What Atomic Bombs Taught Us About Our Brains

​Another one of my favorite science stories: the tale of how Jonas Frisen of the Karolinska Institute enlisted the help of nuclear physicists to unravel one of the most pernicious problems in modern neurobiology.

As I've written in the past, people who dislike animal research have little idea just how difficult it is to do effective studies in people. While it's possible to do more descriptive studies in animals (under proper regulations, of course), when human subjects become involved things become much more limited. As a researcher, I can get permission to give people surveys and tests, perform various scans, collect blood and maybe the occasional biopsy. Other than that, you start hitting walls. In addition to being frustrating, it slows down research progress dramatically.

To illustrate this, I'll give you an example straight out of one of the research fields I specialize in. Since 1965, we've known that the mouse brain has the capacity to generate new brain cells. Two researchers showed that rodents retained radioactive thymidine (one of the four nucleotides that make up DNA), suggesting that DNA synthesis (and accompanying cell division) occurred in rodents. Despite broad dissemination this information, it took an additional thirty years before researchers could show something similar happens in the human brain. The reason: No one could think of a way to do the experiment that would show it. University research review boards were somewhat hesitant to allow the injection of radioactive nucleotides into people. 

But life finds a way. In the late nineties, doctors were conducting a clinical trial where patients with head and neck cancer were injected with a compound called bromodeoxyuridine (Which I will henceforth call BrdU).  BrdU is nearly identical to thymidine , save for the substitution of a single bromine atom. The addition of this halogen made a DNA strand containing BrdU more likely to kink and malfunction when exposed to high energy radiation. The idea was that rapidly-dividing tumor cells would incorporate injected bromodeoxyuridine in place of thymidine, rendering them (relatively) more sensitive to subsequent radiation treatment*.

Well, someone who had no business noticing this clinical trial noticed. Neurobiologist Fred "Rusty" Gage, one of the big names in neural stem cell research, gathered a number of the brains of patients who'd received this type of treatment (after they'd died from natural causes) and looked for brain cells that had incorporated BrdU. And he found some. It was huge news, the first-ever evidence that the brain is capable of regenerating itself to some degree. We made a daisy chain for Rusty when he visited us at the University of Florida (full disclosure: I just looked up 'daisy chain' to make sure it wasn't dirty and... and it kind of is. Fuck it, I'll leave it up to your imagination how we feted him). 

Problem was, the BrdU thing was kind of a one-off deal. Once the clinical trial failed*, no one had permission to inject any more BrdU into people, and we all kind of hit a wall again. Although we knew some cells were dividing, we really had very little idea what was dividing, how often, and what kinds of cells were being replaced or added. This was really, really important if you wanted to use the endogenous capacity of the brain to repair itself. People were doing all kind of elegant studies in mice and monkeys, but had no way to relate it to human biology because there was simply nothing to compare it to. 

Enter Frisen. He was a fairly big guy in the field of neurogenesis, but he was coming off a major fuck-up where his group had misidentified the major stem cell of the subventricular zone (a part of the brain few people care about)**. I like to imagine him sulking in a castle in Stockholm (although I would later learn during a visit that the Karolinska Institute where he works is actually quite modern), trying desperately to think of a way to redeem himself.

And then he did. Frisen absolutely hit one out of the park by - as Gage did - paying attention to something he had no business paying attention to and using it to his advantage. 

Frisen's solution to the neurogenesis dilemma involved nuclear fission. Towards the end of World War 2, we started doing above ground nuclear testing. Then Russia started. For like ten years we were like two kids with a bunch of fireworks, popping off nukes wherever and whenever we damn well pleased. 

Every time an above-ground nuclear detonation occurrs, a heavier-than-normal isotope of carbon (called C14) was produced and flooded into the atmosphere. C14 is pretty harmless, and slowly decayed to normal carbon over time. Nevertheless, we had 'contaminated' our atmosphere with this unique material that was previously not present. 

Very simply, Frisen more-or-less made C14 into his BrdU. Using recorded levels of atmospheric C14, he estimated how much C14 should be present in brain cells of patients, comparing that number to what he actually found. To use a very simplistic example, if postnatal neurogenesis didn't occur, a person born before nuclear testing would have virtually no C14 in their brain. However, if neurogenesis was occuring, new cells would have to incorporate some of the atmospheric C14, labeling them bright as day (to a nuclear physicist, at least). The result was the first birth-dating technology for human brain cells. A simple way to think of this would be carbon dating, not for paleolithic bones, but for cells in our bodies. 

In addition to confirming the Gage result, Frisen et al were able to use their approach over and over, in different brains of different ages, sex and lifestyle to determine in much more detail what was happening within our craniums. For example, Frisen et al discovered that something like 700 new neurons are born each day in the hippocampus (the part of the rodent brain where neurogenesis was discovered so long ago) and, as observed in nonhuman primates, this rate decreases as we get older. Moreover, they determined that, over time, a huge fraction of the neurons in the memory centers of our brains are replaced.

This result was, truly, a breakthrough for the medical community: for the first time, we knew that the brain has an endogenous capacity to regenerate itself. Moreover, we received some idea of the kinetics of this process and, in subsequent projects, we learned what sort of cells were being produced. All of this is crucial information to anyone attempting to harness or enhance endogenous neurogenesis for purposes of recovery after, say, a stroke or traumatic brain injury. 

Since the original report, Frisen and Co have come up with an even more detailed dynamic characterization of neurogenesis. 

It's not all wine and roses though. Retrospective carbon dating has some significant downsides. It can only be done on postnatal (or, in theory, biopsied) tissue. Also, if the publication rate of Frisen's guys is any indication, it's painstaking and/or expensive to get reliable data (eight years elapsed between papers). That tells me that this is a tedious process, a stopgap measure that will advance our knowledge until better methods emerge. Which leads us to... 

Where do we go from here?

So far as I'm aware, no one has managed to develop a better way for measurement of neurogenesis in humans than this method. There's a PET imaging study that claims to recognize the faint pattern of brain energetics associated with neurogenesis, but the integrity of that report has been tangled up in suspicions (I think the authors may have tried to commercialize the technology, making the algorithm they used secret, which is a big no-no in peer-reviewed research). 

What else might the future hold? One of the best ideas I've come up with is to use our knowledge about the gene-level changes in neurogenesis in an attempt to identify similar gene changes in peripheral tissue (i.e., nucleated blood cells) that may act as a proxy measure for neurogenesis. Another option is the identification of a small molecule (protein or other metabolite) that is primarily produced in the brain that subsequently leaks into the blood. We call these biomarkers, the smoke that accompanies the fire. We've done work on both these topics in my laboratory, and hope to publish the results sometime soon, maybe even this year. If you're really interested, I'll be making a presentation on some of our cutting-edge research at this year's Society for Neuroscience meeting in San Diego this November. 


*For those interested, this was one of those "looks good on paper/didn't work in practice" ideas. Bromodeoxyuridine has some rather pernicious effects, and is completely nonspecific in which tissues it targets. If you're reading this and have cancer, trust me, you're not missing anything good. 

**Side note: Frisen published his findings in Cell, arguably the most rigorous scientific journal out there. So don't believe everything you read. I should also mention that I'm not 100% sure he ever admitted being wrong.