Retrograde Tuning of Tuning

One way to localize sounds is to measure differences in sound intensity at the two ears. This comparison is made in the lateral superior olive, where signals from both ears converge. Magnusson et al. in this issue of Neuron show that dendritic GABA release can regulate this comparison, which may allow animals localizing sounds to adapt to listening conditions. One way to localize sounds is to measure differences in sound intensity at the two ears. This comparison is made in the lateral superior olive, where signals from both ears converge. Magnusson et al. in this issue of Neuron show that dendritic GABA release can regulate this comparison, which may allow animals localizing sounds to adapt to listening conditions. The lateral superior olive (LSO) is a brainstem nucleus that plays an important role in sound localization. Mammals use three major cues to localize sounds: the intensity and timing differences of sounds at the different ears and spectral properties of sounds as they are filtered by the external ear. The LSO is the first place in the auditory pathway where cell responses are affected by the intensity difference between the two ears (called interaural level difference or ILD) (Boudreau and Tsuchitani, 1968Boudreau J.C. Tsuchitani C. J. Neurophysiol. 1968; 31: 442-454PubMed Google Scholar). LSO neurons fire spikes at a rate that roughly corresponds to azimuthal position: spike rates are highest in response to sounds originating on the ipsilateral side, they are intermediate for sounds produced directly behind or in front, and there is little response to sounds originating on the contralateral side. Different LSO cells show changes in firing rate at different azimuthal positions. These differences in tuning allow the LSO in total to cover a wider range of sound locations. The response properties of LSO neurons appear to result from a simple, feed-forward circuit. Auditory nerve fibers from the cochlea excite bushy cells in the anteroventral cochlear nucleus (AVCN), which in turn form excitatory synapses onto neurons in the ipsilateral LSO; while bushy cells in the contralateral AVCN excite principal cells of the medial nucleus of the trapezoid body (MNTB), which send inhibitory inputs to the LSO (Sanes and Friauf, 2000Sanes D.H. Friauf E. Hear. Res. 2000; 147: 46-58Crossref PubMed Scopus (83) Google Scholar, Schwartz, 1992Schwartz I.R. Webster D.B. Popper A.N. Fay R.R. The Mammalian Auditory Pathway: Neuroanatomy. Springer-Verlag, New York1992: 117-167Google Scholar; Figure 1A). The interaction of ipsilateral excitation and contralateral inhibition yields the response characteristics of the LSO cells. This circuit has stood as an example of how the nervous system can perform a calculation that has behavioral relevance (Tollin, 2008Tollin D.J. Dallos P. Oertel D. Audition. Academic Press, San Diego, CA2008: 631-654Google Scholar). Thus, the LSO provides a useful model to understand how specific adaptations in cells and synapses underlie perception. In this issue of Neuron, Magnusson et al., 2008Magnusson A.K. Park T.J. Pecka M. Grothe B. Koch U. Neuron. 2008; 59 (this issue): 125-137Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar investigated how the ILD response characteristics of LSO neurons are influenced by neuromodulation, specifically through metabotropic GABAB receptors (GABABR; Figure 1B). In in vivo experiments, they found that LSO neurons decreased their firing rate upon local activation of GABABRs using the exogenous agonist, baclofen (Figure 1C, left). In addition, there was a shift in ILD tuning, such that LSO neurons became less responsive to ipsilateral sounds (Figure 1C, right). This effect could be accounted for by a larger impact of GABABR activation on excitatory (ipsilateral) inputs than inhibitory (contralateral) inputs. The authors confirmed this in in vitro experiments. The involvement of GABABRs in regulating binaural sensitivity is not surprising in that GABABR activation can suppress firing by opening potassium channels and can reduce synaptic strength by modulating presynaptic calcium channels (Hille, 2001Hille B. Ion Channels of Excitable Membranes.Third Edition. Sinauer, Sunderland, MA2001Google Scholar). What is of particular interest in the adjustment of binaural sensitivity is that GABA appears to be released from the dendrites/soma of LSO neurons, and then it acts retrogradely to suppress neurotransmitter release from presynaptic terminals arising from neurons in the AVCN and the MNTB. Thus, LSO neurons can specifically regulate their own synaptic inputs and thereby regulate their activation. Retrograde signaling is a widespread form of communication (Ludwig, 2005Ludwig M. Dendritic Neurotransmitter Release. Springer, New York2005Crossref Scopus (11) Google Scholar, Zilberter et al., 2005Zilberter Y. Harkany T. Holmgren C.D. Neuroscientist. 2005; 11: 334-344Crossref PubMed Scopus (33) Google Scholar). Neurons can release many types of chemical messengers from their dendrites and somas. Retrograde messengers include endocannabinoids, peptides, and conventional neurotransmitters. Usually they are released in a calcium-dependent manner and activate presynaptic receptors to reduce synaptic strength. Retrograde signaling can lead to either a transient reduction of synaptic strength lasting seconds or a long-term depression (LTD) of synaptic strength. LTD mediated by retrograde signaling is known to have numerous behavioral consequences, but less is known about the physiological consequences of transient suppression of synaptic strength. Magnusson et al., 2008Magnusson A.K. Park T.J. Pecka M. Grothe B. Koch U. Neuron. 2008; 59 (this issue): 125-137Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar showed that transient synaptic suppression mediated by retrograde GABA signaling can adjust binaural sensitivity. Although dendritic GABA release clearly leads to retrograde inhibition, many aspects of dendritic release are poorly understood. Within the LSO, dendritic GABA release is calcium-dependent and appears to be a result of vesicle fusion because it is blocked by botulinum neurotoxin D light chain. Beyond these basic properties, there is much still to learn of the specializations that control dendritic GABA release in the LSO. The identity of the calcium sensors mediating release, their calcium sensitivity, and their proximity to calcium channels are not known. Dendritic release of GABA can occur either in dendrites with specializations that resemble those in presynaptic boutons, as in the olfactory bulb, retina, and thalamus (Ludwig, 2005Ludwig M. Dendritic Neurotransmitter Release. Springer, New York2005Crossref Scopus (11) Google Scholar). In other neurons that release GABA from their dendrites, such as LSO neurons or cortical bitufted cells (Zilberter et al., 1999Zilberter Y. Kaiser K.M. Sakmann B. Neuron. 