A cholinergic spinal pathway for the adaptive control of breathing

Abstract

The ability to amplify motor neuron (MN) output is essential for generating high intensity motor actions. This is critical for breathing that must be rapidly adjusted to accommodate changing metabolic demands. While brainstem circuits generate the breathing rhythm, the pathways that directly augment respiratory MN output are not well understood. Here, we mapped first-order inputs to phrenic motor neurons (PMNs), a key respiratory MN population that initiates diaphragm contraction to drive breathing. We identified a predominant spinal input from a distinct subset of genetically-defined V0_C cholinergic interneurons. We found that these interneurons receive phasic excitation from brainstem respiratory centers, augment phrenic output through M2 muscarinic receptors, and are highly activated under a hypercapnia challenge. Specifically silencing cholinergic interneuron neurotransmission impairs the breathing response to hypercapnia. Collectively, our findings identify a novel spinal pathway that amplifies breathing, presenting a potential target for promoting recovery of breathing following spinal cord injury.

Figure 1.. Distribution of monosynaptic inputs to phrenic motor neurons (PMNs)

(A) Tracing strategy for mapping monosynaptic PMN inputs. (B) Examples of mCherry-labeled brainstem (rVRG) and spinal cord interneurons projecting to PMNs. moXII: hypoglossal motor nucleus, NA: Nucleus Ambiguus, cc: central canal. Scale bar = 200 μm. (C) Distribution of PMN inputs in the brainstem and spinal cord (n = 7). (D) Quantification of the rostrocaudal distribution of spinal inputs to PMNs. (E) Spinal cord PMN inputs are largely localized to the ipsilateral side. (F) Dorsoventral distribution of PMN inputs in the spinal cord. (G) Cholinergic interneurons (ChAT+ INs) around the central canal (cc) in the cervical spinal cord directly project to PMNs. Scale bar = 200 μm (top) and 50 μm (bottom). (H) Rostrocaudal distribution of ChAT+ INs projecting to PMNs (ChAT+ INs → PMNs). (I) ChAT+ INs comprise ~10% of total PMN inputs. (J) Quantification of contralateral and ipsilateral ChAT+ INs projecting to PMNs. (K) Quantification of the rostrocaudal distribution of ChAT+ INs projecting to PMNs. (L) ChAT+ INs from the brachial and thoracic spinal cord directly project to PMNs. Scale bar = 200 μm (top) and 50 μm (bottom).

Figure 2.. ChAT+ INs that project to PMNs are morphologically and topographically distinct.

(A) Transsynaptic retrograde labeling of ChAT+ INs projecting to PMNs (ChAT+ INs → PMNs) and limb (biceps)-innervating MNs (ChAT+ INs → LMNs). (B-C) Representative images of contralateral (B) and ipsilateral (C) ChAT+ INs → PMNs. Scale bar = 100 μm. (D) Representative image of ChAT+ INs → LMNs. Scale bar = 100 μm. (E) Topographical distribution of ChAT+ INs → PMNs (n = 86 cells from 7 mice) and ChAT+ INs → LMNs (n = 69 cells from 5 mice). Rectangular region is enlarged to the right. (F) Quantification of ChAT+ INs → PMNs and ChAT+ INs → LMNs horizontal distance to the central canal. (G-H) Reconstruction of ChAT+ INs → PMNs (G) and ChAT+ INs → LMNs (H) morphology. Scale bar = 50 μm. (I) Sholl analysis of ChAT+ INs → PMNs (n = 36 cells) and ChAT+ INs → LMNs (n = 35 cells). (J-N) ChAT+ INs → PMNs have higher maximum Sholl intersections (J), greater overall dendritic length (K), cover a larger area (L), and have higher maximum branch level (M) and greater maximum branch depth (N) compared to ChAT+ INs → LMNs. ** p < 0.01, **** p < 0.0001

Figure 3.. Pitx2-derived cholinergic synapses on PMNs.

(A) Transverse section of the cervical spinal cord showing retrograde CTB labeling (green) in PMNs (squared region). Scale bar = 100 μm. (B) Enlargement of CTB-labeled PMNs (green) shown in (A). Square region indicates a PMN shown in (C). Scale bar = 20 μm. (C) Enlargement of a single CTB-labeled PMN (CTB, green) shown in (B) from a Pitx2^tdTom adult mouse with synapses derived from Pitx2^Tdtom+ interneurons (red) and cholinergic synapses (VAChT, blue). Numbers indicate labeled synapses on the PMN soma. Scale bar = 5 μm (D) Magnification of numbered synapses from (C) showing colocalization of tdTomato and VAChT. Scale bar = 1 μm. (E-F) Representative images of VAChT+ puncta on PMNs in P4 Pitx2^tdTom mice. Square region in (E) is enlarged in (F). Scale bar = 100 μm (E) and 20 μm (F). (G) Enlargement of the square region in (F). Numbers indicate examples of synapses on the PMN soma. Scale bar = 5 μm. (H) Magnification of numbered synapses from (G). Scale bar = 1 μm. (I) Percentage of VAChT+ puncta that are Pitx2^tdTom+ on PMN somas at P4. (J-K) Representative images of VAChT+ puncta on PMNs in adult (P60) Pitx2^tdTom mice. Square region in (J) is enlarged in (K). Scale bar = 100 μm (J) and 20 μm (K). (L) Enlargement of the square region in (K). Numbers indicate examples of synapses on the PMN soma. Scale bar = 5 μm. (M) Magnification of numbered synapses from (L). Scale bar = 1 μm. (N) Percentage of VAChT+ puncta that are Pitx2^tdTom+ on PMN somas at P60.

