Breakthroughs, and new crises, in the lab
Lynch Lab sits between a toll road and the UC Irvine main campus, in an office park of indistinguishable low-rise, beige-on-beige stucco buildings. Neuroscientist Gary Lynch had moved his lab and office -- for a while, just a desk in a hallway -- numerous times during his Irvine career, often as the result of some feud or slight. He ended up in the office park largely because everybody -- including him -- concluded all parties would be better-served if there were physical distance between Lynch and his university peers.
The lab is at 101 Theory Drive, a developer’s idea of a scientific street name that Lynch found presumptuous.
It is a mark of the difficulty of life sciences -- biology and its many descendants -- that to call something a theory is to honor, not slight it. Theory, evolutionary biologist P.Z. Myers has written, is what scientists aspire to. Lynch, for all of his bombast, was respectful of the intellectual protocols of his science.
“I would have called it Hypothesis Drive,” he said.
The hypothesis is the fundamental organizing principle in scientific research. Its “if this, then that” structure underlies almost all scientific experiments. The work in Lynch’s lab has been driven by a single overriding hypothesis Lynch first published in 1980.
Lynch proposed that the fundamental act by which a memory was encoded involved a nearly instantaneous physical restructuring of portions of brain cells, called neurons. That restructuring allowed neurons to be built into small networks. Each small network would be a memory, he thought.
Lynch’s research focused on a particular area of the brain, a structure called the hippocampus, long thought to be involved in memory. Most neurons in the hippocampus have roughly triangular bodies. Slender fiber extensions called dendrites sprout from the top and bottom. The branches coming out of the top are called apical dendrites. Those coming from the bottom are called basal dendrites.
Also coming out of the bottom is a single larger extension called an axon. All along their lengths, the dendrites are marked by microscopic nubs called spines, thousands of them per dendrite. The axons of one neuron extend to meet the dendritic spines of other neurons. These dendrite-axon junctions are the synapses.
Lynch proposed that the dendritic spines at these junctions changed shape during a process known as long-term potentiation (LTP), which resulted in the strengthening of the bond between a dendrite and an axon. The remodeled dendrites, he said, were the base elements of memory.
Lynch acknowledged that the details of the biochemical interactions that caused the shape change were complex and not well-understood -- at the time he originally proposed it, in fact, not understood at all.
But the crux of the hypothesis was that human interaction with the environment -- a glimpse of blue ocean, the touch of a silk scarf -- resulted in an actual physical change in cells in the brain, and that those changes were the underpinning of memory.
Two notable properties of memory are its vast size and that it can be made in a moment, yet last essentially forever. Any attempt to describe the physical components of memory had to account for those properties.
There are about 100 billion neurons in the human brain. Each neuron has dozens of dendrites, and each dendrite has thousands of potential synapses. So the synapses offered immense storage capacity. But how could storage be so long-lasting? “For me it was really, really obvious it had to be structural, but beyond that, what could I tell you?” Lynch said.
Aside from the pure scientific achievement of understanding the proposed memory mechanism, the value to ordinary people has become more apparent almost every day since Lynch proposed it.
We are in the midst of a brain failure epidemic. Worldwide, it is estimated that by 2040, more than 100 million people will suffer some form of dementia. The physical mechanisms of memory break, and do so with frightening frequency. Lynch was fond of saying that you had no hope of fixing it if you hadn’t first figured out how it was supposed to work.
To understand why it can take so long to figure out, imagine a vast pile of broken plates. A hypothesis is what someone, after surveying the pile, might say about putting the pieces back together.
If the hypothesis holds for a while, survives challenge and criticism, much of it improbably hostile, it might eventually come to some rough, general acceptance and be joined to other hypotheses to form something more far-reaching. Neuroscientists habitually use a particular word to categorize such a body of thought -- or collected wisdom -- in a part of their science. They don’t, as a layperson might, refer to it as a theory. Instead, they call it a story.
Lynch was perilously close to believing he knew the story of human memory -- why it exists, how it works, how it fails.
Eric Kandel, a Nobel laureate for his own memory research and a competitor, said of Lynch’s work, “The current view of LTP is Gary Lynch’s view” meaning that much of what was known about the process was what Lynch had discovered.
“He was the first person to fully appreciate the significance of LTP as a physiological phenomenon -- very prescient,” said Richard Morris of the University of Edinburgh.
At the time it was published, Lynch’s hypothesis was a lonely view. By January 2005, after he and others around the world had spent decades collecting evidence, the hypothesis was widely shared but by no means proven.
Lynch and his colleagues then embarked on a series of experiments intended to prove or disprove his hypothesis once and for all. And, as if to taunt Lynch, almost everything the lab did for a month didn’t work. Computers crashed, fundamental experiments backfired, grad students went AWOL. The whole research program seemed imperiled. The lab was not a happy place.
