With Stephanie Brassen, PhD, and Caroline Apovian, MD
Since a clearer understanding of the reasons that drive people to overeat may provide breakthrough therapies or preventative measures for type 2 diabetes (T2D) and obesity, researchers used fMRI to study the effect of central insulin on the value of food cues in individuals with normal insulin sensitivity and those with insulin resistance.1 The findings indicated that intranasal insulin reduced preference ratings for food cues in individuals with normal insulin resistance, but not in individuals with poor insulin resistance,1 which was published in Nature Communications.
In an attempt to curb the rise of obesity and type 2 diabetes, Stephanie Brassen, PhD, of the University Medical Centre Hamburg-Eppendorf in Germany and colleagues, sought to better understand the role of central insulin in the neural control of food processing in non-homeostatic pathways.1 The hope is that by investigating the neural cross-talk between homeostatic and reward-related feeding systems researchers and clinicians will be able to determine the processes behind pathological eating behavior and how to stop it.1
“Much of our hypothesis in this study was based on previous findings in animals, making it important to see if we could replicate and extend these findings in humans,” Dr. Brassen told EndocrineWeb.
Focused specifically on the mesolimbic system, which deals with reward processing, the researchers examined the ventral tegmental area (VTA) and the nucleus accumbens (NAc), both of which impact dopamine release.1 Importantly, insulin receptors are present throughout the mesolimbic system.2,3 Previous research has shown that direct injection of insulin into the VTA reduced hedonic feeding in sated conditions and reduced somatodendritic dopamine.4
To test the role of central insulin on the preference for food-related cues in the VTA-NAc circuits, the researchers used pharmacological fMRI with a food/non-food valuation paradigm on participants who had fasted overnight.1 Participants were separated into normal insulin resistance (NIR, 28 participants) or insulin resistance (IR, 20 participants) using the Homeostatic Model Assessment (HOMA) for insulin resistance.5
Subjects then had 2 fMRI scans, at least 1 week apart, following a 10-hour fast. They were given either 160 IU of intranasal insulin or a placebo. Intranasal insulin has been shown to deliver insulin to the central nervous system without systemic absorption.6 While in the MRI scanner, participants were asked to rate their preference for food or non-food objects with a yes or no answer, and then asked to rate their answer for food items on a 4-point scale.
Only NIR Individuals with higher insulin levels had lowered food preference scores
The results of the study are striking. Insulin levels were indeed correlated with food preference scores, but only in the normal insulin resistance group. The higher the plasma insulin concentration, the lower the preference for food items (r = -0.43, P = 0.02, n = 28, Pearson’s correlation).1 In contrast, in the insulin resistant group, no correlation was observed between insulin level and food preference (r = 0.17, P = 0.47, n = 20, Pearson’s correlation).1
Using rmANOVA analysis, a significant three-way interaction was revealed when the following criteria were analyzed: item (food/non-food), session (placebo/insulin), and group (NIR/IR) (F(1,46) = 5.13, P = 0.028, n2 = 0.1, nNIR = 28, nIR = 20, rmANOVA).1 The preference for food/non-food was significantly reduced in the intranasal insulin group compared with placebo, but only in the normal insulin resistance group (F(1,27) = 7.37, P = 0.011, h2 = 0.22, n = 28, rmANOVA).1
In the insulin resistant group, food preference scores after intranasal insulin actually increased (T(19) = 1.77, P = 0.09, d = 0.16, n = 20, t-test). Confirming these results, HOMA scores directly correlated with changes in food preference driven by insulin.1 In other words, subjects with normal HOMA values showed a greater reduction in food values following intranasal insulin (r = -0.30, P = 0.04, n = 48, Pearson’s correlation).1
Imaging showed insulin-dependent activation in reward-related NAc and VTA during food valuation tasks Examining the fMRI data, the researchers measured the blood oxygenation level-dependent activity (BOLD).1 They found significant activation in a network of metabolic and reward-related brain regions to food vs. non-food items in the placebo group.1 These areas included the bilateral hypothalamus, VTA, amygdala, insula and orbitofrontal cortex.
Looking for regions that showed a correlation between food-specific valuation signals and BOLD response, the researchers found that the bilateral NAc was engaged during food value encoding in normal insulin resistance individuals, but saw no difference in insulin resistance individuals. A significant group interaction was observed in the left NAc (P<0.05, FWE corrected, nNIR = 28, nIR = 20, factorial design).1
“We were astonished by the high specificity of insulin action in the reward system,” remarked Dr. Brassen. “The participants in the insulin-resistant sample were still in a pre- or non-diabetic state and their mean BMI was only 29. Nevertheless, data in this subsample already pointed to pathological conditions which might trigger further overeating and weight gain.”
Caroline Apovian, MD, who was not associated with the study, told EndocrineWeb, “it’s great research, and I agree with the main finding, but the conclusions should be looked at with caution. Insulin does seem to influence food valuation through modulation in this area of the brain, and the authors find that there were differences between insulin resistant and sensitive subjects, but the use of HOMA here must be taken with caution. Other methods should be used, such as clamp studies or frequently sampled IVGTT with the Bergman minimal model to ensure a correct diagnosis of insulin sensitivity or resistance.”
The researchers also evaluated the effects of central insulin on food-specific value signals in the NAc and VTA. They found a significant group interaction in the NAc and left VTA, indicating that intranasal insulin reduced the food-specific valuation signal in the normal insulin resistance group, but the signal actually increased in the insulin resistant group.
“I think the most important takeaway message from our study is that insulin—besides its role in the regulation of blood sugar—can reduce the hedonic value of foods and thereby probably helps people with normal insulin functioning to resist overeating,” said Dr. Brassen, “This function is already disturbed in overweight, pre-diabetic individuals with increased insulin resistance that might hinder them to resist overeating. Reducing and monitoring insulin resistance may, therefore, be an important factor for successful weight management.”
Dr. Apovian told EndocrineWeb, “the takeaway message from the paper is that there is a difference in how central insulin modulates feeding behavior in insulin sensitive vs. resistant subjects, at least based on HOMA measurements and subject preference for foods.
Studies in the future will need to corroborate these findings via better methods to assess insulin sensitivity and risk of type 2 diabetes.” She adds, “it would be interesting to do this study in those who are not yet insulin resistant but have a genetic predisposition.”