Therapeutic targeting of chronic kidney disease-associated DAMPs differentially contributing to vascular pathology

Abstract

Chronic Kidney Disease (CKD) is associated with markedly increased cardiovascular (CV) morbidity and mortality. Chronic inflammation, a hallmark of both CKD and CV diseases (CVD), is believed to drive this association. Pro-inflammatory endogenous TLR agonists, Damage-Associated Molecular Patterns (DAMPs), have been found elevated in CKD patients' plasma and suggested to promote CVD, however, confirmation of their involvement, the underlying mechanism(s), the extent to which individual DAMPs contribute to vascular pathology in CKD and the evaluation of potential therapeutic strategies, have remained largely undescribed. A multi-TLR inhibitor, soluble TLR2, abrogated chronic vascular inflammatory responses and the increased aortic atherosclerosis-associated gene expression observed in nephropathic mice, without compromising infection clearance. Mechanistically, we confirmed elevation of 4 TLR DAMPs in CKD patients' plasma, namely Hsp70, Hyaluronic acid, HMGB-1 and Calprotectin, which displayed different abilities to promote key cellular responses associated with vascular inflammation and progression of atherosclerosis in a TLR-dependent manner. These included loss of trans-endothelial resistance, enhanced monocyte migration, increased cytokine production, and foam cell formation by macrophages, the latter via cholesterol efflux inhibition. Calprotectin and Hsp70 most consistently affected these functions. Calprotectin was further elevated in CVD-diagnosed CKD patients and strongly correlated with the predictor of CV events CRP. In nephropathic mice, Calprotectin blockade robustly reduced vascular chronic inflammatory responses and pro-atherosclerotic gene expression in the blood and aorta. Taken together, these findings demonstrated the critical extent to which the DAMP-TLR pathway contributes to vascular inflammatory and atherogenic responses in CKD, revealed the mechanistic contribution of specific DAMPs and described two alternatives therapeutic approaches to reduce chronic vascular inflammation and lower CV pathology in CKD.

Keywords: anti-inflammatory intervention strategies; chronic kidney disease; damage-associated molecular patterns (DAMPs); toll-like receptors (TLRs); vascular inflammation.

Figure 1

Soluble TLR2 administration inhibits AAN-induced systemic inflammatory and pro-atherosclerotic responses in vivo. C57BL/6J mice (n=5 per group) were injected intraperitoneally 4 times at 3-day intervals with AA (2.5 mg/kg) or PBS, in the presence or absence of sTLR2 (12.5µg/kg) (A). Blood and aortas were obtained at Day 10 or Day 21, Day 0 being the day of the first injection. Cytokine plasma levels were determined by ELISA (B) and innate leukocyte proportions by flow cytometry (C, percentage of gated single cells shown). Open circles denote individual animals, horizontal bars indicate the median value. */#, p<0.05; ***/###, p<0.005 (*, AAN vs PBS; #, AAN +sTLR2 vs AAN), Mann-Whitney U test. The changes between AAN+ Paquinimod and Control groups were not significant. Volcano plots (D, E) compare the effect of AAN and AAN + sTLR2 on blood inflammation and immune responses-associated (D) or aortic atherosclerosis-associated (E) gene expression at Day 21. Red (upregulated, fold change ≥ 2) and green (downregulated, fold change ≤ 0.5) circles represent single genes significantly affected (p value < 0.05, represented by the horizontal line) compared to PBS control. Heatmap (F) displays experimental group hierarchical clustering, as determined according to the aortic expression levels of the 84 genes tested. Each column represents a sample; each row represents a gene; the relative gene expression scale is depicted on the right. At Day 21 (G) AAN mice were i.p injected with live S.epidermidis (5x10⁸ cfu/mouse), in the presence or absence of sTLR2 (12.5µg/kg). Peritoneal lavages and blood were obtained at the indicated time points and bacterial numbers determined by colony counting after growth.

Figure 2

Plasma levels of TLR DAMPs in healthy individuals and CKD patients. Concentrations of TLR DAMPs in plasma from healthy donors and Stage 5 CKD patients. Horizontal red bars denote the median value, open circles denote individual donors. Hsp70, Hsp60, HA, Calprotectin, Fibronectin: Healthy n=30, CKD n=35, CKD + CVD n=38; Decorin: Healthy n=20, CKD n=25. HMGB-1, Histone-DNA complexes: Healthy n=12, CKD n=15. *, p<0.05; ***, p<0.005 (CKD vs Healthy), Mann-Whitney U test.

