Respiratory disturbances and high risk of sudden death in the neonatal connexin‐36 knockout mouse

Figures

• 

  FIGURE 1. Breathing instability in the neonatal Cx36KO mice at rest. (a) Breathing of most Cx36KO mice at P14 (86%) exhibited a slower and more irregular eupneic respiration at rest; (i‐iii) example of raw chest motion records with the corresponding raster and histogram of autocorrelation of breathing cycles, BC; (iv) average of the normalized histogram of breathing periods, BP (N = 10 per group, 10 min; *p < 0.0001, ANOVA). (b) The remaining Cx36KO mice (14%) showed a periodic Cheyne–Stokes breathing with recurrent episodes of hypopneas (h) and crescendo‐decrescendo hyperpneas (H)

• 

  FIGURE 2. Exacerbated chemoreflexes to hypoxia (a), normoxic and hypoxic hypercapnia (b and c) in the neonatal Cx36‐knockout mice. Superimposed responses of wild type and Cx36‐KO mice (in blue and red): (i) breathing rate (BR); (ii) index of tidal volume (∆VT), (iii) arterial O2 saturation (SpO2), and (iv) transcutaneous CO2 partial pressure (PtcCO2). Cx36‐KO mice responded to hypoxia with a sustained increase of BR instant of a fully adapted response of WT mice, and to hypercapnia in normoxia and hypoxia with greater increases in BR and ∆VT. Note that while hypoxia (8% O2) induced similar O2 desaturation in both genotypes, hypoxic hypercapnia (8% O2 and 8% CO2) provoked a more pronounced O2 desaturation in Cx36KO than WT mice. PtcCO2 decreased during hypoxia and increased during hypercapnia in normoxia and hypoxia equally in the two groups (N = 24 per group; in A‐i ♦, p = 0.0215‐<0.0001; in B‐i ♦, p = 0.0495–0.0045; in C‐i ♦, p = 0.0457–0.0098; in B‐ii ♦, p = 0.0448–0.0078; in C‐ii ♦, p = 0.0484–0.0140, and in C‐iii ♦, p = 0.0492–0.0015; t‐Student test)

• 

  FIGURE 3. Increased CO2 sensitivity of neonatal Cx36KO. Wild type and Cx36KO mice (in blue and red) were submitted to increasing concentrations of CO2 in inhaled air (4%, 8%, and 12%) in normoxia (a) and 8% O2 hypoxia (b); (i and ii) breathing rate (∆BR) and index of tidal volume (∆V T; last 2 of 10 min stimulus) normalized relative to basal values. The slopes of normalized breathing rate (∆HR) and index of tidal volume (∆V T) as function of CO2 concentration were ≈twofold greater in Cx36KO than WT mice for both normoxic and hypoxic hypercapnia (N = 15 per group; in A‐i *p = 0.0478; **p = 0.0005 and ***p = 0.0125; in A‐ii *p = 0.0024; in B‐i *p = 0.0216; **p = 0.0216; ***p = 0.0417; in B‐ii, *p = 0.0291; t‐Student test)

• 

  FIGURE 4. Respiratory motor pattern of Cx36KO versus WT mice during hypoxia (a), and normoxic and hypoxic hypercapnia (b and c). (i and iii) Representative raw records of respiratory chest motion (Vent), diaphragm, and intercostals activities (EMGD and EMGI), (ii and iv) averages (20 respiratory cycles) of Vent and rectified EMGs (∫EMGD and ∫EMGI) from WT and Cx36KO mice (in blue and red), (v) averaged duration of inspiratory phase (I) and expiratory phase (E), subdivided in one active (E 2) and two passive post‐inspiratory and pre‐inspiratory expiratory components (E 1 and E 3), (vi–viii) peak‐intensity of rectified inspiratory activity of the diaphragm (∆∫EMGD) and external intercostal muscles (∆∫EMGEI), and expiratory activity of internal intercostals (∆∫EMGII) normalized relative to their values at rest. Cx36KO mice responded to hypoxia with a greater inspiratory activity and shorter breathing cycles and expiratory phases, and to normoxic and hypoxic hypercapnia with more intensive and phasic inspiratory and expiratory motor activities in anti‐phase (N = 15 per group in A–C v; in A‐v *p < 0.0001, and ***p = 0.0082; in B‐v *p = 0.0002, and **p = 0.0008; and in C‐v *p = 0.0001 t‐Student…

• 

  FIGURE 5. Cardiovascular disturbances during chemoreflexes to hypoxia (a), and normoxic and hyperoxic hypercapnia (b and c) in the neonatal Cx36‐knockout mice. Superimposed responses of wild type and Cx36‐KO mice (in blue and red): (i) heart rate (HR), (ii) heart rate variability normalized relative basal value (∆VHR), and (iii) pulse distention (PD). Cx36KO mice responded to hypoxia with a similar initial tachycardia followed by a smaller slow building‐up bradycardia and early trough post‐hypoxia (arrows a‐c); hypercapnia in normoxia provoked similar sharp bradycardia in WT and Cx36KO mice but cardiac variability was less pronounced in Cx36KO mice; during hypoxic hypercapnia, the initial small tachycardia was similar in both genotypes but the sharp bradycardia was damped and the subsequent partial recovery of HR increased more in Cx36KO mice (arrows a‐c); PD decreased more during hypoxia and increased less during hypoxic hypercapnia in Cx36KO than WT mice (N = 25 per group; in A‐i ♦, p = 0.0471–0.0012; in A‐ii ♦, p = 0.02280–0.0004; in A‐iii ♦, p = 0.0485–0.0422; in B‐ii ♦, p = 0.0389–0.0005; in C‐i ♦, p = 0.0491–0.0168; in C‐ii ♦, p = 0.0369‐ < 0.0001, and in C‐iii ♦, p = 0.0497‐ <…

