Soluble uric acid inhibits β2 integrin-mediated neutrophil recruitment in innate immunity

Abstract

Neutrophils are key players during host defense and sterile inflammation. Neutrophil dysfunction is a characteristic feature of the acquired immunodeficiency during kidney disease. We speculated that the impaired renal clearance of the intrinsic purine metabolite soluble uric acid (sUA) may account for neutrophil dysfunction. Indeed, hyperuricemia (HU, serum UA of 9-12 mg/dL) related or unrelated to kidney dysfunction significantly diminished neutrophil adhesion and extravasation in mice with crystal- and coronavirus-related sterile inflammation using intravital microscopy and an air pouch model. This impaired neutrophil recruitment was partially reversible by depleting UA with rasburicase. We validated these findings in vitro using either neutrophils or serum from patients with kidney dysfunction-related HU with or without UA depletion, which partially normalized the defective migration of neutrophils. Mechanistically, sUA impaired β2 integrin activity and internalization/recycling by regulating intracellular pH and cytoskeletal dynamics, physiological processes that are known to alter the migratory and phagocytic capability of neutrophils. This effect was fully reversible by blocking intracellular uptake of sUA via urate transporters. In contrast, sUA had no effect on neutrophil extracellular trap formation in neutrophils from healthy subjects or patients with kidney dysfunction. Our results identify an unexpected immunoregulatory role of the intrinsic purine metabolite sUA, which contrasts the well-known immunostimulatory effects of crystalline UA. Specifically targeting UA may help to overcome certain forms of immunodeficiency, for example in kidney dysfunction, but may enhance sterile forms of inflammation.

Graphical abstract

Figure 1

Leukocyte recruitment is impaired in sterile inflammation in hyperuricemic mice. (A) Alb-creERT2;Glut9^lox/lox mice and Glut9^lox/lox control mice were injected intraperitoneally with tamoxifen. Both groups were fed either an acidogenic diet enriched with inosine or a standard chow diet with inosine for 22 days. On day 21, mice received a subcutaneous injection of monosodium urate (MSU) crystals (5 mg) or vehicle into a preexisting air pouch and were euthanized 12 hours later. (B-C) Serum uric acid (B), blood urea nitrogen (BUN), and creatinine (C) levels of Glut9^lox/lox mice with chow diet and inosine (healthy), Alb-creERT2;Glut9^lox/lox mice with chow diet (HU), or acidogenic diet with inosine (HU plus CKD) on day 22 (n = 7 mice per group) (using 1-way ANOVA). (D) Representative images of vehicle and MSU crystals injected into the air pouch. (E-G) Gating strategy (E) and number (No.) of CD45⁺ leukocytes (F) and neutrophils (G) in air pouch per µL from mice with or without MSU crystals determined by flow cytometry (n = 3-6 mice per group). (H-J) Concentrations of IL-1β (H), IL-6 (I), and CXCL1 (J) measured in the air pouch fluid via ELISA (n = 3-6 mice per group). Data are mean plus or minus SD. *P < .05, **P < .01, ***P < .001. ns, not significant by 1-way ANOVA; ELISA, enzyme-linked immunoassay.

Figure 1

Leukocyte recruitment is impaired in sterile inflammation in hyperuricemic mice. (A) Alb-creERT2;Glut9^lox/lox mice and Glut9^lox/lox control mice were injected intraperitoneally with tamoxifen. Both groups were fed either an acidogenic diet enriched with inosine or a standard chow diet with inosine for 22 days. On day 21, mice received a subcutaneous injection of monosodium urate (MSU) crystals (5 mg) or vehicle into a preexisting air pouch and were euthanized 12 hours later. (B-C) Serum uric acid (B), blood urea nitrogen (BUN), and creatinine (C) levels of Glut9^lox/lox mice with chow diet and inosine (healthy), Alb-creERT2;Glut9^lox/lox mice with chow diet (HU), or acidogenic diet with inosine (HU plus CKD) on day 22 (n = 7 mice per group) (using 1-way ANOVA). (D) Representative images of vehicle and MSU crystals injected into the air pouch. (E-G) Gating strategy (E) and number (No.) of CD45⁺ leukocytes (F) and neutrophils (G) in air pouch per µL from mice with or without MSU crystals determined by flow cytometry (n = 3-6 mice per group). (H-J) Concentrations of IL-1β (H), IL-6 (I), and CXCL1 (J) measured in the air pouch fluid via ELISA (n = 3-6 mice per group). Data are mean plus or minus SD. *P < .05, **P < .01, ***P < .001. ns, not significant by 1-way ANOVA; ELISA, enzyme-linked immunoassay.

