Glutamatergic synaptic inputs and ICAN: the basis for an emergent property underlying respiratory rhythm generation?

Abstract

The mammalian respiratory control system in the brainstem produces a robust rhythm that drives respiratory muscles and maintains blood gas homeostasis from birth to death. Unfortunately, there is a great deal that is still not known with respect to how key cellular and synaptic mechanisms contribute to this fundamental rhythm. Among the models currently being debated (Fig. 1), the pacemaker-network hypothesis proposes that some respiratory neurons have intrinsic membrane conductances that cause the membrane potential to spontaneously oscillate (i.e. pacemaker properties). Synaptic inputs modulate pacemaker neuron excitability, but ultimately the breathing rhythm is derived from pacemaker properties (Smith et al. 2000; Ramirez & Viemari, 2005). In contrast, the network hypothesis proposes that the breathing rhythm is due to neurons that are reciprocally coupled by inhibitory synaptic connections, thereby establishing a straightforward oscillatory neural network (Richter et al. 1992). Finally, the group-pacemaker hypothesis proposes that inspiratory bursts of action potentials are an emergent property of non-pacemaker neurons that are interconnected by chemical and electrotonic excitatory synaptic interactions; pacemaker neurons are not required (Rekling & Feldman, 1998; Feldman & Del Negro, 2006). The pacemaker-network and network hypotheses incorporate the interaction between synaptic inputs and intrinsic membrane conductances into their models, but these interactions are at the heart of the group-pacemaker model. Figure 1 Conceptual models for respiratory rhythm generation Shortly after the group-pacemaker model was first proposed (Rekling & Feldman, 1998), Del Negro et al. (2002) tested the hypothesis that one type of pacemaker neuron was necessary for rhythm generation by pharmacologically blocking persistent sodium (INaP) currents with riluzole in rhythmically active medullary slices. Since respiratory burst frequency was not altered in their experiments, the authors concluded that pacemaker neurons with INaP were not required for rhythm generation. However, other laboratories obtained the opposite results with similar experiments (e.g. Ramirez & Viemari, 2005). In addition, a separate class of pacemaker neurons that express Ca2+-activated non-specific cation (ICAN) currents may also contribute to rhythm generation (Ramirez & Viemari, 2005). To test whether both types of respiratory pacemaker neurons are necessary for rhythm generation, Del Negro et al. (2005) blocked INaP with riluzole and ICAN with flufenamic acid. Although the rhythm was initially abolished under these conditions, rhythmic activity could be restored by increasing network excitability with substance P application. Again, the authors concluded that respiratory pacemaker neurons (at least the types that are known to exist so far) are not required for rhythm generation, and that rhythm generation must therefore be an emergent property of the network. In this issue of The Journal of Physiology, Pace et al. (2007) extend this line of research by investigating potential mechanisms underlying the group-pacemaker hypothesis at the cellular level. With whole-cell recordings of inspiratory neurons in rhythmically active medullary slices, Pace et al. (2007) show that large inspiratory drive potentials (i.e. 10–30 mV depolarizations occurring during the inspiratory phase) are due to activation of ICAN. They also show that ICAN activation requires activation of NMDA or group I metabotropic glutamate (mGluR1) receptors. The authors conclude that glutamatergic excitatory synaptic inputs are required to evoke the ICAN-dependent inspiratory drive potentials, rather than ICAN alone being the main source of inspiratory drive potentials (as suggested by the pacemaker hypothesis). The importance of this study is that, once again, it demonstrates the plausibility of the group-pacemaker hypothesis for respiratory rhythm generation. Does this mean that pacemaker neurons are not involved in rhythm generation? Of course not! Other investigators suggest that pacemaker neurons may still play a critical role in rhythm generation (Ramirez et al. 2007). Also, the respiratory control system is multifunctional, state-dependent, highly modulated, and capable of significant plasticity. Thus, the relative contribution of pacemaker properties and synaptic inputs to rhythm generation is likely to be very dynamic. During normal quiet breathing, the respiratory rhythm may be predominantly due to inhibitory synaptic transmission (network hypothesis) and glutamatergic–ICAN interactions (group-pacemaker hypothesis). During hypoxia, however, the respiratory control network may shift to a different operating mode that is more dependent on pacemaker properties and produce gasping behaviour (Paton et al. 2006). To an outside observer, the debate regarding the cellular and synaptic mechanisms underlying respiratory rhythm generation may appear to be an exercise in hair-splitting semantics. However, understanding precisely how the brain produces the breathing rhythm is not only a fascinating biological problem, but it may also lead to critical insights for treating respiratory-related diseases and pathological conditions, such as sleep-disordered breathing, spinal cord injury, sudden infant death syndrome, Rett syndrome and congenital hypoventilation.