What brain scientists love about the hippocampus, apart from its presumed significance, is that it is largely a good neighborhood in which to work. It is, compared with much of the rest of the brain, orderly, neatly layered and segregated. It is well-mapped; its regions named with straightforward simplicity -- CA1, CA3, etc. If you chose to poke around there, you generally knew what you were poking.
There is a place in the hippocampus, however, where the naming scheme as well as all ideas about the function of the thing ran aground. The place is called, in an appropriately Middle Earthian way, the mossy fibers.
“It’s the strangest connection in the brain -- the strangest thing in the mammalian brain -- right in the middle of the memory structure,” Lynch said. “There’s no hypothesis as to why these things are sitting in the middle of the memory structure.”
“These things” are very long, faintly furry axons extending from neurons in an area of the hippocampus known as the dentate gyrus. That they are there bugged Lynch to no end. Here sat something dead-center in the thing he had studied for decades, and he was utterly perplexed by it. Clues were gathering, however.
Laura Colgin, a postdoc in the lab, was intrigued by weak electric pulses that apparently originated in the same area as the mossy fibers. Other researchers had reported similar low-frequency waves occurring elsewhere in the brain during sleep and periods of wakeful rest. They called them sharp waves. No one knew what the sharp waves did until Colgin discovered that in the right circumstances, they seemed to erase LTP. In other words, if LTP was the mechanism for memory, sharp waves could be a mechanism for forgetting.
As counterintuitive as it seemed, the idea that there might be an active forgetting mechanism made sense. No one remembered or would want to remember everything. There had to be a way to get rid of stuff.
“The brain is set up to detect patterns and causality, even when they don’t exist,” Lynch said. Because the brain does this, Lynch thought, it would be useful for the brain to have a way to rid itself of patterns that turned out to be misleading.
“The assumption that contiguity means causality probably flows from LTP, and controlling it provides a good excuse for the erasure process,” Lynch said.
Say, for example, you scratched your nose at the same time a dog barked. The brain might read the fact that these two events occurred at the same time as evidence they were related. There was a need to rid yourself of these sorts of false associations.
While Eniko Kramar was leading the troubled effort to validate Lynch’s broader LTP hypothesis, Colgin’s sharp wave research prompted Lynch to begin a broader investigation of ordinary, non-disease-related memory decline, which had been documented in the psychology literature for decades. The general consensus of that research was that memory started to decline not long after humans reached physical maturity. You peaked at about 20 years of age. After that, in terms of your ability to remember new things, it was a long, slow, steady slide. Some studies indicated the rate of decline was pretty much a straight line. You lost as much ability to memorize between the ages of 20 and 30 as you did between 50 and 60.
To neuroscientists, forgetfulness was largely seen as a behavioral curiosity, and was not much studied as a biological phenomenon. When it was, it was generally regarded less as a process itself than as the failure of the memory process.
In Lynch’s work, it was clear that LTP was greatly diminished in elderly rats, but there were no experimental data to support the notion of middle-aged decline. If LTP were the underpinning of memory, and memory began declining in middle age, then LTP’s decline should mirror that of memory.
Lynch decided to do a more systematic search. He assigned the project to a young graduate student, Chris Rex, who was a student of Lynch’s collaborator, Christine Gall. Rex had come to Lynch Lab to learn physiology, and never left. He fell in love, he said, with the direct contact to the biology that he felt doing physiology experiments. “It’s more like having a conversation where the challenge is really asking the right questions,” he said.
Lynch had a vague idea that one reason no one had ever found LTP decline in middle-aged animals was that everyone had looked in the wrong places. That they did so was largely a matter of convenience. The apical dendrites are more numerous and easier to study, so most research focused there.
Lynch sent Rex off into the basal dendrites of middle-aged rats, where -- wonder of wonders -- Rex found distinct failure of LTP.
Since Colgin’s discovery that forgetting could be the result not of a failure but of an active process, Lynch had begun formulating a broader view of LTP. He began to see it as the result of an exquisitely balanced set of inputs. Some of the inputs encouraged LTP. Others inhibited it. Such systems are common in mammalian biology. They can fail from either direction -- too little incitement or too much inhibition. Maybe the failure of memory was simply the result of this mechanism getting slightly out of kilter. It could be as simple, Lynch thought, as too much of one protein or too little of another.
Lynch had known since the early 1990s that too much of a molecule called adenosine outside a neuron interfered with LTP. He suggested Rex administer a drug known to block adenosine to brain slices of middle-aged rats where LTP was inhibited.
