Nr4a1-dependent Ly6C(low) monocytes monitor endothelial cells and orchestrate their disposal

Abstract

The functions of Nr4a1-dependent Ly6C(low) monocytes remain enigmatic. We show that they are enriched within capillaries and scavenge microparticles from their lumenal side in a steady state. In the kidney cortex, perturbation of homeostasis by a TLR7-dependent nucleic acid "danger" signal, which may signify viral infection or local cell death, triggers Gαi-dependent intravascular retention of Ly6C(low) monocytes by the endothelium. Then, monocytes recruit neutrophils in a TLR7-dependent manner to mediate focal necrosis of endothelial cells, whereas the monocytes remove cellular debris. Prevention of Ly6C(low) monocyte development, crawling, or retention in Nr4a1(-/-), Itgal(-/-), and Tlr7(host-/-BM+/+) and Cx3cr1(-/-) mice, respectively, abolished neutrophil recruitment and endothelial killing. Prevention of neutrophil recruitment in Tlr7(host+/+BM-/-) mice or by neutrophil depletion also abolished endothelial cell necrosis. Therefore, Ly6C(low) monocytes are intravascular housekeepers that orchestrate the necrosis by neutrophils of endothelial cells that signal a local threat sensed via TLR7 followed by the in situ phagocytosis of cellular debris.

Graphical abstract

Figure 1

Characterization of Ly6C^low Patrolling Monocytes in a Steady State (A) Left, isovolume-rendered blood vessels (TRITC dextran, magenta) and monocytes (GFP) from the dermis (ear), kidney, and mesentery. The scale bar represents 100 μm. Right, number of crawling CX3CR1^high Ly6C⁻ monocytes per μl in the dermal (ear), kidney, and mesentery blood vessels (left) and circulating CX3CR1^high Ly6C⁻ monocytes per μl (right). Geometric mean, 95% confidence interval, n ≥ 10 fields over ≥ 6 mice per condition. (B) Crawling monocytes (GFP) in a kidney peritubular capillary (left) labeled with CD11b PE Ab (inset) and a glomerulus (right). Capillaries are magenta (TRITC-labeled 70 kD dextran). Shown in the right inset is the TRITC channel alone. The scale bars represent 10 μm. (C) CX3CR1-GFP monocytes in a mesenteric blood vessel (top) and Gr1Ab staining (top, bottom); the white arrow follows a CX3CR1⁺ GR1⁻ cell. Time, min:s. The scale bar represents 40 μm. (D) Fluorescence signal summed over time in the mesenteric blood vessels for CX3CR1-GFP, CD11b Ab, and Gr1Ab. The scale bar represents 100 μm. Data are representative of n = 6 mice. (E) Deconvolved intravital imaging of CX3CR1-GFP-labeled monocyte in a dermal blood vessel (TRITC-dextran; magenta). Data are representative of >10 mice. (F) Intravital imaging of 2 μm latex beads (TRITC, magenta) uptake in peritubular capillaries. The bead associates with endothelium (dotted circle) and is phagocytosed by CX3CR1^high monocyte. The scale bar represents 20 μm; time, min:s. (G) Uptake of 2 MDa dextran by a crawling monocyte (GFP) in a kidney peritubular capillary. The bottom shows an isovolume rendering of the same cell. The scale bars represent 10 μm. (H and I) Representative transmission electron micrograph (TEM) of a mononuclear cell (black arrow) in peritubular capillaries in a Cx3cr1^+/gfp;Rag2^−/−;Il2rg^−/− mouse. The black arrows in (I) indicate endosomes. The scale bars represent 1 μm. Also see Figure S1 and Movie S1. Monocyte Intravascular Crawling in the Kidney, Related to Figure 1, Movie S2. Attachment, Related to Figure 1, Movie S3. Detachment, Related to Figure 1, Movie S4. Intravital Labeling and Cell Tracking, Related to Figure 1, Movie S5. Mac 1 Expression, Related to Figure 1, Movie S6. Filopodia In Vivo, Related to Figure 1, Movie S7. Monocyte Scavenging, Related to Figure 1.

