Chris J. McBain, PhD, Head, Laboratory of Cellular and Synaptic Neurophysiology
J. Joshua Lawrence, PhD, Staff Scientist
Tue Banke, PhD, Visiting Fellow
Kenneth Pelkey, PhD, Visiting Fellow
Christine Torborg, PhD, Visiting Fellow
Brian Jefferies, BS, Biologist
Xiaoqing Yuan, MSc, Biologist
Tsz-wan Michelle Ho, BS, Graduate Student
Katherine Travis, BS, Postbaccalaureate Fellow
Joseph Churchill, BS, HHMI Scholar

GABAergic inhibitory interneurons consist of a population of hippocampal cells whose high degree of anatomical and functional divergence make them suitable candidates for controlling the activity of large populations of principal neurons. GABAergic inhibitory interneurons play a major role in the synchronization of neuronal activity and are involved in the generation of large-scale network oscillations. Thus, many interneurons function as a clock that dictates when principal cells fire during suprathreshold excitatory drive. Interneurons receive strong excitatory glutamatergic innervation via numerous anatomically distinct afferent projections, and recent evidence has demonstrated that the molecular composition of both classes of AMPA-preferring glutamate receptors expressed at interneuron synapses are often distinct from those found at principal cell synapses. Furthermore, single inhibitory interneurons can synthesize distinct AMPA receptors with defined subunit composition and target them to synaptic domains innervated by different afferent inputs. Using high-resolution whole-cell patch-clamp recording techniques in brain slices of the rodent hippocampus, we investigate differential mechanisms of synaptic transmission onto hippocampal inhibitory interneurons and the role of intrinsic voltage-gated channels in regulating interneuron excitability.

Compartmentalized Ca2+ channel regulation at functionally divergent release sites of single mossy fibers underlies target cell-dependent plasticity

Pelkey; collaboration with Lacaille, Roche, Topolnik

Activity-dependent alterations in synaptic efficacy are considered to represent the cellular substrate underlying learning and memory formation and have been implicated in neurodegenerative disorders such as epilepsy and chronic pain. Despite rapid progress in elucidating cellular mechanisms responsible for long-term plasticity at excitatory synapses between principal cells, there is relatively little insight into the plasticity of excitatory transmission onto inhibitory interneurons. Indeed, bi-directional plasticity of excitatory drive onto any identified interneuron population has not previously been observed. Moreover, excitatory transmission typically displays target cell-specific regulation, indicating that the same rules governing plasticity at principal cell synapses cannot be applied to inputs onto interneurons. For instance, in contrast to MF-pyramid (PYR) synapses, where long-term potentiation (LTP) occurs, hippocampal mossy fiber (MF) inputs to CA3 stratum lucidum interneurons (SLINs) undergo long-term depression (LTD) following high-frequency stimulation (HFS). Furthermore, activity-induced potentiation of MF-SLIN transmission has not been previously observed. We have now shown that presynaptically located metabotropic glutamate receptor subtype 7 (mGluR7) is a metaplastic switch at MF-SLIN synapses and that the receptor's activation and surface expression govern the direction of plasticity. In naive slices, mGluR7 activation during high-frequency stimulation generates MF-SLIN LTD, depressing presynaptic release through a PKC-dependent mechanism. Moreover, following agonist exposure, mGluR7 undergoes internalization, unmasking the ability of MF-SLIN synapses to undergo presynaptic potentiation in response to the same HFS that induces LTD in naive slices. Thus, selective mGluR7 targeting to MF terminals contacting SLINs, but not PYRs, provides target cell-specific plasticity and bi-directional control of feedforward inhibition.

