An exciting decade of two-photon fluorescence imaging has revealed that the shape and turnover of dendritic spines are dynamic and modulated by synaptic plasticity (Matsuzaki et al., 2004). Such structural changes can be induced via focal glutamate uncaging at individual spines, but they can also be driven by experience (Holtmaat et al., 2006). Indeed, the size of both the head and neck of dendritic spines can be modulated by neuronal activity (Matsuzaki et al., 2004; Bloodgood & Sabatini, 2005). In turn, the geometry of spines will affect the compartmentalization of signals generated at the spine head and the extent to which ions and second messengers can spread via the neck to the parent dendrite and nearby spines. Interestingly, Santamaria et al. (2006) demonstrated in an earlier study that diffusion in spiny dendrites cannot be described by the standard diffusion equation. Instead, spines temporarily ‘trap’ molecules causing anomalous diffusion in the parent dendrite (Fig. 1).
Uncaging dextran-conjugated fluorescent probes and imaging the fluorescence changes over time along a stretch of dendrite enabled calculation of the spatial variance of the diffusing molecules and estimation of the anomalous exponent dw. In smooth dendrites diffusion had an anomalous exponent of dw = 2, whereas in spiny dendrites dw turned out to be high (dw > 2) resulting in slowed diffusion. The importance of this finding is that the diffusion of second messengers within dendrites could be controlled by changes in spine shape or density. It is known that spine density in CA1 pyramidal neurons varies considerably across their dendritic arbors, with a progressive increase in spine density and spine size in the distal segments of apical dendrites (Konur et al., 2003). Thus, different dendritic compartments may exhibit different rates of diffusion of postsynaptic signaling molecules.
In their study published in this issue of EJN, Santamaria et al. (2011) show that, in both Purkinje cells and CA1 pyramidal neurons, the anomalous exponent dw scales linearly with spine density. Anomalous diffusion in dendrites appears to be a hallmark of all neurons with spines. In addition, these authors provide empirical evidence that neither branching geometry nor dendrite diameter account for anomalous diffusion, confirming their previous theoretical models (Santamaria et al., 2006).
The mechanistic studies by Santamaria et al. (2011) might be relevant to explain, for example, the functional impact of estradiol on memory formation (Andreescu et al., 2007) or of the pathological impact of a lack of fragile X mental retardation protein on motor behavior (Koekkoek et al., 2005). In both cases, altered geometry of spines and/or their inputs are likely to affect the diffusion of second messengers in the dendritic network. Therefore, it will be of general interest to identify the molecules for which anomalous diffusion plays a relevant role under physiological and pathophysiological conditions. For instance, active extrusion of calcium is likely to supersede the effects of spine geometry on diffusion of this ion in dendrites, whereas diffusion of inositol triphosphate (IP3) is slowed down (Santamaria et al., 2006) and that of GTPase signaling proteins varies among their family (Yasuda & Murakoshi, 2011). Anomalous diffusion imposed by spine geometry may thus represent a mechanism modulating the function of a selected set of signaling molecules.