Until the late 20th century, scientists believed the adult brain was hardwired – that is, after a critical period in early childhood, the brain would stop changing and become static. We now know that adult brains are actually quite plastic, capable of generating new neurons and remodeling after experience or injury. In fact, connections between neurons, called synapses, are constantly being assembled and removed, a process thought to be essential for our brain’s plasticity.
In addition to being structural connections between neurons, synapses are also functional connections through which neurons communicate to each other. This communication underlies information processing in the brain. For example, input to a neuron through an excitatory synapse makes the neuron more likely to send an output signal to another neuron. Conversely, input to a neuron through an inhibitory synapse makes the neuron less likely to produce an output. Changes in synapses are thought to underlie processes such as learning and memory. However, we still don’t understand fully how the dynamics of excitatory and inhibitory synapses relate to each other and how quickly these synapses change in a living brain.
To address this question, a recent study by Villa et al., from Elly Nedivi’s lab at the Picower Institute for Learning and Memory at MIT, simultaneously visualized inhibitory and excitatory synapses in a living adult mouse brain. Villa et al. focus on a newly discovered class of inhibitory synapses that coexist with excitatory synapses on neuronal structures called dendritic spines. To be able to visualize individual synapses, Villa et al. tagged an excitatory synapse protein, PSD-95, and an inhibitory synapse protein, gephyrin, with fluorescent proteins, and then used a microscopy technique called two-photon microscopy to visualize these proteins. While conventional fluorescence microscopy uses a single photon of high energy light to excite fluorescent proteins, two-photon microscopy instead uses two photons of low energy light (e.g. infrared light), which helps decrease background signal and increases the resolution of the image. Villa et al. used this technique to image synapses once every day for nine consecutive days.
First, Villa et al. observed three structural classes of dendritic spines: spines without excitatory synapses, spines with just an excitatory synapse (termed SiS), and spines with both excitatory and inhibitory synapses (DiS). Excitatory synapses on spines were very stable, with nearly 100% of excitatory synapses on DiS remaining visible for the entire experiment. In stark contrast, inhibitory synapses were much more dynamic, with ~64% of inhibitory synapses on DiS appearing, disappearing, or reappearing at the same location on the spine over the course of the experiment.
To test how sensory experience could change synapse dynamics, Villa et al. then performed a monocular deprivation (MD) experiment, followed again by two-photon microscopy. The researchers permanently closed one eye of adult mice to deprive that eye of visual input, a procedure which is known to induce plasticity in the part of the brain’s visual cortex that receives input from both eyes. After MD, Villa et al. found that while excitatory synapse dynamics remained the same, inhibitory synapses became more dynamic. Specifically, inhibitory synapses disappeared and reappeared more often, and overall remained present for shorter amounts of time and absent for longer amounts of time. In other words, MD caused inhibitory synapses to enter a more dynamic state, with the net effect of decreasing inhibition. This decrease in inhibition could be the brain’s way of regulating the amount of input a neuron is receiving—after losing information from one eye, a neuron can indirectly compensate for the loss of excitatory input by reducing inhibitory input. This mechanism allows the neuron to maintain a balance of excitatory and inhibitory inputs, much in the way that humans can maintain a constant body temperature through homeostasis.
Overall, this study is the first time researchers have directly imaged excitatory and inhibitory synapses side-by-side, allowing the authors to compare the dynamics of excitatory and inhibitory synapses sharing the same dendritic spine. Villa et al. suggest that the repeated insertion and removal of spine inhibitory synapses at the same site could be a mechanism for reversibly modulating the input of neighboring, static excitatory synapses. Indeed, the observation that changing sensory experience via MD caused a shift to a more dynamic synapse state, instead of a one-time loss of inhibitory synapses, supports the idea that dynamic synapses are important for the brain’s plasticity in response to experience.
Because Villa et al. only visualized one protein as a proxy for an entire inhibitory synapse, their study does not address whether other components of synapses remain intact (or not) during inhibitory synapse removal – an interesting question for future research would be whether an inhibitory synapse uses “residual” synapse components to reinsert itself at the same location on the spine, and if so, which components are important in the reinsertion process? By observing the dynamics of excitatory and inhibitory synapses and how they change in response to sensory experience, Villa et al.’s study sheds light on how inhibitory synapses can regulate excitatory inputs to neurons, and provides a foundation for future studies into the mechanism and function of rapid and reversible inhibitory synapse dynamics.