This differential distribution of plasticity indicates that divergent presynaptic elements of neighboring terminals along the same axon serve as autonomous computational elements capable of modifying release independently of each other. Kenneth Pelkey, in collaboration with Lisa Topolnik of Jean-Claude Lacaille's laboratory, examined whether differential Ca2+ channel regulation within MF terminals innervating principal cells (PCs) and SLINs contributes to disparate plasticity within these divergent MF inputs. Using a combination of electrophysiological recording with two-photon imaging of individual mossy fiber boutons (MFB)s and filopodia, we found striking compartmentalization of presynaptic Ca2+ channel regulation between MFB and filopodia extension (Fil) release sites in response to both chemically and activity-induced forms of long-term synaptic plasticity. Despite similar contributions of N- and P/Q-type VGCCs to transmission at each terminal type, we demonstrated that MF-SLIN LTD is ultimately expressed as a reduction in P/Q-type presynaptic Ca2+ channel function while expression of MF-PC long-term plasticity is independent of presynaptic Ca2+ channel alterations. Thus, specific mGluR7 localization to MF-SLIN presynaptic elements, but not to MFBs, allows for MF-SLIN LTD expression via PKC-dependent depression of presynaptic VGCC function, whereas MF-PYR plasticity proceeds independently of VGCC alterations.

Rapid internalization of mGluR7 in response to activation by the allosteric agonist AMN082

Pelkey, Yuan; in collaboration with Roche

Experiments described above highlighted a role for mGluR7 internalization as a bi-directional control for MF-interneuron plasticity. We carried out our previous experiments investigating mGluR7 trafficking in pyramidal cells overexpressing a myc-tagged mGluR7. While the experiments were pivotal in implicating receptor internalization in response to agonist binding, they were not, by their design, suitable for exploration of the temporal relationship of mGluR7 internalization and the receptor's subsequent surface re-expression. In addition, from a general point of view, a lack of specific pharmacological tools for probing mGluR7 function has hampered the unequivocal assignment of specific functions to mGluR7. All orthosteric ligands for mGluR7 lack clear specificity across the Group III mGluRs. Recently, however, an mGluR7-specific allosteric agonist, AMN082, was described. The arrival of a selective allosteric modulator of mGluR7 function provides a potentially useful tool for probing such function. Moreover, accumulating evidence indicates that, in addition to G-protein activation, G protein-coupled receptors trigger critical intracellular signaling cascades during agonist-induced internalization. Thus, to determine if the allosteric modulator AMN082 would be useful for evaluating signaling events related to mGluR7 internalization and receptor activation, Xiaoqing Yuan and Kenneth Pelkey examined whether AMN082 induces mGluR7 endocytosis. First, in collaboration with Katherine Roche's laboratory, we used the traditional immunofluorescence assay to demonstrate that AMN082 induces robust internalization of mGluR7 overexpressed in dissociated hippocampal neurons. AMN082-induced mGluR7 internalization was resistant to inhibition by a competitive antagonist, consistent with the distinct binding site of the allosteric agonist from the glutamate-binding pocket used by conventional orthosteric ligands. Xiaoqing Yuan then engineered an N-terminally pHluorin-tagged mGluR7 in neurons to permit live imaging of surface receptors in real time. Using the NICHD Imaging Core, we demonstrated that AMN082 treatment produced a rapid loss of surface mGluR7 as indicated by decreased fluorescence, thereby confirming the ability of allosteric receptor activation to trigger mGluR7 endocytosis. Thus, AMN082 will be effective for investigating physiological processes related to both mGluR7 activation and internalization, such as control of bi-directional plasticity at mossy fiber-st. lucidum interneuron synapses.

Depolarization-induced long-term depression at hippocampal mossy fiber-CA3 pyramidal neuron synapses

Ho, Pelkey; in collaboration with Huganir

Classically, long-term plasticity at mossy fiber-CA3 pyramid synapses (MF-CA3) has been considered to result from alterations in presynaptic release via cAMP/PKA-dependent signaling cascades. However, depolarization of CA3 pyramids in young animals (P6-P20) induces a persistent depression (DID) of MF-CA3 transmission attributable to decreased post-synaptic AMPAR function (Lei et al., J Neurosci 2003;23:9786). This novel form of long-term depression (LTD) is independent of NMDARs, mGluRs, cannabinoid receptors, opioid receptors, or coincident synaptic activity but is dependent on post-synaptic Ca2+ elevation through L-type Ca2+ channels and release from InsP3 receptor-sensitive intracellular stores. Ca2+ imaging of both proximal and distal CA3 pyramidal neuron dendrites demonstrated that the depolarizing induction paradigm differentially elevated intracellular Ca2+ levels. L-type Ca2+ channel activation is observed only at the most proximal locations where mossy fibers make synapses. Depolarization-induced LTD did not occlude the conventional 1 Hz-induced LTD or vice versa, suggesting that independent mechanisms underlie each form of plasticity. The paired-pulse ratio and coefficient of variation of synaptic transmission were unchanged after LTD induction, suggesting that the expression locus of LTD is post-synaptic. Moreover, peak-scaled nonstationary variance analysis indicated that depolarization-induced LTD correlated with a reduction in post-synaptic AMPA receptor numbers without a change in AMPA receptor conductance.