After a couple of weeks of false starts, including another computer crash and some difficulty administering the drug, Rex decided to wait until after LTP was induced to block the adenosine. Everybody else in the lab thought that was a weird idea that would never work. Lynch shook his head. “Crazy kid,” he said.
At 1 a.m. on a Saturday in February 2005, Rex erased the LTP deficit. It was completely, utterly gone. Full LTP was restored. In brain science, even many successful experiments have vague results that could be read in various ways. Seldom is anything this clear-cut.
“It should never have worked,” Lynch said.
Rex, a second-year grad student, grinned and shrugged his shoulders. “Pretty lucky,” he said.
The result was astonishing. What Rex seemed to have discovered was a major cause -- if not the major cause -- of one of the most persistent, widespread real-world effects of aging: forgetfulness. And it seemed to be caused mainly by too much of a single molecule, adenosine.
After weeks of repeated failures on almost every other front, Lynch was ecstatic. “You mean this crap actually works?” he said. “You don’t expect to see a result this black-and-white. You expect ambiguity. Aging does not occur uniformly even across a single neuron. It’s an instant default explanation for memory loss. It’s getting to the point where we might have to start believing we were right.”
A good rain
Rex’s long-shot success dislodged some karmic plug in the LTP universe. Things started to work throughout the lab.
Kramar had struggled for weeks with the experiment that was intended to definitively validate Lynch’s hypothesis that the dendrites of neurons were physically reorganized during LTP. She had endured a weird barrage of difficulties: problems with her experimental apparatus, outdated chemical reagents, improperly prepared specimens.
Suddenly, the difficulties all disappeared. The experiment started working precisely as planned, and her results were almost too good to be true.
“The data are perfect,” Lynch said.
Kramar, still shaken by the preceding bleak weeks, spent hours alone in the imaging room, studying her results; she was afraid to believe what she was seeing. She produced stunning images showing the cellular reorganization Lynch had hypothesized as the end stage of LTP, the step that locked a memory in.
She did a series of experiments to block LTP, and the cellular reorganization disappeared. She incorporated Rex’s adenosine findings. The results were clear-cut. Adenosine blocked the reorganization. Take it away, and the process worked perfectly. She blocked the integrins, the molecules that stitched everything else into place. The LTP disappeared. Every crank of the wheel churned out another supporting result.
After decades of struggle, all of the pieces were falling into place. Lynch’s long-standing hypothesis was being borne out to the smallest detail. He could hardly believe it.
“I have to say, I’m flabbergasted,” Lynch said. “I genuinely believe that what we’re staring at is the exact thing that occurs in adult mammals as they lay down memories. The exact thing. That’s it. That’s what I wanted. I wanted to see the thing itself. . . . It’s just colossal. It’s a very hard thing to believe.”
The mundane nature of the molecules involved made the findings more convincing, Lynch thought. No divine intervention, no magic gene -- “just another lift from the parts bin,” as he termed it.
“You give up grandeur, but in return you get confidence,” he said.
One morning, not long after, Lynch woke up and could barely get out of bed. He had no balance. He couldn’t walk down the hall. Within the week, he developed an acute respiratory infection. He had an attack of gout. Another unrelated ancient ailment -- caused by a chronic spinal condition -- recurred. It was like a perverse illustration of his constant complaint that you never knew what was about to go wrong.
The lack of balance was thought to be the result of a viral infection of the inner ear; his doctor sent him to get an MRI to rule out problems deeper inside.
Each image of a typical MRI shows a very thin slice of whatever body part is being examined. A brain MRI produces in digital form what you would get if you were able to take a very large kitchen mandoline and work your way down, slice by slice, from the top of a skull to the bottom. The resulting stack presents a digital photo album of the inside of the head. After the exam, Lynch asked for copies of the images.
He left the clinic that day with a CD-ROM containing the interior images of his head in the pocket of his black cotton jacket. He hopped in his brand-new cobalt-blue 400-horse Chevrolet Corvette convertible and headed toward Irvine.
Lynch is a torque man. He drives very fast, especially in the lower gears, where the experience of speed is visceral. Unless he’s on the freeway, he seldom gets out of third gear. Of course, in the Corvette, third gear can mean flying 100 mph down a blind alley, giggling like a schoolgirl.
Lynch took the CD-ROM back to his lab, where he fed the images into a computer program that allowed him to scroll from top to bottom -- like riding the Magic School Bus with Ms. Frizzle -- through his brain. Upon first sight of his own brain, Lynch began to make some not altogether happy noises. There were low whistles, smacked lips and much muttering. He shook his head a couple of times. He grew uncharacteristically somber.
