Date Published: March 8, 2019
Publisher: Public Library of Science
Author(s): Julia P. Dunn, Naji N. Abumrad, Bruce W. Patterson, Robert M. Kessler, Robyn A. Tamboli, Ines Armando.
We aim to identify physiologic regulators of dopamine (DA) signaling in obesity but previously did not find a compelling relationship with insulin sensitivity measured by oral-minimal model (OMM) and DA subtype 2 and 3 receptor (D2/3R) binding potential (BPND). Reduced disposition index (DI), a β-cell function metric that can also be calculated by OMM, was shown to predict a negative reward behavior that occurs in states of lower endogenous DA. We hypothesized that reduced DI would occur with higher D2/3R BPND, reflecting lower endogenous DA. Participants completed PET scanning, with a displaceable radioligand to measure D2/3R BPND, and a 5-hour oral glucose tolerance test to measure DI by OMM. We studied 26 age-similar females without (n = 8) and with obesity (n = 18) (22 vs 39 kg/m2). Reduced DI predicted increased striatal D2/3R BPND independent of BMI. By accounting for β-cell function, we were able to determine that the state of insulin and glucose metabolism is pertinent to striatal D2/3R BPND in obesity.
Diminished dopamine (DA) signaling and food reward are associated with obesity and are postulated to contribute to and/or perpetuate obesity. The mechanisms of the anorexic effects of insulin include regulating food reward through DA signaling, and thus, impaired insulin sensitivity (i.e. insulin resistance) is expected to dysregulate DA signaling  as occurs in rodent models of diet induced obesity (DIO) . We previously reported that higher body mass index (BMI) and lower fasting acyl ghrelin concentrations were associated with increased striatal DA subtype 2 and 3 receptor (D2/3R) binding potential (BPND), which we interpreted to reflect lower levels of endogenous DA competing with the displaceable radioligand, [18F]fallypride . Reduced striatal DA levels occur in DIO rodents  and one human report had trend level data demonstrating reduced pharmacologically-induced DA release in obesity . Using the same radioligand we utilized, [18F]fallypride, others found positive relationships between BMI and D2/3R BPND in the dorsal striatum (caudate and putamen) with conflicting findings in the ventral striatum [6, 7]. In our previous report, insulin resistance also predicted higher striatal D2/3R binding, (i.e. lower endogenous DA), but this effect was not independent of BMI . Eisenstein et al used β-cell function and [11C](N-methyl)benperidol([11C]NMB), a non-displaceable, D2 receptor-selective radioligand, to examine the relationship between DA signaling, obesity, and insulin. They did not find associations between striatal D2R levels and BMI or β-cell function (determined by disposition index, DI). This lack of relationship between BMI and receptor levels with a non-displaceable radioligand supported our interpretation that differences in endogenous DA levels were a predominate factor defining the relationship we identified with BMI and receptor levels measured with a displaceable radioligand. Eisenstein et al did report that β-cell function was associated with increased delayed discounting, a detrimental reward behavior which reflects impaired inhibitory control and is attenuated by agents that increase extracellular DA levels. Essentially, they found that impaired β-cell function occurred with a behavior that is present in a state of decreased endogenous DA . This finding prompted us to re-examine our data to determine the relationship of β-cell function measured by DI to striatal D2/3R BPND estimated with a displaceable radioligand. Further we sought to determine if any identified relationships were independent of BMI as our primary aim is to define physiologic regulators of DA signaling.
The study protocol was approved by the Vanderbilt University Institutional Review Board and all participants gave written informed consent. We studied 26 weight-stable females, 8 non-obese (22±3 kg/m2) and 18 obese (39±6 kg/m2) of similar age (Table 1), 22 who were included in our prior report . Screening included history and physical exam, laboratory testing including urine drug screen, and magnetic resonance imaging (MRI) of the brain. Exclusion included pregnancy, significant current psychiatric, neurologic or medical condition. One participant had diet-controlled type 2 diabetes mellitus. Individuals were also excluded if current tobacco use, substance abuse or heavy alcohol use, or if treated with central acting medications or insulin sensitizing agents in the preceding six months.
D2/3R BPND was 10–15% higher in the obese (caudate 29 vs. 33, p = 0.006; putamen 34 vs. 38, p = 0.013; ventral striatum 19 vs. 22, p = 0.030) (Table 1). We again found that higher BMI, lower insulin sensitivity (SI), and lower fasting acyl ghrelin levels predicted increased D2/3R binding throughout the striatum (Table 2, Fig 1). DI also had significant negative relationships with striatal D2/3R BPND. Neither age (p-values = 0.11–0.60) nor ϕ total (p-values = 0.42–0.98) were associated with regional receptor binding in our cohort. When adjusted for BMI, SI did not maintain relationships with D2/3R BPND (p-values = 0.1–0.2) but DI maintained robust negative relationships with D2/3R BPND (Table 3). DI explained all variance in D2/3R BPND attributed to BMI without significantly modifying parameter estimates for DI. Since in our previous report fasting acyl ghrelin concentrations were associated with striatal D2/3R BPND independent of BMI, we explored the relationship of both DI and fasting acyl ghrelin concentrations to receptor binding. When both were included as covariates to explain striatal D2/3R BPND, DI continued to explain the majority of the variance (p values = 0.088–0.028) while acyl ghrelin concentrations did not maintain independent (p values≥0.15) relationships with striatal D2/3R BPND; however, the parameter estimates for DI changed by 19–31%, demonstrating acyl ghrelin concentrations as a relevant modifier (Table 3).
We determined that impaired β-cell function (lower DI) is associated with higher striatal D2/3R binding. Again, this is interpreted as lower endogenous DA as we used a displaceable radioligand (Fig 2). DI allows insulin secretion to be interpreted in light of the prevailing insulin sensitivity, providing a quantitative index that estimates an individual’s state of glucose tolerance based on a physiologic challenge . Our regressions demonstrate that this physiologic measure can describe over twice the variance in D2/3R binding as the non-physiologic BMI measure. The effect size we detail for DI with D2/3R binding is substantially greater than what others have reported for BMI with D2/3R binding using the same radioligand . We did not find age to be a modifier of striatal D2/3R binding as others have reported; however it is important to note the DI declines with aging and can be modified by weight loss . BMI is commonly used to explore the effect of obesity on central reward signaling, but the current work demonstrates its limitations. The state of insulin and glucose metabolism are pertinent to design and interpretation of studies of DA signaling in obesity. Further, if aspects of insulin and glucose metabolism are influencing central reward signaling, this presents the potential for personalized therapeutic targets in diseases of reward signaling and obesity.