HIPPOCAMPAL INTERNEURONS
AND THEIR ROLE IN CONTROLLING EXCITABILITY
D. Ieuan Evansm, Ph.D., Postdoctoral Fellow
J. Joshua Lawrence, Ph.D., Postdoctoral Fellow
Saobo Lei, Ph.D., Postdoctoral Fellow
Andre Fisahn, Ph.D., Guest Researcher
Emily Phillips-Tansey, M.S., Biologist
Xiaoqing Yuan, Biologist
Zachary Grinspan, HHMI Scholar
Harrison Walker, HHMI Scholar
Electrotonic Analysis of Strontium-Induced Asynchronous Events at Mossy Fiber–CA3 Interneuron Synapses
Lawrence
Network excitability in CA3 is determined in part by the capacity of mossy fibers to engage interneurons in feed-forward inhibition. We have previously used variance-mean (VM) analysis to estimate the probability of release, the number of release sites, and the quantal amplitude (q) for mossy fiber (MF) transmission onto CA3 interneurons. We have corroborated our estimate of q at MF synapses by evoking asynchronous events in the presence of 4-8 mM strontium. Rapidly rising EPSCs, presumably from proximal synapses, were evoked at low stimulus intensities. To distinguish strontium-induced asynchronous events from background spontaneous activity, we separated events with rise times similar to the synchronous component from slower-rising, presumably more distally occurring, events. Five to 20 milliseconds (ms) after the synchronous component, the frequency of these asynchronous events increased to about nine times the baseline frequency and then decayed exponentially (with a t of the order of 125 ms) to a rate not significantly different from that seen in the absence of strontium. In contrast, the frequency of the slower events remained unchanged at all latencies. We calculated the average q by collecting fast-rising asynchronous events at latencies of between 5 and 260 ms after the synchronous component, at which time the probability of contamination by fast-rising spontaneous events was less than 20 percent. The average q was about 22 pA, consistent with that obtained independently through VM analysis of single-fiber inputs (about 27 pA). Combining this level of analysis of mossy fiber synaptic transmission with compartmental models of CA3 interneurons, we are currently estimating the attenuation in quantal amplitude due to electrotonic location at MF synapses. Such analyses will provide powerful insights into the computational abilities of interneurons within the CA3 hippocampal network.
Differential Expression of NMDA Receptors at Mossy Fiber–Stratum Lucidum Inter-neuron Synapses
Lei
MF-pyramidal neuron synapses differ from most central synapses in that they contain few NMDA receptors (NMDAR). The ratio of NMDAR- to AMPAR-mediated EPSCs at MF-principal neuron synapses is one-third of that found at CA3-associated-commissural-principal cell synapses. Moreover, one of the hallmarks of the MF-principal cell synapse is its apparent lack of NMDAR-dependent synaptic plasticity. While the nature and function of AMPARs at MF–interneuron synapses are well described, there has been no confirmation of the existence of NMDA receptors at either synapse type. Furthermore, the potential for differential distribution and properties of NMDARs at synapses containing Ca2+-permeable (CP) or Ca2+-impermeable (CI) AMPARs has not been explored.
We demonstrated that the extent of philanthotoxin (PhTx) inhibition of synaptic AMPA receptors varied significantly between synapses, indicating a heterogeneous level of GluR2 expression. The data suggested that MF–interneuron synapses consist of a continuum that ranged from CP-AMPAR to CI-AMPAR–containing synapses. The amplitudes of evoked NMDAR-mediated EPSCs were negatively correlated with the degree of PhTx block of the AMPA EPSCs. The NMDAR/AMPAR ratio at CP synapses was about one-fourth of that at CI synapses. In addition, the decay time constant of the NMDAR-mediated component was markedly lower at CP than CI synapses. The sensitivity to ifenprodil, a specific blocker of NR2B-containing NMDARs, was much higher, and the open probability, probed with the irreversible open channel blocker MK-801, was smaller at CP than at CI synapses. Taken together, these data likely account for the larger amplitude and more rapid NMDAR EPSCs at CI synapses. In contrast, estimates of the single-channel conductance using nonstationary noise analysis revealed no difference in the single channel conductances of NMDARs (about 50pS) at either CI or CP synapse.