MRIs, as useful as they can be, remain crude tools. They allow one to see larger structures inside the body. Unfortunately, the work of the brain occurs mainly at the micro scale. The MRI would give Lynch a flyover from 35,000 feet; what he really wanted was to blow down through the anatomical weeds, low to the ground in first gear in the Vette.
Even so, the MRI revealed cause for concern. The human brain contains in each hemisphere large cavities -- literal holes in the head -- called ventricles, where cerebrospinal fluid is produced. The ventricles in Lynch’s brain were enlarged. This in itself came as no great surprise. Ventricle enlargement often accompanies aging. The crucial questions were how much expansion and from what cause.
As he sat in his office, looking at his brain blown up to quadruple scale on his giant Mac monitor, he exhaled, shook his head, pointed, and said, “Boy howdy. That doesn’t look very good.”
Lynch and a company he co-founded were that very month trying to get Federal Drug Administration approval to begin testing the drugs he had invented, called ampakines, in humans. The ampakines were intended to help alleviate a wide variety of brain malfunctions.
He slumped back in his chair and said, “You better hope we come up with something on them ampakines. Normal or not, you don’t want this.”
Despite the siege of illnesses, Lynch continued going to the lab every day. One Friday, he realized he had forgotten to submit to the National Institutes of Health crucial data supporting his request for renewed funding. The data were long past due. Lynch was stricken.
That afternoon, everybody in the lab gathered to celebrate the incredible run of good fortune. Somebody dug out three bottles of Sierra Nevada Pale Ale and divided it up among a dozen or so people. With no drinking glasses at hand, they poured the beer into test tubes and tiny chemical beakers.
Lynch, standing in the middle of the celebration, raised his beaker:
“We do adenosine, Eni’s integrin experiments. We ran the table. We ran the table. Then I realize: I forgot my grant. I forgot to send the supplemental material. I’m a chronic screw-up. I promise you, I’m the only neuroscientist in history who forgot his grant. This is a screw-up of biblical proportions. Even I have to say it -- that’s a screw-up. This grant is cursed.”
Lynch stood there, swaying back and forth. His face, expressive even when becalmed, now seemed about to stretch beyond the bones beneath it. His jaw worked slowly from side to side, his grin shifting with it. He leaned on a lab bench for support, to keep standing.
He stood with his little beaker of beer and grinned. He stood there like that, grinning and quiet and swaying, for a long time.
“It doesn’t matter,” he said. “I can barely walk a straight line, and I blew my grant. I’m a chronic screw-up. Who cares? I have this beautiful science raining down all around me.”
(BEGIN TEXT OF INFOBOX)
Glossary of terms
Adenosine: A molecule that exists throughout mammalian biology. In the brain, it appears to perform a specific function in the memory process -- erasure.
Ampakines: A class of drugs designed to enhance communication between brain cells. The drugs, still in development, are envisioned to enhance almost all cognitive activities.
Axon: A fiber that extends from a neuron and sends signals to other fibers called dendrites. Axons and dendrites meet at the synapse.
Dendrite: A fiber that extends in bunches from a neuron. Dendrites receive signals from another sort of fiber called an axon. Dendrites and axons meet at the synapse.
Dentate gyrus: Part of the hippocampus; Lynch Lab found that sharp waves, an electrical rhythm, originated here. The lab hypothesized that the waves were a means by which the brain erased things it did not want to put into long-term memory.
Integrin: One of the most common types of molecules in mammalian biology, integrins tie things into place. For example, they cause blood cells to clot, allowing wounds to heal. Lynch Lab hypothesized that integrins solidified LTP, locking molecular changes into place.
LTP, or long-term potentiation: The strengthening of connections between brain cells that occurs once they have communicated, making subsequent communication more efficient. The communication consists of electrochemical exchanges between two neurons at the synapse, which is where they meet.
Neuron: The most common type of cell in the brain (numbering in the hundreds of millions); LTP occurs between two neurons.
Sharp waves: A naturally occurring brain rhythm that originates in the hippocampus and seems to erase LTP -- or to cause forgetting.
Spine: The point on a dendrite where it contacts an axon.
Synapse: The point at which two neurons communicate in the brain. It is actually not a structure but a gap of about 20 nanometers (20-billionths of a meter) across which one neuron sends chemical signals to the other. The chemicals set off cascades of events inside the receiving neuron. There are estimated to be 100 trillion to 10 quadrillion synapses in a human brain, allowing for immense memory capacity.
About this series
Sunday: After decades of work, Gary Lynch, a UC Irvine neuroscientist, prepares a series of experiments he hopes will show how memories form.
Monday: Working the hypothesis. Things in the lab go haywire.
Today: The lab begins an unparalleled run of success.
Wednesday: The culmination. Can an actual memory inside the brain be seen?