Glucose Ingestion Fails to Inhibit Hypothalamic Neuronal Activity in Patients With Type 2 Diabetes

We studied 7 type 2 diabetic men who were treated with either oral glucose-lowering medication or diet alone and 10 healthy men without family history of type 2 diabetes. The characteristics of the subjects are given in Table 1. Subjects were excluded if they had poorly controlled hypertension (>160/100 mmHg), chronic renal failure or leg ulcers, or any chronic disease and if they smoked or had a history of alcohol or drug abuse.

Glucose solution (75 g in 300 ml water) or 300 ml tap water were ingested after a 10-h overnight fast in random order on 2 separate days with an interval of at least 1 week. Beginning 8 min before drinking, the hypothalamic BOLD signal was recorded for 30 min by fMRI on a 3.0 Tesla scanner (Philips Achieva; Philips Medical Systems, Best, The Netherlands). A T₂-weighted echo-planar imaging sequence (repetition time 120 ms, echo time 30 ms, flip angle 30°, image matrix 256 × 231, field of view 208 × 208 mm, 12 signal averages per scan) was used to scan a 12-mm–thick midsagittal slice. A T₁-weighted anatomical scan was made of the same slice (repetition time = 550 ms, echo time 10 ms, field of view 208 × 208 mm). The images were preprocessed and analyzed as described previously (2), except for the use of a thalamic reference area (10 × 10 pixels, Fig. 1A) instead of a reference area in the frontal cortex (which had poor image quality). For every subject, the hypothalamus was manually segmented and divided into four regions (Fig. 1A), i.e., the upper anterior hypothalamus, upper posterior hypothalamus, lower anterior hypothalamus, and lower posterior hypothalamus, by two orthogonal axes according to predefined criteria (6). This was done to make the data comparable with those reported in previous publications. Figure 1B sketches relevant nuclei contained within each hypothalamic area by approximation.

SPSS 12.0 (SPSS, Chicago, IL) was used for data management and statistical analysis. The baseline BOLD signal was defined as the average of signals recorded during 8 min before ingestion of test solutions. Signal changes in response to intervention were calculated per minute as a percentage of baseline values. Results are presented as means ± SEM. Differences within groups were assessed by differential regression analysis (7). A Bonferroni-corrected threshold of significance (P = 0.0013) was applied, since 38 t tests were performed per comparison.

On average, it took 5.5 ± 2.4 min (both groups pooled) to ingest the glucose solution or the water. Figure 2 shows BOLD signal changes in the four hypothalamic regions defined by our zoning strategy. After a considerable drop occurring during ingestion of the test solutions in both groups, the signal returned to baseline. The signal drop is an artifact caused by swallowing and slight movements of the head during drinking. Thereafter, the BOLD signal steadily, consistently, and significantly decreased in response to glucose ingestion in all regions in the control group. In contrast, glucose ingestion did not impact the BOLD signal in type 2 diabetic patients in any hypothalamic area (Fig. 2). Notably, the BOLD signal increased significantly in the lower anterior hypothalamus and lower posterior hypothalamus after water ingestion in type 2 diabetic patients but not in control subjects.

The diminution of hypothalamic BOLD signals in response to an oral glucose load that we observed in control subjects corroborates earlier reports of similar studies in healthy humans (2,8). The identity of the neurons, whose activity is suppressed by glucose ingestion, remains to be established. Hypothalamic nuclei harbor a host of distinct neuronal cell types that are involved in the control of postprandial satiety and fuel flux. These neurons respond to various circulating nutrients and hormones to guide homeostatic adjustments of food consumption and metabolism (1). For example, inhibition of neuropeptide Y neurons in the arcuate nucleus (located in the lower posterior hypothalamic region as defined in the current study) by meal-related hormones such as glucagon-like peptide 1 (GLP-1) and peptide YY induces satiety (9,10). Moreover, this component of the gut-brain axis may reinforce insulin action after meals to effectively inhibit endogenous glucose production and promote storage of ingested nutrients (11,12). The mechanism driving these adaptations involves other hypothalamic areas, including the paraventricular nucleus and lateral hypothalamus (Fig. 1B) (1).