Figure 4.. A subset of cervical Pitx2+ interneurons are integrated within respiratory circuits.

(A) Schematic of the experimental setup showing extracellular ventral root recordings and intracellular whole-cell patch clamp recordings from individual Pitx2^tdTom+ interneurons (red) in hemisected brainstem-spinal cord preparations from Pitx2^tdTom neonatal mice. (B) Pie charts showing the relative proportion of respiratory-related (red) and non-respiratory-related (gray) Pitx2+ interneurons within and below the C3–4 spinal segments. (C) Example trace of voltage-clamp recording from a respiratory-related Pitx2+ interneuron (top) during ongoing respiratory burst (bottom). Red dotted box showing zoomed-in traces during a respiratory burst. (D) as in (C) but regarding current-clamp recording.

Figure 5.. Effect of methoctramine on the respiratory motor output.

(A) Brainstem-spinal cord preparation from neonatal mice. (B) Raw (top) and integrated/rectified (bottom) traces from the C4 ventral root during baseline, methoctramine (10 μM) and washout. a-c boxes indicate 40 seconds of recording at the end of each condition (black = baseline; red = methoctramine; grey = washout), expanded at the bottom. (C) Average respiratory-burst amplitude over the last 10 minutes during baseline (black), 10 μM methoctramine (red), and washout (grey); black lines show mean and SD (n = 7). (D) as in (C) but showing respiratory-burst frequency. (E) Experimental design to block M2 receptors at spinal levels only in the brainstem-spinal cord preparation from neonatal mice. (F) as in (B) but showing the effect of methoctramine at spinal level only. (G) and (H) as in (C) and (D) but showing the effect of methoctramine at spinal level only (n = 10). (I) Working heart-brainstem preparation from adult rats. (J) as (B) and (F) but showing the phrenic neurogram trace from the working heart-brainstem preparation. Square boxes indicate 7 seconds of recording at the end of each condition (black = baseline; red = methoctramine; grey = washout), expanded at the bottom. (K) and (L) as in (C) and (D) but showing the effects of methoctramine in the adult preparation. Data analyzed with mixed-effect model and Holm-Šídák’s multiple comparisons test. *p < 0.05, **p < 0.001

Figure 6.. ChAT+ INs are highly activated under a hypercapnic gas challenge

(A) ChAT+ MNs and INs are labeled by GFP in ChAT::eGFP mice. Regions including PMNs and ChAT+ INs around the central canal are shown in (B) and (C), respectively. (B-C) Both PMNs (B) and ChAT+ INs (C) are highly activated, as indicated by high c-Fos expression (red), after exposure to 10% CO₂. White arrows indicate c-Fos+ ChAT+ INs. Scale bar = 40 μm. (D) PMN activation is positively correlated to ChAT+ IN activation. (E, G, and I) Percentage of c-Fos+ PMNs after 10% CO₂ for 1 hour (E), 10% CO₂ for 15 minutes (G), and 5% CO₂ for 15 minutes (I). (F, H, and J) Number of c-Fos+ ChAT INs after 10% CO₂ for 1 hour (F), 10% CO₂ for 15 minutes (H), and 5% CO₂ for 15 minutes (J). * p < 0.05, ** p < 0.01, *** p < 0.001

Figure 7.. Cholinergic interneuron silencing impairs the response to hypercapnia

(A) Experimental design for whole body plethysmography and schematic of a representative breath. A Pitx2^ΔChAT or Dbx1^ΔChAT mouse was paired with a sex-matched control littermate and the mice were exposed to normal air conditions (21% O₂, 79% N₂) for 45 minutes, followed by 15 minutes of hypercapnia (5% CO₂, 21% O₂, 74% N₂ or 10% CO₂, 21% O₂, 69% N₂). (B-C) Examples of breath traces under normal air and 10% CO₂ in Pitx2^ΔChAT mice and their control littermates. A single breath was enlarged in (C). (D, H, L, and P) Breath frequency (D), minute ventilation (H), tidal volume (L), and PIF (P) distribution under normal air in Pitx2^ΔChAT and control mice (n = 14–17 per group). (E, I, M, and Q) Mean and normalized frequency (E), minute ventilation (I), tidal volume (M), and PIF (Q) under normal air in Pitx2^ΔChAT and control mice. (F, J, N, and R) Breath frequency (F), minute ventilation (J), tidal volume (N), and PIF (R) distribution under 10% CO₂ in Pitx2^ΔChAT and control mice (n = 14–17 per group). (G, K, O, and S) Mean and normalized frequency (G), minute ventilation (K), tidal volume (O), and PIF (S) under 10% CO₂ in Pitx2^ΔChAT and control mice.