead, I couldn’t believe my eyes!” Tsai recalls. On the film, she saw strong kinase activity associated with only one tissue—the brain. “I experienced what was beyond joy. I had believed in my pursuit, and I had found my answer.” Cdk5 appeared to be active only in the adult murine brain and in the embryonic mouse nervous system, Tsai found.

I learned to think outside the box and not to be constrained by any dogma.

To figure out what was special about Cdk5 kinase activity in the brain, in 1994 Tsai identified and cloned the gene for the regulatory subunit of Cdk5, p35, which is expressed exclusively in brain tissue—but only in mature, nondividing neurons and not in dividing neuronal progenitors—and is responsible for the tissue-specific activation of Cdk5.

As she finished her postdoc and began to look for faculty positions, Tsai decided to leave cancer research behind and start her own lab concentrating on brain development and neuroscience. “In my job talks, I proposed to make transgenic mouse models and knock out p35 in mice,” she says. She was offered a position as an assistant professor at Harvard Medical School and set up her own lab there in 1994.

“Only recently did my postdoc colleagues tell me that, at the time, they thought I was crazy to give up my cancer research and venture into an entirely new area in which I had limited experience,” Tsai says.

Here, Tsai recalls the stress of being a child processing her grandmother’s dementia; dealing with snow for the first time; and her lab’s recent results showing that Alzheimer’s disease–associated brain plaques can be dispersed by exposing neurons to flickering light.

Tsai Tackles

Jarring childhood memory. Tsai was born in Taipei, Taiwan, and raised by her maternal grandmother in Keelung, a small fishing village north of the capital, while her parents remained in Taipei working for the customs department. When she was five, her parents moved to Keelung. Tsai retains a vivid memory of her grandmother that has stayed with her for life: “When I was about three, we were walking home from our daily trip to the market, and there was a thunderstorm. We took shelter at a bus stop, and after the storm was over, I urged her to get us home. She looked at me and said, ‘Home, where is home?’ The startled look she gave me has never left my memory. She looked completely lost,” Tsai recalls. “She was diagnosed with dementia around then, although we don’t know if it was Alzheimer’s or another form.”

Drawn to science. Tsai’s parents invested in her education and that of her younger brother and sister, filling their house with books. “I gravitated towards the science books—biology, astronomy, physics—and I assumed everyone had those interests,” she says. “I was predetermined to be a scientist, even though I didn’t know it yet.”

Cold shock. Because of her love of biology and animals, Tsai decided to train as a veterinarian after high school. In Taiwan, a four-year college education was merged with professional school, and she entered a veterinary program at the National Chung Hsing University in Taichung in 1978. “I learned a lot of biology but no molecular biology, of course, and didn’t know about the possibility of doing laboratory research.” Toward the end of school, her classmates were choosing what type of veterinary practice to join—zoo, farm, or pet-based practices. “All of a sudden it dawned on me that I didn’t want to do any of that.” Through friends at other universities, Tsai learned about graduate school opportunities abroad and applied to universities that had veterinary schools. She received a fellowship to do a master’s program at the University of Wisconsin–Madison and arrived there in January 1984. “It was an interesting experience. I grew up on a subtropical island and had never before experienced snow. I stepped out of the airplane, and there was snow everywhere. It was so cold! There was nothing that I could have purchased in Taiwan that would have prepared me for a Wisconsin winter,” Tsai says.

A turning point. At the University of Wisconsin, Tsai joined the lab of veterinary microbiologist Michael Collins. She studied a genus of bacteria, Pasteurella, that infects dairy cattle. “I realized that I loved laboratory research and took a molecular biology course that was eye-opening. I knew then that I wanted to do a PhD in molecular biology. It was when I found the direction of my life.” After completing the two-year master’s program, Tsai began her PhD studies at the University of Texas Southwestern Medical Center in Dallas. She joined Bradford Ozanne’s tumor biology and virology laboratory. “I learned to think outside the box and not to be constrained by any dogma,” says Tsai. Because the lab did not have much grant funding, obtaining resources for research was a struggle. Yet Tsai, who was studying c-fos, a proto-oncogene, published a paper characterizing its mRNA and protein expression in a leukemia-derived cell line.

Tsai Tinkers

Hub of productivity. After being told by her graduate committee that she could complete her PhD earlier than she expected, Tsai quickly decided she wanted to study tumor suppressor genes and applied to only one lab to do a postdoc. In 1990, she began her postdoc in Harlow’s lab at Cold Spring Harbor Laboratory in New York. “I arrived and lived in student housing that was five minutes from the lab. The cafeteria was in the same building as the lab, and I realized that I could just work almost nonstop,” she says. Within a few months, however, she and the rest of the lab packed up and moved to the Massachusetts General Hospital Cancer Center in Boston. “Almost right away I realized that with the proper resources and support, I could be incredibly productive, getting interpretable and beautiful experimental results,” she says. Tsai published her first Nature paper only one year later, in 1991: she identified the human cyclin-dependent kinase 2 (cdk2) gene, originally identified in Saccharomyces cerevisiae, then followed this work with the Cdk5-p35 story.

