Our scientific program is at the interface of Developmental and Integrative Neurosciences.
Breathing is an essential behavior that offers a unique opportunity to understand how the nervous system functions normally, how it balances robustness with regulated lability, how it adapts to changing conditions, and how dysfunctions result in disease. We base our research program on recent advances concerning the two essential sites for respiratory rhythmogenesis : the RTN and preBötC oscillators, bilateral collectives of about 1000 excitatory glutamatergic neurons that constitute unique developmental and physiological objects in the CNS whose vital importance rests on a small number of genetically and optically accessible neurons.
Function of the RTN
We have provided molecular and developmental signatures for the RTN : Phox2b and Atoh1 (Dubreuil et al. 2009)(two transcriptional determinants of these cells), the NKR1 receptor, the VgluT2 transporter. Moreover, these cells derived from Egr2+ progenitors(in rhombomeres 3 or 5)(Thoby-Brisson et al. 2009), and Lbx1+ (i.e. dB2) precursors. We have physiologically analyzed a mouse model of Congenital Central Hypoventilation Syndrome (CCHS), a disease caused by a mutation in PHOX2B that entails an abrogation of CO2 sensitivity (the paramount drive to breathe) and neonatal respiratory arrests : these mice reproduce the human respiratory syndrome and have an atrophic RTN (Ramantsoa et al. 2011). We have analyzed two other genetic backgrounds that lesion the RTN and phenocopy the respiratory syndrome, further strengthening the case for a dual chemosensory/oscillatory role of the RTN in vivo (Goridis et al. 2011).
Despite these advances, evidence for the RTN roles is still tentative, because none of the genetic lesions is strictly specific. Exquisite specificity is essential since many other neurons, including serotonergic and other Phox2b+ interneurons, are implicated in the still popular theory of “distributed chemosensitiviy”. Furthermore, the role of the RTN in respiratory rhythm generation is still a matter of heated debate. In order to gain foolproof genetic evidence for the requirement, absolute or partial of the RTN, we will assay gain and loss of function on ex-vivo preparations of the embryonic medulla, in which one can measure parameters of the respiratory phenotype and investigate lesions that are lethal at birth. The “gain of function” will consist in exciting the RTN by optogenetics after expression of ChannelRhodopsin2 under the control of the promoter of VgluT2, in a transgenic line, or after cre-mediated recombination increasingly selective for the RTN (that combine the criteria of Phox2b and, Atoh1 expression).
The capacity to entrain the preBötC and/or expiratory pre-motor neurons will be evaluated by electrophysiological recordings and calcium imaging. The “loss of function” of the RTN will be achieved by intersectional transgenic lesions that combine two, or even three, molecular criteria will be combined (their history of expression of Phox2b and Atoh1 (and either Egr2 or Lbx1) to conditionally kill RTN cells with a toxic allele of Phox2b (Phox2b27Ala), or to acutely and reversibly inhibit them, with allatostatin or by optogenetics using an Arch-EGFP-ER2 floxed line.
The respiratory CPG features both robustness and regulated lability to ensure metabolic homeostasis : it is permanently active in mammals, in all arousal and behavioral states and even following loss of cortical function or decerebration, and also adapt to non homeostatic uses of breathing such as vocalization, and will generate about half a billion breaths in a human lifespan. However, the functional connectivity underlying the cyclical intercellular spread of activity within the oscillator resulting in coherent population activity (respiratory burst) remains largely unknown.
The combination of light microscopy with functional reporters, caged compounds and, more recently, optogenetics offers the possibility to control activation and inhibition of neuronal activity and monitor functional responses in a noninvasive manner enabling the analysis of well-defined neuronal population in intact neuronal circuits. We are using dynamical reconfiguration of the optical wavefront that allows efficient multiple-scale excitation from single to multiple diffraction limited spots, to shapes covering a single subcellular process, or a population of sparse neurons.
