Hypoglycemia activates arousal-related neurons and increases wake time in adult rats

Abstract

Hypoglycemia resulting from excess of exogenous or endogenous insulin elicits central nervous system activation that contributes to counterregulatory hormone secretion. In adult humans without diabetes, hypoglycemia occurring during sleep usually produces cortical activation with awakening. However, in adult humans with type 1 diabetes, hypoglycemic arousal appears blunted or absent. We hypothesized that insulin injection sufficient to produce hypoglycemia would induce awakening in adult male rats. Polysomnographic studies were carried out to characterize the effect of insulin injection on measures of sleep and waking during a circadian time of increased sleep. Compared to a baseline day, insulin treatment more than doubled the time spent awake, from 18.4+/-2.6% after saline injection to 48.0+/-5.5% after insulin. Insulin injection also reduced rapid eye movement sleep (REMS) from 27.3+/-1.8% to 5.6+/-1.3%. The percent of time in non-REM sleep (NREMS) sleep was not different between saline and insulin days, however, NREMS after insulin was fragmented, with increased number and decreased duration of episodes. These electrophysiological data indicate that insulin-induced hypoglycemia is an arousing stimulus in rats, as in nondiabetic adult humans. We also studied the effect of insulin on activation of selected arousal-related neurons using immunohistochemical detection of Fos. Fos-immunoreactivity increased in orexin (OX) neurons after insulin, from 8.7+/-4.9% after saline injection to 37+/-9% after insulin. Basal forebrain cholinergic nuclei also showed increased Fos-immunoreactivity after insulin. These correlated behavioral and histological data provide targets for future studies of the neural pathways underlying hypoglycemic arousal.

Figure 1

Diagrams of atlas levels used for counting Fos staining in selected basal forebrain cholinergic nuclei: medial septum (MS), diagonal band of Broca (DBB), magnocellular preoptic nucleus (MCPO), based on ref. . Diagrams are modified from a rat brain atlas [48].

Figure 2

Fos labeling in brain arousal-related neurons after saline injection (panels A and C) or after insulin injection (panels B and D). Panels A and B show the region of the perifornical hypothalamus in tissue prepared with double-immunohistochemical labeling for OX (blue/gray cytoplasmic stain) and Fos (brown nuclear stain). Arrows indicate double-labeled Fos+/OX+ neurons after insulin treatment. Panels C and D show the region of the magnocellular preoptic nucleus in tissue prepared with double-immunohistochemical labeling for choline acetyltransferase (ChAT - blue/gray cytoplasmic stain) and Fos (brown nuclear stain). Arrows indicate double labeled Fos+/ChAT+ neurons after insulin treatment. The calibration bar in B indicates 30 μm in panels A and B, and 18 μm in panels C and D.

Figure 3

Total numbers of ChAT+/Fos+ cells in forebrain (A) and brainstem (B) cholinergic cell groups after saline or insulin treatment. In 6 saline-treated and 8 insulin-treated rats, quantification was carried out on Fos and ChAT double-labeled cells in 3 forebrain cholinergic nuclei as shown in figure 1. In 3 saline-treated and 6 insulin-treated rats, quantification was carried out in brainstem laterodorsal tegmental (LDT) and pedunculopontine tegmental (PPT) cholinergic nuclei. Insulin treatment was associated with significantly greater numbers of Fos+ cholinergic neurons only in forebrain nuclei.

Figure 4

The effect of afternoon saline or insulin treatment on percent time in non-rapid eye movement sleep (NREMS), rapid eye movement sleep (REMS) and awake is shown as % time in stage. Data were collected from 30 minutes to 120 minutes after injection of saline or insulin and are expressed as means ± SEM. During this time after insulin injection, blood glucose is generally low and stable at 2.4-2.6 mM.

Figure 5

Representative printout of delta power tracing and hypnogram from a polysomnography study. On the baseline day, the rat was given saline 1 ml/kg subcutaneously at 1 PM and sleep was recorded for 120 minutes. A typical sleep pattern consisting of multiple cycles between waking, NREMS, and REMS was observed. There are several episodes of prolonged increases in delta power (indicated as dark bars underneath the delta power graph), associated with a predominance of consolidated NREMS in those time periods. These are interspersed with times of greatly reduced delta power associated with REM sleep and waking. Below, the panel from the hypoglycemia day shows the effect of insulin 3 U/kg given subcutaneously. There are fewer consolidated episodes of elevated delta power and fewer REMS episodes. Cycling between W and NREMS is increased, although time in NREMS is preserved. The differences between baseline and insulin days are more pronounced during the second hour of recording, which should be the time of maximum hypoglycemia after the insulin injection.

Figure 6

The time course of changes in W and NREMS episodes after saline or insulin injection. Repeated measures analysis of variance indicated a significant interaction of treatment (saline or insulin) and time (30-60; 60-90; 90-120 minutes after injection) in the following sleep measures: A. W episode number (F=7.087; p=.026), B. NREMS episode number (F=9.25; p=.015) and C. NREMS episode duration (F=6.97; p=.027).