Mesenchymal Stromal Cells Accelerate Epithelial Tight Junction Assembly via the AMP-Activated Protein Kinase Pathway, Independently of Liver Kinase B1

Abstract

Background: Mesenchymal stromal cells (MSC) are fibroblast-like multipotent cells capable of tissue-repair properties. Given the essentiality of tight junctions (TJ) in epithelial integrity, we hypothesized that MSC modulate TJ formation, via the AMP-activated kinase (AMPK) pathway. Liver kinase-β1 (LKB1) and Ca²⁺-calmodulin-dependent protein kinase kinase (CaMKK) represent the main kinases that activate AMPK.

Methods: The in vitro Ca²⁺ switch from 5 μM to 1.8 mM was performed using epithelial Madin-Darby canine kidney (MDCK) cells cultured alone or cocultured with rat bone marrow-derived MSC or preexposed to MSC-conditioned medium. TJ assembly was measured by assessing ZO-1 relocation to cell-cell contacts. Experiments were conducted using MDCK stably expressing short-hairpin-RNA (shRNA) against LKB1 or luciferase (LUC, as controls). Compound STO-609 (50 μM) was used as CaMKK inhibitor.

Results: Following Ca²⁺ switch, ZO-1 relocation and phosphorylation/activation of AMPK were significantly higher in MDCK/MSC compared to MDCK. No difference in AMPK phosphorylation was observed between LKB1-shRNA and Luc-shRNA MDCK following Ca²⁺ switch. Conversely, incubation with STO-609 prior to Ca²⁺ switch prevented AMPK phosphorylation and ZO-1 relocation. MSC-conditioned medium slightly but significantly increased AMPK activation and accelerated TJ-associated distribution of ZO-1 post Ca²⁺ switch in comparison to regular medium.

Conclusions: MSC modulate the assembly of epithelial TJ, via the CaMKK/AMPK pathway independently of LKB1.

Figure 1

Role of the AMPK kinases, LKB1 and CaMKK, in AMPK activation and ZO-1 relocation following a Ca²⁺ switch in MDCK cells. Representative immunoblotting (a) and quantifications (b) of phospho-acetyl-Coa carboxylase (pACC), phospho-AMP-activated protein kinase (pAMPK), and total AMPK (AMPKt) in low Ca²⁺ conditions (S-MEM) and following Ca²⁺ switch using MDCK cells or LKB1-shRNA MDCK cells. Compounds STO-609 and C were used as CaMKK and AMPK inhibitors, respectively. Quantifications of immunoreactive signals were performed by stain-free method after normalization to total protein content of each lane. Quantifications of phospho-ACC, phospho-AMPK, and AMPKt signals following Ca²⁺ switch were calculated and expressed by the ratio to the immunoreactive signal of SMEM condition in each individual experiment (a). For the sake of bar-graph clarity (b), SMEM values of all experiments were normalized to 1 in order to represent mean ratios of phospho-ACC/total protein content and phospho-AMPK/AMPKt in different experimental conditions (b). Data are presented as mean ± SD; ns: not significant, ^∗∗p ≤ 0.01, ^∗∗∗p ≤ 0.001. Representative immunofluorescence (c) and quantifications (d) of ZO-1 deposits at increasing time points following Ca²⁺ switch in similar conditions as in (a) and (b) (scale bar: 16 μm). No statistically significant difference was observed between MDCK and LKB1-shRNA MDCK (ns: not significant). MDCK exposed to compound C (^∗p ≤ 0.01) or STO-609 (^§p ≤ 0.01) showed a significant reduction of ZO-1 relocation in comparison to control MDCK. Data are presented as mean ± SD.

Figure 2

Impact of mesenchymal stromal cells (MSC) on AMPK activation and ZO-1 relocation following a Ca²⁺ switch in MDCK cells. Representative immunoblotting (a) and quantifications (b) of phospho-acetyl-Coa carboxylase (pACC), phospho-AMP-activated protein kinase (pAMPK), and total AMPK (AMPKt) in low Ca²⁺ conditions (S-MEM) and following Ca²⁺ switch using MDCK cells or LKB1-shRNA MDCK cells, with versus without MSC. Compounds STO-609 and C were used as CaMKK and AMPK inhibitors, respectively. Quantifications of immunoreactive signals were performed by stain-free method after normalization to total protein content of each lane. Data are presented as mean ± SD; ns: not significant, ^∗∗p ≤ 0.01, ^∗∗∗p ≤ 0.001. Representative immunofluorescence (c) and quantifications (d) of ZO-1 deposits at increasing time points following Ca²⁺ switch in similar conditions as in (a) and (b) (scale bar: 16 μm). MSC/MDCK (i.e., MDCK (^§p ≤ 0.01) or MDCK LKB1-shRNA (^∗p ≤ 0.01)) cocultures show significantly increased ZO-1 deposits at 1-hour post Ca²⁺ switch in comparison to MDCK alone. At 2 hours post Ca²⁺ switch, no significant (ns) difference in ZO-1 lengths is observed between MDCK and MSC/MDCKs. Data are presented as mean ± SD.

Figure 3

Impact of mesenchymal stromal cell- (MSC-) conditioned medium (CM) on AMPK activation and ZO-1 relocation following a Ca²⁺ switch in MDCK cells. Representative immunoblotting (a) and quantifications (b) of phospho-AMP-activated protein kinase (pAMPK) and total AMPK (AMPKt) in low Ca²⁺ conditions (S-MEM) and following Ca²⁺ switch using MDCK cells exposed to regular versus MSC-preexposed medium. Quantifications of immunoreactive signals were performed by stain-free method after normalization to total protein content of each lane. Representative immunofluorescence (c) and quantifications (d) of ZO-1 deposits at increasing time points following Ca²⁺ switch in similar conditions as in (a) and (b) (scale bar: 16 μm). Data are presented as mean ± SD; ^∗∗∗p ≤ 0.001.