How does CO₂ activate the neurons of the retrotrapezoid nucleus?

In the realm of biology, CO₂ usually operates via the proxy of pH and virtually all proteins, ion channels and neurons display some degree of pH sensitivity. The notion that the respiratory chemoreflex (the activation of breathing by increases in CNS partial pressure of CO₂ ()) could rely on a few specialized neurons is therefore counterintuitive. Yet, this possibility is becoming increasingly plausible and the lynchpin of the system seems to be the retrotrapezoid nucleus (Guyenet et al. 2010). This nucleus (RTN) consists of around 2000 glutamatergic neurons in rats (600–800 in mice) that selectively target the pontomedullary regions implicated in generating the respiratory rhythm and pattern (Guyenet et al. 2010). RTN neurons are strongly activated by hypercapnia in vivo and by CO₂ or protons in slices and their selective activation in vivo increases breathing (Guyenet et al. 2010). Their inhibition attenuates the chemoreflex in resting conscious rats without changing basal respiration (Marina et al. 2010). Finally, the selective and complete elimination of RTN neurons by genetic manipulations produces animals that survive but lack a respiratory chemoreflex for weeks after birth (Ramanantsoa et al. 2011).

Two non-mutually exclusive mechanisms may account for the sensitive response of RTN neurons to CO₂, excluding inputs from other acid-sensitive neurons that may exist but are off topic here. Protons generated by increases in brain could excite RTN neurons directly (Mulkey et al. 2004). These effects could be mediated by the intrinsic acid sensitivity of subsets of potassium channels or they could be mediated indirectly via specialized proton receptors and intracellular messengers. Alternately, CO₂ may activate RTN neurons by causing the surrounding glia to release ATP (Gourine et al. 2010). The latter view is backed by considerable evidence such as the demonstrated pH sensitivity of subsets of glia, the ability of glia to release ATP and the fact that ATP is released from the ventral medullary surface by hypercapnia (Gourine et al. 2010; Mulkey & Wenker, 2011). In addition, juxtacellularly applied ATP excites most lower brainstem neurons and RTN glia depolarization with channelrhodopsin activates breathing in vivo (Gourine et al. 2010). These findings are intriguing but they have not yet established the causality link between brain , glia, RTN neurons and breathing. For example, ATP receptor antagonists do not attenuate the CO₂ sensitivity of RTN neurons in slices, unless extreme levels of (15%) are applied and, given the virtually ubiquitous sensitivity of lower brainstem respiratory neurons to ATP and the weak response of RTN neurons to ATP, the glial theory does not satisfactorily explain why the central respiratory chemoreflex is so highly dependent on the integrity of RTN neurons (Mulkey et al. 2006; Guyenet et al. 2010; Ramanantsoa et al. 2011).

In a recent issue of The Journal of Physiology, Onimaru et al. (2012) reinvestigated the potential contribution of ATP to the CO₂ response of RTN neurons in vitro. They used an en bloc preparation of the lower brainstem and identified RTN neurons by the presence of Phox2b and the characteristic biphasic discharge pattern that these neurons exhibit in such a preparation. They lowered pH by increasing to consider the possibility that molecular CO₂ rather than protons could be the key determinant of the regulation of RTN neurons by the glia (Huckstepp et al. 2010). Like prior investigators who worked with slices, they found that a reduction in potassium conductance contributes to the excitatory effect of CO₂ and that the effect of CO₂ persists after blocking action potentials with TTX (Mulkey et al. 2004). Also, like prior investigators, Onimaru et al. (2012) report that ATP-receptor antagonists have no detectable effect on the response of RTN neurons to moderate changes in (Mulkey et al. 2006; Guyenet et al. 2010). The most important new result from the study by Onimaru et al. is that the depolarization of RTN neurons by CO₂ is unaffected by concentrations of cadmium that should block vesicular release. This finding is interpreted as further evidence against the possibility that regular transmitters from other neurons or gliotransmitters such as ATP could contribute to the CO₂ response of RTN neurons.

In summary, the present study favours the possibility of a direct depolarizing action of CO₂ and or protons on neonatal RTN neurons. However, a glial contribution is not entirely ruled out by the data, even in this in vitro system, because CO₂ could conceivably trigger the release of gliotransmitters by a calcium-independent (cadmium-resistant) mechanism such as the proposed opening of connexin channels (Huckstepp et al. 2010) and the gliotransmitter responsible for activation of RTN neurons by CO₂ could be something other than ATP. Onimaru et al. (2012) did rule out the contribution of substance P but the list of possible gliotransmitters is extensive. This list includes glutamate which could be acting on metabotropic receptors. Also, central respiratory chemosensitivity is notably weaker during the neonatal period, a fact that could conceivably be explained by the lesser ability of immature glial cells to release ATP upon acidification. In the final analysis, much remains to be done to understand how CO₂ activates the RTN neurons at the molecular and cellular level. These mechanisms, glial or otherwise, need not be unique to the retrotrapezoid nucleus, however. The wiring of the RTN neurons and the strength of their synaptic connections could be the main reasons behind their importance to respiratory chemoreception.