Salk Institute for Biological Studies The world stopped to rethink what was known about the structure and the workings of the brain when the American neuroscientist Fred Gage published, in 1998, in Nature Medicine, the first solid evidence that the human central nervous system continues to generate new cells after adulthood. The result of years of work of the teams under Gage and other researchers, this finding underscored a phase of discoveries that would shake the concept of brain structure and evolution put forth almost one century before by Santiago de Ramón y Cajal.
A Spanish physician and histologist, Ramón y Cajal identified the microscopic architecture of the central nervous system and stated that once the development stage reached its end, the brain was fixed and immutable, given that the “source of growth and regeneration” of brain cells had permanently dried up. Twelve years ago, however, Gage won his place in the history of western science when he demonstrated that this idea was no longer valid, at least not for the entire brain.
Ever since he confirmed the proliferation of cells in the adult brain, a phenomenon known as neurogenesis and described jointly with the Swedish scientist Peter Eriksson, Gage has not stopped creating new experiments in order to identify the function of these young neurons.
One of the most influential neuroscientist of our days, Gage coordinates at the Salk Institute in California a laboratory with some 40 people, which has released more than 600 articles, cited in 57 thousand other works.
During a visit in September to the town of Caxambu, Minas Gerais state, where he attended the 34th Congress of the Brazilian Society of Neurosciences and Behavior, Gage told us how he confirmed neurogenesis among adults and talked about projects under way.
Did your interest in the capacity of the adult brain to create new cells appear while you were still an undergraduate student ?
It was a bit later. When I was an undergraduate, I was interested in discovering how the adult brain reacts to lesions. This is the so-called adult neuroplasticity, which I began researching in the 1970’s, when we still believed that the brain was practically immutable after development. At that time, experiments started showing that the brain might have some recovery capacity after being damaged.
As far back?
Yes, that far back. It wasn’t yet the capacity to produce new neurons. However, if a neuron were cut, it seemed to be able to grow again, to sprout. As this was very limited growth, we thought: “If they can grow a little, perhaps it’s possible to make them grow more.” As soon as this started to be studied in more depth, we found that plasticity was even greater. I was 18 or 19 when I began working in a laboratory and it became clear that the brain had a lot more recovery capacity than we had previously imagined. The discovery that new neurons might appear occurred much later.
Did you reach these conclusions with your 1970’s work as the starting point?
No, things were not that linear. I was busy trying to understand what the regeneration capacity of the different brain areas was. Before I got involved with this subject, there was already some evidence or, at least, published articles stating that there had to be cells that were dividing in the adult brain. However, lots of people didn’t believe this. The researcher that discovered this phenomenon, Joe Altman, is still alive, but he abandoned the field early on, because nobody believed him.
Did he also believe that the proliferation of cells took place in the hippocampus, the brain region connected with forming long-term memory and with spacial localization capabilities?
Yes, in a way. And also in the cerebellum. However, he used a different technique that does not allow one to quantify matters and that wasn’t consistent enough. Today, when we analyze what he did, we see it differently. It was truly outstanding. Later, in the early 1980’s, another researcher [Fernando Nottebohm] showed that there was cell division in the brain of adult birds.
In connection with learning new songs?
That was what he said. The neurons died during one season and new neurons appeared when the birds learnt new songs. But this was highly controversial. This researcher used the same methods that the other one had used and this gave rise to a battle because the method was unconvincing. Other groups, using the same techniques, were unable to reproduce the results. However, I remember having paid attention to what he had said and it was truly impressive. There were still people trying to repeat the experiments, but I came into this story in a different way. I was working with a protein, FGF, the fibroblast growth factor. I had been in Sweden and then went to California in the mid-1980’s, taking along with me what I had learnt about molecular biology and virology.
That was the start of everything.