The lack of GluR2 in AMPARs of CP synapses was correlated with the presence of NR2B subunits in NMDARs and vice versa. Within many CNS regions, NR2B and GluR2 expression is developmentally regulated, with high NR2B and low GluR2 predominating early in development. To address whether the absence of GluR2 reflected developmentally immature MF–interneuron synapses, we examined four age groups spanning stages at which MF–inter-neuron synapses appear (Postnatal day (P) 9–10), mature (P15–16), and then fully developed (P20–21 and P30–40). At P9–10, CP synapses were the predominant synapse type observed (61.3 percent) while CI synapses predominated at P30–40 (45.6 percent). The intermediate age group of P15–16 contained the highest percentage of synapses comprised of “mixed” AMPARs, suggesting that P15–16 may represent a period of transition from CP to CI synapses. Therefore, although both CI and CP AMPAR synapses were developmentally regulated, both synapse types were observed in considerable numbers at all developmental stages, suggesting that the two synapse types correspond to functionally distinct MF-interneuron synapses. Of particular interest, the subunit composition of AMPARs and NMDARs at mossy fiber synapses is not independent. One possible explanation is that the expression of the two subunits is coregulated within the framework of a common genetic program. Alternatively, selective protein-protein interactions may constrain the assembly and subunit composition of the two receptor types in the post-synaptic density.
Activation of Kinetically Distinct Synaptic Conductances on Inhibitory Interneurons by Electrotonically Overlapping Afferents
Lawrence, Walker
We have shown previously that single CA3 str. lucidum inhibitory neurons target AMPA receptors with molecular compositions distinct from those on synapses formed by specific afferent inputs. Thus, single interneurons receive input from MF of dentate gyrus granule cells onto either CP or CI AMPA receptors while recurrent collaterals of CA3 pyramidal neurons synapse exclusively onto CI AMPA receptors. While these afferent-specific populations of AMPA receptors are pharmacologically distinct, the spatial distribution of these synapses along the dendritic arbor is unknown. Traditional methods such as paired recordings have been problematic because of the exceedingly low probability of connected pairs in dual dentate granule cell–CA3 interneuron recordings; a given granule cell axon contacts only 40 to 50 inhibitory neurons in the CA3 hippocampus, separated over a considerable distance. Consequently, due to the uncertainty in electrotonic distances and distortion of somatically recorded synaptic currents attributable to dendritic filtering, it was not known whether the underlying afferent-specific synaptic conductances were associated with different time courses. Given that the time course of the synaptic conductance is a major determinant of how neurons integrate afferent input, an accurate estimate of this parameter is required to understand fully how either feed-forward or feedback excitation engages the post-synaptic inhibitory interneuron.
We circumvented the problem of dendritic filtering by adopting a somatic “voltage jump” technique that involves applying hyperpolarizing voltage jumps from the somatic recording electrode at various times after onset of an evoked synaptic current. If the receptors of interest are open when the voltage jump reaches the site of the synapse, the increase in the driving force draws additional charge through the channels. Conversely, if conductance has ended and the receptors are closed, no additional charge will be recovered. The time course of the recovered charge yields the “true” time course of the conductance occurring at the synapse. Moreover, because the time course of the change in driving force that reaches the synapse is related to the distance between soma and synapse, the voltage jump method also provides valuable information regarding the electrotonic distri-bution of synapses along the somatodendritic axis.
We found that two afferent-specific populations of AMPA receptors were associated with different excitatory synaptic conductance time courses but overlapped in electrotonic location. Mossy fiber synaptic conductances were on average nearly twice as rapid as recurrent collateral synapses (1.5 versus 2.7 msec). Interestingly, electrotonic information revealed that mossy fiber inputs tended to be clustered at more proximal locations to the somata while recurrent collateral inputs were diffusely distributed across the entire dendritic tree. Thus, afferent-specific conductance time courses allow single interneurons to integrate feed-forward and feedback information differentially without the need to segregate distinct AMPA receptor subunits to different electrotonic domains.
GABAB Receptor–Mediated Regulation of Excitatory and Inhibitory Synaptic Transmission
Lei
CA3 stratum radiatum inhibitory interneurons are a heterogeneous population that receive both excitatory synaptic innervation from the collateral recurrent fibers of CA3 pyramidal neurons and inhibitory synaptic transmission from other GABAergic interneurons. GABAB receptor–mediated post-synaptic IPSP(C)s have been described in interneurons throughout the hippocampus, including CA1 stratum pyramidale, CA1 lacunosum-moleculare, and dentate-hilus border. However, to our surprise, GABAB receptor–mediated modulation of presynaptic glutamate and GABA release at hippocampal interneuron synapses has not been elucidated. Since activation of presynaptic GABAB receptors may be an important step in modulation of short- (and long-) term synaptic plasticity, we investigated whether GABAB receptors were present on either excitatory or inhibitory synaptic terminals onto CA3 interneurons and asked whether their activation altered the dynamics of transmission at either of these synapses.
We demonstrated that activation of GABAB receptors by baclofen inhibited the amplitudes of both evoked AMPA receptormediated EPSCs and GABAA receptormediated IPSCs with IC50s of 8.5 and 1.7 mM, respectively. Baclofen enhanced the paired-pulse ratio and coefficient of variation of evoked EPSCs and IPSCs but reduced the frequencies of spontaneous and KCl-evoked miniature EPSCs and IPSCs, consistent with a presynaptic mechanism. Baclofen blocked the frequency-dependent depression of both EPSCs and IPSCs but was ineffective at blocking frequency-dependent facilitation of EPSCs. Whereas N- and P/Q-types of Ca2+ channels contributed equally to GABAB receptor–mediated inhibition of EPSCs, more P/Q-type Ca2+ channels were involved in GABAB receptor–mediated inhibition of IPSCs. Taken together, our data demonstrate that presynaptic GABAB receptors are expressed on the terminals of both excitatory and inhibitory synapses onto CA3 interneurons and that their activation modulates essential components of the release process to modulate transmission at these two synapse types.
