Synaptic activity in X-linked mental retardation: a thorny issue

Excitatory synaptic contacts onto glutamatergic neurons are in large part localized on dendritic spines, which are specialized post-synaptic structures that allow functional segregation of individual inputs; consequently, spine density is used as an indirect measure of synapse number. Yet, the fundamental reason for spine existence is not completely clear, as inhibitory neurons can generate strict functional segregation of synaptic inputs even in aspiny dendrites (Goldberg et al. 2003). Several hypotheses have been put forward to explain the potential advantages of spine-located excitatory synaptic transmission, including allowing a more effective sampling of the space surrounding the dendrite and providing electrical and biochemical segregation that result in linear integration of inputs in distributed networks (Yuste, 2011). Independently of the detailed role, however, the idea of a direct correlation between spine number and synaptic transmission appears undisputable. In keeping with this idea, major neuropsychiatric disorders are normally accompanied by decreases in spine number and/or spine morphological alterations although it is not always clear whether abnormal spines lead to dysfunctional circuits or vice versa. In this issue of The Journal of Physiology, Powell et al. report the results of a study on oligophrenin-1-lacking mice, a rodent model of X-linked mental retardation that involves a protein that is expressed both pre- and postsynaptically.

The authors show that in hippocampal slices from these mice both the excitatory and inhibitory components of synaptic transmission are disrupted; additionally, lack of oligophrenin-1 is also associated with decreased spine density. Intriguingly, they found that a short (20 min) pharmacological treatment of the slices with inhibitors of the Rho signalling cascade restored the frequency of spontaneous inhibitory and excitatory currents but did not reverse the decrease in spine density. Because a previous report on these mice found inhibition of Rho kinase activity for 48 h to effectively rescue the decrease in spine length (Govek et al. 2004), a likely explanation for this finding is that the pharmacological treatment was too short to allow effective spine remodelling. Alternatively, it can be assumed that the restored synaptic activity was completely dependent on effects at the presynaptic sites. In any case, the mechanism(s) for the restored activity and the location of the synaptic contacts remain unresolved. Is it conceivable that in oligophrenin-1-lacking mice some of the synapses are maintained in a sort of ‘dormant’ mode where the physical connection is still present and only needs to be awakened? This could represent a case of extreme reduction in spine length so that, at the optical microscope level, the size reduction appears as a decreased spine density. This hypothesis is supported by the finding of Khelfaoui et al. (2007) who hypothesized that in the hippocampus of mice lacking oligophrenin-1 more synapses are located on dendritic shafts (rather than on spines) than in control animals, suggesting that loss of oligophrenin-1 somehow results into a reversal of the maturation process of dendritic spines and contacts.

This interpretation provides a basis for the prompt recovery in synaptic activity upon pharmacological treatment, as some form of postsynaptic structure would be present and ready to operate. Connectivity may be maintained at sites where spines disappear, possibly by contacts located on the dendritic shaft, while spines could still be present onto which the amount of synaptic transmission is dramatically reduced by presynaptic mechanisms.

Independently of the precise functional and morphological cascade of events, the observation by Powell and coworkers that the frequency of excitatory synaptic currents was restored in the absence of any detectable change in spine morphology seems to offer some insight for establishing the temporal order of the events leading to the re-establishment of the synaptic function because, at least in the case of this specific model, the presynaptic changes lead the way so that functional synaptic activity guides the change in postsynaptic morphology. In this framework, a conditional oligophrenin-1 knockout mouse would prove extremely interesting as it may shed light on the time course of spine regression compared to disappearance of synaptic activity.

On a more practical aspect, these observations are particularly intriguing as they suggest that even morphological changes that seem ‘hardwired’ because they depend on a genetic alteration can be reversed by relatively simple pharmacological treatments. Even more exciting is the observation that Rho-kinase inhibitors, while restoring functional and morphological properties in oligophrenin-lacking mice, did not show any detectable effect in wild-type controls, thus opening the door to potential clinical use of such compounds. Certainly, the idea of a ‘magic pill’ that could treat mental retardation might still seem quite far-fetched, but the proof of concept is now there. Indeed, being able to find such clear and relatively simple anomalies in animal models of disease, even when they reproduce only part of more complex syndromes, is a welcome outcome to envisaging a cure for early stages of mental retardation even at a time when we still don't know how the brain works in its entirety. Indeed several synaptic proteins, including oligophrenin-1, are expressed ubiquitously and thus wouldn't require drug targeting to specific brain sites. As such, the hope for some effective therapy may be closer than otherwise conceivable.