Figure 2

Rasburicase treatment reverses the impaired leukocyte recruitment and inflammation in mice with hyperuricemia and kidney dysfunction. Alb-creERT2;Glut9^lox/lox and Glut9^lox/lox control mice were injected with tamoxifen and placed on an acidogenic diet with inosine (HU plus CKD) or a chow diet with inosine (healthy) for 22 days. HU plus CKD mice received vehicle or rasburicase treatment prior to injection of vehicle (PBS) or MSU crystals (5 mg) into a preexisting air pouch and were euthanized after 12 hours. (A-B) Serum uric acid (A) and BUN (B) levels of HU plus CKD mice with rasburicase or vehicle treatment (n = 5 per group, Student t test). (C) Serum hydrogen peroxide (H₂O₂) levels measured in HU plus CKD mice treated with rasburicase or vehicle (n = 5 per group, Student t test). (D) Representative images of MSU crystal injection into the air pouch with or without rasburicase. (E-F) Flow cytometry analysis of infiltrating CD45⁺ leukocytes (E) and neutrophils (F) into the air pouch from mice with or without MSU crystals and/or rasburicase treatment (n = 4-6 per group). (G-I) Concentrations of IL-1β (G), IL-6 (H), and CXCL1 (I) from air pouch fluid measured via ELISA (n = 4-6 per group). Data are mean plus or minus SD. *P < .05, **P < .01, ***P < .001. ns, not significant by 2-way ANOVA; PBS, phosphate-buffered saline; ELISA, enzyme-linked immunoassay.

Figure 3

Decreased neutrophil adhesion and extravasation in crystal-stimulated cremaster muscle venules in hyperuricemic mice. (A) Alb-creERT2;Glut9^lox/lox and Glut9^lox/lox control mice were injected with tamoxifen and placed on a chow or acidogenic diet with inosine (n = 5-7 mice per group). HU plus CKD mice received vehicle or rasburicase treatment. After 22 days, mice were injected intrascrotally with MSU crystals (0.5 mg in 200 µL NaCl per mouse) or vehicle (200 µL NaCl per mouse) 4 hours prior to intravital microscopy. (B) Quantification of adhesion efficiency (n = 5-7 mice per group). (C-D) Cumulative distribution (C) and mean leukocyte rolling velocity (µm per second) (D). (E) Quantification of leukocyte rolling depicted as rolling flux fraction. (F) The total number of perivascular neutrophils and mononuclear cells per mm² after MSU crystal injection (n = 5 mice per group). (G) Quantification of the adhesion efficiency (n = 5 mice per group). (H-I) Cumulative frequency over leukocyte rolling velocity (H) and mean of rolling velocity (µm per second) (I). (J) The total number of perivascular neutrophils and mononuclear cells per mm² after MSU crystal injection in vehicle- and rasburicase-treated HU plus CKD mice (n = 4-5 mice per group). Data are mean plus or minus SD. *P < .05, **P < .01, ***P < .001. ns, not significant by Student t test or 1- or 2-way ANOVA.

Figure 4

Soluble uric acid affects β₂ integrin activation and prevents human neutrophil migration and phagocytosis in vitro. (A) Human neutrophils were isolated from healthy individuals and preincubated with or without 10 mg/dL sUA for 30 minutes prior to stimulation with human CXCL8 or SARS-CoV2–derived DAMP/PAMP supernatant. (B) Schematic of the 3 β₂ integrin conformations that reflect the different affinity stages: (1) the bent form, inactive, (2) the extended form with a closed ligand-binding head of intermediate affinity, and (3) the extended form enabling the ligand-binding for mAB24 with high affinity (adopted from Evans et al). (C-D) mAB24 binding illustrated as a histogram (C) and the expression levels of mAB24 (D) determined by flow cytometry (n = 6-7 per group). (E-F) The expression levels of LFA-1 (E) and MAC-1 (F) shown as mean fluorescence intensity (MFI) relative to isotype control (n = 6-7 per group) were quantified by flow cytometry. (G-H) Transwell migration assays were carried out, and the total number of neutrophils that migrated toward the chemoattractants human CXCL8 and IL-1β (G) and fMLP (H) was determined after 3 hours by flow cytometry (n = 6-12 per group). (I-J) The expression levels of LFA-1 (I) and MAC-1 (J) shown as MFI relative to isotype control (n = 4 per group) were quantified by flow cytometry. (K) Transwell migration assays were carried out, and the total number of neutrophils that migrated toward the SARS-CoV2 supernatant (20%) was determined after 3 hours by flow cytometry (n = 4 per group). (L) Percentage (%) of phagocytosed IgG FITC-latex beads in untreated or CXCL8-stimulated neutrophils with or without sUA preincubation was determined by flow cytometry (n = 4 per group). Data are mean plus or minus SD. *P < .05, **P < .01, ***P < .001. ns, not significant by 2-way ANOVA; Ctrl, control; IgG, immunoglobulin G.

