Correlated plasticity of synaptic structures and its relationship to the stabilization of synaptic enlargement

The ability to adapt to environmental changes, to learn and to
memorize information is one of the brain’s most extraordinary
features. One important process underlying this ability is
considered to be synaptic plasticity, i.e. the structural and
functional modification of synaptic connections. Synaptic
plasticity can occur either by genesis or elimination of synaptic
connections, or at existing connections by modifications in the
strength of synaptic transmission. Synaptic connections are complex
entities consisting of different functional structures: The
majority of hippocampal and cortical excitatory synapses are made
up of a postsynaptic compartment called dendritic spine and a
presynaptic compartment called bouton. Within the spine and the
bouton dense molecular structures, which serve the synaptic
transmission between pre‐ and postsynapse, exist, namely the
postsynaptic density (PSD) in the spine, and the active zone (AZ)
in the bouton. All these structures are correlated in size and with
synaptic strength. The function of this correlation serves the
efficient and fast transmission of neuronal signals. During
synaptic plasticity, a coordinated change in the size of all
synaptic structures is expected, for the maintenance of their
correlation. However, to date, such coordinated modifications have
not been examined in detail. Furthermore, the mechanisms underlying
the maintenance of structural and functional changes after synaptic
plasticity remain poorly understood. The aim of this thesis was to
explore these questions. To achieve this I carried out two
complementing experimental approaches: In a first set of
experiments, I studied changes in spine and PSD size by twophoton
time‐lapse imaging to explore correlated modifications in these two
synaptic structures. To induce structural spine plasticity I
stimulated single dendritic spines of Schaffer collateral synapses
in cultured hippocampal slices by two‐photon glutamate uncaging.
This was shown previously to be accompanied by an increase in spine
size and synaptic strength. To visualize structural plasticity of
spines and their PSD, the cytosolic marker tdTomato and EGFP‐tagged
structural proteins of the PSD, namely PSD‐95 and Homer1c, were
co‐expressed. PSD‐95 and Homer1c are important and abundant
scaffolding proteins of the PSD, which have been used previously as
markers for PSD size. I found that both PSD‐95 and Homer1c levels
increased after spine stimulation. Homer1c increased rather rapidly
whereas PSD‐95 did so in a delayed manner relative to the increase
in spine volume. Thus, the naïve correlation between PSD protein
level and spine volume was only transiently disrupted after
plasticity induction, but was reestablished over a time course of 3
hours. Furthermore, PSD‐95 level only increased significantly in
spines with persistent enlargement, but not in spines with
non‐persistent enlargement. On the other hand, Homer1c level
initially increased both in spines with and without persistent
enlargement, and then decayed back to original level in spines with
non‐persistent enlargement. Because the increase in PSD‐95 level
was delayed, I investigated whether the application of the PKA
activator forskolin, which supports an increased and persistent
enlargement of spines after glutamate uncaging, might promote and
therefore accelerate an increase in PSD‐95 level. However, these
experiments led to unexpected results: forskolin application
neither had an effect on spine volume nor on PSD‐95 level increase.
Although PSD‐95 and Homer1c are important and abundant PSD
scaffolding proteins, they represent only two out of a multitude of
proteins which form the PSD. Consequently, an increase in the PSD
marker proteins does not necessarily represent an increase of the
PSD as a whole. Therefore, in a second experimental approach, I
applied electron microscopy to stimulated spines which displayed a
stable enlargement over 3 hours after stimulation. Hereby, I was
able not only to reconstruct the spine and the entire PSD, but also
the bouton at the stimulated spine: I found that spine, PSD and
bouton displayed matching dimensions 3 hours after stimulation,
similar to naïve, unstimulated spines. In summary, by combining
two‐photon glutamate uncaging with time‐lapse imaging and electron
microscopy, I found that spine, the PSD and bouton increase during
structural plasticity, and that the correlation between these
structures is reestablished after stimulation on a time scale of 3
hours. Furthermore, an increase of synaptic structures correlates
with the stabilization of synaptic modifications after plasticity.
This suggests a model where the balancing of synaptic structures is
a hallmark for the stabilization of structural modifications during
synaptic plasticity.