24 October 2011. There are two hotspots for neurogenesis in the adult human brain—the hippocampus and the olfactory bulb. These regions churn out new neurons constantly, but scientists aren't exactly sure why. "Consensus is growing that these cells play some kind of a role [in memory], but the exact role remains unknown," said Henriette van Praag of the National Institute on Aging (NIA) in Baltimore, Maryland. Two studies from the October 19 issue of the Journal of Neuroscience—one looking at the hippocampus and one at the olfactory bulb—tackle that question in unconventional ways. They both find that new neurons seem to be essential for the long-term retention and expression of newly formed memories.
Previous studies suggested that newborn neurons are integrated into memory circuits (see ARF related news story on Kee et al., 2007), are necessary for spatial memory (see ARF related news story on Clelland et al., 2009), or are responsible for time-stamping memories (see ARF related news story on Aimone et al., 2009). But in all of these studies, the scientists blocked neurogenesis—by genetic manipulation, chemicals, or radiation—before memories formed, then looked to see what happened a few weeks later. "A drawback of that approach is that there's huge potential for compensation by developmentally generated neurons in the hippocampus," said Paul Frankland, The Hospital for Sick Children, Toronto, Canada. Frankland is senior author on one of the papers.
Frankland and colleagues took a different tack to probe the role of neurogenesis in memory. They let memories form first, then ablated newly formed neurons to see if that altered the memory. The group specifically wiped out adult-generated neurons immediately after training mice on one of three hippocampus-dependent tasks. The team reported that killing those new neurons, most of which had reached a mature stage, degrades those recently formed memories. To specifically target these new neurons, first authors Maithe Arruda-Carvalho and Masanori Sakaguchi and colleagues created a mouse that expresses an inducible diphtheria toxin receptor—normally harmless to mouse cells—in adult-generated neurons. The researchers turned on the gene in four-week-old mice, so that receptors were expressed only in the neurons born at that time. Seven weeks later, the researchers trained mice on one of three different tasks—contextual fear conditioning, and spatial navigation in a water maze with or without visual cues. One day after training ended, the team administered diphtheria toxin, which induced cell apoptosis in receptor-bearing newborn neurons. For comparison, a second group of mice got diphtheria toxin treatment during the week before training to wipe out newborn cells before training began.
Mice exhibited memory loss only when researchers delivered the toxin after training. Animals given the toxin one week before training formed new memories as well as control mice—they froze only in the trained shocking chamber, found the platform in a water maze, and associated a visual cue with a water maze platform just as well as normal littermates. But those given diphtheria toxin one day after training lost adult-generated neurons and their memory was impaired. They froze in both trained and similar but untrained chambers, had more trouble finding the water maze platform, and were not sure which visual cue signaled a platform location. One group of mice got diphtheria toxin 35 days after water maze training and still showed memory deficits. In all of these cases, the memories were degraded but not erased completely. Mice knew to freeze or to look for a visual cue, but they couldn't distinguish between similar cues or contexts.
"If you remove new neurons from the memory trace, then subsequent expression of that memory is impaired," said Frankland. "Our study shows that if new neurons are born into the adult brain, that they mature sufficiently, they become incorporated into memory circuits and then they constitute an essential component of that memory circuit."
Anne Didier and a team from the University of Lyon, France, published a similar paper in the same issue of the Journal of Neuroscience. They explored the role of neurogenesis in the olfactory bulb in a new way, modulating memory rather than neurogenesis. Knowing that learning supports, in turn, the survival of newborn cells (Alonso et al., 2006), Didier and first authors Sebastien Sultan and Nolwen Rey asked if forgetting would compromise those cells' survival. For five days, the researchers trained mice to associate an odor with a cereal reward. On the sixth day, the team began training to break that memory trace by teaching the same mice to associate the cereal reward with a visual cue instead, with the scent randomly presented in the environment. They trained a control group of mice on both the odor and later the visual cues minus the randomly presented scent. On day 14 the control mice still remembered the odor association, but the experimental group had forgotten it, suggesting their odor memory trace was broken.