1999; 24: 979-988Abstract Full Text Full Text PDF PubMed Scopus (115) Google Scholar), less is known about the ultrastructure of the dendrites. Further studies could provide insight into what mechanisms are used by LSO neurons to release GABA and whether that release is restricted to specific sites. The mechanisms of GABA release are likely to have important consequences for the relationship between postsynaptic firing and the extent of retrograde inhibition. Can dendrites sustain GABA release, or does vesicle depletion eventually limit the GABA signal that results in retrograde inhibition? Determining the number of available vesicles and their ability to recycle will help to clarify this issue. In addition, there is likely to be some delay between increases in LSO activity, the release of GABA, and activation of metabotropic receptors. How might this delay affect responses to ongoing sounds and perception of their location? In considering the manner in which retrograde signaling alters the ILD responses of LSO neurons, it is important to realize that Magnusson et al., 2008Magnusson A.K. Park T.J. Pecka M. Grothe B. Koch U. Neuron. 2008; 59 (this issue): 125-137Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar show that both ipsilateral excitation and contralateral inhibition are suppressed by retrograde GABA signaling. The effects they observe on ILD tuning in vivo can be explained entirely through suppression of ipsilateral excitation. This is consistent with their in vitro observation that higher GABA levels are needed to suppress the inhibitory synapses. This raises the interesting possibility that LSO neurons firing at even higher frequencies, as might occur with high intensity sounds, could strongly suppress both inhibitory synapses and excitatory synapses. This suggests that the differential sensitivity of the two synapses could lead to ILD tuning that depends upon sound intensity. The obvious question is, why regulate ILD tuning? One possibility, pointed out by the authors, is that this helps compensate for stimulus changes and keeps the LSO neuron in a useful operating range. The need for homeostasis is quite compelling for sensory pathways, which must recognize stimulus features under varied conditions. For example, a potential complication in sound localization is that while LSO neurons are primarily viewed as responding to ILD, they are also influenced by the overall sound intensity (Park et al., 2004Park T.J. Klug A. Holinstat M. Grothe B. J. Neurophysiol. 2004; 92: 289-301Crossref PubMed Scopus (49) Google Scholar). As the total intensity of a sound is increased, the maximal response of many LSO cells also increases, even though the ILD remains constant. This causes the point of half-maximal activation to shift toward the ipsilateral side. Would this cause changes in perception, say, changing the apparent location of loud versus quiet sounds? That depends critically on how the higher auditory centers interpret LSO firing rate, which is not clear. Perhaps such effects would be mitigated by retrograde regulation of synapses described by Magnusson et al., 2008Magnusson A.K. Park T.J. Pecka M. Grothe B. Koch U. Neuron. 2008; 59 (this issue): 125-137Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar. Considering the effects of retrograde suppression on excitatory inputs, during loud sounds, higher activity in LSO neurons would cause them to release GABA onto their presynaptic inputs, thereby reducing levels of excitation and keeping LSO firing rates from increasing too much. During quiet sounds, less GABA would be released, so excitatory ipsilateral inputs could still influence LSO firing rates. An obvious prediction is that blocking GABABRs should cause firing properties of LSO neurons to be even more different for loud versus quiet sounds. Furthermore, the effect of GABABR agonists should be occluded during loud sounds, while the effect of antagonists should be occluded for quiet sounds. In this paper, both agonists and antagonists had significant effects, perhaps because the sound intensity was intermediate. In addition, there were hints that the dynamic range of LSO neurons also changed with manipulations of GABABR activation. While the effects were variable between cells, tuning appeared sharper when GABABRs were activated and broader when GABABRs were suppressed. This raises the interesting possibility that LSO neurons adjust the breadth of their tuning, depending on how active they find themselves. It will be very interesting if this finding is replicated in future studies, as it could have implications for how signal detection is optimized without sacrificing acuity or efficient coding. The regulation of synaptic strength by retrograde GABA signaling in the LSO is likely to influence behavior. It will be interesting to extend the results found in the electrophysiological studies described here to behavioral paradigms of sound localization. The prediction would be that disruption of the GABA signaling system in the LSO, by genetically preventing retrograde GABA release or pharmacologically blocking GABABRs, should reduce acuity. These effects should be particularly significant under conditions where high firing rates of LSO neurons are expected during localization of loud sounds. Thus, this system presents many advantages for identifying the behavioral effects of retrograde neurotransmitter release. Retrograde GABA Signaling Adjusts Sound Localization by Balancing Excitation and Inhibition in the BrainstemMagnusson et al.NeuronJuly 10, 2008In BriefCentral processing of acoustic cues is critically dependent on the balance between excitation and inhibition. This balance is particularly important for auditory neurons in the lateral superior olive, because these compare excitatory inputs from one ear and inhibitory inputs from the other ear to compute sound source location. By applying GABAB receptor antagonists during sound stimulation in vivo, it was revealed that these neurons adjust their binaural sensitivity through GABAB receptors. Using an in vitro approach, we then demonstrate that these neurons release GABA during spiking activity. Full-Text PDF Open Archive