GABAergic Neurons: Gate Masters of the Amygdala, Mastered by Dopamine

Previews 877

Mark R. Cookson1 and Jordi Clarimon1 1 Cell Biology and Gene Expression Unit Laboratory of Neurogenetics National Institute on Aging National Institutes of Health Bethesda, Maryland 20892

afferents to both the major input and output station of the amygdala, are inhibited during action of dopamine via D1 receptors, and are thus likely to represent important cellular players during dopaminergic disinhibition related to increased affective behavior.

Selected Reading

Almost everybody has experienced a stressful or threatening event that has influenced their emotional behavior, in particular in relation to fear and anxiety. Accumulating evidence indicates that inhibitory GABAergic transmission in the amygdala and modulation through the transmitter dopamine (DA) are critically involved in this. First, the amygdala plays a central role in imbuing sensory events with an affective value and transducing it into behavioral responses via projections into relevant brain regions (Maren and Quirk, 2004). Second, activity of amygdaloid projection neurons is under inhibitory control through GABAergic mechanisms (Royer et al., 1999; Szinyei et al., 2000). Third, DA is increased under stress in the amygdala (Inglis and Moghaddam, 1999) and enhances amygdala-related behavior, most likely through a reduction of inhibition (Rosenkranz and Grace, 1999). While this line of evidence seems straightforward, there is a major caveat to this when it comes to correlating cellular effects of DA to behavioral outcome. Specifically, recordings from identified GABAergic interneurons in the amygdala demonstrated a depolarizing response to DA (Kro¨ner et al., 2005), which increased rather than reduced inhibitory signals and which is difficult to reconcile with a disinhibitory effect of DA on amygdala-related behavior. In this issue of Neuron, Marowsky and colleagues (Marowsky et al., 2005) present an elegant piece of work in which they identify the cellular substrates of DA-induced disinhibition in the amygdala. To do this, they used mice described first by Tamamaki et al. (2003), in which enhanced green fluorescent protein (EGFP) is expressed under the control of the promoter for GAD67, a key enzyme for GABA synthesis. The GAD67GFP mice display reliable GFP labeling of GABAergic neurons, and they allowed Marowsky et al. to visualize subsets of GABAergic neurons in slice preparations of the amygdala in vitro. In a meticulous series of experiments, the basic properties of GABA neurons, the synaptic connectivity and the responsiveness to DA, were characterized through whole-cell recording from single and synaptically connected pairs of neurons combined with microstimulation. With the focus on the basolateral amygdaloid complex (BLA), which comprises the major afferent input station of the amygdala, two major populations of GABAergic neurons were discerned: ‘‘local’’ GABAergic interneurons scattered in the BLA, and paracapsular intercalated cells (pcs) organized in two clusters. One cluster of intercalated cells (the lateral subdivision, lpc) is located along the external capsule. A second cluster of intercalated cells (the medial subdivision, mpc) is located at the border between the BLA and the central amygdaloid nucleus (CeA) and is thereby situated between the major input and output station of the amygdala. The lpcs, whose properties and connectivity are described here for the first time, turned out to make a most interesting contribution. Their properties include a small size, a poorly developed dendritic tree,