I knew that this was the path that neuroscience should tread. So we took the gene that codes this growth factor and inserted it into cells. The idea was to implant these cells in the brain and see what the growth factor would do. The cloning of genes and gene therapy were only incipient. The gene had just been discovered at Salk by Roger Guillemin [who was awarded the Nobel Prize for Medicine in 1977 for his neuropeptides studies]. We obtained the gene’s clone, inserted it into fibroblasts and placed the brain cells in a culture with the fibroblasts to see whether any growth would materialize. From one minute to the next, young neurons starting proliferating like crazy. The plate was overrun with cells. Then I thought: “Wow, we must have discovered something new!” At first, we thought that something in the fibroblasts might have triggered a reaction and started secreting an unknown growth factor. We turned to the chemistry of proteins to discover what had happened. The vector we used produced a lot of protein; never before had the effect of high concentrations of FGF upon young neurons been observed. Low concentrations make the neurons grow while high concentrations made them divide. Gerd Kempermann and Malcolm Schinstine were at my lab and we said: “Look there’s this crazy thing about neurons splitting in the brain. If it’s true, perhaps we can insert fibroblasts into the brain and produce new neurons.”
Salk Institute for Biological Studies So what happened?
Well, then we read what had been published on the subject and tried to repeat the experiment, but we were unable to replicate the data. I realized that the problem concerned marking the cells. The way one identified if a cell was dividing was to mark it with a radioactive element and to count the points, which were recorded on a photographic emulsion. It was completely manual. Later, a chemical analog with a high affinity for antibodies was developed. When one administers this analog to the cells, it integrates itself into the DNA. Antibodies can then adhere to those cells and locate their nucleus precisely. We used this compound, called BrdU [bromodeoxyuridine] to identify neurogenesis in adults. At the time, we started using confocal microscopy. Olympus gave us a machine and we managed to see the BrdU marking the antibodies. We also discovered that somebody had used an antibody to mark the nuclei of mature neurons. So I went to Michigan, got the antibody against the neuronal marker known as NeuN and we marked the cells doubly, with BrdU and NeuN. With the confocal microscope we were able to dissect each neuron, rebuild it three-dimensionally and prove that these cells were indeed dividing.
How were you able to see this if it was happening in vivo?
We had to inject BrdU into animals. Four or five techniques were being developed and we started using them as early as possible.
Were you already looking for neurogenesis in the hippocampus at that time?
No. I was trained to work with the hippocampus. But I wasn’t looking for neurogenesis. Nobody paid any attention to this. I was researching the connections between the septus and the hippocampus, trying to reconstruct this pathway. We conducted the experiments, we persuaded ourselves about it and published several scientific articles. The most important thing we did was to show that neurogenesis occurs and that in certain environments the experiment can change the number of cells. In general, laboratory mice share a small cage with other animals. This environment is restricted in that it doesn’t enable mice to explore anything beyond it. One of the experiments we conducted was to move the animals that lived in small cages into larger environments with lots of toys and other objects that might stimulate their curiosity and exploratory capabilities. In these animals, we observed an increase in the number of hippocampus cells. However, there were still doubts about the quantification of the phenomenon. So a new technique was developed, stereology, which allows one to count the exact number of cells. This technique, coupled with confocal microscopy and double marking, enabled us to show not only that there are more cells splitting in the hippocampus of an adult animal, but also that the number of cells increases by as much as 15 percent just with changes in the environment. In general, the dentate gyrus of a mouse’s hippocampus has 300 thousand cells. After one month, the animals in the control group still maintained 300 thousand cells in this region, whereas the animals that spent this time in an environment with more objects had 350 thousand cells. So we managed to convince everybody that it wasn’t marking the cells with BrdU that was giving rise to these results, but that there really were a larger number of neurons there. The subsequent question was to find out whether this was also the case in humans. At the time, we realized that some people had been treated with BrdU, which marks cells that are dividing, in order to evaluate the progression of cancer. When these people died, their brains went for pathological examination. So I called colleagues who were pathologists in cancer laboratories and inquired whether they had patients who had been treated with BrdU and their brains. We got a few samples, but they were fixed in paraffin. Nevertheless, we found BrdU in their hippocampuses three years before publishing the 1998 article. But this was insufficient, because we were unable to conduct the double marking of the cells, so we thought: “Nobody will believe in our work.” Two of my post-doctoral students who had returned to their own countries, one to Finland and the other to Sweden, tried to join teams that were conducting clinical trials with people with cancer in a peripheral organ and that were being treated with BrdU. We would wait for them to die and nurses, along with these post-doctors, would extract the fresh brain and send it to Salk. We would examine this under a microscope and see the cells dividing. So I started calling people from my own lab and from others who were not involved in the research to look at that. Theo Palmer, who is not one of the article’s authors, was a really critical colleague at the time. He was pointing things out and saying: “I trust this one, not that one.” We worked until we got a number of samples sufficient for us to trust.