The fact that the BOLD signal did not decline in response to ingestion of glucose in the type 2 diabetic patients suggests that the hypothalamus of the diabetic patients inappropriately perceives and/or processes signals reflecting encroachment of fuel homeostasis in response to a nutrient load. In view of the critical role of the hypothalamus in the regulation of food intake and fuel flux described above, such a maladaptive response may have important ramifications. Thus, erroneous hypothalamic processing of metabolic and/or endocrine signals mirroring energy availability may compromise postprandial satiety in type 2 diabetic patients. Also, hypothalamic dysfunction may hamper postprandial insulin action and thereby contribute to inadequate suppression of glucose production and diminution of nutrient disposal after meals, metabolic anomalies that mark type 2 diabetes. Finally, postprandial thermogenesis and energy expenditure may be abnormal in diabetic humans, as the hypothalamus is critically involved in the regulation of energy metabolism (13).

Conversely, the metabolic anomalies that typify type 2 diabetes may have led to the aberrant hypothalamic response reported here. The mechanistic link between oral glucose ingestion and hypothalamic neuronal activity remains to be established. Various circulating metabolic and endocrine cues that change in response to glucose intake, including glucose, fatty acids, insulin, and a number of gut peptides, may be involved (13). Since type 2 diabetes is marked by abnormal postprandial plasma concentrations of virtually all of these cues, it is conceivable that the metabolic state of our patients has contributed to the failure of hypothalamic neurons to properly respond to an oral glucose load. Although both circulating glucose and insulin can modulate neuronal firing in the hypothalamus (14,15), we have some evidence that neither of these cues alone contributes significantly to the hypothalamic response observed here. Indeed, maltodextrin intake does not impact hypothalamic activity determined by fMRI, despite ensuing plasma glucose and insulin profiles similar to those in response to glucose ingestion (16). As GLP-1 secretion in response to meals is diminished in type 2 diabetes (17), and GLP-1 inhibits hypothalamic neuronal activity (18), this peptide may be one of the aberrant cues involved. However, whatever the cause of the phenomenon we have observed, given the important role of the hypothalamus in the control of metabolism, improper relay of signals by (circulating) cues supposedly compromises postprandial metabolic flux.

At least two alternative explanations for the lack of BOLD signal response to glucose ingestion in type 2 diabetes patients should be considered here. First, neurovascular coupling, i.e., the mechanism by which changes in neuronal activity regulate local cerebral blood perfusion (19), could be compromised in patients with type 2 diabetes. There are no data on the influence of diabetes on neurovascular coupling to date. Importantly, normal aging, often associated with insulin resistance, did not affect neurovascular coupling in a recent study (20). Second, microvascular disease, which often accompanies diabetes, may hamper hemodynamic adaptations to neural activity. However, the patients participating in our study had diabetes for only 4.7 ± 2.8 years, were in good glycemic control, and had virtually no signs of microvascular complications, as only one of seven patients had microalbuminuria, and none had retinopathy.

We report that the BOLD signal in the hypothalamus does not properly subside in response to an oral glucose load in moderately obese patients with type 2 diabetes, indicating that glucose ingestion fails to inhibit hypothalamic neuronal activity in these patients. This suggests that hypothalamic neurons of type 2 diabetic patients erroneously perceive and/or process endocrine cues reflecting nutrient availability. Since the hypothalamus coordinates metabolic and behavioral adaptations in response to these cues to maintain energy balance, hypothalamic dysfunction may hamper postprandial insulin action and compromise satiety in patients with type 2 diabetes.

This study was supported by the Dutch Diabetes Foundation (grant 2002.01.005).