Bragg, D.C., Camp, S.M., Kaufman, C.A., Wilbur, J.D., Boston, H., Schuback, D.E., Hanson, P.I., Sena-Esteves, M., and Breakefield, X.O. (2004). Perinuclear biogenesis of mutant torsin-A inclusions in cultured cells infected with tetracycline-regulated herpes simplex virus type 1 amplicon vectors. Neuroscience 125, 651–661. Breakefield, X.O., Kamm, C., and Hanson, P.I. (2001). TorsinA: movement at many levels. Neuron 31, 9–12. Bressman, S.B., Sabatti, C., Raymond, D., de Leon, D., Klein, C., Kramer, P.L., Brin, M.F., Fahn, S., Breakefield, X., Ozelius, L.J., and Risch, N.J. (2000). The DYT1 phenotype and guidelines for diagnostic testing. Neurology 54, 1746–1752. Clarimon, J., Asgeirsson, H., Singleton, A., Jakobsson, F., Hjaltason, H., Hardy, J., and Sveinbjornsdottir, S. (2005). Torsin A haplotype predisposes to idiopathic dystonia. Ann. Neurol. 57, 765–767. Dang, M.T., Yokoi, F., McNaught, K.S., Jengelley, T.A., Jackson, T., Li, J., and Li, Y. (2005). Generation and characterization of Dyt1 DeltaGAG knock-in mouse as a model for early-onset dystonia. Exp. Neurol. 196, 452–463. Gonzalez-Alegre, P., and Paulson, H.L. (2004). Aberrant cellular behavior of mutant torsinA implicates nuclear envelope dysfunction in DYT1 dystonia. J. Neurosci. 24, 2593–2601. Goodchild, R.E., and Dauer, W.T. (2004). Mislocalization to the nuclear envelope: an effect of the dystonia-causing torsinA mutation. Proc. Natl. Acad. Sci. USA 101, 847–852. Goodchild, R.E., and Dauer, W.T. (2005). The AAA+ protein torsinA interacts with a conserved domain present in LAP1 and a novel ER protein. J. Cell Biol. 168, 855–862. Goodchild, R.E., Kim, C.E., and Dauer, W.T. (2005). Loss of the dystonia-associated protein torsinA selectively disrupts the neuronal nuclear envelope. Neuron 48, this issue, 923–932. Gruenbaum, Y., Margalit, A., Goldman, R.D., Shumaker, D.K., and Wilson, K.L. (2005). The nuclear lamina comes of age. Nat. Rev. Mol. Cell Biol. 6, 21–31. Klein, C. (2005). Movement disorders: classifications. J. Inherit. Metab. Dis. 28, 425–439. Naismith, T.V., Heuser, J.E., Breakefield, X.O., and Hanson, P.I. (2004). TorsinA in the nuclear envelope. Proc. Natl. Acad. Sci. USA 101, 7612–7617. Ozelius, L.J., Hewett, J.W., Page, C.E., Bressman, S.B., Kramer, P.L., Shalish, C., de Leon, D., Brin, M.F., Raymond, D., Corey, D.P., et al. (1997). The early-onset torsion dystonia gene (DYT1) encodes an ATP-binding protein. Nat. Genet. 17, 40–48. DOI 10.1016/j.neuron.2005.12.006

GABAergic Neurons: Gate Masters of the Amygdala, Mastered by Dopamine A hyperdopaminergic state, such as stress, is associated with an increase in affective behavior. In this issue of Neuron, Marowsky and colleagues identify two clusters of paracapsular intercalated GABA neurons in amygdala slice preparations of GAD67-GFP mice. These GABA neurons mediate inhibition from cortical

