The lower brainstem has been implicated in the monitoring of brain glucose levels, but the mechanisms employed are unclear. In order to investigate whether brainstem glucose sensing involves similar mechanisms to that of pancreatic β-cells, we have tested for the presence of glucokinase and the glucose transporter Glut-2 in the rat brainstem. Additionally, hexokinase I was used as a control.

Sprague Dawley rats (150 g) were anaesthetised (Sagatal; 60 mg kg^-1, I.P.) and transcardially perfused with 4 % paraformaldehyde. The brain was removed and 40 µm coronal brainstem sections cut using a vibratome. Sections were incubated with polyclonal antibodies against glucokinase (1:50; Santa Cruz Biotechnologies), hexokinase I (1:50; Santa Cruz), or Glut-2 (1:500, 1:100; Biogenesis), then with biotin-labelled secondary antibodies and visualized by horseradish peroxidase staining (Ni-enhanced DAB). For mRNA isolation, rats were decapitated under anaesthesia, the brainstem removed, sliced, further dissected and mRNA isolated using a kit (Quiagen).

Strong glucokinase immunoreactivity was observed in ependymocytes lining the central canal and the fourth ventricle. Slightly weaker signals were obtained from large cell bodies in the hypoglossal and dorsal vagal nucleus and from small cells in the nucleus of the solitary tract (NTS) and the area postrema. Discrete staining was also found in the ventral medulla within the raphe pallidus, and the pyramidal tract. Hexokinase I immunoreactivity was seen in hypoglossal and dorsal vagal neurones, within the nucleus of the solitary tract, and in the area postrema. Ependymocytes did not show hexokinase I immunoreactivity. Furthermore, staining was observed within the reticular nucleus of the ventrolateral brainstem.

Glut-2 immunoreactive cells included ependymocytes lining both the central canal and fourth ventricle. In contrast to the results for glucokinase, the cilia of these cells stained for Glut-2. Additionally, hypoglossal, NTS, and dorsal vagal neurones exhibited Glut-2 immunoreactivity. Within the ventral part of the medulla, raphe pallidus and pyramidal tract cells were immunoreactive.

RT-PCR confirmed that mRNA encoding glucokinase, Glut2 and hexokinase is present in the central, lateral and ventrolateral brainstem.

In summary, our results suggest that proteins involved in glucose sensing are present in discrete areas of the lower brainstem. Immunoreactivity for Glut-2 and glucokinase is largely overlapping Surprisingly, also hexokinase I immunoreactivity shows a similar distribution. It remains to be verified if all cells expressing glucokinase and Glut-2 actually act as central glucose sensors and if glucokinase and hexokinase I are co-expressed in some cells.

This work was supported by an MRC Career Development Award to ST.