The Olfactory System
The olfactory bulb is the first site of processing olfactory information after transduction by receptor cells in the nasal epithelium. Each receptor cell appears to express only one type of receptor protein, resulting in about a thousand different types of receptor cells (roughly the number of receptor gene products made). Furthermore, as shown by Richard Axel and colleagues, axons from similar receptor cells generally converge on the same glomerulus in the olfactory bulb where they synapse on the distal primary dendrites of mitral cells. The specificity of these excitatory inputs to the olfactory bulb would therefore appear to produce a map of "olfactory" space. However individual receptor cells respond to many different odorants, so that even single compounds result in excitatory synaptic inputs to mitral cells in widespread regions within the bulb. Rather than functioning as a relay station in a conventional sensory labeled line, it seems that the role of the olfactory bulb is to begin the decoding the various patterns of synaptic input associated with different odorants. We are interested in the synaptic interactions in the olfactory bulb and how inhibitory interneurons are involved in information processing in the olfactory system.
The granule cell is the most common type of interneuron in the olfactory bulb. Since they lacks an axon, granule cells are only able to produce inhibition very locally--which they do by releasing GABA from specialized dendritic spines that contact long horizontal (secondary) dendrites of mitral cells. These inhibitory synapses are typically associated on the same spine with an excitatory synapse made between the mitral cell and interneuron, resulting in a reciprocal dendrodendritic synaptic connection. As part of his Ph.D. dissertation, Craig Jahr directly demonstrated that in a turtle, this dendritic microcircuit could generate recurrent inhibition onto mitral cells. Together with Jeff Isaacson in 1998, we showed that a similar circuit functioned in the mammalian olfactory bulb and that dendritic transmitter release could underlie lateral as well as recurrent inhibition. We also demonstrated that activation of NMDA receptors on granule cells are required to cause GABA release. This finding was unexpected since granule cells express function AMPA receptors as well as NMDA receptors, with AMPA receptors typically associated with fast excitatory synaptic transmission in most brain regions.
Our laboratory has examined the requirement for NMDA receptors in dendrodendritic inhibition using a variety of techniques. While GABA release from granule cells can be controlled by calcium influx through conventional voltage-dependent calcium channels (VDCC), we recently showed that calcium influx through NMDA receptors also can initiate transmitter release. Interestingly, calcium entering granule cell spines through NMDA receptors is able to cause GABA release even after blockade of all VDCC normally coupled to release sites. We now view the NMDA receptor on granule cell spines as providing both electrical and chemical signals that can independently lead to GABA release (see diagrams to the right). The electrical signal--part of the conventional EPSP response--interacts with intrinsic conductances in the spine and can lead to GABA release via activation of VDCC. The calcium influx through the NMDA receptor itself represents a direct chemical stimulus for secretion. The enhanced GABA release activated by the direct effect of calcium influx through NMDA receptors may contribute to the critical role NMDA receptors play in dendrodendritic inhibition.
We recently completed a study on Blanes cells, one of a group of seven non-granule cell types located in the granule cell layer of the olfactory bulb described by Cajal and his contemporaries (including Blanes). Blanes cells are large spiny GABAergic interneurons that inhibit granule cells and appear to excited by mitral cells. We found that Blanes cells can fire persistently following transient depolarizing stimuli. Using 2-photon microscopy we found that persistent activity in Blanes cells evokes prolonged barrages of inhibitory synaptic potentials onto monosynaptically-coupled granule cells. Several lines of evidence suggest that the non-selective cation current Ican mediates the afterdepolarization in Blanes cells and contributes to their persistent firing mode. The persistent firing state we found in Blanes cells is similar to the persistent firing mode reported by Alonso and colleagues in entorhinal cortical cells, which also appears to depend on Ican. However unlike entorhinal neurons, Blanes cells can fire persistently under normal phamacological conditions whereas entorhinal cortical neurons can only enter the persistent firing state when mACh receptors are activated.
Mechanisms of mitral cell spike patterning
In many brain regions, neurons express specialized ion channels that shape their responses to depolarizing and hyperpolarizing stimuli. Different combinations of these intrinsic channels can produce a staggering diversity in neuronal behavior that have profound effects on information processing and network activity. Mitral cells have unique intrinsic currents that generate clusters of action potentials interspersed with periods of fast (40 Hz) subthreshold voltage oscillations in response to depolarizing stimuli. We’ve shown that a single type of voltage-sensitive potassium channel that inactivates slowly after opening (known as I-D) is responsible this behavior. Interestingly, when driven by trains of depolarizing stimuli that mimic trains of EPSPs that occur during sniffing, these K channels produce clusters of APs that are precisely phase-locked to the stimulus train. We are currently investigating other intrinsic mechanisms that regulate how mitral cells respond to hyperpolarizing, inhibitory stimuli.