Ndufs4 inactivation in glutamatergic neurons reveals swallow-breathing discoordination in a mouse model of Leigh Syndrome

Abstract

Swallowing, both nutritive and non-nutritive, is highly dysfunctional in children with Leigh Syndrome (LS) and contributes to the need for both gastrostomy and tracheostomy tube placement. Without these interventions aspiration of food, liquid, and mucus occur resulting in repeated bouts of respiratory infection. No study has investigated whether mouse models of LS, a neurometabolic disorder, exhibit dysfunctions in neuromuscular activity of swallow and breathing integration. We used a genetic mouse model of LS in which the NDUFS4 gene is knocked out (KO) specifically in Vglut2 or Gad2 neurons. We found increased variability of the swallow motor pattern, disruption in breathing regeneration post swallow, and water-induced apneas only in Vglut2 KO mice. These physiological changes likely contribute to weight loss and premature death seen in this mouse model. Following chronic hypoxia (CH) exposure, swallow motor pattern, breathing regeneration, weight, and life expectancy were not changed in the Vglut2-Ndufs4-KO CH mice compared to control, indicating a rescue of phenotypes. These findings show that like patients with LS, Ndufs4 mouse models of LS exhibit swallow impairments as well as swallow-breathing dyscoordination alongside the other phenotypic traits described in previous studies. Understanding this aspect of LS will open roads for the development of future more efficacious therapeutic intervention for this illness.

Keywords: Hypoxia; airway protection; dysphagia; mitochondrial disease.

Figure 1.

Alteration in, and variability of swallow motor pattern in Vglut2-Ndufs4-KO mice exposed to room air. A) Representative traces of water triggered swallow in control (blue), Vglut2-Ndufs4-KO (green), and Gad2-Ndufs4-KO (pink). Total swallow duration: f-a; submental complex (SC) duration: c-a; laryngeal complex (LC) duration: f-d; and swallow sequence: e-b. B) Graphical depiction of the preparation with location of all recorded muscles and nerves. The F test to compare variances indicated a statistically significant variance between Vglut2-Ndufs4-KO and control room air (RA) mice in C) the total swallow duration, D) SC duration, and E) LC duration indicated with the bracket and asterisk to the right of the bar graphs. F) Overlay of representative trace of a swallow from each animal in control RA (blue, top) and Vglut2-Ndufs4-KO RA (green, bottom) to illustrate the variable swallow motor patten in the Vglut2-Ndufs4-KO mice but not the control mice. Ramp duration was calculated as h-g in both SC and LC; and decay duration from i-h. LC in Vglut2-Ndufs4-KO mice have a significantly longer duration (E), likely due to the significantly longer ramping of the muscle activity (F) compared to the control RA mice.

Figure 2.

Disruption of respiratory rhythm regeneration following a swallow in Vglut2-Ndufs4-KO room air mice. Respiratory rhythm regeneration was determined by calculating the inter-burst-interval (IBI) for 50 cycles following a swallow. A) Plot of the average diaphragm IBI following a swallow for control (blue), Vglut2-Ndufs4-KO (green), and Gad2-Ndufs4-KO (pink) exposed to room air. Vglut2-Ndufs4-KO mice have significantly longer IBI compared to control mice from breaths 1–28 and increased variability (shaded area). The red arrow indicates the first breath post swallow termed B) swallow related (SR) inspiratory delay. Vglut2-Ndufs4-KO mice have a significantly longer and variable (asterisk on the side) duration in which breathing regenerates following a swallow. C) Representative traces of control (blue) and Gad2-Ndufs4-KO (pink) of breathing regeneration following a swallow induced by water (down arrow). D) Representative traces of 4 Vglut2-Ndufs4-KO mice indicating the variability of breathing regeneration and 4 cycles following a swallow. Schluckatmung, a less known behavior describes diaphragm activity during swallow (55, 56). SR expiration duration d-a, SR inspiratory delay c-b, diaphragm IBI d-c.

Figure 3.

