Modeling the distribution of insulin in pancreas

TL;DR

This paper develops a mathematical model to estimate and analyze the distribution of insulin concentration along the pancreatic vein, providing insights into insulin dynamics relevant for understanding diabetes.

Contribution

It introduces a novel modeling approach to estimate insulin distribution in the human pancreas using physiological and imaging data.

Findings

Insulin concentration increases nearly linearly with glucose levels.

Distribution varies along the pancreatic vein, increasing faster in the tail.

Modeling provides new insights into insulin dynamics in the pancreas.

Abstract

Maintenance of adequate physical and functional pancreatic $\beta$-cell mass is critical for the prevention or delay of diabetes mellitus. It is well established that insulin potently activates mitogenic and anti-apoptotic signaling cascades in cultured $\beta$-cells. Loss of $\beta$-cell insulin receptors is sufficient to induce type 2 diabetes in mice. However, it remains unclear whether the {\em in vitro} effect in human islets and the {\em in vivo} effects in mice can be applied to human physiology. The major obstacle to a complete understanding of the effects of insulin's feedback in human pancreas is the absence of technology to measure the concentrations of insulin inside of pancreas. To contextualize recent {\em in vitro} data, it is essential to know the local concentration and distribution of insulin in pancreas. To this end, we continue to estimate the local insulin concentration within pancreas. In this paper, we investigate the distribution of insulin concentration along the pancreatic vein through a novel mathematical modeling approach using existing physiological data and islet imaging data, in contrast to our previous work focusing on the insulin level within an islet. Our studies suggest that, in response to an increase in glucose, the insulin concentration along the pancreatic vein increases nearly linearly in the fashion of increasing quicker in tail area but slower in head area depending of the initial distribution.