Figure 5

Soluble uric acid affects β₂ integrin internalization and recycling in human neutrophils. (A) Schematic illustrating the process of integrin internalization and recycling in neutrophils (detailed information is described in supplemental Methods). (B) Human neutrophils were isolated from healthy individuals and preincubated with or without 10 mg/dL sUA for 30 minutes prior to stimulation with CXCL8. The integrin internalization assay for anti-human CD18 in CXCL8-activated neutrophils was performed using flow cytometry. The percentage (%) of internalized CD18 in CXCL8-activated neutrophils was determined after 10 and 30 minutes (n = 7 per group). (C) Immunoblotting for CD18 and β-actin was performed, and the protein CD18/β-actin ratio to medium quantified (n = 4 per group). (D) The integrin internalization assay was performed and the percentage of internalized MAC-1 after 10 and 30 minutes determined by flow cytometry (n = 4 per group). (E) The integrin recycling assay was performed and the percentage of recycled CD18 after 10 and 30 minutes determined by flow cytometry (n = 4 per group). (F-H) Cytoskeletal dynamics were determined in CXCL8-activated neutrophils (n = 4 per group) via flow cytometry using the forward scatter (FSC-H) for quantifying cellular size (F-G) and the dye DiBAC₄(3) for quantification of membrane potential (as MFI) (H). (I-J) Intracellular pH was determined as MFI of the dye SNARF-1 (MFI) (I) by flow cytometry and the intracellular pH calculated (J) (n = 4 per group). Data are mean plus or minus SD. *P < .05, **P < .01. Ctrl, control; ns, not significant by 2-way ANOVA.

Figure 6

Intracellular uptake of uric acid via the urate transporter SLC2A9 in neutrophils. (A-B) Human neutrophils from healthy subjects (A, A', and A'') and primary human TECs (B, B', and B'') were stained with the anti-human SLC2A9 antibody (red) and DAPI (white, nuclei), and confocal microscopy was performed. (C-D) MFI of intracellular and surface SLC2A9 expression in human neutrophils (C) and TECs (D) compared with isotype control was determined by flow cytometry (n = 3 per group). (E) Neutrophils from healthy subjects were cultured with or without 10 mg/dL sUA for 10 and 30 minutes and the intracellular UA levels determined using an UA assay kit (n = 7-8 per group). (F) Intracellular UA levels from CXCL8-activated neutrophils in the presence or absence of 10 mg/dL sUA determined by UA assay kit (n = 6-8 per group). Data are mean plus or minus SD. *P < .05, **P < .01, ***P < .001. ns, not significant by 1-way ANOVA; DAPI, 4′,6-diamidino-2-phenylindole.

Figure 7

Soluble uric acid impairs neutrophil migration in patients with kidney dysfunction. (A) Schematic of study design. Out of 51 individuals, 8 patients were excluded due to the intake of immunosuppressive drugs or xanthine oxidase inhibitors. Of the remaining, 10 patients represented with CKD (CKD stage G2-4), 18 patients with ESKD that were on hemodialysis (CKD stage G5), and 15 healthy individuals without any renal pathologies (healthy, CKD stage 0). (B) Human neutrophils were isolated and identified as CD15⁺CD66b⁺ by flow cytometry (with gating strategy) with a purity of ∼99%. (C) Number of human neutrophils that were isolated from patients with kidney dysfunction (CKD and ESKD) as well as healthy individuals. (D) Expression (MFI) of MAC-1 relative to isotype determined by flow cytometry (healthy, n = 15; CKD, n = 10; ESKD, n = 18; 1-way ANOVA). (E-F) The total number of neutrophils isolated from healthy individuals and ESKD patients that migrated toward the chemoattractants human CXCL8 (E) and fMLP (F) (healthy, n = 6-7; ESKD, n = 6; 2-way ANOVA) was determined by flow cytometry after 3 hours. (G-H) Neutrophils were incubated with medium and serum from healthy individuals or CKD/ESKD patients and AnnexinV (AnV)-PI staining performed to quantify cell death by flow cytometry. Representative images of dot plots (G) and the percentage (%) of live (AnV-PI⁻), apoptotic (AnV+PI⁻), late apoptotic/early necrotic (AnV+PI⁺), and necrotic (AnV-PI⁺) neutrophils (H). (I) Percentage (%) of phagocytosed IgG FITC-latex beads in untreated or CXCL8-stimulated neutrophils with or without preincubation of sera from healthy individuals or CKD/ESKD patients was determined by flow cytometry (n = 4 per group; 2-way ANOVA). (J) Healthy neutrophils were incubated with serum from healthy individuals or patients with kidney dysfunction in the absence or presence of rasburicase, and the total number of neutrophils that migrated toward fMLP was determined by flow cytometry after 3 hours (n = 5-6 per group; 2-way ANOVA). Data are mean plus or minus SD. *P < .05, **P < .01, ***P < .001. ns, not significant; Ctrl, control.