Figure 3

CKD-associated DAMPs decrease transendothelial resistance and induce limited pro-inflammatory responses by human arterial endothelial cells. (A-C) Triplicate cultures of human umbilical arterial endothelial cells were grown to confluence prior to stimulation or not (NS) with single (A) or combined CKD-associated DAMPs (B, C) at the indicated concentrations (A, B) or 100 ng/ml (C). Trans-endothelial electrical resistance (TER) was measured at the indicated time points (A, B) and ZO-1 expression (C) was quantified at cell-to-cell contacts 5h following DAMP stimulation. White arrows in (C) indicate preferential ZO-1 localisation at cell-to-cell in resting conditions (NS) or in the cytoplasm following DAMP treatment. Individual data points from 3 independent experiments, (mean +/- SD) are shown (A, B) or 1 experiment representative of 3 (C) *, p<0.05; **, p<0.01; ***, p<0.005, Stimulation vs no stimulation, unpaired Student’s t test. D-F. Triplicate cultures of human aortic endothelial cells were stimulated (18h) with LPS (10 ng/ml), Pam₃CSK₄ (P3C, 500 ng/ml) or the indicated concentrations of Calprotectin, Hsp70, HMGB-1, or HA, alone (D, F) or combined (E, F, 1 µg/ml). Cytokine levels in culture supernatants (D, E) are shown as individual data points from one experiment representative of 4 (mean +/- SD). *, p<0.05; **, p<0.01; ***, p<0.005, Stimulation vs no stimulation, unpaired Student’s t test. ICAM-1 and VCAM-1 surface expression levels are shown (F, 20,000 cells/condition), from 1 experiment representative of 3.

Figure 4

CKD-associated DAMPs increase the migratory capacity of monocytes towards MCP-1 in a TLR-dependent manner. Triplicate cultures of Mono-Mac 6 monocytes were stimulated (18h, 37°C) with LPS (10 ng/ml) or the indicated DAMPs (1 µg/ml), alone or in combination after pre-exposure (1h, 37°C, B, C) or not (A) to a combination of anti-TLR2 and anti-TLR4 blocking antibodies (5 µg/ml) or the relevant isotype control (10 µg/ml). (A, B) Cells were subsequently starved in serum-free medium for 1 h prior to seeding (200,000 cells, in triplicates) in the top chamber of 8 µm pores trans-wells. The bottom compartment was filled with RPMI + 10% serum + MCP-1 (50 ng/ml). Cell numbers were counted in the bottom compartment after 2h, 4h, 6h (A) and 24h (A, B). Results shown are the mean +/- SD from one experiment representative of 7 (A, LPS and all DAMPs combined) or 3 (A, single DAMPs), or from 3 experiments (B, open circles denote individual experiments). *, p<0.05; **, p<0.01; ***, p<0.005; *, Stimulation vs no stimulation or ^#, anti-TLR antibodies vs isotype, unpaired Student’s t test (A), paired Student’s t test (B). Cells were then analysed by flow cytometry for the levels of CCR2 (C). Results are shown as mean fluorescence intensity (MFI) for 20,000 cells/condition from 3 independent experiments. Identical symbols identify paired results. *, p<0.05, All DAMPs vs no stimulation (NS), paired t-test.

Figure 5

CKD-associated DAMPs induce pro-inflammatory mediator production in a TLR-dependent manner. (A) Triplicate cultures of primary monocyte M-CSF-derived macrophages were stimulated (18h, 37°C) with LPS (10 ng/ml), Pam₃CSK₄ (P3C, 500 ng/ml) or the indicated concentrations of Calprotectin, Hsp70, HMGB-1, or HA. (B) Triplicate cultures of macrophages were pre-exposed (1h, 37°C) to anti-TLR2 or anti-TLR4 blocking antibodies (10 µg/ml) or the relevant isotype control (I.C), prior to stimulation (18h, 37°C) with all DAMPs combined (each at 1 µg/ml). Levels of the indicated cytokines in culture supernatants were determined by ELISA. Results shown in (A) are from at least 6 experiments performed with cells from different donors, each represented by an open circle. Horizontal lines in boxes denote the median value. (B) shows the average cytokine production relative to control for 3 experiments, each depicted by an open circle. */#, p <0.05; **/##, p <0.01; ***/###, p <0.005, (A) Stimulation vs no stimulation, Wilcoxon signed-rank test (paired samples), (B) *, Stimulation vs no stimulation, ^#, αTLR antibody vs isotype control (I.C.), unpaired Student’s t test.