Sultan and colleagues then looked at the newborn cells in the olfactory bulbs after having injected bromodeoxyuridine (BrdU)—a marker of proliferating cells—13 days before behavioral testing. The researchers saw fewer BrdU-labeled cells in the animals that had forgotten the association than in the control group that remembered it.
"Newborn neurons that were saved during learning were killed by memory suppression," said Didier, who adds that these findings complement those reported by Frankland for the hippocampus. "It's a new way of showing that these newborn neurons support olfactory memory and are involved in networks that contain the memory trace."
To be sure that the dying neurons were the ones that had supported the odor association, the researchers injected the pan-caspase inhibitor benzyloxycarbonyl-Val-Ala-Asp (ZVAD) to halt apoptosis in a separate group of mice. These mice underwent the same training to make and erase the odor association, but they retained the odor-association. ZVAD-treated olfactory bulbs also kept more of their newborn neurons. Results suggest that apoptosis of newborn cells was responsible for the memory erasure in the absence of ZVAD.
"To my knowledge, this paper from Didier's lab is one of the first to provide evidence that these newborn neurons that have risen from stem cells are critical for olfactory memory," said Mark Mattson of the NIA. The field is still a long way away from being able to apply this information to disease prevention. Meanwhile, both papers lend support to the idea that "if you can increase the survival of stem cells and newly generated neurons, perhaps that would have an impact on decline in learning and memory with aging or maybe in Alzheimer's disease," Mattson speculated.
On a related note, researchers interpreting neurogenesis experiments might want to take heed of a third paper, also in the October 19 Journal of Neuroscience. Paskco Rakic and first author Alvaro Duque of the Yale University School of Medicine, New Haven, Connecticut, caution researchers to be aware of adverse effects that BrdU—which substitutes for the base thymidine in DNA replication—can have on newly generated cells. Unlike radioactively labeled thymidine, BrdU introduces an entirely foreign atom into the DNA strand that may have wide-ranging and long-lasting effects on DNA transcription and translation, the authors wrote. Duque and Rakic looked at newly generated cells in the cerebral cortex of developing macaque monkeys, whose mothers had been injected with the markers intravenously during gestation. Compared to cells labeled with 3H-thymidine, cells labeled with BrdU were fewer in number, and were more dispersed in the cortex. It may be that translation problems caused those neurons to die or become lost during migration.
BrdU has many practical advantages over the 3H-thymidine, Duque told ARF, but "people need to be aware of the limitations and take them into consideration when interpreting the results of their studies," he said.
The findings "should prompt some control studies where people determine whether BrdU is affecting whatever endpoint they're focusing on," said Mattson. But he agrees that BrdU still has advantages as a marker.
Meanwhile, Didier, who used the BrdU label in her study, doesn't believe it affected the outcome because control mice were also treated with BrdU. Nevertheless, she added that "it's an important paper because it sheds light on the limits of this BrdU method that sometimes we use without being really aware."—Gwyneth Zakaib.
Arruda-Carvalho M, Sakaguchi M, Akers KG, Josselyn SA, Frankland PW. Posttraining Ablation of Adult-Generated Neurons Degrades Previously Acquired Memories. J Neurosci. 2011 Oct 19. Abstract
Duque A, Rakic P. Different Effects of Bromodeoxyuridine and [3H]Thymidine Incorporation into DNA on Cell Proliferation, Position, and Fate. J Neurosci. 2011 Oct 19. Abstract
Sultan S, Rey N, Sacquet J, Mandairon N, Didier A. Newborn Neurons in the Olfactory Bulb Selected for Long-Term Survival through Olfactory Learning Are Prematurely Suppressed When the Olfactory Memory Is Erased. J Neurosci. 2011 Oct 19. Abstract