More recently, we examined the molecular details of DID, focusing on the scaffold protein PICK1 and the kinase PKC because of their importance in other post-synaptically expressed LTDs. In PICK1 knockout mice (/) mice, EPSCs remained at about 95 percent of control 5 minutes following the DID induction paradigm. In addition, a role for PKC in DID is implicated because EPSCs failed to depress after DID induction in the presence of PKC inhibitors. Emerging evidence indicates that PICK1 and PKC control surface GluR2 expression, thereby regulating AMPAR subunit composition. Thus, we next examined whether GluR2-lacking, Ca2+permeable (CP)-AMPARs participate in MF-CA3 transmission. Surprisingly, in PICK1+/+ slices, CP-AMPARs contributed significantly to MF-CA3 transmission as determined by inwardly rectifying I-V relations of MF-CA3 EPSCs. In contrast, MF-CA3 synapses in PICK1/- slices yielded significantly more linear I-V relations, indicating that GluR2-containing AMPARs dominate MF-CA3 transmission in the absence of PICK1. Our working hypothesis posits that depolarization-induced depression of mossy fiber-pyramidal cell transmission results from conversion of GluR2-lacking MF-CA3 synapses to GluR2-containing synapses in a PICK1 and PKC dependent manner.

Transient incorporation of GluR2-lacking AMPA receptors mediates an initial component of hippocampal LTP

Pelkey; in collaboration with Collingridge, Isaac

Over the last few years, our work has focused on the roles played by both Ca-permeable and Ca-impermeable AMPA receptors in mechanisms of synaptic transmission and plasticity. Study of the latter receptors has generally concentrated on their role in the inhibitory interneuron population, largely because something of a dogma asserted that postnatal glutamatergic principal neuron synapses express only Ca-impermeable, GluR2-containing AMPARs under physiological conditions. However, over the course of the last few years, a scan of the literature revealed certain inconsistencies within the glutamate receptor field, particularly in experiments involving targeting of overexpressed recombinant AMPA receptors during plasticity. Specifically, many highly influential laboratories repeatedly demonstrated that homomeric GluR1-containing AMPA receptors could traffic into synapses following activity, yet the same laboratories discounted the idea that native homomeric channels could do the same. The prevailing current hypothesis for principal neuron AMPAR trafficking during NMDAR-dependent LTP holds that AMPARs composed of GluR1 and GluR2 subunits are incorporated at synapses during LTP and, through constitutive recycling, are subsequently replaced by AMPARs containing GluR2 and GluR3 subunits. This scheme is based on the premise that all principal neuron AMPARs contain GluR2. However, recent work suggested that principal neurons contain substantial intracellular reserve pools of GluR2-lacking AMPARs, which may be synaptically incorporated under specific conditions. Therefore, we investigated whether hippocampal NMDAR-dependent LTP drives changes in the GluR2 content of native AMPARs at Schaffer collateral-CA1 pyramidal cell synapses.

Kenneth Pelkey, in collaboration with the laboratories of John Isaac and Graham Collingridge, demonstrated that induction of the most conventional and widely studied form of LTP in CA1 hippocampal pyramidal neurons caused a rapid but transient incorporation of native GluR2-lacking Ca-permeable AMPARs, which were replaced by more conventional GluR2-containing AMPARs about 25 minutes after LTP induction. Mechanisms that interfered with Ca-permeable AMPA receptor function during this period resulted in decremental potentiation that returned to baseline. The data indicate that Ca entry through this transient population of receptors may serve as an important signaling mechanism for conversion into stable LTP. Thus, our experiments revealed that Ca-permeable AMPARs are physiologically expressed at CA1 pyramidal cell synapses during LTP and may be required for LTP consolidation.

