Decoding the Brain

"Joe has been one of the pioneers who has really pushed the power of mouse molecular genetics to get at the fundamental questions in learning and memory," says Rob Malenka of Stanford University. "He's not the only person in the world trying to do this," he adds, "but several of the others are Nobel laureates" - particularly Eric Kandel and Susumu Tonegawa, both of whom Tsien worked with as a postdoc. "So I give him credit for trying to take them on. It's a pretty ballsy thing to do, because these are very smart, powerful people."

TACKLING THE BRAIN

Tsien started life as Zhou Qian but changed his name when he became a US citizen, because the original was too hard for Americans to pronounce. He didn't set out to compete with neuroscience's best and brightest; he simply wanted to study the brain. After picking up electrophysiologic recording techniques as an undergraduate in China, Tsien joined the lab of Lester Drewes at the University of Minnesota in Duluth. While there he explored the signaling pathways that become activated when acetylcholine binds to its receptor which gave him a solid background in molecular biology.

By the early 1990s, Tsien was ready to tackle something larger, something more cognitive. "I began to wonder, when you move up to higher levels, like learning and memory, what is involved in such amazing cognitive processes." So he went to work with Kandel. During his three-and-a-half years at Columbia, Tsien identified a handful of rat genes whose expression is regulated by brain activity - work Kandel calls "some very beautiful science."

Tsien still wasn't satisfied, however. "When you see genes whose expression goes up or down in response to stimulation, what does that really mean?" What role do such genes play, if any, in the brain's activity or the animal's behavior? "That led me to find a place where I could use genetics to manipulate genes, to try to knock them out, to see what happens," says Tsien. "So I went to work with Tonegawa at MIT."

That's where the real fun began. For the next three-and-a-half years, in the early to mid-90s, Tsien labored over the development of a technique to produce a region-specific gene knockout in mouse brain. Adapting the Cre/loxP system that phage use to insert their genes into bacterial chromosomes, Tsien was able to shut down the activity of the NMDA receptor in the CA1 region of the hippocampus, an area critical for learning and memory. Like patients who have lesions in this part of the brain, the mutant mice showed profound memory deficits, providing, says Tsien, "strong evidence that the NMDA receptor is indeed important for memory."

Presented in three back-to-back papers in a December 1996 issue of Cell, the work was a "classic set of studies," says Kandel. "It proved a hypothesis that had been around and had a lot of support, but none that was as direct as this." It also showcased Tsien's technical prowess and his ability to think a bit outside the box. "People in other fields had used the Cre/loxP system," says Ya-Ping Tang of the University of Chicago, Tsien's former postdoc, "but Joe Tsien was the first person to use it in neuroscience research, successfully producing a brain region-specific knockout."

BUILDING A SMARTER MOUSE

The same sort of lateral logic led Tsien to the study that caught the attention of David Letterman: the smart mouse. "People spent a lot of time thinking about how learning and memory could be impaired," recalls Tang, who also worked on the project. "But no one thought about how learning and memory could be enhanced."

The study, which involved ramping up the numbers of functional NMDA receptors in the hippocampus, "was the logical thing to do, and he did it," says Richard Thompson of the University of Southern California. "Most people make a knockout to figure out what a particular gene or protein is doing. Here he sort of knocked something in. Now that's not new, either. What was new is that he was first to do it with the NMDA receptor in the hippocampus."

"In a way it was a fairly obvious experiment to do," says Malenka of the smart mouse, described in Nature in 1999. "But Joe's the one who did it. And it was an important complement to the older work: First he asked, 'Ok, if I knock out this receptor, is memory impaired?' Yes it is. Then he asked, 'Well, how about if I make the receptor better?' Lo and behold, he reported that learning and memory [were] enhanced, which was pretty cool."

With all the media hoopla, a bit of a circus atmosphere ensued. "It was pretty crazy," laughs Tsien. "It makes you realize that there are a lot of things beyond your control" - such as whether the world press will descend on your lab and spend three months asking you questions about the enhancement of intelligence and the potential dangers of breeding supersmart mice. "When we set out doing the experiments, we did not think about all those things," says Tsien, who was then a faculty member at Princeton University, where he had moved in 1997. "We simply tried to prove experimentally that the NMDA receptor is indeed a molecular switch for memory formation. I never thought about anything else."

Tsien soon recovered from the circus, and he continued to probe the functioning of the NMDA receptor using two additional types of inducible knockout techniques he developed in lab: one that allowed him to turn on and off the gene that encodes the receptor at will, and another to control directly the activity of one of the proteins that lies downstream in the NMDA receptor signaling pathway. These methods revealed that the NMDA receptor is not only involved in the initial acquisition of information, but also plays a role in consolidating and storing memories as well: The receptor must be activated again and again to keep that synapse strong and prevent the memory from disappearing.

"That was surprising," says Malenka, "because people had been thinking of the NMDA receptor as a trigger for synaptic modification, but that once the modifications had been triggered, they thought maybe you didn't need NMDA receptors anymore." Malenka says that this paper, published in Science in 2000, is his favorite because, in addition to yielding intriguing results, "it was just a very clever, sophisticated use of reversible molecular knockout approaches" - one of Tsien's fortes.

A QUEST FOR THE NEURAL CODE

Although he still continues to do molecular studies, after moving to Boston University in 2004, Tsien took up a whole new avenue of investigation: searching for the neural code, the patterns of neuronal activity that define who we are. "The brain is not a piece of liver," says Tsien. "But if all you focus on is the molecular side of the process, when you get deeper and deeper it seems that everything is the same. You have a receptor on the cell surface and then a signaling cascade with kinases, phosphatases, transcription factors, and so on. So if you only look at the molecular side you're never going to crack the brain code."

So he and his colleagues have begun to explore what, exactly, a memory is. "We know a lot about how memory is laid down in the brain: by modification of NMDA receptor activity. And we know where memories are laid down. But what is a memory? At the neural network level, what kind of pattern is associated with the formation of memory?"

In addition to designing a device that allows the researchers to monitor simultaneously the activity of hundreds of neurons in the peanut-sized brain of an awake mouse, Tsien and his colleagues have come up with a novel behavioral approach to eliciting the formation of robust memories: for example, dropping the mice from a miniaturized murine version of Disney's Tower of Terror. "It seems that you need to experience that sort of thing only once and you remember it for a long, long, long, long time," laughs Tsien.

"I think they were trying to do something else and they shook the box the mouse was in and saw this huge increase in activity pattern," recalls Remus Osan, a postdoc who helps perform the mathematical analyses of these "startling" experiments. "I think that was Joe's inspiration."

It was an inspired choice. Using this approach, Tsien finds that he can look at the patterns of neural activity and predict whether an animal experienced a Tower-of-Terror type plummet versus, say, a simulated earthquake (placement on a vortex mixer) or a sudden loud noise. Approaching problems from a slightly different angle seems to work for Tsien. "When he got into the electrophysiology [in his recent work], most people were looking at place cells," says Osan. Place cells increase their firing rate when animals visit specific locations in their environment. "That was a well-established paradigm, yet we did things a bit differently. And I think that may be where his strength is. He's trying to address problems that are of interest to everyone from a slightly different point of view, thus bringing something different to the table."

"He's really trying to get at the big questions and I give him credit," says Malenka. "Through the sheer force of his intellect and drive and motivation, he's garnered the resources to do these really challenging and expensive kinds of experiments, and he's trying to move from molecules to network activity to understand one of the most fundamental functions of the mammalian brain."

khopkin@the-scientist.com