A cell-based computational modeling approach for developing site-directed molecular probes

Abstract

Modeling the local absorption and retention patterns of membrane-permeant small molecules in a cellular context could facilitate development of site-directed chemical agents for bioimaging or therapeutic applications. Here, we present an integrative approach to this problem, combining in silico computational models, in vitro cell based assays and in vivo biodistribution studies. To target small molecule probes to the epithelial cells of the upper airways, a multiscale computational model of the lung was first used as a screening tool, in silico. Following virtual screening, cell monolayers differentiated on microfabricated pore arrays and multilayer cultures of primary human bronchial epithelial cells differentiated in an air-liquid interface were used to test the local absorption and intracellular retention patterns of selected probes, in vitro. Lastly, experiments involving visualization of bioimaging probe distribution in the lungs after local and systemic administration were used to test the relevance of computational models and cell-based assays, in vivo. The results of in vivo experiments were consistent with the results of in silico simulations, indicating that mitochondrial accumulation of membrane permeant, hydrophilic cations can be used to maximize local exposure and retention, specifically in the upper airways after intratracheal administration.

Figure 1. General methodology of integrative, cell based transport modeling.

A) For computational simulations at the cellular level, a monobasic compound diffuses across a phospholipid bilayer and undergoes ionization and partition/binding in each compartment. The neutral form of the monobasic molecule is indicated as [M], and the protonated, cationic form of the molecule is indicated as [MH⁺]. B) For computational simulations at the histological level, each airway generation is modeled as a tube lined by epithelial cells; as molecules are absorbed over time, the drug concentration in the lumen decreases accompanied by an increase in drug concentration in the circulation C) For computational simulations at the organ level, the lung is modeled as a branching tree, with airway generation modeled as a cylinder, from the trachea to the alveoli. D) Experimental design of insert system with patterned pore arrays on membrane support for viewing lateral transport of fluorescent molecules along the plane of a cell monolayer, away from a point source. E) Transmitted light image of a 5×5, 3 µm diameter pore array (20 µm spacing) on a polyester membrane. F) Transmitted light image of an MDCK cell monolayer above a membrane support with 3×3, 3 µm diameter pore array (40 µm spacing). Scale bar: 40 µm. G) 3D reconstruction of confocal images of the distribution of three fluorescent probes added to the uppermost surface of NHBE cell multilayers grown on air-liquid interface cultures on porous membrane support. Each 3D plane is composed of the image with the fluorescent channel; red (MTR), blue (Hoe), and green (LTG). H) Illustration of the tiling algorithm used to visualize and quantify the distribution of Hoe and MTR in lung cryosections, after IT and IV coadministration of the probes.

Figure 2. Virtual screening of monobasic compounds based differential tissue distribution in the airways and alveoli.

The combinations of logP_n and pK_a were used as input. For simulations, the initial dose was set to 1 mg/kg for airways and alveoli. Contour lines indicate: A) The calculated AUC (unit: mg/ml*min) in airways; B) The AUC (unit: mg/ml*min) in alveoli; C) The AUC contrast ratio of airways to alveoli; D) The mass percentage (%) in alveoli relative to the total mass in lung; E) The mass percentage (%) in airways relative to the total mass in lung; F) The mass ratio of alveoli to airway. Matlab scripts used to generate plots A–C (Text S1) and D–F (Text S2) are included in the supplementary materials.

Figure 3. Simulations of local pharmacokinetics of MTR and Hoe after IV an IT administration.

A) The simulated tissue concentration in airways (dash line) and alveoli (solid line) of MTR administered by IT instillation (Matlab script used to generate this plot is included as Text S3 in the supplementary materials); B) The simulated tissue concentration in airways (dash line) and alveoli (solid line) of Hoe administered by IT instillation (Matlab script used to generate this plot is included as Text S4 in the supplementary materials); C) The simulated tissue concentration in airways (dash line) and alveoli (solid line) of MTR administered by IV injection (Matlab script used to generate this plot is included as Text S5 in the supplementary materials); D) The simulated tissue concentration in airways (dash line) and alveoli (solid line) of Hoe administered by IV injection (Matlab script used to generate this plot is included as Text S6 in the supplementary materials).

Figure 4. Probing the intracellular retention of Hoe along the plane of a cell monolayer.

For the experiments, Hoe was added to the basolateral compartment and incubated for 3 hrs, with cell monolayers sitting on top of patterned pore arrays. Red spots indicate the location of pores; Scale bar: 80 µm. Cells were imaged using the DAPI channel of an epifluorescence microscope. A) 5×5 array of 3 µm pores with 20 µm spacing; B) 3×3 array of 3 µm pores with 40 µm spacing; C) 3×3 array of 3 µm pores with 80 µm spacing; D) 3×3 array of 3 µm pore array with 160 µm spacing; E) Fluorescent images of a cell monolayer incubated for 3 hours in the presence of Hoe in the basolateral compartment; F) corresponding measurements of fluorescence intensity of cells in A), showing the average fluorescence of each nucleus normalized by the average fluorescence of the nucleus closest to the pore at the 3 hr time point, and plotted as mean ± s.d. (n = 6). G) Fluorescence image of cell monolayer on a 3×3 array of 3 µm pores with 40 µm spacing after 2 hr incubation with Hoe and BCECF-AM in the basolateral compartment; H) FITC channel corresponding to BCECF staining of the same cells as in; I) Image overlays of C and D.

Figure 5. Probing the intracellular retention of MTR along the plane of a cell monolayer.

Cell monolayers on pore arrays were incubated for 2 hr with Hoe and MTR in the basolateral compartment. White spots indicate the location of pores; Scale bar: 20 µm. A) Fluorescent image acquired with the DAPI channel showing Hoe diffusing on a cell monolayer sitting on top of a single pore of a 3×3 array of 3 µm pores with 160 µm spacing; B) Same field as in A, visualized with the TRITC channel to show the staining of MTR; C) Overlay of A and B showing the overlapping Hoe (blue) and MTR (red) staining patterns. D) Plots of the fluorescence intensity of Hoe and MTR, separated by 0, 1, 2 or 3 layers of cells from a pore, and normalized by the fluorescence intensity of the cell closest to the pore; asterisk indicates a statistically significant difference using Student's T-test; p<0.05; n = 6).

Figure 6. Fluorescent confocal images of NHBE cell multilayers on the porous membrane with Z-stacks.

Cell multilayers were stained with MTR, Hoe and LTG. Each compartment (membrane inserts (bottom), inner cell layers, surface cell layer, and apical compartment (top)) through z-axis were indicated with the red arrows in x–z planes while cell nuclei and cytoplasm in x–y planes. The panel to the left shows an x, y cross section through the apical surface layer of the cell multilayer. The panel to the right shows an x, y cross section through the inner cell layer of the cell multilayer. Scale bar: 20 µm.

Figure 7. Tiled fluorescent micrographs of coronal cryosections obtained from the left lungs of mice.

Mice received either an IV (A, C, E, G) or IT (B, D, F, H) dose of a mixture of Hoe and MTR. A) DAPI channel fluorescence image showing Hoe distribution following IV administration; B) DAPI channel fluorescence image showing Hoe distribution following IT administration; C) High magnification view of the boxed region in A; D) High magnification view of the boxed region in B; E) TRITC channel fluorescence image showing MTR distribution following IV administration; F) TRITC channel fluorescence image showing MTR distribution following IT administration; G) High magnification view of the boxed region in E; H) High magnification view of the boxed region in F. Scale bar = 1 mm. Asterisks mark the cross-sections of the airways, apparent as ellipsoids at high magnification.