Podocytes regulate neutrophil recruitment by glomerular endothelial cells via IL-6-mediated crosstalk

Abstract

Stromal cells actively modulate the inflammatory process, in part by influencing the ability of neighboring endothelial cells to support the recruitment of circulating leukocytes. We hypothesized that podocytes influence the ability of glomerular endothelial cells (GEnCs) to recruit neutrophils during inflammation. To address this, human podocytes and human GEnCs were cultured on opposite sides of porous inserts and then treated with or without increasing concentrations of TNF-α prior to addition of neutrophils. The presence of podocytes significantly reduced neutrophil recruitment to GEnCs by up to 50% when cultures were treated with high-dose TNF-α (100 U/ml), when compared with GEnC monocultures. Importantly, this phenomenon was dependent on paracrine actions of soluble IL-6, predominantly released by podocytes. A similar response was absent when HUVECs were cocultured with podocytes, indicating a tissue-specific phenomenon. Suppressor of cytokine signaling 3 elicited the immunosuppressive actions of IL-6 in a process that disrupted the presentation of chemokines on GEnCs by altering the expression of the duffy Ag receptor for chemokines. Interestingly, suppressor of cytokine signaling 3 knockdown in GEnCs upregulated duffy Ag receptor for chemokines and CXCL5 expression, thereby restoring the neutrophil recruitment. In summary, these studies reveal that podocytes can negatively regulate neutrophil recruitment to inflamed GEnCs by modulating IL-6 signaling, identifying a potential novel anti-inflammatory role of IL-6 in renal glomeruli.

FIGURE 1.

Illustration of the coculture system in the transwell. (A) Coculture of GEnCs and podocytes with or without being in juxtaposition. Podocytes and GEnCs on the opposite side of the porous insert formed cocultures in juxtaposition. Cocultures without being in juxtaposition were formed allowing the cell types to communicate through release of soluble mediators. Neutrophil recruitment was assessed to represent all cells that had bound to the GEnCs (i.e., firmly adherent and transmigrated). (B) Fluorescent micrographs showing neutrophil binding to and transmigration through GEnC/podocyte cocultures: (i) Apical surface of transwell insert with neutrophils adherent to GEnCs in coculture, (ii) basal side of transwell insert with transmigrated neutrophils adjacent to podocytes in coculture, (iii) apical surface of transwell insert with neutrophils adherent to GEnCs in monoculture, and (iv) basal side of transwell insert with transmigrated neutrophils in GEnC monocultures. Original magnification ×40.

FIGURE 2.

Recruitment of neutrophil to TNF-α–treated GEnC mono- and cocultures. (A) Primary or (B) immortalized cell line cultures of GEnCs and podocytes were cultured in juxtaposition (open bars) and stimulated with TNF-α (0-100 U/ml). GEnC monocultures (filled bars) were assessed as controls. Neutrophils were allowed to adhere and migrate through the constructs. Neutrophil recruitment was assessed at 1 h and taken to represent all cells that had bound to the GEnCs (i.e., firmly adherent [hashed bars] and transmigrated [plain bars]). Data are means ± SEM. (A) n = 4 and (B) n = 11–32. **p < 0.01, ***p < 0.0001 for the effect of cytokine treatment as determined by a paired Wilcoxon signed-rank test between treatments within cultures. ^δp < 0.05, ^δδδp < 0.0001 for difference between cocultures and monocultures as determined by a Mann–Whitney U test.

FIGURE 3.

Role of soluble mediators in regulating neutrophil recruitment to GEnCs. GEnCs were seeded onto the apical surface (i.e., inside) a porous filter. Podocytes were seeded onto the bottom of the well. Cocultures were formed by placing the GEnC-coated filter above the podocyte-coated well, allowing the cell types to communicate through release of soluble mediators. Cytokine-induced neutrophil recruitment was assessed at 1 h. Data are means ± SEM (n = 4). A Friedman test shows a significant effect of cytokine treatments on neutrophil recruitment (p < 0.001). *p < 0.05 comparing the effects of TNF-α treatment to untreated cultures by a Dunnett post hoc test, ^δp < 0.05 comparing cocultures (open bars representing cocultures) to monocultures (filled bars) by an unpaired t test.

FIGURE 4.

Role for IL-6 in negative regulation of neutrophil recruitment in cocultures. (A) Detection of IL-6 in supernatants of GEnC and podocyte cocultures (▴) and monocultures of GEnC (▪) and podocytes (●) that were treated with and without TNF-α. Data are means ± SEM (n = 4). Two-way ANOVA shows a significant effect of cytokine treatment on IL-6 production (p < 0.05). *p < 0.05 comparing IL-6 production between cocultures and GEnC monocultures by a Bonferroni post hoc test. (B) Neutralization of IL-6 in cultures by incorporation of 5 μg/ml anti–IL-6 Ab (labeled as IL-6 block) into TNF-α (100 U/ml)–stimulated cultures for 24 h starting from the time of establishing the cocultures. Neutrophil recruitment was assessed at 1 h and is expressed as percentage of cells recruited. Data are represented as minimum to maximum levels of recruitment (n = 5). The effects were compared with an isotype-matched mouse IgG in these experiments. *p < 0.05, **p < 0.001 comparing the effect of IL-6 neutralization with controls by a paired t test. (C) Recapitulation of coculture by addition of exogenous IL-6 in GEnC monocultures (n = 4). *p < 0.05 by Mann–Whitney U test when comparing treatments, ***p < 0.0001 comparing the effect of cytokine treatment by a paired Wilcoxon signed-rank test between treatments within cultures, ***p < 0.0001 comparing the effect of cytokine treatment by a paired Wilcoxon signed rank test between treatments within cultures.

