Dendritic spines-tiny protrusions that mediate synaptic contacts between neurons-are present in vast numbers in the mammalian brain (up to ten thousand per neuron). For this reason, spine growth and retraction could hold the key to the making and breaking of neuronal connections that are thought to be essential for learning and memory. So just how stable are these spines? Two reports in today's Nature provide quite different answers.

Both papers describe the use of transgenic mice to image, in vivo, spines expressing fluorescent proteins. In work directed by Karel Svoboda, Cold Spring Harbor Laboratory, Long Island, New York, first authors Joshua Trachtenberg and Brian Chen, together with colleagues at the University of Lausanne, Switzerland, and Washington University, St. Louis, used green fluorescent protein to track the appearance and disappearance of spines in the barrel cortex of young adult mice. Trachtenberg et al. found three types of spines-transient, semistable, and stable-which appeared to last about one day, two-three days, and longer than eight days, respectively. About 20 percent of spines were transient, while about 50 percent were stable. Even the latter, however, were found to turn over, with a projected half-life of about 120 days.

Trachtenberg et al. also demonstrate, using the electron microscope, that the waxing and waning of spines is accompanied by synapse formation and loss, suggesting that spine turnover is functionally significant. In support of this, the authors found that sensory perception could influence the turnover of spines. To investigate this possibility, Trachtenberg et al. first determined which neurons in the barrel cortex were stimulated by tweaking specific whiskers. They then plucked, chessboard fashion, selected whiskers from one side of the face only, and found that spine turnover increased by almost 50 percent in corresponding neurons. Dendrites associated with the intact whiskers showed no change in spine dynamics.

In a remarkably similar approach, researchers in Wen-Biao Gan's lab at New York University used transgenic mice expressing yellow fluorescent protein to reveal, surprisingly, a totally different dynamic in the primary visual cortex of one-month-old mice. Jaime Grutzendler et al. found that spines in this region of the brain are extremely stable. At most, six percent of spines appear to turn over within three days, and 27 percent over a month. At four months of age, spines were even more stable. In these animals, 96 percent of spines remained unchanged after a month; spine half-life was calculated to be just over one year.

Why these two groups should obtain such different results is unclear, write Ole Ottersen and Johannes Helm, University of Olso, in a companion news and views article. They suggest that the different cortical regions selected for study-barrel vs. visual cortex-may offer some explanation, and they conclude that these first successful attempts at imaging live spines should provide the necessary impetus for further experimentation.—Tom Fagan


  1. In this week’s issue of Nature, two different groups report a technique to observe neurons in the brains of living, intact animals. Both groups demonstrated the ability to image various neuronal structures with two-photon microscopy in the brains of transgenic mice that express fluorescent proteins in neurons. Particular attention was paid to the dynamic behavior of dendritic spines, small protrusions along dendrites that are the primary sites of post-synaptic contacts. Trachtenberg, et al. report that about 60 percent of spines in the barrel cortex of six- to 10-week-old mice are stable when observed over an eight-day period, with the remainder of spines being highly dynamic in regards to their formation and elimination. In their study, the number of stable spines dropped to about 50 percent when tracked over a duration of one month, with larger spines appearing to be the most stable. The authors conclude that the high degree of turnover in the number of dendritic spines may contribute to the adaptive remodeling of complex neural circuits. Grutzendler, et al. report that dendritic filopodia in the visual cortex of one month-old animals are highly dynamic structures, but dendritic spines are highly stable (99 percent of spines were stable over a three-day period). This apparent stability did not change appreciably over a one-to-two month period, and many spines persist during a significant portion of an animal’s life. The authors believe that the stability they see in dendritic spines could potentially provide a structural basis for the storage and maintenance of long-term memories.

    On the surface, the two manuscripts appear to present opposing findings on the stability of dendritic spines. However, the differences in both the age of the animals, the areas of the brain that were analyzed and the criteria of differentiating between a spine and a filapodium could explain some of the conflict. Furthermore, the different transgenic mice (2 independent lines) used by each group may not be equivalent in their ability to label all dendritic spines. In spite of these apparent discrepancies, both groups conclude that axons and dendrites are extremely stable structures in vivo, that dendritic spines are dynamic structures during development, and at least a subset of theses structures become stable during adulthood. These are both very interesting papers that demonstrate the power of these types of fluorescent mice in allowing one to image living neurons in vivo over extended periods of time. Application of these mice to disease biology will also likely yield important future results in regard to brain plasticity under different conditions.

Make a Comment

To make a comment you must login or register.


No Available References

Further Reading

No Available Further Reading

Primary Papers

  1. . Long-term in vivo imaging of experience-dependent synaptic plasticity in adult cortex. Nature. 2002 Dec 19-26;420(6917):788-94. PubMed.
  2. . Long-term dendritic spine stability in the adult cortex. Nature. 2002 Dec 19-26;420(6917):812-6. PubMed.
  3. . How hardwired is the brain?. Nature. 2002 Dec 19-26;420(6917):751-2. PubMed.