Research Highlights

Brain Synapse Size Grows Larger, Then Smaller in Regular Cycle Between Wakefulness and Sleep


ASI size declines in sleep according to a scaling relationship.  (A) (Left) Visualization of ASIs in one dendrite. Scale bar, 2.5 μm. (Right) Effect of condition. ASI size decreases in sleep (blue) relative to both spontaneous wake (orange) and enforced wake (red). ASI size is shown for all synapses, each represented by one dot. **P < 0.01. (B) Log-normal distribution of ASI sizes in the three experimental groups. (Inset) Same on a log scale. (C) The decrease in ASI size during sleep is due to scaling. (D) Monte Carlo simulations comparing different models of scaling. Size-dependent selective scaling (green) fits the actual data better than uniform scaling (asterisk) or selective scaling independent of size (brown) (supplementary materials, materials and methods).

03-02-2017 UW-Madison

Imagine that your brain, like your lungs, breathes in and out, breathing in during periods of wakefulness, then out during periods of sleep. This analogy holds if you imagine the brain and lungs getting bigger on the inbreath, then smaller on the outbreath. It’s this mechanism that a team led by drs. Chiara Cirelli and Giulio Tononi at UW-Madison, including researchers at NCMIR, studied and published recently in Science magazine.

Depending on the species, the cerebral cortex is populated by millions to billions of neurons: 16 billion in humans, 14 million in mice. Each neuron, in turn, contains thousands of synapses.

When an electrical signal arrives at the synapse, it can stimulate the release of neurotransmitters. These neurotransmitters are released into the synaptic cleft to bind onto the receptors of the postsynaptic membrane to influence another cell, either to inhibit or excite it. When a synapse is repeatedly excited, it can grow in strength and size, a mechanism called synaptic potentiation.

Of the cortical synapses in adult mice, ~80% are excitatory, and most are located on dendritic spines. The size of the spine strongly correlates with synaptic strength. Importantly, changes in strength mediate learning and memory retention. But at the same time, according to the synaptic homeostasis hypothesis (SHY) of sleep proposed by the team at UW-Madison, this process of synaptic potentiation needs to be balanced to avoid saturating the synapse or obliterating neural signaling and memories recently set down.

According to SHY, during periods of wakefulness, learning occurs primarily through synaptic potentiation, leading to an increase in synaptic strength. Synaptic weakening instead takes place during sleep, when we and other animals, in effect, are disconnected from the environment. This allows spontaneous neural activity to sample memories in a comprehensive and fair manner when the brain is “offline,” unstimulated by inputs from the environment. Sleep, thus, can consolidate and integrate new information (memory), while at the same time restoring cellular function.

Because stronger synapses are larger and weaker ones smaller, SHY strongly predicts that the cortical excitatory synapses should increase in size after wakefulness and decrease after sleep, independently of circadian time. Furthermore, while this process of synaptic renormalization (increasing and decreasing in strength) should affect the majority of synapses, it should also be selective to support both brain stability and plasticity.

In this context, researchers from UW-Madison, Università Politecnica delle Marche, (Ancona, Italy), and UCSD/NCMIR used serial block-face scanning electron microscopy (SBEM) to obtain high-resolution 3D volume measurements of synaptic size in two regions of mouse cortex across thousands of synapses during the wake/sleep cycle. Brains were collected from three groups of mice: S (for “sleep”) mice spent at least 75% of the first seven hours of the light period asleep (normal behavior as mice are nocturnal), EW (enforced wake) mice were kept awake during that time by exposure to novel objects, and SW (spontaneous wake) mice spent at least 70% of the first seven hours of the dark period spontaneously awake. S mice were compared to both SW and EW mice to distinguish sleep/wake effects from potential confounding factors: time of day, light exposure, and stimulation/stress associated with enforced wake.

The researchers focused on the axon-spine interface (ASI) as a structural measure of synaptic strength because, in SBEM images, its borders are relatively easy to demarcate. First they investigated whether ASI size changes as a function of wake vs. sleep. Just as SHY predicted, they found that ASI size decreases with sleep on average by ~18%, independently of time of day. Moreover, this decrease in ASI appeared to adhere to a scaling relationship.

But, interestingly, the researchers found that the change in ASI sizes between wake and sleep does not scale uniformly across all synapses. Rather, it’s selective: A fraction of all synapses scales, but the remaining fraction does not. In fact, they found that the majority of all synapses (>80%) scaled but a minority (<20%) was less likely to do so.

The researchers wondered whether distinguishing between small-medium synapses (the smallest 80%) vs. large synapses would predict scaling vs. not scaling. They found that ASI size scales down after sleep in small- and medium-sized synapses (~80% of the total population) but is less likely to do so in large synapses (~20%). In addition, synapses with vesicles and endosomes inside the spine head (~80%) are more likely to scale down after sleep compared to synapses in spines without these organelles (~20%).

The non-uniform scaling of synaptic size is consistent with the requirement that learning during wakefulness activates synapses specifically and with the hypothesis that selective renormalization during sleep favors memory consolidation and “smart” forgetting. Wakefulness may lead to selective upscaling of a smaller proportion of synapses because learning is limited to a particular portion of the brain. Downscaling during sleep, however, may be broader to allow the brain to sample its memories comprehensively.

Because the volume of the spine head (HV) also strongly correlates with synaptic strength, the researchers investigated changes in HV as a function of wake and sleep. The results were consistent with those in the ASI studies.

Luisa de Vivo and Michele Bellesi (the first two authors of the published work; see citation below) performed all experiments, collected most of the data, and coordinated the work of seven people who, blind to experimental conditions, segmented all the spines and ASIs. Says de Vivo, “It took more than three years to collect and analyze the data, which we believe represent the biggest dataset of 3D reconstructed dendrites and synapses from mouse neocortex. We plan to expand our analysis to other brain regions and specific paradigms of learning.”

This work supports SHY’s main hypothesis that wakefulness leads to a net increase in synaptic strength, while sleep renormalizes synaptic strength through a net decrease. The researchers are now investigating whether synaptic scaling across the wake/sleep cycle might be a general phenomenon, irrespective of species, brain region, and specific brain plasticity mechanisms.

Funding Source: This work was supported by NIH grants DP 1OD579 (GT), 1R01MH091326 (GT), 1R01MH099231 (GT, CC), 1P01NS083514 (GT, CC), and P41GM103412 for support of NCMIR (MHE).

Relevant Publication: de Vivo, L., Bellesi, M., Marshall, W., Bushong, E. A., Ellisman, M. H., Tononi, G., Cirelli, C., (2017), "Ultrastructural evidence for synaptic scaling across the wake/sleep cycle", Science, 2017/02/06, 3556324: pg: 507-510, Feb 03, 0036-8075, (DOI: 10.1126/science.aah5982).