Somatodendritic Kv7/KCNQ/M channels control interspike interval in hippocampal interneurons

Lawrence, Churchill, Travis; in collaboration with Saraga, Skinner

The M-current (IM) is a voltage-activated K+ conductance that plays a key role in the control of cell excitability. In hippocampal principal cells, IM controls action potential (AP) accommodation and contributes to medium-duration afterhyperpolarization. Kv7 subunits, the molecular correlates of IM, have been detected in interneurons, but the role of IM in the control of interneuron excitability is unclear. In a follow-up study to the experiments described above, Joshua Lawrence, Katherine Travis, and Joseph Churchill investigated the function and cellular localization of IM in hippocampal stratum oriens interneurons by using an immunocytochemical, electrophysiological, and computational approach. Somatodendritic expression of both Kv7.2 and Kv7.3 subunits was revealed on SO interneurons, including O-LM cells. Upon deactivation from -30 to -50 mV, a relaxation characteristic of IM was present; it was attenuated by low concentrations of TEA, XE-991, and linopirdine and potentiated by retagabine, features consistent with a Kv7.2-containing phenotype.

When strongly depolarized in voltage clamp, the Kv7-mediated outward current activated rapidly and constituted up to 20 percent of total outward current. Recordings in loose-patch and whole-cell configurations revealed that inhibition of IM increased AP frequency without influencing half-width or first spike latency. In collaboration with Fernanda Saraga and Frances Skinner, we then generated a multicompartment O-LM interneuron model that incorporated IM. Somatodendritic placement of Kv7 channels best reproduced experimentally measured IM, and the models indicated that, although IM and Kv3-mediated delayed rectifier K+ conductances are both activated during single APs, they play fundamentally distinct roles. While Kv3 channels mediate rapid repolarization of the AP, the slower deactivation of Kv7 channels make IM the predominant conductance during the interspike interval, affecting dendritic excitability for hundreds of milliseconds following an AP.

TASK-like conductances present within hippocampal CA1 interneuron subpopulations

Torborg, Jefferies; in collaboration with Bayliss

The resting membrane potential of a neuron is determined in part by the potassium leak channels it expresses. TASK-1 (KCNK3) and TASK-3 (KCNK9) are members of a family of two-pore-domain potassium channels and form either homomeric (only TASK-1 or TASK-3) or heteromeric (TASK-1 and TASK-3) open rectifier (leak) channels. A recent study suggested that these channels contribute to the resting potential in several types of neurons, including hippocampal CA1 pyramidal cells. However, it was also reported that TASK-like conductances were present in only a small proportion of inhibitory interneurons. At press time, we had data suggesting otherwise. Using a combination of immunohistochemistry and electrophysiology, Christine Torborg explored the distribution of TASK-like channels in specific populations of inhibitory interneurons. First, using immunocytochemistry, we demonstrated that TASK-3 channels were predominantly expressed in parvalbumin-positive and somatostatin-positive interneurons. Second, even though no specific inhibitors of TASK channels exist, TASK currents are sensitive to local anesthetics and pH. In addition, we detected TASK-like currents (modulated by both pH and bupivacaine) in 46 percent of CA1 stratum oriens interneurons of various morphological classes. In most neurons, basic shifts in pH had a larger effect on the TASK-like current than acidic shifts, indicating that the current is mediated by TASK-1/TASK-3 heterodimers. The data suggest that TASK-like conductances are more prevalent in inhibitory interneurons than previously thought.

COLLABORATORS

Douglas A. Bayliss, PhD, University of Virginia, Charlottesville, VA
Graham Collingridge, FRS, FMedSci, University of Bristol, Bristol, UK
Richard Huganir, PhD, Howard Hughes Medical Institute, The Johns Hopkins University, Baltimore, MD
John Isaac, PhD, Developmental Plasticity Unit, NINDS, Bethesda, MD
Jean-Claude Lacaille, PhD, University of Montreal, Montreal, Canada
Katherine Roche, PhD, Basic Neurosciences Program, NINDS, Bethesda, MD
Fernanda Saraga, PhD, University of Toronto, Toronto, Canada
Frances Skinner, PhD, University of Toronto, Toronto, Canada
Lisa Topolnik, PhD, University of Montreal, Montreal, Canada

For further information, contact mcbainc@mail.nih.gov.