Computational Modeling of Glucose Uptake in the Enterocyte

Abstract

Absorption of glucose across the epithelial cells of the small intestine is a key process in human nutrition and initiates signaling cascades that regulate metabolic homeostasis. Validated and predictive mathematical models of glucose transport in intestinal epithelial cells are essential for interpreting experimental data, generating hypotheses, and understanding the contributions of and interactions between transport pathways. Here we report on the development of such a model that, in contrast to existing models, incorporates mechanistic descriptions of all relevant transport proteins and is implemented in the CellML framework. The model is validated against experimental and simulation data from the literature. It is then used to elucidate the relative contributions of the sodium-glucose cotransporter (SGLT1) and the glucose transporter type 2 (GLUT2) proteins in published measurements of glucose absorption from human intestinal epithelial cell lines. The model predicts that the contribution of SGLT1 dominates at low extracellular glucose concentrations (<20 mM) and short exposure times (<60 s) while the GLUT2 contribution is more significant at high glucose concentrations and long durations. Implementation in CellML permitted a modular structure in which the model was composed by reusing existing models of the individual transporters. The final structure also permits transparent changes of the model components and parameter values in order to facilitate model reuse, extension, and customization (for example, to simplify, or add complexity to specific transporter/pathway models, or reuse the model as a component of a larger framework) and carry out parameter sensitivity studies.

Keywords: CellML; GLUT2; OpenCOR; SGLT1; computational modeling; glucose uptake.

Figures

Figure 1

Schematic of enterocyte showing the relevant transporters in the apical and basolateral membrane along with the apical (lumen) and basolateral (interstitium) extracellular domains.

Figure 2

Figure shows the modularity of CellML model. Encapsulation hierarchy (purple), the CellML model imports (blue) and the other key parts (units, parameters, components and mappings) of the top level CellML model.

Figure 3

Dynamic response of the model to an extracellular glucose stimulus. The stimulus consists of a step increase followed by an exponential decay (A). Apical and basolateral membrane potentials (B), transepithelial potential (C), and intracellular concentrations of glucose (D), sodium (E), potassium (F), chloride (G), and pH (H) are shown.

Figure 4

Comparison of model responses against the model of Thorsen et al. (2014). Each variable has been normalized against the corresponding steady value from the Thorsen model. (A) Apical glucose stimulus. (B,C) Apical and basolateral membrane potential respectively. (D) Transepithelial potential. (E–H) Sodium, potassium, chloride and glucose intracellular concentration.

Figure 5

Intracellular glucose concentrations for a range of extracellular glucose concentrations in Caco2 and IEC6 cells and exposure times (A: 30 s, B: 60 s, C: 300 s, D: 600 s). Experimental data points and error bars were digitally extracted from Zheng et al. (2012). Strips for the model predictions represent the range of values generated by setting Vb = mVc, m = 0.1, 1, 10, 100, ∞.

Figure 6

Intracellular glucose concentration vs. extracellular glucose concentration in Caco2 in the presence/absence of Apical GLUT2 with different number of SGLT1 transporter (A) Output of model with apical GLUT2 (B) Model does not have apical GLUT2 (C) model does not have apical GLUT2 and the number of SGLT1 is doubled (D) model does not have apical GLUT2 and the number of SGLT1 is 3-fold higher. Experimental data points and error bars were digitally extracted from Zheng et al. (2012). Strips for the model predictions represent the range of values generated by setting Vb = mVc, m = 0.1, 1, 10, 100, ∞.

Figure 7

Glucose Flux through SGLT1 and GLUT2 in 600 s of simulation along with total apical flux of glucose.

Figure 8

Normalized steady state basolateral glucose flux vs. different stimulus of glucose in the lumen when number of SGLT1 is 3-fold higher(green), number of GLUT2 is 3-fold higher(blue) and number of both SGLT1 & GLUT2 are 3-fold higher(red).

Figure 9

(A) Ratio of sodium flux and Chloride flux through NHE3 and AE1 compare to NaCl co-transporter flux in the Thorsen model. (B) Chloride flux through CFTR and chloride diffusion ratio compare to the ration of sodium flux through ENaC and sodium diffusion.

Figure 10

Model output—Intracellular glucose concentration vs. Extracellular glucose concentration in (A) Caco2 cell line and (B) IEC6 cell line.