Characteristics of water induced apneas seen only in Vglut2-Ndufs4-KO mice exposed to room air. Apnea was defined as quiescence of the diaphragm two times the average duration of the respiratory IBI across 10 respiratory cycles. A) representative trace of type 0 (T0) where injection of water (arrow) did not induce apnea but a swallow. B) Representative trace of type 1 (T1), defined as a water induced diaphragm quiescence one time the average IBI duration prior to a swallow, occurred 3 times in 2 of the 7 mice. C) Duration of diaphragm quiescence before and D) after a swallow. E) Representative trace of type 2 (T2), defined as water induced apneas with the absence of swallow or any upper airway behavior, occurred 4 times in 3 of 7 mice. F) Representative trace of type 3 (T3), defined as diaphragm quiescence following water injection with tonic laryngeal activation, occurred 14 times in 3 out 7 mice. G) Bar graph representing the percentage of Vglut2-Ndufs4-KO mice exposed to room air that had each type of water induced apnea. H) A one-way ANOVA indicated no statistical difference in the duration of water induced apneas across all three types, which ranged from 0.65s to 28s with an average of 4 ± 7s. I) Average heart rate (HR) 10s prior to and during the water induced apnea. In 19 of the 21 apnea bouts (green), we found a slight but significant decrease in HR during the apnea bout (540 ± 38 bpm, 534 ± 35bpm, p = 0.001) compared to 10s prior to apnea with an average percent change of −10 ± 12%. Two of the 21 apneas (black) lasted longer than 20s and had on average a 174% decrease in HR. J) A Pearson correlation found no significant correction between length of apnea and % change in HR.

Figure 4.

Exposure to chronic hypoxia (CH) increases life expectancy, maintains body weight, and prevents alteration and variability in swallow motor pattern. A) A Mantel-Cox Log-rank test revealed a significant increase in survival for Vglut2-Ndufs4-KO mice exposed to CH (p <0.0001). Purple bar at the top indicated time in CH. B) We found no significant difference in body weight of control mice exposed to room air (RA), control mice exposed to CH, or Vglut2-Ndufs4-KO mice exposed to CH. C) CH prevents variability, seen in RA (Figure 1F) of swallow motor patterning with no change in ramp or decay of the submental and laryngeal complex activity. D) CH prevents the increase in total swallow, submental and laryngeal complex duration seen in Vglut2-Ndufs4-KO mice exposed to room air.

Figure 5.

Chronic hypoxia (CH) rescues delayed inspiratory regeneration following a swallow but does not fully restore diaphragm inter-burst-interval (IBI) variability. A) Plot of the average diaphragm IBI following a swallow for control (gold) and Vglut2-Ndufs4-KO (purple) mice exposed to CH and Vglut2-Ndufs4-KO (green) exposed to room air. Vglut2-Ndufs4-KO mice exposed to RA have significantly longer IBI compared to control and Vglut2-Ndufs4-KO mice exposed to CH from breaths. There is no difference in IBI duration between control and Vglut2-Ndufs4-KO mice exposed to CH. The red arrow indicates the first breath post swallow termed B) swallow related (SR) inspiratory delay which is significantly longer in Vglut2-Ndufs4-KO RA mice compared to control CH mice, but there is no difference between Vglut2-Ndufs4-KO and control CH mice. Figure 2 and C) depicts the variability of IBI duration in Vglut2-Ndufs4-KO RA mice (green) which is only partially rescued by CH (purple). We see little differences in the variability between control RA (blue) and CH (gold) mice with the Gad2-Ndufs4-KO RA mice (pink). D) Representative traces of control and Vglut2-Ndufs4-KO mice exposed to CH showing no change in IBI and inspiratory delay in water-triggered swallows.

Figure 6.

Microglia in breathing and swallow- related medullary centers. Using Iba-1 we stained for microglia in 1) the ventral respiratory column (VRC), 2) postinspiratory complex (PiCo), and 3) the nucleus of the Solitary Tract (NTS) of A) control room air (RA), B) Vglut2-Ndufs4-KO RA, C) control chronic hypoxia (CH), and D) Vglut2-Ndufs4-KO CH mice. A one-way ANOVA with Tukey’s multiple comparison test revealed a significant decrease in E) the density of Iba-1 positive cells in Vglut2cre-Ndufs4-KO mice in the VRC area compared to RA control mice (p = 0.05). F) There was a decrease in the number of Iba-1 positive cells in Vglut2cre-Ndufs4-KO CH mice in PiCo compared to Vglut2cre-Ndufs4-KO RA (p = 0.01). G) There were no changes to the number or density of microglia in the NTS. Abbreviations: cAmb, nucleus ambiguus pars compacta; Sp5, spinal trigeminal nucleus; AP, area postrema; IO, inferior olive; cc, central canal; XII, hypoglossal nucleus.