Ca1

. ■ ■ • •

FIGURE 4 Axonal sprouting after Schaffer collateral transection. Images of hippocampal slice cultures stained with an antibody against the growth associated protein, GAP-43, a marker for growing axons. A: Low-power charge-coupled device (CCD) image of the lesion site in a culture in which the Schaffer collateral pathway was transected 5 days earlier. Two higherpower confocal images from the labeled regions are shown below. Note the large number of immunoreactive fibers leaving the CA3 side of the lesion and growing toward area CA1. Numerous growth cones are visible at the tips of the growing axons. B: Confocal image of area CA3 in an unlesioned culture maintained in vitro for 14 days. Scale bars = 100 mm in upper panel of A, 25 mm in A1, A2, and B. These results illustrate the lack of growing axons in cultures at the time the lesion is made, and the strong upregula-tion of GAP-43 in newly sprouted axons originating in area CA3 and growing toward area CA1 after axonal injury.

receptor antagonist, bicuculline, by producing pronounced epileptiform discharges, whereas unlesioned cultures never did.

Dual intracellular recordings made 14 to 21 days after the lesion revealed that the probability that any two CA3 pyramidal cells were connected by a monosynaptic excitatory synapse was increased by 50% (compared with unlesioned sister cultures). In addition, complex polysynaptic unitary EPSPs were readily elicited in lesioned cultures in response to pairs of presynaptic action potentials. This hyperex-citability outlasted the presence of immunoreactive sprouting axons, suggesting that the newly generated axons remain after GAP-43 expression has returned to undetectable levels. We concluded that the connectivity of the CA3 pyramidal cell population had increased after the axonal injury as a result of the sprouting of new axonal collaterals. Excitation was able to spread more readily among the population of CA3 pyramidal cells along these numerous, newly sprouted connections.

Is this axonal sprouting a "culture artifact," or is it relevant to "real" brains in situ? There are two reasons to believe that hippocampal slice cultures do indeed provide a useful model of hyperexcitability after brain injury. First, the low level of GAP-43 expression at 14 days in vitro indicates that the normal developmental phase of axonal elongation and synaptogenesis is complete at the time we make the lesion; the induced axonal sprouting thus represents plasticity of a mature system. Second, there is electron microscopic evidence of a reinnervation of area CA1 following Schaffer collateral transection in vivo (Goldowitz et al., 1979). Although the source of the regenerated presynaptic terminals in this study was not identified, it is reasonable to suspect that the presynaptic terminals were formed by regenerated Schaffer collaterals of ipsilateral CA3 pyramidal cells.

We are currently taking advantage of the ability to control the extracellular environment of these cultures to explore the role of injury-induced neurotrophin secretion as trigger of axonal sprouting. We have shown that exogenous neuro-trophins acting at trkB neurotrophin receptors promote axonal sprouting by pyramidal cells (Schwyzer et al., 2002) and are now examining injury-induced axonal sprouting in slice cultures from transgenic mice with reduced levels of trkB and in injured wild-type cultures exposed to immunoadhesion molecules that bind to and inactivate endogenously secreted neurotrophins.

Postsynaptic Changes after Injury

Although much of the current focus is on changes in synaptic circuitry and function as the cause of epilepsy, changes in intrinsic postsynaptic excitability or receptor expression are attracting increased attention (Bernard et al., 2004; Chen et al., 2001; Shah et al., 2004). There are several examples of injury-induced hyperexcitability in which amplification of intrinsic excitability has been implicated. First, axotomy increases the excitability of several cell types through alterations in the number and properties of intrinsic voltage-dependent ion channels (e.g., Abdullah and Smith, 2001). Second, denervation supersensitivity to acetylcholine is observed in denervated muscle as the result of an increased expression of receptors at both extrajunctional (Axelsson and Thesleff, 1959) and junctional sites (Davis and Goodman, 1998). A similar phenomenon also occurs in neurons. Decreasing postsynaptic activity in cultured cortical neurons with tetrodotoxin (TTX) results in increased postsynaptic receptor expression (Rutherford et al., 1998).

Finally, intrinsic excitability can also be increased following periods of low levels of cell discharge. Desai and colleagues (1999a) demonstrated that inactivation of cortical cell cultures with TTX leads to a very selective down-regulation of voltage-dependent Na+ and K+ currents, perhaps including Ca2+-activated K+ current. It is noteworthy that the effects of TTX on the expression of both glutamate receptors and intrinsic conductances have been attributed to decreases in the activation of postsynaptic trkB receptors by endogenously released BDNF (Desai et al., 1996b; Rutherford et al., 1998). The possibility that injuries to brain regions susceptible to epilepsy induce similar changes in postsynaptic excitability via axonal damage, partial denervation, and decreased neuronal discharge has not yet been tested.

