How sweet it is: sensing low glucose in the carotid body

The paper by Zhang et al. (2007) in this issue of The Journal of Physiology suggests that the mammalian carotid body may, in addition to sensing hypoxia, function as a peripheral glucosensor. The carotid body might therefore join a group of other extracranial, glucose-sensitive areas including the pancreas, the liver and the hepatic portal vein, and play a role in initiating the reflex, counter-regulatory responses to hypoglycaemia that are initiated in vivo when blood glucose levels fall below the normal range of 4–6 mm (ca 70–110 mg dl⁻¹). This is an intriguing possibility and confirms findings made previously both in vivo and in vitro, but is not without some controversy. Thus, insulin-induced hypoglycaemia does initiate a carotid body-dependent, counter-regulation via elevations in blood glucagon and cortisol (Koyama et al. 2000), but a direct role for glucose sensing has been questioned (Bin-Jaliah et al. 2004) and the stimulus sensed in vivo has been suggested to be hypermetabolic rate rather than hypoglycaemia per se– a suggestion that maybe borne out by a recent, preliminary study on man (Ward, 2006).

Zhang et al. (2007) contend that the carotid body is a glucososensor on the basis of their finding that postsynaptic, afferent discharge, in a co-culture preparation of type I cells and petrosal ganglion cells, is augmented by reduction in superfusate glucose levels over a physiological range. Such excitation was, however, not observed by Almaraz et al. (1984) or Bin-Jaliah et al. (2004) and both groups found no stimulation by decreases in glucose to as low as 2 or 0 mm for periods of up to 2 h. Given that catecholamine release (Pardal & Lopez-Barneo, 2002) and ATP release (Zhang et al. 2007) in response to hypoglycaemia have been reported in thin slice preparations of the carotid body, it is unclear why there is, as yet, no consensus on this important topic. The onus is upon finding the reason why low glucose doesn't cause excitation in all preparations. It is worth recalling, however, that for the carotid body to function also as a sensor of hypoxia, it is essential that it is able to increase its metabolism during periods of reduced oxygen supply and it does not seem to be limited in its ability to utilize substrates other than glucose for this purpose. This may suggest, perhaps, that a lack of glucose, rather than some product of metabolism, is being sensed.

If glucose is being sensed directly, it appears to share, ultimately, the same Ca²⁺-dependent, neurosecretory pathways as hypoxia but not via the same mechanism, as Zhang et al. (2007) describe a low glucose-mediated depolarization in type I cells that is associated with a reduction in type I cell input resistance rather than the hypoxia-mediated depolarization that occurs consequent to K⁺ channel inhibition and thus increased input resistance.

Zhang et al. (2007) suggest that a reason that Bin-Jaliah et al. (2004) might have failed to see a response to low glucose in vitro is that these other authors utilized a hyperoxic P_O2 of 400 mmHg for their in vitro carotid body preparation, whilst they used an arterial P_O2 of 90 mmHg. The question arises, therefore, as to what is the correct P_O2 at which to conduct these and similar experiments. It appears that a normal tissue P_O2 of ca 50 mmHg (Wilson et al. 2000) might be expected for such a highly vascular and highly metabolic tissue as the carotid body during arterial normoxia. Thus for single cell (or co-culture) studies, a superfusate P_O2 of 50 mmHg might be preferable, and the 90 mmHg utilized by Zhang et al. (2007) might therefore also be considered hyperoxic. However, this cannot explain why these authors found a decreased sensitivity to glucose when a P_O2 of 140 mmHg was used as it is difficult to see how the cells might perceive this P_O2 difference. Further confusing the issue is the correct level of P_O2 to use for slice work. Pepper et al. (1996) calculated that P_O2 might be expected to fall by ca 1 mmHg per micrometre depth from the surface in superfusate preparations and they therefore have commonly used a P_O2 of 400 mmHg to superfuse the 200 μm diameter carotid body preparations that they use, to ensure adequate oxygenation to the core. Most brain slice preparations are thicker than this and so routinely use 95% O₂ (a P_O2 of ca 700 mmHg) for this very reason. There have been relatively few direct measures of tissue P_O2 profiles in slice preparations, but an elegant study by Wilson et al. (2003) shows how this profile is steep and can become steeper with increased metabolic activity. However, despite these concerns, the novel and exciting finding by Zhang et al. (2007) is that concurrent hypoxia is not a requirement for the glucose response.

A final question that can be raised is, do different preparations variably alter cellular metabolic processes? Thus, cultured cells or slice preparations of the carotid body may become more reliant upon glycolysis than less reduced preparations and perhaps are demonstrating a Pasteur-like effect where glucose consumption is increased with increased production of lactate. As described above, this possibility does not appear to be driven, here, by anaerobic conditions and so one may consider the integrity or otherwise of the mitochondria as an explanation. As this question may be linked with O₂ sensing by the carotid body, studying glucose sensing appears to have an exciting future as the polymodal sensing processes of this organ become more understood.