Most of the work consisted in developing tools to be able to see the reproduction of cells within the brain.
We knew that we had to persuade extremely skeptical people. In the article, we presented three or four different cell marking techniques that we had used in our work. Nature and Science refused to review the article, but the editor of Nature Medicine at the time took a risk and said: “This is truly astonishing!” and sent the article to a bunch of people for evaluation, including our chief rival. To this day, we continue to worry that someone might say that it was an artifact generated by the technique used. But nobody has said this so far.
Does neurogenesis occur in two regions of the brain only?
Well, there are two regions in which it occurs regularly and is easy to detect. Some people have already stated that it also takes place in other parts. However, the data is unreliable. When one analyzes the spinal cord, one sees that there are stem cells everywhere, but they don’t produce new neurons. They are dormant. My impression is that these cells find themselves in this state because the environment is unsuitable for them to turn into neurons. Part of the challenge is to discover what to do for neural stem cells to generate new neurons.
Does neurogenesis take place with regularity in the hippocampus and the olfactory bulb?
Yes. There is one more region within the brain, called the subventricular zone. That is where the cells are splitting. Afterwards, they migrate over a long distance until they get to the olfactory bulb. It’s an interesting pathway as it’s a long way. As for the hippocampus, it lies two or three cell layers below and the migration is far faster and easier to study.
What triggers this migration?
It took a long time to persuade people that it happens. The next step was to ensure that these neurons matured, established connections, became active. Now, we’re trying to discover why this happens.
Salk Institute for Biological Studies Is the function of these new neurons already known?
One idea that has reappeared is that this region of the hippocampus is involved in discriminating different objects, in determining what differentiates them. One of the tests to identify this is to show two similar objects in sequence. First you show only one and then, the two together. These situations activate an area of the hippocampus called the dentate gyrus. One seeks the image of the object that was presented previously, which is then compared with the current image of both. One must have a degree of recollection of the first object to evaluate their similarity. It is the young cells that are activated again when one the new pattern is observed [two objects shown together]. The old cells are busy looking for differences between the previous pattern and the new one, because there was only one object to start with and now there are two. This is a difference, but what one is really trying to identify is what is different between the two objects. We are working together with neuropsychologists to create tests that try to show this empirically. First, we show volunteers an image with complex drawings, then, two similar images and finally, we ask them which is similar to the first. At the initial level this is easy, because the difference is major. However, we show about 50 pairs of images with patterns that become more and more similar amongst them. We’re trying to see up to what point people are able to tell them apart.
Could such knowledge be used to understand diseases that affect memory, such as Alzheimer’s?
We need to create tests for humans because in all these diseases the neurogenesis rate drops. And there only very generic clinical tests. The truth is that we don’t have to have a hippocampus to tell apart things that are fairly different from each other. We can use just the visual cortex for that. When we talk, we analyze a whole lot of information. We try to understand everything that is said, to translate back into the mother tongue and analyze the gestures in order to make sense out of what is being said, based on personal references. The cells that appear in neurogenesis seem to be particularly important for making these associations. There are three parts to this question. First, this separation of patterns. Second, when they are young, these cells help to form memories. Last, after they mature, they help us to distinguish between images with similar patterns. This is what our mathematical models indicate and what our experiments are showing. The other part is that we tested whether what the cells learn when they are young is stored in them. Later, when the same event is presented to that cell, it recalls it. This is very exciting and led us to conduct experiments with genetically modified animals in such as way that the neurons that remember the initial experiences are marked a different color from those that recall more recent experiments, which take place after a certain amount of time. We plan to map the pattern of gene expression of the cells that recall the event and of those that do not.
