New ways to test beta cell functionality in health and diabetes

Abstract

Beta cell functionality is often characterised by indices describing different phases of insulin secretion. The typical biphasic insulin secretion pattern observed with a square wave glucose stimulation has laid the foundation for most modelling work regarding quantification of beta cell function. Within the context of control theory, the beta cell functionality is usually modelled as versions of a classic Proportional-Integral-Differential (PID) controller, and the different phases of insulin secretion are described in relation to the different control component, with the first phase of insulin secretion being related to the differential control component, and the second (late) phase to the integral control component. This is, of course, a phenomenological description. We propose a model of the glucose sensing mechanisms in the beta cell describing the timedependent physiological processes underlying the different insulin secretion phases. The model results show that glucokinase is the key regulatory step in the glucose sensing mechanisms. We argue that it is not glucose per se, but some signal(s) downstream of the glycolytic pathway that controls the activity of glucokinase, and hence the final insulin secretion pattern. We show that the first phase of insulin secretion is related to the rate of change of glucose in a non-linear saturable fashion, and that the second phase is due to translocation of glucokinase from an inactive to an active state. Hence, the glucose sensing mechanisms in the beta cell can, in some sense, be regarded as working as a classic PID controller, with intrinsic non-linearities in the sensing machinery. A meal tolerance test (MTT) is the best test for assessing beta cell indices as well as indices for insulin action in a physioloigcal relevant setting. In that context we have analysed MTT data from a large population of healthy subjects and from subjects with type 2 diabetes displaying a wide range of fasting plasma glucose (FPG) concentrations. Due to the heterogeneity in the FPG values of the subjects with diabetes, we stratified them according to their FPG and divided the subjects into five groups. Interestingly, when correcting for the FPG, the mean plasma glucose concentration profiles from each of the five groups are strikingly similar, despite quite large differences in the corresponding mean plasma insulin profiles. From the graphs of the means of the differentiated individual glucose profiles within the respective groups of subjects with diabetes, this similarity of the glucose profiles is even more evident. Same results are obtained when analysing the data from the database, where the subjects with type 2 diabetes have been followed throughout years, and where different standard treatments are also present. The graph of the mean healthy glucose profile shows a clear distinction from the corresponding graphs from subjects with type 2 diabetes. Of special interest is the observation that for the healthy persons the plasma insulin is still high even though plasma glucose has returned to fasting values, hence secretion of insulin continues despite glucose has returned to fasting values, and the increased insulin does not lead to hypoglycaemia. Hence in healthy subjects it appears that the glucose uptake is controlled in such a way as to follow the rate of glucose appearance from the meal. Similar conclusions cannot be drawn from the meal profiles of the subjects with type 2 diabetes. The glucose profiles for the subjects with type 2 diabetes seem similar despite different insulin profiles, and it appears that insulin merely follows the glucose profile without controlling it. However glucose undershoot is observed, probably due to elevated insulin concentration at the end of the meal test. The analysis of the MTT data provides a new tool to distinguish the healthy after-meal responses from responses of people with diabetes. Furthermore our analysis indicates that a mechanism that works more or less independent of insulin is activated in healthy persons after a meal and that this mechanism apparently is damaged and/or diminished in persons with type 2 diabetes. We argue, by referring to literature, that this mechanism is a result of the brain participating in the overall control of glucose concentration and fluxes of glucose equivalents. Hence, neural effects seem to be an important component that needs to be added to models that are set up to describe beta cell functionality as well as glucose uptake in a physiological relevant setting.