Neuron 878

spike forms and patterns in between those of ‘‘classical’’ interneurons and projection neurons, and they lack the calcium-binding protein and neuropeptide expression typical of other types of GABAergic neurons in the amygdala. Recording from lpcs-BLA cell pairs and afferent input stimulation revealed that lpcs receive excitatory input from cortex and lateral inhibition from other interneurons and mediate feedforward inhibition to BLA projection neurons. Compared to feedforwardly connected local BLA interneurons (Szinyei et al., 2000), the lpcs seem to provide the dominant (70%) contribution to overall feedforward inhibition. A further important message is that the lpcs network functionally complements that of the mpcs, with the former enabling feedforward control of signal flow from cortex to the BLA, and the latter providing a feedforward inhibitory gate for signals between the BLA and the central amygdala (Royer et al., 1999). Altogether, available data now allow us to construct a GABAergic network architecture in which powerful inhibitory feedforward systems control the signal flow at the entry and the exit of the amygdala and in which prominent inhibitory influences are associated with cortical input pathways (Figure 1). Future studies are needed to unravel more detailed aspects of the network. Some open questions remain: Are lpcs topographically organized within a functional interface for specific gating of cortical afferent impulse traffic, as mpcs are between the BLA and CeA (Royer et al., 1999)? Will lpcs couple via BLA neurons to the mpcs, thereby forming a disinhibitory synaptic loop between cortical inputs and the output of the amygdala? Are lpcs equipped with specific sets of membrane channels enabling them to control surrounding networks, for instance through bistable behavior as in mpcs (Royer et al., 2000)? More specifically, can GABAergic signals function as time windows enabling theta synchronization in amygdalohippocampal networks during conditioned fear behavior (Seidenbecher et al., 2003), and to what extent are the different subsets of GABAergic neurons involved in the plasticity of GABAergic synaptic transmission (Bauer and LeDoux, 2004)? For the time being, further exciting news from the Marowsky et al. paper is that the different subsets of GABA neurons in the amygdala show qualitatively different responses to DA. Both lpcs and mpcs clusters are packed with tyrosine hydroxylase positive plexus, indicating a dense dopaminergic innervation. Therefore, it does not come as a surprise that both mpcs and lpcs neurons respond vigorously to application of DA. The unexpected finding was that pcs neurons show a depressant response to DA, thereby sharply contrasting with the excitatory dopaminergic response in local GABAergic interneurons and projection neurons (Kro¨ner et al., 2005). More specifically, DA evokes a membrane hyperpolarization associated with a decrease in input resistance and a marked drop in spike firing in pcs. These responses are mediated via the D1 subtype of DA receptors, coupled via a G protein to activation of an inwardly rectifying K+ channel of the GIRK 1/4 subtype. In an important experimental step, Marowsky and colleagues went on to verify the functional significance of DA action in the synaptic network. Paired recordings from lpcs and BLA neurons, and recordings of evoked synaptic responses in pcs target neurons in the BLA and CeA

Figure 1. Dopaminergic Modulation of Feedforward Inhibitory Circuits in the Amygdala BLA and CeA indicate the basolateral and central amygdaloid nucleus, respectively; LPC and MPC represent lateral and medial paracapsular intercalated GABAergic neurons; D1 and D2 denote dopamine receptor subtypes; up and down arrows denote increase and decrease in activity.

revealed that DA via D1 receptors strongly reduces the excitability and GABAergic output of the pcs system, thereby relieving feedforward inhibition to the BLA and the CeA. In the case of mpcs, but not lpcs, a concomitant decrease was observed at the excitatory cortical input. The consequence for network activity is a strong shift in the synaptic balance toward excitation in both the major input (BLA) and the output station (CeA) of the amygdala (Figure 1). This overall disinhibition will be boosted by the depolarizing action of DA on BLA projection neurons (via a D1 receptor-mediated decrease in K+ conductance) and the decrease in GABA release from local interneurons, whereas the D1 and D2 mediated increase in excitability of local GABA neurons can be assumed to limit excessive excitation in local networks (Figure 1; Bissie`re et al., 2003; Kro¨ner et al., 2005). The mutual interconnections of lpcs and mpcs may provide a further target of DA modulation, similar to the situation in local GABA neurons (Bissie`re et al., 2003). The overall message is that conditions under which DA is increasingly released will gate the signal flow through the amygdala by an increase in excitability at both the entry and the exit gate. This disinhibitory effect is most heavily linked to the cortical input system acting on pcs. However, the role of subcortical thalamic input signals to the amygdala remains largely unknown in this scenario. Their potential contribution is suggested by previous findings indicating that synaptic responses to thalamic inputs are long-term potentiated in relation to conditioned fear behavior (Maren and Quirk, 2004) and during DA-mediated suppression of local interneuronal activity (Bissie`re et al., 2003), and feedforward GABA interneurons receive input from both thalamic and cortical fibers (Szinyei et al., 2000). In any case, a hallmark feature of the pcs is their innervation from the medial prefrontal cortex (mPFC), which has been demonstrated by both anatomical and physiologic studies (for the mpcs, but not yet for the lpcs). Much data