How is the sensitivity of postsynaptic cells to synapti-cally released glutamate affected by axonal transection? Is there "denervation suprasensitivity"? We have taken advantage of the excellent optical properties of slice cultures to use laser microphotolysis, in which a brief pulse of ultraviolet light is focused within the preparation to cause release of chemically "caged" neurotransmitters, to study post-synaptic changes in cells in injured tissue. These experiments have yielded two exciting observations. First, we found that distal CA1 cell dendrites in control hippocampal cultures display all-or-none spike-like responses, called plateau potentials (Wei et al., 2001). Distal, terminal dendritic branches of CA1 cells responded to weak photolysis of glutamate with passive subthreshold depolarizations. Stronger glutamate applications, however, triggered all-or-none Cd2+-sensitive responses. Calcium imaging revealed that the subthreshold responses generated only small local Ca2+ transients, whereas all-or-none responses resulted in Ca2+ signals throughout the stimulated dendritic segment. Intra-cellular 1,2-bis(o-amino-5-bromophenoxy) ethane-N,N. N', N'-tetraacetic acid (BAPTA), apamin, and high concentrations of TEA prolonged the responses twofold to fourfold, suggesting that Ca2+-activated K+ conductance, mediated by the small conductance (SK) channel subtype, is responsible for their termination (Cai et al., 2004). We concluded that terminal dendrites respond to strong stimuli with an active spike, mediated by voltage-gated Ca2+ channels, that is confined to a single dendritic compartment. Such binary behavior could enable distal dendritic segments to perform parallel nonlinear processing of synaptic inputs.

Second, we discovered that 7 days after Schaffer collateral transaction, CA1 cells display pathologically prolonged plateau potentials and an abnormal ability to produce action potentials in response to dendritic glutamate photolysis or synaptic stimulation. When we compared the effectiveness of photolysis of caged glutamate in the distal dendrites of CA1 cells in control unlesioned cultures with CA1 cells that had been deafferented 7 days earlier, we found no change in the amplitude or kinetics of the responses to the smallest

FIGURE 5 Dendritic hyperexcitability in denervated CA1 cells. Membrane potential changes elicited with microphotolysis of caged glutamate to single distal apical dendrites of control cells (left column) and cells that were denervated 7 days earlier as the result of a Schaffer collateral transaction (right column). The responses in the upper row were obtained in the presence of tetrodotoxin (TTX) to block fast-action potentials, whereas the responses in the bottom row were obtained in the absence of TTX. In each panel a series of responses was elicited using a series of photolysis steps of increasing duration (see Wei et al., 2001). Note that dendritic plateau potentials are greatly prolonged after denervation and that this facilitates the triggering of fast-action potentials.

FIGURE 5 Dendritic hyperexcitability in denervated CA1 cells. Membrane potential changes elicited with microphotolysis of caged glutamate to single distal apical dendrites of control cells (left column) and cells that were denervated 7 days earlier as the result of a Schaffer collateral transaction (right column). The responses in the upper row were obtained in the presence of tetrodotoxin (TTX) to block fast-action potentials, whereas the responses in the bottom row were obtained in the absence of TTX. In each panel a series of responses was elicited using a series of photolysis steps of increasing duration (see Wei et al., 2001). Note that dendritic plateau potentials are greatly prolonged after denervation and that this facilitates the triggering of fast-action potentials.

glutamate pulses, indicating that transmitter supersensitivity does not occur in this model. We did observe a striking prolongation of the responses in the deafferented cells, however. In deafferented cells, the duration of the plateau potentials was increased sevenfold compared with control cultures (Figure 5). We also observed a marked increase in the ability of dendritic excitation to trigger action potential discharge after Schaffer collateral transaction; whereas photolysis of caged glutamate on terminal dendrites never elicited action potential discharge in control cells, the prolonged plateau potentials always elicited prolonged trains of action potentials in denervated cells. We also asked whether this prolongation of plateau potentials might contribute to the hyperexcitability seen with synaptic stimulation after Schaffer collateral transection. Indeed, unusual evoked synaptic responses were recorded from CA1 cells in the vicinity of a lesion made 7 to 21 days earlier. These data suggest that increased postsynaptic excitation may contribute to lesion-induced hyperexcitability.

Was this article helpful?

0 0

Post a comment