Figure S1

Phenotype of “Patrolling” Monocytes, Related to Figures 1 and 2 (A) In vitro staining of a human CD14^dim monocyte spread on an ICAM and CSF1 coated coverslip. (Gray – DIC, Cyan– Phalloidin, Magenta– α-CD11a, Blue – DAPI). (B) Left - Leukocyte counts per ml for 11-wk old C57BL6 mice treated with isotype ctrl, or anti-CD11a Ab i.v. for 15 min. Right - representative dot plots (giving mean and SEM for each quadrant frequency) Ly6Clow, I-A- cell frequency is highlighted in red n = 4 mice per group.

Figure 2

CCR2-Independent, NR4A1-Dependent Ly6C^low Monocytes Require LFA1 and ICAM1 or ICAM2, but Not Gαi or CX3CR1, for Intravascular Crawling in a Steady State (A) Number and percentages of circulating monocyte subsets per ml of blood in Itgal^+/+ and Itgal^−/− littermates quantified by flow cytometry. Mean ± SEM, n = 3 mice per genotype. (B) Number and representative tracks and vectors of crawling monocytes per hour per field in the mesenteric blood vessels of Itgal^+/+ and Itgal^−/− littermates. Mean ± SEM; ^∗, p ≤ 0.05; n = 4 mice per genotype. The scale bars represent 60 μm. Blue arrows indicate blood flow direction. (C) Data idem as in (B) for Icam1^−/−, Icam2^−/−, and Icam1^−/− and Icam2^−/− mice. (D and E) Circulating and crawling monocyte subsets and PMNs in Ccr2^+/+ and Ccr2^−/− mice. ^∗, p ≤ 0.05; mean ± SEM; n = 3 mice per genotype. (F) Representative tracks, vectors, and confocal micrograph of crawling monocytes in mesenteric blood vessels of Cx3cr^−/+ and Cx3cr1^−/− mice (white, CX3CR1-GFP; magenta, TRITC-70kD dextran). The scale bars represent 10 μm. n = 5 mice. (G) Data idem as in (B) for mice treated with pertussis toxin (PT) 100 μg i.v. Mean ± SEM, n = 2 mice per condition. (H and I) Data idem as in (D) for Nr4a1^+/+ and Nr4a1^−/− mice. Mean ± SEM, n = 6 mice per genotype. (J) Schematic representation of the monocyte subsets. The x axis represents I-A expression, and the y axis represents Ly6C expression divided by Nr4a1 and Ccr2 requirement. Also see Figure S1 and Movie S4.

Figure S2

Response to R848 and LPS, Related to Figure 3 (A) Application of R848 or LPS to the kidney capsule induces inflammatory cytokines. qPCR for Il1b or Tnf mRNA in sub-capsular kidney cortex of Cx3cr1^−/+; Rag2^−/−; Il2rg^−/− mice tissue treated in vivo by painting the kidney capsule with PBS, R848 (200 μg) or LPS (200 μg) for 5 hr n = 2-4 mice per condition. mRNA quantity was normalized to Gapdh and is expressed as fold change over PBS. (B) CD11b expression on circulating Ly6C^low monocytes after kidney painting with either R848 (top) or LPS (bottom). Representative histograms are shown (left) and cumulative data mean ± SEM right, n = 3 for R848, n = 2 for LPS.

Figure 3

Retention of Crawling Monocytes in the Kidney Vasculature in Response to TLR7 Agonist (A) Representative monocyte tracks and vectors in the kidney cortex after painting with PBS or R848 in Tlr7^+/+ and Tlr7^−/− mice over 5 hr. n = 3 or 4 mice per condition The scale bar represents 40 μm. (B) Track length and speed for monocytes from the experiments described in (A). ^∗, p ≤ 0.05; mean ± SEM. (C) Mean track duration, track displacement, and confinement ratio of crawling monocytes from the experiments described in (A). ^∗, p ≤ 0.05; mean ± SEM. (D) Left, cumulative number of crawling monocytes per frame from experiments described in (A). Middle and right, the same experiment split over two graphs for clarity after PBS, LPS, R848 painting, or i.v. injection of PBS or R848. Data points for the R848 painting are shown twice. ^∗, p ≤ 0.05; n = 3-5 mice per condition. (E and F) Intravital imaging of peritubular capillaries in Cx3cr1^+/gfp mice after i.v. injection of CD11b-PE (magenta), 4.5 hr after R848 painting, and quantification of GFP⁺ cells in the kidney cortex and capillaries at t0 and 4.5 hr after R848 painting. n = 4, mean ± SEM. The scale bar represents 10 μm. Also see Figure S2 and Movies S8 and S9.

