Healthcare and Medicine Reference
correct number of synaptic boutons, but these termi-
nals remain spread over a larger portion of the tono-
topic axis, compared to controls (Sanes and Takács,
1993). In the MSO, inhibitory terminals fail to be elim-
inated from the MSO dendrite (Figure 9.30). Interest-
ingly, raising gerbils in omnidirectional white noise to
reduce the binaural cues that MSO neurons respond to
also leads to significantly higher density of glycine
receptor puncta on the dendrites (Kapfer et al., 2002).
Finally, the segregation of inhibitory afferents into
stripes within the rat auditory midbrain is prevented
by functionally denervating the inhibitory projection
neurons (Gabriele et al., 2000). Thus, inhibitory arbors
do not go through their normal period of anatomical
refinement when they are deprived of synaptic input
from the cochlea, suggesting that inhibitory activity
also plays a role in the maintenance or stabilization of
inhibitory synaptic contacts.
Since the refinement of excitatory terminals has
been associated with their activity, a question that
arises is whether the physical elimination of inhibitory
synapses is associated with a weakening in the
strength of inhibitory transmission. For example, we
learned that inhibitory synapses elicit a depolarizing
postsynaptic response in neonatal animals (Chapter 8).
Therefore, it has been proposed that inhibitory termi-
nals may use an “excitatory” calcium-depend mecha-
nism, similar to glutamatergic terminals (Kullman
et al., 2002; Kim and Kandler, 2003).
The gain of inhibitory synapses can also display
activity-dependent adjustment (i.e., LTP and LTD),
similar to glutamatergic synapses (Gaiarsa et al., 2002).
Do MNTB synapses display an activity-dependent
form of long-term depression that could support
synapse elimination? In fact, low-frequency stimula-
tion of MNTB afferents produces a profound depres-
sion of the evoked inhibitory synaptic response.
Furthermore, this inhibitory LTD is age-dependent,
being prominent during the period of synapse elimi-
nation and declining during the third postnatal week
(Kotak et al., 2000).
Since IPSPs are no longer depolarizing at this time,
it seems likely that alternative mechanisms mediate
inhibitory LTD. One intriguing possibility for the sig-
naling pathway is suggested by the observation that
MNTB-evoked inhibition is primarily GABAergic at
first and gradually becomes glycinergic during the first
two postnatal weeks (Kotak et al., 1998). These results
have been confirmed anatomically using immunohis-
tochemical staining against GABA, glycine, the glycine
receptor anchoring protein (gephyrin), and a GABA A
receptor subunit (Kotak et al., 1998; Korada and
Schwartz, 1999). Individual terminals actually display
a developmental decrease in GABA content and a com-
mensurate increase in glycine (Nabekura et al., 2004).
Co-release of GABA and glycine has now been demon-
strated directly in other developing systems and may
be a general principle of inhibitory synapse maturation
(Jonas et al., 1998; O'Brien and Berger, 1999; Russier
et al., 2002; Keller et al., 2001: Dumoulin et al., 2001).
Whereas glycine receptors appear to be solely
ionotropic, the presence of GABAergic transmission
permits inhibitory terminals to communicate through
a metabotropic pathway. LTD in the LSO can be
blocked with a GABA B receptor antagonist (Kotak et
al., 2001). Furthermore, focal application of GABA to
the postsynaptic LSO neuron is sufficient to depress
the inhibitory response; in contrast, glycine application
has no effect (Chang et al., 2003). Thus, inhibitory
synapses undergo a period of refinement, much as
excitatory systems do, during which they attain a
precise pattern of innervation. A growing number of
studies suggest that spontaneous or sensory-evoked
inhibitory transmission can influence this process.
SYNAPTIC INFLUENCE ON
Synaptic activity plays an extremely important role
in regulating postsynaptic neuron morphology.
During early development, even if denervation does
not result in cell death (see Chapter 7), then it certainly
leads to shrinkage of cell body size, atrophy of den-
dritic processes, or loss of dendritic spines in most
areas of the central nervous system (Globus and
Scheibel, 1966; Valverde, 1968; Rakic, 1972; Harris and
Woolsey, 1981; Vaughn et al., 1988). Furthermore,
changes in the amount of sensory experience given to
an animal during development (which presumably
affects neural activity) also lead to measurable alter-
ations in nerve cell morphology. Thus, young rats that
are reared in an enriched, social environment have
more dendritic branching than rats reared alone in an
impoverished environment (Fiala et al., 1978). More
precise manipulations of the sensory environment,
such as sound attenuation or vertical stripe rearing,
have also been associated with specific changes in
auditory or visual neuron morphology, respectively.
Although it is common to perform a manipulation
and then wait days or weeks to look for a change in the
central nervous system, the effects of denervation occur
at a surprising rate. In the chick nucleus laminaris (NL), it
is possible to denervate the ventral dendrites while
leaving the dorsal dendrites untouched (Figure 9.31).
The entire NL dendritic arborization is then visualized
with a Golgi stain, beginning one hour after the affer-