E. Berry

Modeling approaches to the dynamics of a living cell are presented that are strongly based on its underlying physical and chemical processes and its hierarchical spatio-temporal organization. Through the inclusion of a broad spectrum of processes and a rigorous analysis of the multiple scale nature of cellular dynamics, we are attempting to advance cell modeling and its applications. The presentation focuses on our cell modeling system, which integrates data archiving and quantitative physico-chemical modeling and information theory to provide a seamless approach to the modeling/data analysis endeavor. Thereby the rapidly growing mess of genomic, proteomic, metabolic, and cell physiological data can be automatically used to develop and calibrate a predictive cell model. The discussion focuses on the Karyote cell modeling system and an introduction to the CellX and VirusX models. The Karyote software system integrates three elements: (1) a model-building and data archiving module that allows one to define a cell type to be modeled through its reaction network, structure, and transport processes as well as to choose the surrounding medium and other parameters of the phenomenon to be modeled; (2) a genomic, proteomic, metabolic cell simulator that solves the equations of metabolic reaction, transcription/translation polymerization and the exchange of molecules between parts of the cell and with the surrounding medium; and (3) an information theory module (ITM) that automates model calibration and development, and integrates a variety of data types with the cell dynamic computations. In Karyote, reactions may be fast (equilibrated) or slow (finite rate), and the special effects of enzymes and other minority species yielding steady-state cycles of arbitrary complexities are accounted for. These features of the dynamics are handled via rigorous multiple scale analysis. A user interface allows for an automated generation and solution of the equations of multiple timescale, compartmented dynamics. Karyote is based on a fixed intracellular structure. However, cell response to changes in the host medium, damage, development or transformation to abnormality can involve dramatic changes in intracellular structure. As this changes the nature of the cellular dynamics, a new model, CellX, is being developed based on the spatial distribution of concentration and other variables. This allows CellX to capture the self-organizing character of cellular behavior. The self-assembly of organelles, viruses, and other subcellular bodies is being addressed in a second new model, VirusX, that integrates molecular mechanics and continuum theory. VirusX is designed to study the influence of a host medium on viral self-assembly, structural stability, infection of a single cell, and transmission of disease.

Fenfluramine improves glucose tolerance in obese subjects independently of its anorectic effect. Increased insulin action has been reported, but such an effect may be secondary to reduced hyperglycemia following augmented insulin secretion. For this reason, we investigated whether the dextro-enantiomer of fenfluramine (dexfenfluramine, dF) has a direct effect on insulin secretion using the glucose infusion test, a technique that can also be used to indirectly evaluate insulin action. Ten lean controls (BMI 21 +/- 0.4 kg/m2), 9 non-diabetic obese subjects (BMI 32.3 +/- 1.1) and 10 obese mild non-insulin dependent diabetics (BMI 36 +/- 2.6, fasting plasma glucose (FPG) 7.9 +/- 0.9 mmol/l) were studied with a random, double blind cross-over protocol. Each subject received 15 mg dF (or placebo) twice daily for 3 days prior to glucose infusion; 30 mg dF (or placebo) were given 90 min before the test. Obese control and diabetic subjects continued treatment with either dF or placebo for one month after which they were retested. There was no significant change in weight or fasting plasma glucose or insulin in any group. Biphasic insulin secretion was demonstrated in both non-diabetic groups, whereas a complete lack of first-phase and a delayed and reduced second phase insulin response was demonstrated in the diabetic subjects. There was no acute or chronic (obese subjects) effect of dF on insulin secretion in any group. In lean control subjects, plasma glucose curves during glucose infusion were unchanged by dF, whereas in non-diabetic obese subjects the glucose slope was improved after both acute and chronic dF, implying augmented glucose disposal. In diabetic patients, no significant acute or chronic effect of dF on glucose response was registered. When all obese subjects were re-grouped according to fasting plasma insulin levels (FPI) and not glucose tolerance, those with relative fasting hyperinsulinemia (> 100 pmol/l, mean +/- SE = 144 +/- 12 pmol/l) showed significant improvement of insulin action after dF, whereas those with low FPI (< 100 pmol/l, mean +/- SE = 65 +/- 7 pmol/l) did not. These findings, together with previously published results of improved insulin action in diabetics during hyperinsulinemic clamps, suggest that dF-mediated improvement in glucose clearance may require substantial plasma insulin concentrations.(ABSTRACT TRUNCATED AT 400 WORDS)