Biomedical Engineering Reference
In-Depth Information
potential changes in a quantitative manner with high temporal and spatial resolution, we recently suc-
ceeded to use SHG to measure action potential invasion into dendritic spines [13].
For those experiments, we loaded neurons in acute brain slices with the SHG dye FM4-64 through
the patch-clamp pipette and illuminated neurons with 1064 nm femtosecond laser. FM4-64 diffused
into the neurons and filled the inner leaflet of plasma membrane from the site of patch (i.e., soma). Over
time, distal structures such as oblique dendrites and basal dendrites became visible by SHG imaging
and later, dendritic spines were observed both by SHG imaging and two-photon fluorescence. After
locating target spines, we illuminated the target spines continuously (point-scan) while inducing action
potential backpropagation by injecting positive current at the soma. The generation of action potential
at the soma was monitored by current clamp recording. When we compared SHG signals from the
target spines in the nonstimulated status and with action potential backpropagation, we observed SHG
signal changes that occurred simultaneously with the somatic action potential. In addition to this kind
of qualitative argument, SHG imaging allows us to draw quantitative conclusions. For instance, a 10%
change in relative SHG (ΔSHG/SHG) meant a 100 mV membrane potential in our hands regardless of
the location or morphology of the target. Therefore, we compared peak SHG signal changes obtained at
soma and dendritic spines to compare the amplitudes of action potentials at these locations. This analy-
sis revealed that when the action potential invades into dendritic spines, it invades with full magnitude,
without significant decay.
One might think that this observation of the lack of voltage attenuation at dendritic spines does not
match with the prediction from cable theory [26]. However, one thing that we should keep in mind is
that the dendritic spines we observed and therefore measured by SHG imaging are located relatively
close to the soma, mostly within a distance of 50 μm. Although we have little knowledge in voltage
dynamics in dendritic spines, previous studies have characterized backpropagation of action potential
into dendritic shafts. These studies show that there is indeed an attenuation of peak voltage as it goes
distant from the soma, but the decay is relatively mild with the length constant of 138 μm even at basal
dendrite, where the attenuation is more severe than in the most studied apical dendrite [27]. If the
attenuation is small enough, the difference in SHG signal change may come short of statistical differ-
ences. Another intriguing possibility is that the dendritic spines are endowed with active conductances
mediated by various voltage-gated ion channels and boost action potential inside. While the presence of
voltage-gated sodium channels in spines remain elusive, it is well known that spines have a rich variety
of functional voltage-gated calcium channels [28]. Furthermore, active roles of these voltage-gated ion
channels in dendritic spines have been suggested [29]. Therefore, it is of particular interest to study back-
propagation of action potentials into dendritic spines and parent dendrites in distal locations by SHG
imaging, in combination with pharmacological and genetic tools to dissect out molecular and cellular
mechanisms of propagation of action potential into dendritic spines.
10.5.3 SHG imaging of Axons
As another example of SHG measurements of voltage in neurons, we will discuss recent SHG imag-
ing of axons (Figure 10.4). Axons are considered to be designed to deliver the sole output of neurons,
action potentials to distal sites. In addition to the known roles of faithfully transmitting action potential
to distal sites, however, more complex roles of axons have recently been proposed [30,31]. In fact, in
some preparations, electrophysiological recordings were successfully obtained from axons of mamma-
lian neurons, which provide clues to such additional analog information processing in axons [32,33].
However, it still remains a challenging task to perform electrophysiological experiments at axons in an
intact preparation. Because of this, we utilized SHG to measure the action potential propagations into
axonal arbors [34].
For some yet unknown reasons, axons cannot be visualized by SHG in brain slices by intracellular
loading of SHG chromophores. Obviously, it is not difficult to imagine that passive diffusions of FM
chromophores are extremely slow in axons that have much smaller diameters compared to dendrites. To