Previews 879

implicate PFC-amygdala interactions in the regulation of affective behaviors. This is evident, for instance, when stressful situations lead to increased fear reactions, or when fear memories are extinguished. How might the mPFC participate in both an increase in fear behavior and an extinction of fear memories? One important target is GABAergic inhibition in the amygdala. Stress attenuates GABAergic inhibition in the BLA and facilitates fear memory (Manzanares et al., 2005), increases DA levels (Inglis and Moghaddam, 1999), and produces disinhibition in the CeA similar to the effect of severing mPFC inputs (Correll et al., 2005). The likely candidate for mediating this effect is the pcs system. Extinction of fear memories, on the other hand, involves activity in the mPFC, which shows in the amygdala as inhibition of CeA output neurons (Quirk et al., 2003). The influence of the mPFC likely involves activation of feedforward GABAergic neurons, because the direct mPFC-CeA projection is sparse and the mPFC-BLA connection is of an excitatory nature (Likhtik et al., 2005). Again, the GABAergic pcs system is the likely candidate for mediating this effect. In conclusion, the study by Marowsky and coworkers suggests a simple resolution for the paradoxical result that DA is disinhibitory in the amygdala in vivo and yet excites local GABAergic neurons: the GABAergic pcs system is inhibited while local GABAergic neurons are excited by DA. The result breaks new grounds in our understanding of DA-mediated disinhibition in the amygdala related to cortical input systems, but also points to some challenges inherent in attempting to relate cellular mechanisms to behavior. Within any given network, the effect of synaptic transmission will depend on a host of factors, including the detailed connectivity, the temporal sequence and plasticity at individual synapses, the signal patterns at the input and output stations, and the adaptations that occur in response to the individual behavior. Future studies are therefore needed to explore the functional impact of the pcs system and related synaptic networks for the modulation of affective behaviors. One important issue relates to the influence of the hippocampus, given the hippocampal contribution to contextual affective memory and the context dependence of extinction (Maren and Quirk, 2004). It will also be important to track abnormal interactions between the mPFC and the amygdala, which have been implicated in a number of psychiatric disorders, including posttraumatic stress disorders, depression, and schizophrenia. The work of Marowsky and colleagues prepares the ground for such studies. Hans-Christian Pape1 Institute of Physiology I Department of Medicine Westfaelische Wilhelms-University D-48149 Muenster Germany 1

Selected Reading Bauer, E.P., and LeDoux, J.E. (2004). J. Neurosci. 24, 9507–9512. Bissie`re, S., Humeau, Y., and Lu¨thi, A. (2003). Nat. Neurosci. 6, 587–592. Correll, C.M., Rosenkranz, J.A., and Grace, A.A. (2005). Biol. Psychiatry 58, 382–391.

Inglis, F.M., and Moghaddam, B. (1999). J. Neurochem. 72, 1088– 1094. Kro¨ner, S., Rosenkranz, J.A., Grace, A.A., and Barrionuevo, G. (2005). J. Neurophysiol. 93, 1598–1610. Likhtik, E., Pelletier, J.G., Paz, R., and Pare´, D. (2005). J. Neurosci. 25, 7429–7437. Manzanares, P.A.R., Isoardi, N.A., Carrer, H.F., and Molina, V.A. (2005). J. Neurosci. 25, 8725–8734. Maren, S., and Quirk, G.J. (2004). Nat. Rev. Neurosci. 5, 844–852. Marowsky, A., Yanagawa, Y., Obata, K., and Vogt, K.E. (2005). Neuron 48, this issue, 1025–1037. Quirk, G.J., Likhtik, E., Pelletier, J.G., and Pare´, D. (2003). J. Neurosci. 23, 8800–8807. Rosenkranz, J.A., and Grace, A.A. (1999). J. Neurosci. 19, 11027– 11039. Royer, S., Martina, M., and Pare´, D. (1999). J. Neurosci. 19, 10575– 10583. Royer, S., Martina, M., and Pare´, D. (2000). J. Neurosci. 20, 9034– 9039. Seidenbecher, T., Rao Laxmi, T., Stork, O., and Pape, H.-C. (2003). Science 301, 846–850. Szinyei, C., Heinbockel, T., Montagne, J., and Pape, H.-C. (2000). J. Neurosci. 20, 8909–8915. Tamamaki, N., Yanagawa, Y., Tomioka, R., Miyazaki, J.-I., Obata, K., and Kaneko, T. (2003). J. Comp. Neurol. 467, 60–79. DOI 10.1016/j.neuron.2005.12.002