Biophysical model of ion transport across human respiratory epithelia allows quantification of ion permeabilities

Abstract

Lung health and normal mucus clearance depend on adequate hydration of airway surfaces. Because transepithelial osmotic gradients drive water flows, sufficient hydration of the airway surface liquid depends on a balance between ion secretion and absorption by respiratory epithelia. In vitro experiments using cultures of primary human nasal epithelia and human bronchial epithelia have established many of the biophysical processes involved in airway surface liquid homeostasis. Most experimental studies, however, have focused on the apical membrane, despite the fact that ion transport across respiratory epithelia involves both cellular and paracellular pathways. In fact, the ion permeabilities of the basolateral membrane and paracellular pathway remain largely unknown. Here we use a biophysical model for water and ion transport to quantify ion permeabilities of all pathways (apical, basolateral, paracellular) in human nasal epithelia cultures using experimental (Ussing Chamber and microelectrode) data reported in the literature. We derive analytical formulas for the steady-state short-circuit current and membrane potential, which are for polarized epithelia the equivalent of the Goldman-Hodgkin-Katz equation for single isolated cells. These relations allow parameter estimation to be performed efficiently. By providing a method to quantify all the ion permeabilities of respiratory epithelia, the model may aid us in understanding the physiology that regulates normal airway surface hydration.

Figure 1

(A) Schematic diagram of the main pathways of ion and water transport in respiratory epithelia, all of which are included in the model. The paracellular pathway allows the diffusion of ions and water between the apical and basolateral compartments. The basolateral membrane expresses the Na-K pump, Cl⁻ channels, K⁺ channels, and the Na-K-2Cl cotransporter. The apical membrane expresses Na⁺ channels, Cl⁻ channels, and K⁺ channels. Both basolateral and apical membranes are permeable to water. (B–F) Parameter fitting was accomplished using analytical expressions for the steady-state short-circuit current (I_SC) and membrane potential at short-circuit condition (V_m) to perform an efficient search of the parameter space. (B and C) Triplets of apical Na⁺ permeability, apical K⁺ permeability, and basolateral Cl⁻ permeability {p^a_Na, p^a_K, p^b_Cl} that reproduce the experimental I_SC and V_m as predicted by Eqs. 10 and 11 (solid lines). Some triplets {p^a_Na, p^a_K, p^b_Cl} are associated with negative values of other ion permeabilities, which are physically unrealistic (dashed lines). The triplet {1.6, 0.09, 1.4} × 10⁻⁸ m/s obtained with a Monte Carlo method (see the Supporting Material) is shown for comparison (symbols). (D and E) By varying the ratios r_K = p^b_K/p^a_K and r_Cl = p^b_Cl/p^a_Cl and simulating various experimental conditions (see Table S3), the pair {r_K, r_Cl} that minimized the score function SSD was identified for each set of paracellular permeabilities {p^p_Na, p^p_Cl}. (F) To estimate the Na⁺ and Cl⁻ paracellular permeabilities, these parameters were varied under the hypotheses p^p_Cl = p^p_Na or p^p_Cl = (D_Cl/D_Na)p^p_Na = 1.53 p^p_Na. Simulations were run for various experimental conditions (see Table S3) and the best fits were obtained for p^p_Na = 2.5–3.0 × 10⁻⁸ m/s.

Figure 2

Model results (solid lines) are compared with experimental data (symbols) from open-circuit Ussing Chamber experiments using cultures of human nasal epithelium (HNE) (18,19). (Dashed lines) Prediction bands (two SDs around ensemble average) obtained with a Monte Carlo method (see the Supporting Material). Addition of drugs or changes in buffer solution occurred at 10 min (see vertical line). (A) Effect of inhibition of apical Na⁺ channels with amiloride. (B) Inhibition of the Na-K-2Cl cotransporter with bumetanide. (C) Reduction of basolateral Na⁺ from 140 mM to 3 mM by choline⁺ replacement. (D) Reduction of basolateral Cl⁻ from 119.6 mM to 3 mM by gluconate⁻ replacement.

Figure 3

Model predictions (solid lines) are compared with experimental data (symbols) from open-circuit Ussing Chamber experiments using cultures of human nasal epithelium (HNE) (18,19). (Dashed lines) Prediction bands (two SDs around ensemble average) obtained with a Monte Carlo method (see the Supporting Material). Addition of drug or changes in buffer solution occurred at 10 min (see vertical line). (A) Inhibition of the Na-K pump with ouabain. (B) Reduction of apical Na⁺ from 140 mM to 3 mM by choline⁺ replacement (18,19). (C) Reduction of apical Cl⁻ from 119.6 mM to 3 mM by gluconate⁻ replacement (18,19).