The vesicular glutamate transporter VGLUT3 contributes to protection against neonatal hypoxic stress

Abstract

Neonates respond to hypoxia initially by increasing ventilation, and then by markedly decreasing both ventilation (hypoxic ventilatory decline) and oxygen consumption (hypoxic hypometabolism). This latter process, which vanishes with age, reflects a tight coupling between ventilatory and thermogenic responses to hypoxia. The neurological substrate of hypoxic hypometabolism is unclear, but it is known to be centrally mediated, with a strong involvement of the 5-hydroxytryptamine (5-HT, serotonin) system. To clarify this issue, we investigated the possible role of VGLUT3, the third subtype of vesicular glutamate transporter. VGLUT3 contributes to glutamate signalling by 5-HT neurons, facilitates 5-HT transmission and is expressed in strategic regions for respiratory and thermogenic control. We therefore assumed that VGLUT3 might significantly contribute to the response to hypoxia. To test this possibility, we analysed this response in newborn mice lacking VGLUT3 using anatomical, biochemical, electrophysiological and integrative physiology approaches. We found that the lack of VGLUT3 did not affect the histological organization of brainstem respiratory networks or respiratory activity under basal conditions. However, it impaired respiratory responses to 5-HT and anoxia, showing a marked alteration of central respiratory control. These impairments were associated with altered 5-HT turnover at the brainstem level. Furthermore, under cold conditions, the lack of VGLUT3 disrupted the metabolic rate, body temperature, baseline breathing and the ventilatory response to hypoxia. We conclude that VGLUT3 expression is dispensable under basal conditions but is required for optimal response to hypoxic stress in neonates.

Figure 1. Lack of Vglut3 does not alter the histological organization and function of respiration-related areas under basal conditions

See Supplemental Fig. 1 for schematic representation of the brainstem sections. A and B, immunolabelling showing that NK1R (red) in the pre-Bötzinger complex (Aa and Ba) and at the midline raphe level (Ab and Bb) and Phox2B (red) at the RTN/pFRG level (Ac and Bc) have similar expression patterns in control and Vglut3^−/− pups. Anatomical limits of the facial nucleus have been superimposed on Ac and Bc section (dotted line, Lazarenko et al. 2009). C, examples of phrenic bursts (PBs) recorded from the C4 root of in vitro brainstem preparations from control (a) and Vglut3^−/− (b) pups. Int C4: integrated C4 activity. D and E, neither phrenic burst frequency (PBf) (D) nor the irregularity score (E) was significantly affected by the Vglut3 mutation under basal conditions (control n = 64; Vglut3^−/− n = 22). F, examples of recordings of respiratory variables using plethysmography in one control (a) and one Vglut3^−/− (b) newborn mouse. G and H, Vglut3^−/− pups (n = 53) had normal weights (G) and body temperatures (H) compared to control pups (n = 164). I–L, breathing patterns were not affected by the Vglut3 mutation under thermoneutral conditions, whatever the breathing variable considered: mean respiratory frequency (R_f) (I), tidal volume (V_T) (J), ventilation () (K), or number of apnoeas per 30 s period (L). ns: non significant (P > 0.05); Student's unpaired t test. Values shown are means ± SEM. See Tables 1 and 2 for full statistical analyses.

Figure 2. Lack of Vglut3 impairs 5-HT metabolism in the brainstem, and the RRG response to 5-HT

See Supplemental Fig. 1 for schematic representation of the brainstem sections. A, immunolabelling showing that Vglut3 (green) colocalizes poorly with VMAT2 (red) in control mice at the level of the dorsal raphe (DR) (a), raphe pontis (RPn) (b), raphe magnus (RMg) (c) and raphe pallidus nuclei (Rpa) (d). B and C, immunodetection of 5-HT (red) and VGLUT3 (green) expression in coronal sections of control mice at the level of the raphe obscurus nucleus (ROb). Arrows in the upper right corner of panels B and C point toward the ventral medullary surface (B) and RMg (C). VGLUT3 was found in soma of cells at the ROb (Bb) and RMg (Cc) levels, and is co-expressed with 5-HT in cells of ROb and at RMg level (stars, yellow cells in Bc and Cc, respectively). D, 5-HT turnover in Vglut3^−/− pups (n = 12) differs from that in controls (n = 6). Endogenous levels of 5-HT (a), its main metabolite 5-hydroxyindoleacetic acid (5-HIAA) (b) and its precursor l-tryptophan (l-Trp) (c) in the brainstem of Vglut3^−/− and control pups, measured by HPLC. The 5-HT/5-HIAA ratio was significantly higher in Vglut3^−/− than in control pups (d). E and F, the spinal response to exogenous 5-HT is impaired in Vglut3^−/− pups. E, examples of phrenic bursts recorded from the C4 root of in vitro brainstem preparations from control (a) and Vglut3^−/− (b) pups, superfused with normal aCSF and aCSF containing 5-HT (25 μm, 5 min). Int C4: integrated C4 activity. F, tonic discharges in response to 5-HT appeared with the same latency in Vglut3^−/− (n = 11) and control (n = 14) preparations (a) but were shorter (b) and of smaller amplitude (c) in Vglut3^−/− preparations compared to control preparations (*P < 0.05; Student's unpaired t test, ns: non significant). a.u.: arbitrary units. Values given are means ± SEM. See Table 1 for full statistical analyses.

Figure 3. Lack of Vglut3 disrupts respiratory control and thermogenesis in the cold

Aa and b, examples of changes in central respiratory drive induced by acidosis in Control (a) and Vglut3^−/− pups (b) in vitro. Int C4: integrated C4 activity (C4). Ac, the effect of acidosis (shaded) on mean phrenic burst frequency (PBf) was significantly increased in Vglut3^−/− pups (n = 4) when compared to control pups (n = 12, P < 0.05; Student's unpaired t test). Ba and b, examples of changes in central respiratory drive induced by anoxia in control (a) and Vglut3^−/− pups (b) in vitro. Bc, the effect of anoxia (shaded) was significantly increased in Vglut3^−/− pups (n = 8) when compared to controls (n = 18, P < 0.001). C, examples of plethysmographic recordings in 6-day-old Vglut3^−/− and control pups during normoxia (a), hypoxia (10% O₂) (b) and post-hypoxia (Cc). Da–c, breathing variables: ventilation (), tidal volume (V_T) and breathing frequency (R_f). Vglut3^−/−pups (n = 63) displayed decreased ventilation during both normoxia and hypoxia, compared to control pups (n = 200). Both groups displayed a biphasic response to hypoxia (shaded), with the initial increase in being followed by a marked decrease (under the control of V_T and R_f). Ea–c, percentage change in breathing variables in response to hypoxia, relative to pre-hypoxic levels (hyperpnoeic response and hypoxic decline); HVD was significantly lower in Vglut3^−/− pups. Dd, oxygen consumption () at 26°C under normoxia and in response to hypoxia (shaded) in 6-day-old Vglut3^−/− (n = 7) and control pups (n = 29). The smaller in Vglut3^−/− pups (P < 0.001) reflected their impaired thermogenesis. Ed, percentage change in relative to pre-hypoxic levels was smaller in Vglut3^−/− pups. *P < 0.05; **P < 0.01; ***P < 0.001, Student's unpaired t test. Values shown are group means ± SEM. See Tables 1 and 2 for full statistical analyses.