CARDIAC COMEBACK
How do you do heart surgery on a fish whose entire body is less than two inches long? Very carefully, of course. And so, working very carefully, a team led by Mark T. Keating, an HHMI investigator at Children's Hospital in Boston and Harvard Medical School, performed surgery on hundreds of adult Danio rerio, better known as zebrafish. The researchers snipped 20 percent off the tip of the lower chamber (ventricle) of the fish's two-chambered heart to see whether the heart would regenerate. Keating already knew that zebrafish could regenerate fins, retinas, and spinal cords. Being a vertebrate like us made this animal a good candidate for researchers "trying to create a model system for heart regeneration," he says.

The bottom line was that after about two months, the hearts did regenerate, but the path to that finding, which was featured in the December 13, 2002, issue of Science, was noteworthy in its own right. The first problem was figuring out how to damage a vital organ about half the size of a pea and beating about 200 times each minute and have the fish survive. "We tried thermal injury, which is hard to control," says Keating. "And we tried radiofrequency ablation, which basically is also thermal injury and is also hard to control. And the injuries were not clean and reproducible." Cutting this Gordian knot, therefore, involved cutting. But what with? "A scalpel isn't a great instrument because you can't really hold on to the heart while you're cutting it," Keating explains. "We decided we needed scissors, and we wanted to get the sharpest, smallest scissors we could." Fine-gauge iridectomy scissors, ordinarily used by eye surgeons doing work on human irises, could do the job.

An incision on the fish's ventral side makes it easy to push the two-chambered heart out of the body while it's still connected to the rest of the cardiovascular system. Then the scissors are deployed to remove 20 percent of one ventricle of the heart—arduous trial-and-error efforts by Lindsay G. Wilson, a research associate in Keating's lab, revealed that about 90 percent of the fish could survive this procedure. The only problem is that a cut to the heart, not surprisingly, leads to profuse bleeding. A little pressure on the heart with gauze stops the bleeding and facilitates clotting. Then, Keating says, "we gently put the heart back in the chest" and, given that they are groggy coming out of anesthesia, "stimulate the fish a little bit to get them to swim, which passes oxygenated water through their gills." Two months later, the heart is full-size again.

"There are really two aspects to this research," Keating notes. "First, we had to show that this beast can regenerate. And now we need to get to molecular mechanisms." The obvious hope is that understanding the genetics that govern regeneration in the zebrafish heart could lead to treatments that would induce damaged human hearts to regenerate. That fishing expedition, however, would be pointless had the zebrafish not cooperated.

IF YOU GIVE A MOUSE A FEZ
Does a mouse live by smell alone? A mouse's brisk sniffing brings pheromones in its environment into contact with the vomeronasal organ (VNO), which sends signals to a distinct brain region known as the accessory olfactory bulb (AOB). Lawrence C. Katz, an HHMI investigator at Duke University Medical Center, and his colleagues were able to get deep inside the AOB of mice, in work that was featured on the cover of the February 21, 2003, issue of Science. They found that a mouse interprets another mouse as a distinct pheromonal pattern, similar to how the human brain recognizes another person via "face neurons" in the brain's visual-processing regions. But doing this research presented unique challenges, considering that a mouse AOB is the size of the head of a pin.

Most previous research focused on the responses, in vitro, of the peripheral sense organ—the VNO—to the application of pheromones or substances containing pheromones, such as urine. Technical obstacles have blocked the way to investigating how these peripheral responses are represented in the brain itself. Katz's team had something a little more true to life in mind. "The basic idea," he says, "is that we wanted to be able to monitor the activity of individual neurons in this unexplored area of the brain as an animal is engaged in natural social interactions with other mice."

But to record individual neurons in the AOB required the ability to target and manipulate incredibly tiny electrodes in behaving animals. "That's a tall order," says Katz, "because you need to advance and retract the electrode and have that electrode be stable enough so that you can actually record from a single nerve cell while the animal is moving."

