Maintaining
the correct balance of inhibition and excitation is extremely
important for normal cortical function. Too little inhibition
can lead to epileptiform activity, whereas too much inhibition
can severely depress cortical responsiveness. This suggests
that the balance of inhibition and excitation in cortical
circuits should be tightly regulated. In visual cortex,
activity has been shown to affect expression of the inhibitory
neurotransmitter GABA in a manner consistent with a role
in balancing excitation and inhibition; blocking activity
in one eye leads to a down-regulation of GABA in the corresponding
ocular dominance columns. These data suggest that the
level of activity is acting trough some feedback signal
to locally adjust the strength of cortical inhibition,
although the mechanisms by which this occurs remains unclear.
Here we use a culture system to explore the role of activity
in the control of cortical inhibition. We have found that
blocking activity in culture leads to a reversible decrease
in the number of neurons immunopositive for GABA, and
that this decrease can be prevented by the coapplication
of brain-derived neurotrophic factor (BDNF). These data
suggest that activity levels can continuously adjust cortical
inhibition in a bi-directional manner through a BDNF-dependent
mechanism.
Primary visual
cortical cultures were prepared from postnatal (P4-P6)
Long-Evans rat pups. Cultures begin to show signs of synaptic
activity after 3-4 days in vitro, and over the next few
days develop spontaneous firing. After 7-13 days in vitro
cultures were fixed and processed for double-label indirect
immunofluorescence against GABA and the neuronal marker.
The percentage of neurons in each culture that were GABA-positive
was then calculated. Blockade of neuronal activity for
two days with TTX resulted in a decrease in the percentage
of GABA-positive neurons in visual cortical cultures,
to about 70% of control values. This reduction was statistically
significant (p<0.01, student's test). The total number
of neurons in these cultures was not reduced by incubation
with TTX, or by the other manipulations described below.
The neurotrophins,
including BDNF, are a class of factors that have been
shown to affect a diverse set of neuronal properties,
including survival, outgrowth, and synaptic strengths.
The expression of BDNF in hippocampal and cortical cultures
has been shown to be activity-dependent; high activity
levels lead to increased BDNF expression in striatal neurons.
These observations suggested to us that BDNF secretion
might be the signal linking changing activity levels to
the expression of GABA in visual cortical cultures. In
support of this hypothesis, we found that BDNF prevented
the TTX-induced reduction in the number of GABAergic neurons.
BDNF + TTX was significantly different from TTX alone
(p<0.01), and was not significantly different from
control. BDNF alone produced no significant increase in
the percentage of GABA positive neurons, suggesting that
with activity at control values endogenous BDNF levels
are saturating for this effect. This effect was specific
to BDNF; Nerve Growth factor failed to prevent the TTX-induced
reduction in GABA. Incubation with K252a, an inhibitor
of neurotrophin receptors, produced a decrease in the
percentage of GABAergic neurons comparable to that produced
by TTX. These data indicate that BDNF can prevent the
activity-dependent decrease in GABA-positive neurons in
cortical cultures, and suggests that activity-dependent
BDNF secretion may be the mechanism by which activity
regulates GABA levels.
The effects
of activity blockade on GABA expression was reversible.
When cultures were treated for 2 days with TTX, then washed
to remove TTX, the percentage of GABA positive neurons
partially revered to about 85% of control values. Concurrent
incubation with BDNF during the wash completely reversed
the decrease. These data have two interesting implications.
First, since in cultures from this age new neurons are
no longer being generated, activity must be reversibly
decreasing GABA expression by interneurons rather than
selectively decreasing interneuron survival. Second, these
data suggest that activity can continuously and bi-directionally
adjust the level of GABA expression in cortical cultures,
through the regulation of BDNF levels. Such a mechanism
may be crucial for allowing cortical circuits to remain
within the correct operating range despite developmental
or learning-related changes in synaptic strengths.
Another factor
that could contribute to the regulation of activity levels
is the strength of inhibitory or excitatory synapses.
The strength and number of synaptic inputs onto neurons
can change dramatically during development or learning,
thus altering the total amount of excitation received
by a neuron. How do neurons adjust their responsiveness
to avoid firing rates that are too high or too low? Here
we provide evidence that ongoing activity can globally
regulate the strength of excitatory synaptic connections
onto cortical pyramidal neurons.
Whole-cell
recordings were obtained from pyramidal neurons from cultures
of P5-6 rat area 17 after 7-9 DIV, using an Axopatch 1D,
in the presence of TTX and bicuculine. Miniature synaptic
currents (minis) were recorded that could be blocked by
the AMPA antagonist CNQX. Recordings were made from control
cultures or sister cultures treated for 48 hrs with TTX
to block all spikes. In each of 7 experiments the quantal
amplitudes from TTX treated cultures were larger than
control cultures (32.8+1.1 pA, respectively; statistically
significant difference, p,0.01, student's T test). Cumulative
amplitude histograms from the two populations showed that
the distribution was shifted to the right for the TTX-treated
population (TTX-treated statistically different from control,
Kolmagorow-Smirnof test, p,0.001). Treatment with bicuculine,
which blocks inhibitory synaptic inputs and thus increases
neuronal firing rates, produced a regulation of mini size
in the other direction. Two days of bicuculine treatment
reduced the average mini amplitude to 10.1 + 0.9pA,
compared to control values of 16.5 + 1.1 pA. No
differences in resting potential, series resistences,
whole cell capacitance, or rise times of minis were found
for neurons maintained under these different conditions.
These data indicate that the level of neuronal activity
produces a long-lasting regulation in the quantal amplitude
of AMPA-mediated synaptic transmission.
