Pharmacometric model of Glucagon-ACTH-Cortisol in hyperinsulineamiceu/hypoglycaemic and hyperglycemicclamps in individuals who are lean andobese

Abstract: Introduction: Type 2 diabetes (T2D) increases in prevalence and generates significant diseaseburden, where impairment in hormonal systems, such as glucagon, is suspected to contribute to T2Ddevelopment. Covariates such as obesity is a known risk factor for T2D development. Additionally,higher fasting insulin and glucagon concentrations, stronger adrenocorticotrophic hormone- (ACTH)and cortisol-response during hypoglycemia have been demonstrated in obese individuals. In contrast,bariatric surgery has shown to result in diabetes remission, independent of weight loss. The effect ofglucose and insulin on glucagon, ACTH and cortisol have been well studied, but not modelledduring hypoglycemia. Aim: To develop a population model of glucagon, ACTH and cortisol response during hypo- andhyperglycemia by characterizing the relationship between glucagon and glucose/insulin as well asACTH and glucose, connected with cortisol and explore impact of covariates. Followed byinvestigating changes in the concentration-response relationships after Roux-en-Y gastric bypass(RYGB). Methods: Data from a cross-over study with one arm being hyperinsulinemic-eu/hypoglycemicclamp, and the second arm being a hyperglycemic clamp was combined with data from two crossoverdesign studies with arms being hyperinsulineamic-eu/hypoglycemic clamps, conducted beforeand after RYGB. The pre-surgical combined data was used to create a population model throughnon-linear mixed effect modelling where additional delay mediating through an effect compartmentwas investigated on all response variables. Various structural models were explored in order todescribe glucose and/or insulins effect on glucagon response. Glucose inhibition on ACTH production was described through a sigmoidal Imax-model, which was combined with a surge modelto describe ACTH production caused by circadian rhythm. Cortisol production was modelled asbeing driven solely by ACTH concentrations and covariate relationships were investigated through astepwise covariate model. The models were validated with statistical and graphical analysis. Results and discussion: Glucagon production was dependent on an individuals’ insulin sensitivityand fasting glucose and insulin concentration. Glucose inhibition on glucagon productioncontributed with 77 % of the total inhibition in comparison to insulin, which contributed with 23 %.Glucose concentration had a steeper exposure-response relationship in comparison to insulinconcentration on glucagon production. Despite insulin concentrations having smaller effect incomparison to glucose on glucagon production, insulin had significant impact on the models fit tothe data. ACTH and cortisol responses were accurately described, despite some misspecificationswhich were independent of glucose concentrations. Production of ACTH was dependent on insulinsensitivity and fasting glucose. Alteration in glucagon and ACTH production, together with insulinsensitivity was sufficient to describe changes after RYGB. This model can be used in optimal designgenerations where clamping technique is used as guidance of appropriate hypoglycemic clamplevels. Conclusion: A population model of glucagon, ACTH and cortisol response was developed using datafrom hypo- and hyperglycemic studies. Glucagon production was described as dependent onglucose and insulin and ACTH production was described as dependent on glucose, with the ACTHconcentration driving the cortisol production. Insulin sensitivity was found to affect baselineglucose and the sensitivity of ACTH to glucose. The model could describe post-bariatric surgerydata when adjusting the estimating of baseline glucagon and ACTH.