Is there any evidence of a difference in the gene expression of these cells?
That is what we are trying to find out. There are differences among neighboring neurons. The genetic material is different as well as the gene expression. Additionally, environmental changes insert information into the DNA. This might be an epigenetic phenomenon, which alters the functioning of the genes. This effect is being studied by a former post-doctoral student of mine [the Brazilian scientist Alysson Muotri, a professor at the University of California in San Diego]. This is at the frontier of knowledge.
Would it be possible to use the knowledge about neurogenesis to help to differentiate stem cells, fix lesions or treat diseases?
In depression, the hippocampus shrinks. The connection with such data is subtle. These aren’t my data, they are out there. For some reason, the brain of someone treated with antidepressants of the Prozac class [a selective inhibitor of serotonin recapture] shrinks less.
But is it questionable?
No. Data are data. But they don’t say much. Experiments that we conducted showed that Prozac induces cell division. However, it’s not enough to divide. The new cells have to turn into neurons. This takes four to six weeks. Psychiatrists saw this and went crazy, saying: “That is why it takes so long for the medication to work.” These drugs increase the availability of serotonin and stem cells placed on a glass plate with serotonin divide like crazy. That’s how it works, by means of the 5HT1A receptors, which induce proliferation. There are now clinical trials with drugs that act upon neurogenesis and that do not interact with 5HT1A. Still, it’s all very incipient. In almost all the neurodegenerative diseases, a reduction of neurogenesis occurs.
Besides cell death, there is lower neurogenesis.
The doubt is whether the disease is causing the neurogenesis reduction directly or, as we believe, whether the disease makes the animal lethargic and this leads to a drop in neurogenesis. When one restricts the movements of an animal, it becomes stressed and neurogenesis falls. It might be that the drop in the pace of neurogenesis is a consequence of the behavioral change caused by the disease. But this is being tested. We know that one of the problems of patients with Parkinson’s or Alzheimer’s is their difficulty in knowing where they are. They get lost all the time. Perhaps one might not be able to halt the disease, but treat some components of it, enabling people to deal with it better. That is my optimistic goal.
What is the rate of neurogenesis?
It depends on age and species. It is far higher when one is young. And then it drops off. It’s very low among the aged.
So doesn’t it ever stop completely?
Neurogenesis is affected by the degree of activity. An old mouse, aged about 18 months, has virtually no neurogenesis. However, if it is given access for one month to a cage with a wheel [on which it can exercise voluntarily], or to a richer environment, it will have the same number of dividing neurons of a young but inactive mouse. Few of the old animals develop new functional neurons, but many do manage this and the cells that turn into neurons present all the ramifications.
So anybody that wants to keep their memory good should run or, at least, walk.
No, but I think it’s necessary to stay active physically or intellectually. The relation between running and neurogenesis is still unclear. In part, this effect is caused by serotonin. When one moves, the raphe nucleus is activated and secretes serotonin in the hippocampus. However, cardiovascular activity also increases when one is moving and a protein called IGF1 [insulin growth factor 1] is released into the bloodstream. At the spot where neurogenesis takes place there are blood vessels that might be releasing IGF1. Other people have shown that when one blocks IGF1 activity, one blocks the proliferation of neurons. It is a multifactor phenomenon. We don’t know much about it yet.
Salk Institute for Biological Studies
Before I got involved with this subject, there was already some evidence indicating there had to be cells dividing in the adult brain after the development stage, but a lot of people didn’t believe this. We got the clone of the fibroblasts growth factor, inserted it into these cells and placed them in a culture with brain cells. The neurons proliferated like crazy.