A ‘group pacemaker’ mechanism for respiratory rhythm generation

Central pattern generator (CPG) networks emit neural rhythms that drive behaviours such as breathing, locomotion and mastication. Understanding their mechanisms of rhythmogenesis is a longstanding problem. One contemporary debate focuses on whether pacemaker neurons that encapsulate network activity in their intrinsic autorhythmicity or emergent network properties are the fundamental basis for CPG function. Pacemakerdriven mechanisms are intuitive and straightforward – examples in the CNS of many animals from invertebrates to mammals abound – but emergent properties can be abstruse and ill-defined. In this issue of The Journal of Physiology, Mironov (2008) presents a breakthrough analysis of emergent network properties – and emphasizes their importance – using the respiratory oscillator in the pre-Bötzinger complex (preBötC) of the ventral medulla in mammals as a model system (Smith et al. 1991; Feldman & Del Negro, 2006).

The major proposal for emergent network properties in the preBötC – dubbed the group-pacemaker hypothesis – posited that periodic inspiratory bursts originated due to intrinsic currents, which are ordinarily latent and unavailable, except when evoked synaptically in the context of network function (Rekling & Feldman, 1998). Ca²⁺-activated non-specific cation current (I_CAN) has been recognized as a predominant inspiratory burst-generating current coupled to metabotropic glutamate receptors (mGluRs) (Pace et al. 2007) that could satisfy the group-pacemaker mechanism.

Mironov now validates the group-pacemaker hypothesis; he shows the subcellular machinery for synaptically evoking I_CAN, and then manipulates this machinery to perturb and/or abolish respiratory rhythms in vitro.

Performing on-cell recordings in the preBötC while monitoring inspiratory motor output from the hypoglossal (XII) nerve, Mironov recorded ion channel activity prior to – and during – the inspiratory burst. TRPM4, and its close relative TRPM5 (subtypes of the transient receptor potential, i.e. TRP, family) are monovalent cation channels activated by intracellular Ca²⁺. Both subtypes were previously suggested to constitute I_CAN (Crowder et al. 2007). Mironov establishes that the underlying channels in the soma are TRPM4 by testing their sensitivity to Ca²⁺ and phosphatidylinositol 4,5-bisphosphate (PIP₂) as well as blockade by flufenamate (FFA) and Gd³⁺. At the systems level, bath-applied Gd³⁺ stopped the respiratory rhythm within several minutes, which suggests that TRPM4-mediated I_CAN is essential for rhythmogenesis, while avoiding the pharmacological caveats associated with FFA.

To examine the role of I_CAN in networks, Mironov performs two-photon imaging combined with on-cell patch recordings in pairs of connected preBötC neurons. Synaptic excitation gives rise to propagating Ca²⁺ waves in dendrites, which then evoke TRPM4 in the soma (Fig. 1A). During endogenous network activity these same propagating Ca²⁺ waves cause vigorous TRPM4 activity in the prelude to, and during, the inspiratory burst. Additionally, the mGluR agonist DHPG, which modulates respiratory frequency and augments TRPM4 activity, here is shown to accelerate the propagation speed of the Ca²⁺ wave. Interestingly, local application of thapsigargin to deplete the Ca²⁺ stores halts the wave in mid-dendrite, and occludes the inspiratory burst altogether. These observations suggest that group 1 mGluRs contribute to inspiratory bursts via inositol 1,4,5-trisphosphate (IP₃) and Ca²⁺-induced Ca²⁺ release, and may influence respiratory frequency by regulating the speed of Ca²⁺ wave propagation.

Figure 1. The cellular mechanisms underlying a ‘group pacemaker’ in the respiratory CPG contained in the preBötC.

A, group 1 metabotropic glutamate receptors influence the production and speed of propagation of Ca²⁺ waves, where Ca²⁺ released from intracellular stores propagates toward the soma to evoke bursts of TRPM4 channel activity during the inspiratory phase of the cycle. Excitatory preBötC neurons, coupled to one another, are thus able to synaptically evoke inspiratory drive currents in the context of network activity. B, small networks of coupled neurons with these properties, Mironov argues, are thus able to generate a respiratory-like rhythm as a ‘group pacemaker’ if the speed of Ca²⁺ waves is optimal.

To test this novel idea regarding frequency modulation Mironov turns to simulations. He creates Ca²⁺ release compartments coupled to soma-like compartments with I_CAN, which he couples in a ring of alternating compartment types (Fig. 1B). This ring system generates coherent rhythmicity only when propagation speed is held near ∼70 μm s⁻¹, matching the measured value. The range of propagation speeds that support coherent rhythmicity can be extended when ring oscillators are sparsely coupled to one another, roughly representing distinct clusters of rhythmic neurons observed in the preBötC (Hartelt et al. 2008). This arrangement may be advantageous to synchronize preBötC neurons, which may be sparsely connected (Rekling et al. 2000), and by ‘clamping’ the frequency of the network rhythm on the basis of wave propagation speed. This innovative new idea suggests that frequency modulation may not necessarily depend on regulating somatic ion channels to hyper- or depolarize baseline membrane potential, but could be tuned by metabotropic receptors that modify dendritic cable properties or Ca²⁺ release mechanisms and influence the speed of Ca²⁺ waves.

Mironov's analysis establishes a viable model for emergent network properties in a CPG. Post-synaptic ion channels become active in the context of network function; bursts of channel activity depend on metabotropic receptors and intracellular signalling. While some uncertainties remain, i.e. regarding the specific role of ionotropic AMPA receptors (AMPARs) and whether TRPM4 is expressed in dendrites, this demonstration of a group pacemaker in the preBötC should cause us to re-examine the underlying mechanism in other important CPGs (e.g. locomotion, swimming and oral-motor activity) as well as rhythmic brain networks in general.