Fortunately, Minmin Luo, a postdoctoral fellow in Katz's laboratory, previously did bird-song research (as did Katz). And Luo was following the publications of a Bell Laboratories bird-song researcher, Michale S. Fee, who had created miniature instruments for studying the brains of zebra finches while they were singing. The three researchers adapted the bird apparatus for use in the similarly sized mice.

"Michale's developments really had two key components," Katz explains. "One is using advanced materials, like titanium, and micromachining techniques to sculpt a headpiece that is tiny and light enough to be mounted on the skull of a very small animal." The headpiece weighs about 2 grams, which even a mouse can tolerate, and looks like a tiny fez. On this fez, however, the traditional tassel is replaced by ultraminiature motors, which in turn control tungsten electrodes, which taper down to a width of only a few microns where their tips terminate in the AOB.

The other component is a feedback system that measures the strain on the cable. Should the mouse turn left, for example, it won't have to drag along the apparatus in that new direction. Instead, the cable's sensors send signals to a motor that tells the electronics residing in a support structure above the mouse to proceed in that direction as well. The unencumbered animal is thus free to act more normally.

Three micromotors allow Katz to advance and retract electrodes by remote control while listening to the activity of single cells as the animal behaves in various ways. "We can listen in on the activity of about half a dozen neurons at the same time if we wish, by recording from those different electrodes," he says. "We can really hear the staccato bursts of individual neurons, and we can continue to record from them while the animal is engaged in the full range of behaviors."

Katz believes this work to be the first that records nerve-cell activity in the brain while two animals are actively engaging in social interactions. ("We have one wonderful recording of the activity of a single cell in the brain while our mouse was actually mating," he says.) Being an HHMI investigator allowed Katz to venture into the uncharted waters of this research more comfortably. "I had no track record in this area," he acknowledges. "I can only imagine what would have transpired had I submitted a grant asking for funds to do recordings from awake behaving animals, never having done it before."

THE DENDRICK DANCE
Karel Svoboda, an HHMI investigator at the Cold Spring Harbor Laboratory, also ponders the workings of brain cells through the window of the mouse—or, more accurately, through a window on the mouse.

Svoboda and colleagues recently discovered that some subtle and fascinating structural alterations take place in neurons as an animal learns. Dendrites and axons, on the input and output ends of a neuron, respectively, had been assumed to be constant in structure over long periods of time. But when Svoboda zoomed in on the dendrites, a more dynamic picture emerged, as detailed in the December 19, 2002, issue of Nature. Getting to that picture, however, required the combination of a designer mouse and some fancy microscopy.

Joshua R. Sanes, at the Washington University School of Medicine in St. Louis and a coauthor of the Nature paper, created a transgenic mouse strain that expresses the gene for green fluorescent protein (GFP) specifically in a small subset of neurons in the cortex—the part of the brain that interested Svoboda's group. GFP expression makes it possible to visualize the neurons, which glow green when hit with the proper wavelength of light. Sanes's original intent was for the protein to be produced at neuromuscular junctions. "But he found that the gene sometimes lands in a part of the genome so that the protein is expressed in a subset of neurons in the cortex—the part of the brain that we're interested in," says Svoboda.

Magnetic resonance imaging (MRI) of the mouse brain would be a straightforward exercise, but using MRI to view a synapse would be like trying to see Pluto with opera glasses. "Synapses are only the size of the wavelength of light," says Svoboda. "The resolution limit of MRI is off by about 9 orders of magnitude." The device that can sneak into these tiny places is a two-photon laser scanning microscope, which not only persuades the GFP in the neurons to glow but captures images of the shimmering structures.

The idea is to throw a photon of a specific wavelength at the GFP, which will absorb that photon and then spit back a green one. But high-energy light scatters in living tissue, bouncing around before it can penetrate to the neurons. Lower energy light can penetrate deeper, but it doesn't pack the punch needed to excite the GFP—ordinarily. Should two low-energy photons arrive at a GFP molecule at the same time, however, they get absorbed together. As far as the GFP is concerned, the two low-energy photons are indistinguishable from a single high-energy photon. This two-photon absorption is normally a rare event. But modern pulsed lasers, which concentrate photons in brief pulses, allow two-photon absorption to be common enough to make deep brain imaging possible.