• 

  FIGURE 6. Cardiorespiratory coupling during normoxic and hypoxic hypercapnia. (A) Raw records from WT and Cx36KO mice (in blue and red) of respiratory motion (Vent), cardiac activity (ECG), and instantaneous breathing and heart rates (BR and HR) selected from the last 2 of 10 min exposition to hypercapnia‐normoxia (i and iii) and hypercapnia‐hypoxia (ii and iv) showing repetitive episodes of transient bradycardia (ETBs; arrows). (B) Raster of ETBs generated by WT and Cx36KO mice (right and left) along 10 min of hypoxic hypercapnia (i and iv), heart rate average of ETBs generated during last 2 min (HR, ii and v), cross‐correlation of respiration (Vent) with the onset of ETBs (iii and vi). HR during ETBs first increased slightly, and then dropped abruptly and recovered slowly (arrows a–c); in comparison, Cx36KO mice generated ETBs of smaller intensity and its number declined rapidly with the time of exposition to hypoxic hypercapnia. ETBs onset was phase‐locked with the expiration phase of the breathing cycle in both genotypes. (C) Probability of ETBs occurrence within the normalized breathing cycles of WT and Cx36KO mice (i and ii); Vent, ventilatory motion, and ∫EMGD and ∫EMGI, rectified…

• 

  FIGURE 7. Effect of CO2 concentration in the generation of ETBs. (A) Number of ETBs generated by WT and Cx36KO mice (blue and red) during 10 min of hypercapnia alone and in combination with hypoxia (i and ii). During hypercapnia‐normoxia, ETBs number increased rapidly after stimulus starting to greater steady‐state values as larger CO2 concentrations were applied in inhaled air; in comparison, Cx36KO mice generated significantly fewer ETBs. During hypercapnia‐hypoxia, the EBTs rate tended to decline with the time of exposition to hypoxic hypercapnia but more markedly in the case of Cx36 (N = 12 per group; in A i and ii *p < 0.0001; ANOVA). (B) Example of interval histograms of ETBs generated by WT and Cx36KO mice (blue and red) in hypercapnia‐normoxia (i and ii) and hypercapnia‐hypoxia (iii and iv). Note that ETBs tended to generate at a fixed interval (broken line) which probability increased with CO2 level

• 

  FIGURE 8. ECG signal‐based Poincaré plots of ETBs as a function of CO2 concentration. Poincaré diagrams (RRn vs. RR n +1 in x and y‐axis) constructed with steady‐state data of ETBs (last 2 min) in normoxic and hypoxic hypercapnia (A and B) from WT and Cx36KO mice (i and ii). Dynamics of interval sequence during ETBs showed an initial small shortening of RR interval (a→b) followed by a longer lengthening of next RR (b→c), and then a progressive recovery toward initial RR value in successive cardiac cycles (c→d→a); this elongation of RR increased similarly with CO2 concentration up to double and triple of the basal RR interval for WT and Cx36KO mice in normoxic hypercapnia and WT mice in hypoxic hypercapnia; note that in the case of the Cx36KO mice this bradycardic effect was almost completely absent during hypoxic hypercapnia

• 

  FIGURE 9. Functional role of ETBs. The impact of ETBs on oxyhemoglobin (OHb) content per cardiac pulse was evaluated in WT and Cx36KO mice (in blue and red) during normoxic and hypoxic hypercapnia (A and B). (i) Raster synchronized with ETB onset of ventilatory motion (Vent), heart beating (ECG), and oxyhemoglobin pulse‐to‐pulse signal (OHb), and (ii) cross‐correlation function of OHb signal with ETBs. ETBs caused a marked transient increase in OHb content per pulse in both phenotypes in hypercapnia‐normoxia (arrows), but during hypoxic hypercapnia only in WT mice

• 

  FIGURE 10. Sudden respiratory arrest induced by hypoxic hypercapnia in the neonatal Cx36‐KO mouse. (A) Survival curves of homozygous (Cx36−/Cx36−) and heterozygous (Cx36−/Cx36+) versus wild‐type (Cx36+/Cx36+) mice submitted to a protocol of severe hypoxia in combination with increasing CO2 concentrations (upper); homozygous and heterozygous Cx36KO but not WT mice were susceptible to succumbed by respiratory arrest (0 out 18 Cx36+/Cx36+, 5 out 9 Cx36−/Cx36+, and 15 out 26 Cx36−/Cx36−; *p = 0.0021 Cx36−/Cx36− vs. Cx36+/Cx36+ and **p = 0.0017 Cx36+/Cx36− vs. Cx36+/Cx36+, log‐rank test). (B) Stages leading to sudden respiratory arrest: stage 1, eupneic depression; stage 2, paroxysmal respiratory motor activity; and stage 3, terminal gasping; BR, breathing rate and Vent, respiratory chest motion. Inset, expanded records of respiratory motion (Vent), diaphragm, and intercostal muscles (EMGD and EMGI), and cardiac activity (ECG) showing the transition from eupnea to gasping mode (arrowheads and arrows) and the sudden cardiac rate drop occurring during the paroxysmal clonic‐tonic activity (a and b in stage 2). (C) Superimposed average responses to 8% O2 and 8% CO2 of survival WT and Cx36KO mice (N…