We want to test a structural hypothesis accounting for the robustness and flexibility of the oscillators, whereby their function depends on the activity of a small subset of their consituent neurons. This is prompted bya remarkable preliminary result : we can reset the activity of the RTN by 1P patterned light targeting as little as a single digit number of ChR2-expressing RTN neurons. We will exploit the possibility, for the first time, to reversibly and selectively stimulate defined numbers of channelrhodopsin-2 (ChR2)- expressing cells and optically record the activity of responding neurons to derive the “links” connecting the “nodes”, thus defining the structure of the oscillators’ network.
This will be done comparatively in the RTN and the preBötC oscillators whose internal activities respectively rely on electrical and glutamatergic synapses. Opsin expression in RTN and preBötC neurons has been validated using the straight VGluT2 ::ChR2-YFP transgenic line and multiple conditional lines previously mentioned recombining a floxed ChR2-tdTomato allele. Conversely, we will also through using an Arch-EGFP-ER2 floxed line test how many oscillator neurons need to be silenced to observe a collapsing of the oscillator rhythm.
If these studies are successful, they lay the foundation for a host of powerful, heretofore impossible, experiments and provide the basis for novel models that have the potential to lead to a major breakthrough in understanding of the neural control of breathing. In particular, if we are right that a shared property of oscillators is their particular sensitivity to minimal inputs through their structural ability to efficiently operate cellular recruitments, then the question is open as to whether this may be a mechanism through which the rhythm may be maintained or actually be at all generated.
Complementary to the physiological analysis of whole networks or individual neurons, novel tools are now available to decipher connectivity matrices. These include rabies viruses deleted for the glycoprotein (G) essential for its transsynaptic spread. Complementation for this defect in a given neuron restores the ability for transsynaptic spread from that neuron to its presynaptic partners only. This technique was successfully used to trace the distribution of premotor neurons in the spinal cord after injection of a neurotrope viral cocktail in skeletal muscles. We are applying this approach to the motor output of respiratory circuits.
A viral cocktail comprising a G-deficient rabies virus encoding a fluorescent protein (mCherry or GFP) and an Adeno-Associated Virus encoding G for complementation is being injected in the diaphragm to mark the premotor descending inspiratory neurons. Characterization of the molecular identity and developmental origin of these neurons is enabled by the use of battery of Cre-recombinases coupled with Cre-sensitive reporters. Respiratory premotor systems have been described previously using many approaches (electrophysiology through in vivo cross correlation analysis, antidromic stimulations in rats and cats, rabies virus (street virus) trans-synaptic tracing, retrograde tracing with horseradish peroxidase etc…) producing to a wealth of qualitatively disparate and hard to reconcile data on descending “bulbospinal” premotor systems linked to inspiratory and expiratory control. Our ongoing effort will remedy this situation and should draw a definitive anatomic picture.
Beyond motoneurons we intend to perform monosynaptic retrograde tracing from interneurons in the CNS. For this we will use additional technical advances. The site of primary infection can be restricted to a selected neuronal population by pseudotyping the G-defective virus with a foreign envelope protein (EnvA) and providing the receptor (TVA, produced by birds, not mammals) to the selected neuronal population only.
Originally, this receptor has been provided virally, selectivity being achieved through the combination of a specific promoter and stereotaxic injection. More recently Martyn Goulding has developed a mouse line (the double HTB line, now in the lab) in which both TVA and the rabies G protein (B19) are provided by a transgene, conditionally to both Cre and flp-mediated recombination. In that way, only neurons defined by one or two molecular criteria will allow retrograde transynaptic spread after stereotaxic injection of the virus.
We should be in a position to selectively identify the presynaptic partners of respiratory oscillator neurons (defined by Dbx1 expression for the preBötC and by Atoh1 and Phox2b expression for the RTN), thereby also documenting their intra- and inter-connectivities. These approaches, using an array of recombinases, should allow us to reconstruct, the anatomy of respiratory oscillators and of the structures impinging on them, in terms of neuronal populations with a defined molecular code but also a defined developmental history.