Co-Culture of Glomerular Endothelial Cells and Podocytes in a Custom-Designed Glomerulus-on-a-Chip Model Improves the Filtration Barrier Integrity and Affects the Glomerular Cell Phenotype

Abstract

Crosstalk between glomerular endothelial cells and glomerular epithelial cells (podocytes) is increasingly becoming apparent as a crucial mechanism to maintain the integrity of the glomerular filtration barrier. However, in vitro studies directly investigating the effect of this crosstalk on the glomerular filtration barrier are scarce because of the lack of suitable experimental models. Therefore, we developed a custom-made glomerulus-on-a-chip model recapitulating the glomerular filtration barrier, in which we investigated the effects of co-culture of glomerular endothelial cells and podocytes on filtration barrier function and the phenotype of these respective cell types. The custom-made glomerulus-on-a-chip model was designed using soft lithography. The chip consisted of two parallel microfluidic channels separated by a semi-permeable polycarbonate membrane. The glycocalyx was visualized by wheat germ agglutinin staining and the barrier integrity of the glomerulus-on-a-chip model was determined by measuring the transport rate of fluorescently labelled dextran from the top to the bottom channel. The effect of crosstalk on the transcriptome of glomerular endothelial cells and podocytes was investigated via RNA-sequencing. Glomerular endothelial cells and podocytes were successfully cultured on opposite sides of the membrane in our glomerulus-on-a-chip model using a polydopamine and collagen A double coating. Barrier integrity of the chip model was significantly improved when glomerular endothelial cells were co-cultured with podocytes compared to monocultures of either glomerular endothelial cells or podocytes. Co-culture enlarged the surface area of podocyte foot processes and increased the thickness of the glycocalyx. RNA-sequencing analysis revealed the regulation of cellular pathways involved in cellular differentiation and cellular adhesion as a result of the interaction between glomerular endothelial cells and podocytes. We present a novel custom-made glomerulus-on-a-chip co-culture model and demonstrated for the first time using a glomerulus-on-a-chip model that co-culture affects the morphology and transcriptional phenotype of glomerular endothelial cells and podocytes. Moreover, we showed that co-culture improves barrier function as a relevant functional readout for clinical translation. This model can be used in future studies to investigate specific glomerular paracrine pathways and unravel the role of glomerular crosstalk in glomerular (patho) physiology.

Keywords: biological barrier; co-culture; crosstalk; glomerular endothelial cells; glomerular filtration barrier; glomerulus; glomerulus-on-a-chip; organ-on-a-chip; podocytes.

Figure 1

Overview of the production of the glomerulus-on-a-chip device. (a) Schematic design of the design of the glomerulus-on-a-chip device. (b) Schematic overview of the experimental steps to manufacture the SU-8 molds and polycarbonate membranes in the microfluidic device. PC: polycarbonate. Figure 1b was created using Biorender.com.

Figure 2

Podocytes and GEnC grow and differentiate on polycarbonate membrane with a polydopamine and collagen A double coating. mGEnC and MPC-5 were seeded on polycarbonate membranes and grown for 1 and 2 weeks, respectively. Membranes were left uncoated, coated with 1 µg/cm² bovine fibronectin, coated with 1 mg/mL collagen A or coated with 2 mg/mL polydopamine and 1 mg/mL collagen A. (a) Representative fluorescent images of mGEnC stained for the actin cytoskeleton and Hoechst 33,342 as nuclear staining. (b) Representative fluorescent images of MPC-5 stained for the actin cytoskeleton and Hoechst 33,342 as nuclear staining (n = 3).

Figure 3

Podocytes and GEnC grow and differentiate in a glomerulus-on-a-chip device. mGEnC were seeded into the top channel of the chip and were allowed to grow for 1 week on the membrane. MPC-5 were seeded in the bottom channel of the chip and were allowed to grow for 2 weeks on the other side of the membrane. (a) Representative fluorescent images of mGEnC grown for 1 week in the chips stained for Hoechst 33,342 (blue), vWF (green), and actin (red). (b) Representative fluorescent images of MPC-5 grown for 2 weeks in the chips stained for Hoechst 33342 (blue), synaptopodin (green), and actin (red). (c) Representative fluorescent image of mGEnC stained with CellTracker Green 72 h after seeding. (d) Representative fluorescent image MPC-5 stained with CellTracker Red 72 h after seeding. (e) 3D reconstruction of the Z-stack made of the mGEnC-MPC-5 co-culture in the chip 72 h after seeding. Prior to seeding, mGEnC were stained with CellTracker Green and MPC-5 with CellTracker Red.

Figure 4

Co-culture of GEnC and podocytes affect podocyte morphology, glycocalyx thickness and barrier function. mGEnC were seeded in the top channel of the chip and were allowed to grow for 1 week on the top of the membrane. One week prior to seeding mGEnC, MPC-5 were seeded in the bottom channel of the chip and were allowed to grow for 2 weeks on the bottom side of the membrane. (a) Representative fluorescent image of synaptopodin staining in MPC-5 in monoculture. (b) Representative fluorescent image of synaptopodin staining in MPC-5 in co-culture. (c) Quantification of filopodia surface area MPC-5 in mono- and co-culture. (d) 3D reconstruction of the WGA staining of mGEnC cultured alone. (e) 3D reconstruction of the WGA staining of mGEnC cultured together with MPC-5. (f) Quantification of the fluorescent intensity of the WGA staining per Z-stack slice. (g) Calculation of the AUC of the fluorescent intensity of the WGA staining. (h) Fold change in fluorescent intensity of 155-kDA dextran-TRITC in chips without cells and chips with a monoculture of mGEnC. (i) Fold change in fluorescent intensity of 155-kDA dextran-TRITC in chips without cells and chips with a monoculture of MPC-5. (j) Fold change in fluorescent intensity of 155-kDA dextran-TRITC in chips with a monoculture of mGEnC and chips with a co-culture of mGEnC and MPC-5. (k) Fold change in fluorescent intensity of 155-kDA dextran-TRITC in chips with a monoculture of MPC-5 and a co-culture of mGEnC and MPC-5. * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 5

Co-culture of GEnC and podocytes activates pathways involved in cell differentiation and cell adhesion. (a) Schematic overview of experimental design. mGEnC and MPC-5 were cultured in chips alone, or the cells were cultured in co-culture whereafter the mGEnCs and MPC-5 were separately isolated. (b) Principle component analysis (PCA), visualizing the variance in gene expression profiles between the different conditions. (c) Venn-diagram showing the total number of genes and significantly up- and down-regulated genes in mGEnC (red) and MPC-5 (blue). Genes were selected using a logFC of <−1.5 and >1.5. (d) The top 10 upregulated GO terms in biological processes in mGEnC and (e) the top 10 downregulated GO terms in biological processes in mGEnC as a result of co-culture. (f) The top 10 upregulated GO terms in biological processes in MPC-5 (g) and the top 10 downregulated GO terms in biological processes in MPC-5 as a result of co-culture. Gene functional annotation was obtained via DAVID tools and the top 10 most upregulated and top 10 most downregulated GO terms are shown based on the calculated -log p-values.