Figure 6

CKD-associated DAMPs promote foam cell formation in a TLR-dependent manner by reducing cholesterol efflux. (A, B) M-CSF-differentiated macrophages were exposed (24h) to LDL (25 µg/ml) in the presence or absence of the indicated DAMPs, alone or in combination (1 µg/ml) and of the indicated anti-TLR blocking antibodies or relevant isotype control (10 µg/ml, B) prior to staining with Oil Red-O for lipid visualisation by light microscopy (representative images shown). Plots show the percentage of foam cells in each condition. (C) M-CSF-differentiated macrophages were starved in medium supplemented with 0.2% fatty-acid free BSA (18h) in the presence or absence or the indicated DAMPs, alone or in combination (1 µg/ml) before addition of Dil-conjugated acetylated LDL (Dil-AcLDL, 10 µg/ml). After 24h, internalised Dil-AcLDL was quantified by flow cytometry (20,000 events/condition, MFI shown) (B) Triplicate cultures of M-CSF-differentiated macrophages were loaded with BODIPY-labelled cholesterol (5 µM, 18h), before medium removal and exposure to the indicated DAMPs, alone or in combination (1 µg/ml). After equilibration (1h), medium was removed and replaced with fresh medium containing the same concentrations of DAMPs in the presence of 10% FCS as a cholesterol acceptor. BODIPY-associated fluorescence was measured in culture supernatants at the indicated time points. Results are mean +/- SD obtained with macrophages prepared from 3 different donors (each donor shown, A, D) or representative of 3 donors tested (B), or MFI from 20,000 macrophages from 3 donors tested (C, each donor shown). Open circles show individual data points for each mean (A, B, 5 individual fields of view per experimental group, (D), individual triplicates) *, p<0.05; **, p<0.01; ***, p<0.005; *DAMP stimulation vs no DAMP (A, B, D) or as indicated (C), unpaired Student’s t test.

Figure 7

Calprotectin blocking inhibits AAN-induced systemic inflammatory and pro-atherosclerotic responses in vivo. (A) Concentrations of TLR DAMPs in plasma from Stage 5 CKD patients without prior CVD diagnosis or CKD patients with prior CVD diagnosis (B). Horizontal red bars denote the median value, open circles denote individual donors. *, p<0.05 (CKD+CVD vs CKD), Mann-Whitney U test. (B) Correlation between plasma levels of Calprotectin and CRP or HA and CRP in CKD+CVD patients. Statistical analysis was done using a Spearman’s rank correlation test. (C–H). C57BL/6J mice (n=5 per group) were injected intraperitoneally 4 times at 3 days intervals with AA (2.5 mg/kg) or PBS, in the presence or absence of Paquinimod (1 mg/kg). Blood and aortas were obtained at Day 10 (blood only) or Day 21, Day 0 being the day of the first injection. DAMP (C) and cytokine (D) plasma levels were determined by ELISA, and innate leukocyte proportions (percentage of gated single cells shown) and PSLG-1 expression levels (E) by flow cytometry. Circles denote individual animals, horizontal bars indicate the median value. */#, p<0.05; **/##, p<0.01; ***, p<0.005 (*, AAN vs PBS; #, AAN +Paquinimod vs AAN), Mann-Whitney U test. The changes between AAN+ Paquinimod and Control groups were not significant. Volcano plots (F, G) compare the effect of AAN and AAN + Paquinimod on inflammation and immune responses in the blood (F) or atherosclerosis-associated gene expression in aortas (G) at Day 21. Red (upregulated, fold change ≥ 2) and green (downregulated, fold change ≤ 0.5) circles represent single genes significantly affected (p value < 0.05, represented by the horizontal line) compared to PBS control. Heatmap in (H) displays experimental group hierarchical clustering, as determined according to the aortic expression levels of the 84 genes tested. Each column represents a sample; each row represents a gene; the relative gene expression scale is depicted on the right.

Figure 8

Schematic comparison of single DAMP and multiple TLR-targeting strategies to reduce vascular inflammation and disease in CKD patients. Single CKD-associated DAMP targeting strategy: for example, Calprotectin blockade with the pharmacological inhibitor Paquinimod at optimal dose is expected to achieve robust inhibition of the single DAMP’s vascular pro-inflammatory activity. Multi-TLR blocking strategy: sTLR2 will provide partial inhibition of the activity of all DAMP ligands, except for TLR2 DAMP ligands, as inhibition will rely just on the ability of sTLR2 to interact with the common co-receptor CD14 and prevent its enhancing activity, while some CD14-independent suboptimal TLR activation may remain. In the case of TLR2 DAMPs, sTLR2 will prevent their interaction with TLR2 due to its decoy receptor activity, in addition to preventing co-receptor activity. This study demonstrated that both strategies for DAMP inhibition were efficient at lowering CKD associated vascular inflammation.