FIGURE 5.

SOCS3-mediated downregulation of neutrophil recruitment in cocultures of GEnCs and podocytes. (A) Detection of SOCS3 mRNA in TNF-α (100 U/ml)–stimulated cultures of GEnCs (filled bars) and podocytes (open bars) by quantitative PCR following SOCS3 knockdown with three different oligomers (labeled A [mixture of three oligomers] and B and C [individual oligomers]) or a scrambled control. Lipofectamine presence is indicated, whereas “untreated” indicates absence of Lipofectamine, scrambled control, or SOCS oligomers. Data are minimum to maximum (n = 3–6). A Friedman test compared the effect of SOCS3 knockdown (p < 0.05). *p < 0.05 comparing controls within cultures by a Dunnett post hoc test. (B) Western blotting confirming the SOCS3 protein knockdown in GEnCs and podocytes compared with controls (untreated cultures, Lipofectimine-treated cultures, and scrambled RNA-treated cultures). β-actin was used as loading control. (C) Neutrophil recruitment in cocultures (open bars) and GEnC monoculture (filled bars) with SOCS3 knockdown (−/−) with three different oligomers (labeled A [mixture of three oligomers] and B and C [individual oligomers]) or a scrambled control. Lipofectamine presence is indicated, whereas “untreated” indicates absence of Lipofectamine, scrambled control, or SOCS oligomers. Neutrophil recruitment in cocultures was also tested with SOCS3 knockdown in only one cell type. Neutrophil recruitment was assessed in TNF-α (100 U/ml)–treated cultures at 1 h. Data are minimum to maximum of recruited cells (n = 3–7). A Friedman test compared the effect of SOCS3 knockdown (p < 0.05). *p < 0.05, **p < 0.01 comparing with controls by a Dunnett post hoc test.

FIGURE 6.

Role of CXC receptors in supporting neutrophil migration. CXCR1 and CXCR2 receptors were blocked either individually or in combination using respective blocking Abs (4 μg/ml) on neutrophils for 15 min prior to addition in both GEnC monocultures (filled bars) and cocultures (open bars) stimulated with TNF-α (100 U/ml). Neutrophil recruitment was assessed at 1 h and expressed as percentage of added cells. The effects were also compared with an isotype-matched mouse IgG in these experiments. Data are means ± SEM (n = 5). ANOVA shows a significant effect of treatment on neutrophil migration (p < 0.05). *p < 0.05, **p < 0.01, ***p < 0.0001 comparing the effect of receptor blocking on neutrophil migration with their respective controls within cultures by a Bonferroni post hoc test. Untreated, absence of isotype control Ab, CXCR1, and anti-CXCR2 blocking Abs.

FIGURE 7.

Detection of CXCL1, CXCL5, and CXCL8 in cocultures and monocultures. Detection of soluble (A) CXCL1, (B) CXCL5, and (C) CXCL8 in supernatants collected from TNF-α–stimulated cocultures (●), GEnC monocultures (▪), and podocyte monocultures (▴). Data expressed as means ± SEM (n = 3). *p < 0.05, **p < 0.01, ***p < 0.0001 comparing the difference in chemokine expression between cocultures and GEnC monocultures by two-way ANOVA.

FIGURE 8.

Neutralization of CXCL1, CXCL5, and CXCL8 in cocultures and monocultures. Cultures were treated with function neutralizing Abs against (A) CXCL1 (2 μg/ml), (B) CXCL5 (5 μg/ml), and (C) CXCL8 (10 μg/ml) for 24 h prior to addition of neutrophils. Data are means ± SEM (n = 5–6). *p < 0.05, **p < 0.01, ***p < 0.0001 comparing the effect of blocking chemokine function in cultures with their controls by a paired t test.

FIGURE 9.

Effect of SOCS3 deletion in the upregulation of genes that support neutrophil recruitment in cocultures. Quantification of expression of genes (A) CXCL5, (B) DARC, (C) IL-6, and (D) TNFR1 that supported the neutrophil recruitment in cocultures, prior to and after depletion of SOCS3, in GEnCs and podocytes from mono- and cocultures compared with monocultures without SOCS3 knockdown. Data are expressed as fold change to TNF-α–stimulated monocultures (n = 5). *p < 0.05, **p < 0.01, ***p < 0.0001 comparing the difference in gene expression with the stimulated monocultures without SOCS3 knockdown by an unpaired t test.