New neurons are needed for new memories

Around 15 years ago, researchers discovered that the adult rodent brain contains discrete populations of stem cells which continue to divide and produce new neurons throughout life. This discovery was an important one, as it overturned a persistent dogma in neuroscience which held that the adult mammalian brain cannot regenerate.

Since then, neural stem cells have been the subject of intensive investigation, in large part because of their potential uses in treating neurological conditions such as Parkinson's Disease, stroke and epilepsy. Even so, the function of newly-generated neurons has so far eluded researchers. Although numerous studies have shown that newborn cells are incorporated into pre-existing neuronal circuits, their precise role remained unclear.

Using an elegant new strategy for engineering strains of mutant mice, Itaru Imayoshi and his colleagues now provide strong evidence that the continuous generation of new neurons is critical for brain function, and that the newly-generated cells perform two distinct roles which are critical for tissue maintenance in the olfactory system and for the formation of two distinct forms of memory. Their findings are reported in the October issue of Nature Neuroscience.

In mice, as in men, neurogenesis persists in adulthood and takes place in two different regions of the forebrain (although other pockets of neural stem cells may yet be discovered). One of these is the subventricular zone (SVZ), which lines the fluid-filled lateral ventricles. Newborn neurons generated in the SVZ migrate along a pathway called the rostral migratory stream to populate the olfactory bulb, a structure which transmits sensory information from the nose to the brain. The other neurogenic region is the subgranular zone (SGZ), which generates cells that migrate a shorter distance to the internal granular layer of the hippocampus, which is crucial for memory. In both cases, the division is asymmetrical - each stem cell divides to produce a neuron and another stem cell. In this way, new cells are generated but the stem cell population is maintained.

Iamyoshi and his colleagues devised a novel genetic method, based on Cre-mediated recombination, which allowed them to label neural stem cells for long periods of time. It also labelled the neurons generated from them, as they inherit the genetic material from the stem cells as they divide. The method involved engineering mutant mice in which the gene encoding Cre recombinase is under the control of a stretch of regulatory DNA which normally activates a gene called nestin in neural stem cells. These animals were then mated with several other mutant strains, each engineered to possess a "reporter" gene encoding a fluorescent protein. In this system, the activity of Cre recombinase is induced by a drug called tamoxifen. Thus, when the offspring of the mutant mice strains are treated with this compound, Cre causes reshuffling of the genes, so that the reporters are activated only in the neural stem cells.   

These experiments provided several insights into the timescales of the production and migration of the newborn cells and their ultimate fate. When the animals were treated with tamoxifen at 2 months of age, fluoresently labelled immature neurons were observed in the SVZ, rostral migratory stream and olfactory bulb. Tests for proteins synthesized at progressive stages of differentiation showed that mature neurons did not appear in the olfactory bulb until after about 1 month of being born.

The researchers then investigated what would happen in the olfactory bulb if they aborted the supply of new neurons from the SVZ. To do so, they used the same genetic approach to eliminate the newborn cells. In this case though, the first strain of mutants was crossed with another whose cells contained the gene encoding diptheria toxin under the control of regulatory DNA which normally switches on a gene during the early stages of neuronal differentiation. When tamoxifen is administered in the offspring produced from this mating, Cre activation causes the toxin to be synthesized as soon as differentiation begins, and the newborn cells are killed.

These experiments successfully eliminated nearly all newborn neurons. Negligible numbers were, however, still observed, because the Cre-mediated DNA reshuffling did not occur in a small proportion of the stem cells. But these neurons were not found in the olfactory bulb olfactory bulb neurons, probably because they migrate together in chains. With most of their number absent, this chain migration does not occur, so the few new cells that were generated in the SVZ remained there. In the olfactory bulb, the old neurons died regardless, suggesting that continuous neurogenesis in the SVZ is required to maintain a constant number of olfactory bulb cells.       

These findings are in line with earlier studies which have shown that cells generated in the SVZ migrate into the olfactory bulb. But because the new method labelled cells for far longer than previous ones, this study provides more details. It shows that neurons in the olfactory bulb have a short lifetime and undergo continuous turnover. The vast majority of cells in the deeper regions of the bulb are replaced by new neurons over a 12-month period, whereas only about half of those in the superficial layers were replaced in that same period. In that time, migration rates were constant for the first 6 months, but then gradually decreased.

Surprisingly, this all had very little effect on the animals' sense of smell. The researchers presented the mice with various odours, but observed no differences between the behaviour of the mutants and that of mice treated with oil instead of tamoxifen. Both groups of animals were able to discriminate between different smells, and to learn an association between a specific smell and a sugar reward. Thus, while neurogenesis was necessary to maintain the structure of the olfactory bulb, it was not in this case required for olfactory function. This is at odds with earlier findings, and so will need to be confirmed further with more tests of olfactory behaviour.

What happened in the hippocampus was altogether different. Here, newborn cells were found to comprise just 10% of the total in the dentate gyrus at 6 months, after which there was no increase in their numbers. Prevention of neurogenesis did not significantly alter the size or structure of the hippocampus - the number of cells in the dentate remained constant for up to 6 months. (By contrast, the hippocampi of the control animals increased in size.) But it did have dramatic effects on spatial memory and associative learning - the mutant mice were unable to learn their way around a maze, and did not exhibit fear behaviour when they heard a sound which indicated that they were about to receive an electric shock.

This study shows that inhibiting neurogenesis has strikingly different consequences in two distinct regions of the brain. In the olfactory bulb, it leads to significant shrinkage but apparently does not alter smell-related behaviour. In the hippocampus, the effect on structure is not so marked, but it is clear that newly-generated neurons are necessary for the processes of learning and memory. Exactly how the new cells contribute to memory formation is still unknown.


Imayoshi, I. et al (2008). Roles of continuous neurogenesis in the structural and functional integrity of the adult forebrain. Nat. Neurosci. 11: 1153-1161 DOI: 10.1038/nn.2185


More like this

This post links so neatly with your previous one on exercise, radiation therapy and neurogenesis. Anything hippocampus related is bound to be interesting, even in mice and rats.
If without new cells we cannot lay down new memories, what happens when there is damage to only one hemisphere? If one side is happily neurogenerating stem cells and neurons but the other isn't, could that point to the experience of some form of deja vu state and the concurrent awareness of it? On the one hand there is a vague feeling of a repeated experience, on the other a certainty that today is a new day and despite similarities, we have not been here before.
Someone with damage to both hemispheres would seem to be a more likely victim of amnesia.