The Role of the Kainate Receptor Subunits GluR5 and GluR6 in Hippocampal Gamma Oscillations
Fisahn; in collaboration with Heinneman, Buhl
Activation of kainate receptors (but not AMPA receptors) induces gamma oscillations in the hippocampus in vitro, oscillations that depend on intact inhibitory GABAergic neurotrans-mission but not on excitatory transmission via NMDA receptors, metabotropic glutamate receptors, or cholinergic receptors. We investigated the connection between gamma oscillations and the kainate receptor subtypes GluR5 and GluR6. Using extracellular field recordings, we demonstrated that gamma oscillations of comparable power were induced by much lower concentrations of kainate in GluR5-/- (100nM) CA3 hippocampus than in wild type (600nM). Increasing concentrations of kainate (>150nM) degraded the gamma oscillation and precipitated electrographic seizure activity in GluR5-/- hippocampus but not in wild type. In contrast, kainate failed to induce gamma oscillations or seizures at concentrations up to 600nM in GluR6-/- hippocampus. Given that gamma oscillations were readily evoked by other oscillogenic drugs such as muscarine (20µM), this lack of effect of kainate was not explained by altered neuronal circuitry in GluR6/- mice. We concluded that activation of the GluR6 kainate receptor subtype is necessary for the induction of gamma oscillations. Next, using whole cell recordings, we demonstrated that kainate increased sIPSC amplitude by about 50 percent in wild type and GluR5-/- but was without effect on sIPSCs in GluR6-/-. Further-more, kainate activated an inward current and depolarized pyramidal neurons in both wild type and GluR5-/- but not in GluR6-/-. These effects of kainate on the hippocampal network are consistent with the expression patterns of GluR5 and GluR6, which are primarily localized on inhibitory interneuron presynaptic terminals and somata and dendrites of interneurons and pyramidal cells, respectively. Consistent with a presynaptic expression pattern of GluR5, the paired-pulse ratio of evoked IPSCs was increased, suggesting that GluR5 may act to depolarize presynaptic inhibitory terminals tonically, providing an elevated basal state of inhibition within the hippocampal network. Similarly, the increased sensitivity to kainate-induced oscillations in the GluR5-/- likely results from a reduction of the inhibitory control over the CA3 network, leading to an enhanced excitability of pyramidal neurons and a reduced ability of the interneuron network to synchronize itself at gamma frequencies.
.
PUBLICATIONS
- Atzori M, Lei S, Evans DI, Kanold P, Phillips-Tansey E, McIntyre
O, McBain CJ. Differential synaptic processing separates stationary
from transient inputs to the auditory cortex. Nat Neurosci. 2001;4:1230-1237.
- Chittajallu R, Chen Y, Wang H, Yuan X, Ghiani CA, Heckman T, McBain
CJ, Gallo V. Regulation of Kv1 subunit expression in oligodendrocyte
progenitor cells and their role in G1/S phase progression of the cell
cycle. Proc Natl Acad Sci USA. 2002;99:2350-2355.
- Fisahn A, Yamada M, Duttaroy A, Gan JW, Deng CX, McBain CJ, Wess
J. Muscarinic induction of hippocampal gamma oscillations requires
coupling of the M1 receptor to two mixed cation currents. Neuron. 2002;33:615-624.
- Lei S, McBain CJ. Distinct NMDA receptors provide differential modes
of transmission at mossy fiber-interneuron synapses. Neuron. 2002;33:921-933.
- McBain CJ, Fisahn A. Interneurons unbound. Nat Rev Neurosci. 2001;2:11-23.
- McBain CJ, Rudy B. Kv3 Channels: Voltage-gated K+ channels designed
for high-frequency repetitive firing. Trends Neurosci. 2001;24:517-526.
- Phillips-Tansey E, McBain CJ. 2002 Developmental expression of potassium
channel subunit Kv3.2 within subpopulations of hippocampal inhibitory
interneuron. Hippocampus. 2002;12:337-348.
- Walker HC, Lawrence JJ, McBain CJ. Activation of kinetically distinct synaptic conductances on inhibitory interneurons by electrotonically overlapping afferents. Neuron. 2002;35:161–171.
COLLABORATORS
Eberhard Buhl, Ph.D., Leeds University, Leeds, UK
Stephen Heinnemann, Ph.D., Salk Institute, La Jolla, CA