Young neurons led astray

It is now well established that the adult mammalian brain – including that of humans – contains at least two discrete populations of neural stem cells which continue to generate new nerve cells throughout life. These newborn neurons are quickly integrated into existing circuits and are essential for proper functioning of the brain.

A new study published in the open access journal PLoS Biology shows that inhibiting a protein called cdk5 impairs the migration of newly generated neurons into the hippocampus, and causes them to form inappropriate connections with pre-existing cells. The findings shed some light on the molecular mechanisms by which new neurons are generated and have implications for cell replacement therapies for the treatment of neurological diseases.

In the adult brain, the generation of new nerve cells (or neurogenesis) takes place in two discrete proliferative niches. One of these is the subgranular zone (SGZ), which is  located deep within the hippocampus. The SGZ contains neural stem cells which divide asymmetrically, to generate more stem cells, as well new neurons, which migrate out to the dentate gyrus of the hippocampus, then differentiate into granule cells which become incorporated into the neuronal circuitry as they mature. For many years after their discovery, the function of these newborn neurons was unknown, but it is now known that they are essential for the memory functions of the hippocampus.

The new study was led by Fred Gage of the Laboratory of Genetics at the Salk Institute in La Jolla, who 10 years ago was the first to show that the adult human brain contains stem cells. Since then, Gage and his colleagues have been investigating the possible functions of these newborn cells, and in the past few years have analysed the human genome to identify genes associated with adult hippocampal neurogenesis. This yielded a set of almost 200 genes with patterns of activity that vary according to the rate at which neural stem cells divide, the numbers of neurons and non-neuronal cells (astrocytes) generated from them and the numbers of newborn cells which survive.

One of the genes identified in this analysis is cyclin-dependent kinase 5 (cdk5). Located on chromosome 5, this encodes an enzyme which is known to be essential for a wide variety of processes, including the migration of newborn neurons, extension and pathfinding of axons and dendrites and nerve cell signalling and degeneration. In order to further investigate the role of cdk5 in neurogenesis, Gage and his colleagues devised a strategy with which they could reduce the activity of the protein selectively in newborn hippocampal neurons. This involved delivering into newborn hippocampal neurons a genetically engineered retrovirus containing the genes encoding green fluorescent protein and a normal or non-functional version of the cdk5 protein.


Newborn hippocampal neurons expressing a non-functional cdk5 protein have altered dendritic morphology, as indicated by white arrows, whereas cells expressing high levels of cdk5 appear normal. Scale bar = 10 micrometers. (From Jessberger et al, 2008.) 

This strategy therefore enabled Gage’s group to not only increase or block cdk5 activity in newborn neurons, but also to identify the cells which had taken up the retrovirus by the green fluorescence they emitted. Analysis of the animals’ brains revealed that cdk5 activity is not needed for neurogenesis. Increasing or inhibiting cdk5 activity had no observable effect on neurogenesis: the numbers of newly generated hippocampal neurons in both groups of animals was no different to that in the controls. Likwise, neuronal differentiation was not affected. The newborn neurons which overexpressed cdk5, as well as those deficient in the protein, regardless expressed other genes which are activated early on in the differentiation, soon after neurons have been generated from stem cells.

The researchers then looked at the structure of the newborn cells, and found that the dendrites of those deficient in cdk5 had grown in the wrong direction. Newborn cells from control animals, and cells overexpressing cdk5, had the same polarized morphology, with a single dendrite which projected up through the granule cell layer (GCL) before branching as it exits. By contrast, cdk5-deficient cells had no such polarity – they had multiple dendrites which extended along the lower edge of the GCL instead of up through it (above). Consequently, these cells formed stable synaptic connections with the wrong targets, which persisted for more than 1 year. Further investigation showed that the aberrant dendritic patterning was associated with a deficit in migration: neurons deificient in cdk5 did not migrate far enough into the GCL after leaving the proliferative zone where they were born.

Cdk5 therefore appears to be crucial only for later stages of the maturation of newborn neurons. The authors suggest that the improper connections formed by these cells could interfere with information processing in the hippocampus. However, they did not carry out any behavioural tests to explore the consequences of blocking cdk5 activity. This would be an interesting thing to do next, given the recent discovery that new neurons are needed for new memories. Nevertheless, the new findings could have important implications for the use of stem cells in cell replacement therapies for neurological diseases, as they suggest that cells would have to be transplanted accurately into specific locations in order to be effective.


Jessberger, S. et al (2008). Cdk5 Regulates Accurate Maturation of Newborn Granule Cells in the Adult Hippocampus. PLoS Biology 6 (11) DOI: 10.1371/journal.pbio.0060272.


  1. #1 Arikia
    November 13, 2008

    Mo, after reading this, I feel compelled to tell you that you have the most flawlessly consistent ability to describe scientific articles that I have quite possibly ever seen. You should maybe quit your day job and pursue big time science journalism. (Read: Your day job being UCL. Stay at ScienceBlogs forever!)

  2. #2 CopperKettle
    November 13, 2008

    Reelin, a protein known for its role in positioning of migrating neuroblasts, could also be interacting with cdk5.. There’s a whole chapter in the book on reelin about it and cdk5.

  3. #3 CopperKettle
    November 13, 2008

    Here’s the diagram.

  4. #4 Alethea
    November 14, 2008

    Hi, Mo!

    Funny enough, I came on your post after someone flagged me to an article about how mutations in the calcium/calmodulin-dependent serine protein kinase CASK cause a severe microcephaly syndrome, possibly because of its known interaction with a transcription factor, TBR1. Together they regulate the transcription of reelin.

    I had been involved in a paper demonstrating that perturbing a transcriptional regulatory element of EOMES (TBR2, another T-box containing transcription factor) also leads to microcephaly. I presume that the element is regulating the late neuroblast expression of this gene because we only saw discrete CNS expression in the first trimester.

    So it would be quite interesting to know how the T-box transcription factors and CASK fit into the puzzle with Cdk5, reelin, LIS1 etc. and regulate late pregnancy and early postnatal brain growth.

  5. #5 Neuroskeptic
    November 15, 2008

    Ahem. I’ve just put up a post about another neurogenesis-related paper, this one looking at the effects of antidepressants: Prozac Made My Cells Spiky.