Effects of antidiabetic drugs on dipeptidyl peptidase IV activity: Nateglinide is an inhibitor of DPP IV and augments the antidiabetic activity of glucagon-like peptide-1

Dipeptidyl peptidase IV (DPP IV) is the primary inactivator of glucoregulatory incretin hormones. This has lead to development of DPP IV inhibitors as a new class of agents for the treatment of type 2 diabetes. Recent reports indicate that other antidiabetic drugs, such as metformin, may also have inhibitory effects on DPP IV activity. In this investigation we show that high concentrations of several antidiabetic drug classes, namely thiazolidinediones, sulphonylureas, meglitinides and morphilinoguanides can inhibit DPP IV The strongest inhibitor nateglinide, the insulin-releasing meglitinide was effective at low therapeutically relevant concentrations as low as 25 mu mol/l. Nateglinide also prevented the degradation of glucagon-like peptide-1 (GLP-1) by DPP IV in a time and concentration-dependent manner. In vitro nateglinide and GLP-1 effects on insulin release were additive. In vivo nateglinide improved the glucose-lowering and insulin-releasing activity of GLP-1 in obese-diabetic ob/ob mice. This was accompanied by significantly enhanced circulating concentrations of active GLP-1(7-36)amide and lower levels of DPP IV activity. Nateglinide similarly benefited the glucose and insulin responses to feeding in ob/ob mice and such actions were abolished by coadministration of exendin(9-39) and (Pro(3))GIP to block incretin hormone action. These data indicate that the use of nateglinide as a prandial insulin-releasing agent may partly rely on inhibition of GLP-1 degradation as well as beta-cell K-ATP channel inhibition. (C) 2007 Elsevier B.V. All rights reserved.

Integrated structure selection and parameter optimisation for eng-genes neural models

The eng-genes concept involves the use of fundamental known system functions as activation functions in a neural model to create a 'grey-box' neural network. One of the main issues in eng-genes modelling is to produce a parsimonious model given a model construction criterion. The challenges are that (1) the eng-genes model in most cases is a heterogenous network consisting of more than one type of nonlinear basis functions, and each basis function may have different set of parameters to be optimised; (2) the number of hidden nodes has to be chosen based on a model selection criterion. This is a mixed integer hard problem and this paper investigates the use of a forward selection algorithm to optimise both the network structure and the parameters of the system-derived activation functions. Results are included from case studies performed on a simulated continuously stirred tank reactor process, and using actual data from a pH neutralisation plant. The resulting eng-genes networks demonstrate superior simulation performance and transparency over a range of network sizes when compared to conventional neural models. (c) 2007 Elsevier B.V. All rights reserved.