The microscope cannot, however, see through fur and the skull. "So you basically replace the skull with a small imaging window," Svoboda explains. "It goes right on top of the dura, the skin that covers the brain. The animals can live out their entire lives with this window in place."

By observing the same transgenic-mouse brains multiple times over weeks with the two-photon microscope, Svoboda and his colleagues could see tiny spines along the dendrites rising or receding. The rate of spine turnover increased as mice were exposed to novel experiences: As the organism learns, the brain clearly rewires itself. "This is the first demonstration," Svoboda says, "that there is synapse formation and elimination in the adult brain at a high rate."

SPYING ON T CELLS
For the immune system to wipe out a tumor, it must succeed at multiple tasks. It has to generate T lymphocytes specific to the tumor, and those T cells must migrate to the tumor and then be able to destroy the tumor. While some treatments aim to enhance T cell migration and activity, the current ways of measuring their effectiveness mostly involve looking at the final outcome—via biopsy, for example. But Owen N. Witte, an HHMI investigator at the University of California, Los Angeles (UCLA), and colleagues have taken the first steps toward observing the T cell response in progress.

"We want to know what happens to populations of educated, active T lymphocytes"—T cells that have already eradicated a tumor in mice, thereby showing their "education"—"when they're infused into a recipient who has a tumor," Witte says. Witte transferred these educated cells to other immunodeficient mice—the only T cells available were thus the injected ones. Then the questions began. "Did the cells get to the site of the tumor?" he asked. "And if so, when did they get there, and how many of them got there? And did they multiply or expand upon their arrival there or elsewhere in the animal? In a sense, it was an analysis of the bookkeeping." Witte and his colleagues published their research in the February 4, 2003, issue of the Proceedings of the National Academy of Sciences (PNAS).

To transform the before and after snapshots into a moving picture of the process, Witte turned to a new technology developed by UCLA colleagues Simon R. Cherry, Sanjiv S. Gambhir, and Michael Phelps: small-scale positron-emission tomography, or microPET. Traditional PET constructs images by capturing the gamma rays produced when positrons emitted from a sample collide into their abundant, negatively charged counterparts, electrons. The resolution is "at best 0.1 cc [cubic centimeter], and more typically 0.5 to 2.0 cc," Cherry and Gambhir wrote in a 2001 paper in the Institute for Laboratory Animal Research Journal. The two researchers were able to tinker enough to get the resolution down to 0.006 cc, suitable for imaging small animals.

That microPET instrument needed something to measure, however, and ucla biological chemist Harvey R. Herschman developed a process using a "reporter gene" to do the trick. The gene (herpes simplex type 1 virus thymidine kinase, which encodes the enzyme HSV1-TK) was incorporated into the T cells that Witte wanted to follow. When the enzyme was produced, it sequestered a fluorine isotope-labeled compound that generated positrons. The result was a concentration of positrons in the T cells, which made it possible to eavesdrop on their activities.

"It takes about a week for the T cells to figure out where to go, and to get to that site and begin to divide," says Witte. "It looks like they actually go to more traditional lymphoid sites, where antigen in the blood or some other set of signals begins to cause them to recirculate and eventually end up at the tumor."

The hope for the future, Witte says, "is that as we improve these technologies, they will find utility in human beings who are undergoing clinical trials to test therapies."

Download this story in Acrobat PDF format.
(requires Acrobat Reader)

Photos: Courtesy of Keating lab (zebraFish); Courtesy of Minmin Luo and Lawrence Katz (mice); Courtesy of Svoboda laboratory (cerebral neurons); From Dubey, P. et al. 2003. Proceedings of the National Academy of Sciences USA 100: 1,232-7. © 2003 National Academy of Sciences, USA (mouse Tumor).

Reprinted from the HHMI Bulletin,
September 2003, pages 24-27.
©2003 Howard Hughes Medical Institute