We found that perfusion of a low glucose (0.5 mM) solution depolarizes the membrane potential of most NTS neurons by activating a conductance with an Erev around − 50 mV, reproducing our earlier results (Murat & Leão, 2019) and from others (Balfour & Trapp, 2007). The drop in glucose was probably sensed directly by the neuron since low glucose-induced depolarization was observed when action potential-dependent synaptic transmission was blocked by tetrodotoxin but absent when glucose was added to the recording electrode solution (De Bernardis Murat and Leão, 2019). We hypothesized that perfusing the low-glucose solution on NTS neurons would decrease ATP production, leading to membrane depolarization by activating an inward conductance. Given that neurons rely on oxidative phosphorylation in mitochondria to produce ATP from glucose (Yellen, 2018), we then decided to test if inhibition of mitochondrial ATP production would reproduce the effects of low glucose. For this, we used the mitochondrial uncoupler CCCP, which dissipates the proton gradient across the mitochondrial membrane and ATP production by the Fo/F1-ATPase (Plášek et al., 2017). We recently reported that CCCP opens KATP channels in a glycinergic neuron from the auditory brainstem hyperpolarizing the membrane (de Siqueira et al., 2022), an effect opposed to what we observed in NTS neurons. However, it shows that CCCP can inhibit ATP production and that this drop can be detected in whole-cell patch-clamp using the same internal solution as used in this study. However, our experiments with CCCP in the NTS produced only a fast hyperpolarization of the membrane, followed by a more substantial depolarization caused by the activation of an inward conductance as observed in low glucose. We suggest that perfusion of NTS neurons with a low-glucose solution decreases mitochondrial ATP production, activating the inward conductance and causing membrane depolarization. Unlike low-glucose, CCCP produced a fast but significant slight hyperpolarization by activating a current with an Erev close to the potassium equilibrium potential. This hyperpolarization could be caused by the unblocking KATP channels (de Siqueira et al., 2022), but in NTS neurons, KATP channels only open several minutes after low glucose perfusion (De Bernardis Murat and Leão, 2019). One hypothesis is that the acidification of the cytoplasm promoted by CCCP activates the acid-sensitive potassium leak channel TREK-1 (Maingret et al., 1999), which is expressed in the brainstem (Fink et al., 1996) or by ATP bound-KATP channels, which can be positively modulated by intracellular acidification in the presence of ATP (Fan et al., 1994).
Balfour & Trapp (2007) suggested that the depolarization produced by low-glucose in NTS neurons could be attributed to the inhibition of the Na-K ATPase. A drop in ATP production by mitochondria could lead to the dissipation of ionic gradients by a reduction in the activity of the Na/K-ATPase. Additionally, because of the electrogenic nature of the transport mediated by the Na/K-ATPase, its inhibition would produce a small depolarization (Glitsch, 2001). We found that oubain (1-100 µM) induced membrane depolarization in NTS subpostremal neurons, but differently from low glucose or CCCP, there was no change in Rinput, accordingly to the inhibition of a pump current (Gulledge et al., 2013). Thus, our conclusion is that the membrane depolarization induced by low glucose is not caused by inhibition of the Na/K pump. However, we cannot rule out entirely that part of the effect could be caused by the inhibition of the pump current due to the heterogeneity of the low glucose-induced currents.
The AMPK enzyme is involved in metabolic processes (Garcia & Shaw, 2017; Herzig & Shaw, 2018) and can affect several ion channels (Andersen & Rasmussen, 2012). Lamy et al. (2014) found that the AMPK inhibitor dorsomorphin (compound C) reversed the depolarization caused by low glucose in GABAergic neurons, suggesting AMPK's role in detecting glucose decrease, indicating that AMPK mediated the effect of low glucose in these neurons. However, caution should be considered in dorsomorphin experiments since it also affects other kinases besides AMPK (Bain et al., 2007; Vogt et al., 2011).
In our hands, dorsomorphin caused a steady depolarization of the membrane of NTS neurons and did not occlude the depolarization by low glucose. Thus, we tested if the AMPK agonist AICAR (Corton et al., 1995) could reproduce the effect of low glucose. AICAR [5-amino-4-imidazolecarboxamide (AICA) ribonucleoside] is a cell-permeant analog of adenosine, which is phosphorylated intracellularly by adenosine kinase resulting in AICA ribotide (ZMP), which is an allosteric activator of AMPK. However, AICAR did not depolarize the membrane nor occluded the effects of low glucose, suggesting that AMPK does not mediate the depolarization induced by low glucose in subpostremal NTS neurons, differently from what was previously observed by Lamy et al. (2004). We conclude from these findings that AMPK is not essential for the depolarization induced by low glucose in subpostremal NTS neurons.
We conclude that the subpostremal NTS neurons directly detect the decrease in glucose as a decrease in the ATP from mitochondrial respiration, which triggers unknown inward currents leading to membrane depolarization. While activation of AMPK does not participate in generating depolarization, we cannot rule out that some of the depolarization is caused by inhibition of the Na/K-ATPase current. Due to the mixed Erev of the low glucose-induced current, we believe that more than one ionic conductance is responsible for the current. A better understanding of the cellular mechanisms of detecting hypoglycemia by NTS neurons can help understand the global mechanisms that the NTS uses to regulate energy balance.