Retinoic acid and hydrocortisone strengthen the barrier function of human RPMI 2650 cells, a model for nasal epithelial permeability

Abstract

The nasal pathway represents an alternative route for non-invasive systemic administration of drugs. The main advantages of nasal drug delivery are the rapid onset of action, the avoidance of the first-pass metabolism in the liver and the easy applicability. In vitro cell culture systems offer an opportunity to model biological barriers. Our aim was to develop and characterize an in vitro model based on confluent layers of the human RPMI 2650 cell line. Retinoic acid, hydrocortisone and cyclic adenosine monophosphate, which influence cell attachment, growth and differentiation have been investigated on the barrier formation and function of the nasal epithelial cell layers. Real-time cell microelectronic sensing, a novel label-free technique was used for dynamic monitoring of cell growth and barrier properties of RPMI 2650 cells. Treatments enhanced the formation of adherens and tight intercellular junctions visualized by electron microscopy, the presence and localization of junctional proteins ZO-1 and β-catenin demonstrated by fluorescent immunohistochemistry, and the barrier function of nasal epithelial cell layers. The transepithelial resistance of the RPMI 2650 cell model reached 50 to 200 Ω × cm(2), the permeability coefficient for 4.4 kDa FITC-dextran was 9.3 to 17 × 10(-6) cm/s, in agreement with values measured on nasal mucosa from in vivo and ex vivo experiments. Based on these results human RPMI 2650 cells seem to be a suitable nasal epithelial model to test different pharmaceutical excipients and various novel formulations, such as nanoparticles for toxicity and permeability.

Fig. 1

Electron micrographs on the intercellular junctions of the RPMI 2650 layers treated with culture medium (a), 300 μg/mL retinoic acid (b), 500 nM hydrocortisone (c); 250 μM 3′–5′-cyclic adenosine monophosphate (d) for 24 h. Arrows indicate intercellular junctions, arrowheads show the length of tight junctions. N Nucleus, m mitochondrion, ER endoplasmatic reticulum, V microvilli; bar 500 nm. Control conditions: RPMI 2650 cells were grown in Eagle’s minimal essential medium supplemented with 10 % foetal bovine serum and 50 μg/mL gentamicin

Fig. 2

Immunohistochemical staining of RPMI 2650 cells for junctional proteins β-catenin and zona occludens-1 (ZO-1) visualized by confocal fluorescence microscopy. Cells were treated with retinoic acid (RA, 300 μg/mL), hydrocortisone (HC, 500 nM) or 3′–5′-cyclic adenosine monophosphate (cAMP, 250 μM) for 24 h; bar 10 μm. Control conditions: RPMI 2650 cells were grown in Eagle’s minimal essential medium supplemented with 10 % foetal bovine serum and 50 μg/mL gentamicin

Fig. 3

Cell index of RPMI 2650 layers treated with different concentrations of fetal bovine serum (panel a and b), and retinoic acid (RA, 0.01–300 μg/mL (c), 300 μg/mL (d)), hydrocortisone (HC, 500 nM), and 3′–5′-cyclic adenosine monophosphate (cAMP, 250 μM) for 24 h measured by real-time impedance monitoring (panel c and d). Data are presented as mean ± SD, n = 4. Cell index is expressed as an arbitrary unit and calculated from impedance measurements between cells and sensors (xCELLigence, Roche). Control conditions: RPMI 2650 cells were grown in Eagle’s minimal essential medium supplemented with 10 % foetal bovine serum and 50 μg/mL gentamicin

Fig. 4

Barrier functions measured by transepithelial resistance (a) and permeability for FITC-labelled dextran 4.4 kDa (b) of RPMI 2650 layers treated with retinoic acid (RA, 300 μg/mL), hydrocortisone (HC, 500 nM) or 3′–5′-cyclic adenosine monophosphate (cAMP, 250 μM). P_app, apparent permeability coefficient. Control conditions: RPMI 2650 cells were grown in Eagle’s minimal essential medium supplemented with 10 % foetal bovine serum and 50 μg/mL gentamicin