Self-educator. Once settled in her lab at Harvard, Tsai made it her first project to create a loss-of-function p35 transgenic mouse and evaluate the consequences for the developing fetal and adult murine nervous system. Using mouse cortical neurons in culture, the lab initially showed that the Cdk5/p35 kinase is essential for developing neurons to produce new projections, called neurite outgrowth. Then Tsai and her colleagues studied the function of the Cdk5/p35 complex in vivo using the p35 knockout mouse. The lab found that the deletion of p35 was not lethal, but did result in a postnatal phenotype of epilepsy, with severe seizures, and sporadic adult deaths. To study whether the histology of the brain is altered in the mutant animals, Tsai collaborated with a Harvard neuropathologist. “Everyone would send him their mouse tissue samples, and after looking at ours, he looked at me with a saddened expression and said, ‘I know you want your mice to have a phenotype, but I’m sorry to tell you that they look normal.’”

In her bold, “I-need-to-see-for-myself” way, Tsai decided to have a second look. She bought a microscope for her office and spent days doing nothing but staring at brain sections. “And I found a difference between our mutant mice and wild-type ones. All of the brain tissues from the mutant mice had a phenotype in which the six cortex layers appeared inverted, with the deeper ones closer to the surface and vice versa. There was also a change in layer 5, which is normally composed of these huge neurons. In the mutants, these cells were more superficial, in a different location. I went back to the pathologist who studied the blinded samples, and he admitted that he had been wrong,” says Tsai. Her lab’s work characterizing mice that lacked p35—which proposed that the cortex in the animals without the protein is abnormal because p35 is needed for proper neuronal migration and proper differentiation into specific cortical neuronal subtypes—was published in 1997 in Neuron.

Ties to human disease. Tsai’s lab found that p35 could be converted to a smaller, 25-kilodalton protein, p25, under conditions of neuronal stress such as oxidative stress or addition of amyloid-β peptides. The team also found that p25 was expressed in postmortem brain samples from Alzheimer’s disease (AD) patients. Accumulation of p25 results in extended Cdk5 activation and mislocalization, resulting ultimately in primary neuron apoptosis. One year later, in 2000, the lab identified the mechanism by which neurotoxicity converts p35 to p25, suggesting its role in the pathogenesis of AD. Influx of calcium into the cell, Tsai’s lab found, activates a calcium-dependent protease, calpain, which cleaves p35.

Tsai Transcends

Pill-popping mice. When postdoc Andre Fischer joined Tsai’s lab in 2002, he began to investigate the behavioral and memory impairments of p25 transgenic mice, finding that transient expression of p25 actually facilitates memory and synapse formation in mice, but that prolonged p25 expression impairs long-term memory retention and retrieval. The researchers then discovered, to their surprise, that some of these memory and learning defects could be ameliorated when these mice were placed in what Tsai calls a “Mouse Disneyland,” a stimulating environment with other mice and a constant stream of new toys for play—even though p25 was still present in the brains of the animals. Even more surprising to Tsai, the same amelioration could also be achieved with an oral pill, an inhibitor of histone deacetylases, which resulted in new dendrite sprouting, an increase in synapse number, and better learning and long-term memory retrieval. “We argued in that paper that the memory is not really lost, but just unable to be retrieved,” says Tsai.

Life-changing results. In December of 2016, Tsai’s lab published a finding that astounded the scientific community: exposing mouse models of Alzheimer’s disease to flashes of light at the frequency of “gamma waves”—a pattern of neural oscillation in mammals between 30 and 80 Hz, detected by electroencephalography (EEG)—could reverse some of the characteristics of neurodegeneration in the animals’ brains. The lab initially targeted interneurons in the brains of mice using optogenetics, and found that the exposure induced gamma oscillations of these cells—an activity linked to higher-order cognitive abilities.

Then, a graduate student in Tsai’s lab, Hannah Iaccarino, wanted to test whether optogenetic stimulation at gamma frequency would be beneficial for mouse models of AD. “After she did the experiment in the amyloid-β mouse model, she ran into my office saying that the amyloid levels in the brains were drastically reduced following one hour of gamma oscillation,” says Tsai. Tsai only began to believe the results when they were repeated; when her team showed that no other frequency than gamma worked; and when the researchers saw major changes in the gene expression and morphology of microglia, the brain’s immune cells. “The microglia just went crazy, their shapes were completely transformed to be much larger with more-elaborate processes,” says Tsai.

The team also saw that the microglia could now more efficiently phagocytose the amyloid-β protein. Tsai’s colleagues at MIT suggested that strobe lighting may be a way to noninvasively stimulate the same gamma waves. The researchers exposed the mice to 40 flashes per second for an hour and found that soon after, their levels of amyloid-β were about half those prior to the strobe-light exposure, along with the other effects seen with the optogenetic approach in a young Alzheimer’s mouse model. Now, the lab is working on studying whether the behaviors of the mice are altered upon recurring strobe exposure and whether entraining the gamma frequency noninvasively through other sensory modalities could work similarly. 

Greatest Hits

• Identified mammalian Cdk5, a novel cyclin-dependent kinase that is only active in the adult and embryonic nervous system
• Found the mechanism by which p35, a regulatory subunit of Cdk5, controls Cdk5’s neuronal tissue–specific activity
• Identified a mechanism by which p35 is converted to p25 in the brains of mammals, including humans; found that p25 accumulation in mice can result in learning and memory impairments and that it accumulates in the brains of Alzheimer’s patients
• Showed in mice that loss of learning and memory behaviors and decreases in synaptic connections in the presence of elevated p25 levels can be reversed with an oral histone deacetylase inhibitor
• With Ed Boyden’s lab at MIT, demonstrated in a mouse model that stimulating neurons to produce normal gamma waves using optogenetic or strobe lighting can reduce the severity of Alzheimer’s disease–linked amyloid-β plaques