Figure 4

Retention of Crawling Monocytes in the Kidney Vasculature in Response to TLR7 Agonist Requires Chemokine Receptor Signaling (A) Experiments performed idem as in Figure 3 for Cx3cr1^+/−Rag2^−/−Il2rg^−/− or Cx3cr1^−/−Rag2^−/−Il2rg^−/− mice. The scale bar represents 40 μm. n = 3 mice per condition. (B) Data idem as in Figures 3B and 3C. ^∗, p ≤ 0.05; mean ± SEM; n = 3 mice. (C) qPCR for fractalkine (Cx3cl1) messenger RNA in kidney cortex tissue from mice with the indicated genotypes 5hr after painting with R848 or PBS (n = 3–5 mice). ^∗,p ≤ 0.05; mean ± SEM. (D) Data idem as in Figure 3D for the stated genotypes and conditions. Right, mice received control or blocking anti-CD11b Ab i.v. injection immediately prior to the experiment. ^∗, p ≤ 0.05; mean ± SEM; n = 3–5 mice per condition. (E) Intravital imaging of peritubular capillaries in Cx3cr1^+/gfp mice 5 hr after R848 painting and immediately after i.v. injection of 10 μg Gr1-APC or Ly6G-PE antibodies. Mean ± SEM, n = 4 mice per condition. (F) GFP⁻ Gr1⁺ and Ly6G⁺ cells forming clusters in capillaries before and after R848 painting and proportion of clusters in contact with a GFP⁺ cells. n = 5 mice per condition, mean ± SEM. (G) TEM of the superficial kidney cortex from mice 5 hr after PBS or R848 painting. Cells/100 TEM grid squares; mean ± SEM; n = 4 Cx3cr1^+/− mice per condition, 2 Cx3cr1^−/− mice per condition, 2 Nr4a1^−/− mice per condition, 2 Ccr2^−/− mice per condition, and 3 Itgal^−/− mice per condition. ^∗, p < 0.05. Also see Figure S3 and Movie S10.

Figure 5

Focal Necrosis of Endothelial Cells (A) Representative electron micrographs of kidney cortex peritubular capillaries 5 hr after painting with R848 or PBS from Figure 4G. Single arrow, basal lamina; double arrow, endothelium; ^∗, fluid in the subendothelial space; M, mononuclear cell; PMN, polymorphonuclear cell; E, endothelial cell nucleus. In the PBS-treated healthy peritubular capillary, endothelial cells are flat and close to basal lamina. The R848-treated peritubular capillary shows a swollen necrotic endothelial cell, expanded subendothelial space, mononuclear cell phagocytosing mitochondria, and blebbing necrotic endothelium (images are from Cx3cr^−/+Rag2^−/−IL2rg^−/− mice). (B) Similar features to those represented in (A) are shown for Ccr2^−/− mice. (C) Representative healthy endothelium in R848- and PBS-treated kidneys from Itgal^−/−, Cx3cr1^−/−, and Nr4a1^−/− mice. Micrographs are representative of experiments analyzed in Figure 4G. The thick scale represents 5 μm, the thin scale represents 1 μm. Initial magnification was 15,000×. Also see Figure S3.

Figure S3

TEM of Kidney Capillaries 5 hr after Direct Application of PBS or R848 In Vivo, Related to Figures 4, 5, 6, and 7 Single arrow, basal lamina; double arrow, endothelium; M, shows mononuclear cell; PMN, polymorphonuclear cell; E, endothelial cell nucleus. White boxes indicate the high magnification areas. (A) Glomerular capillaries from Cx3cr1^−/+; Rag2^−/−; Il2rg^−/− mice. Left, PBS treated control glomerulus, endothelium is intact and close to basal lamina; Middle, R848 treated glomerular capillary containing mononuclear cell. Endothelium is missing and monocnuclear cell containing phagocytosed material sits directly against basal lamina (see higher magnifications); Right, PMN containing glomerular capillary with thickened area of endothelium. (B) Peritubular capillaries from wild-type B6 mice. Left, PBS treated control peritubular capillary; 3^rd from right, peritubular capillary containing mononuclear cell with necrotic and disrupted endothelium; 2^nd from right, peritubular capillary containing mononuclear cell with necrotic and disrupted endothelium and expanded subendothelial space; 1^st from right, peritubular capillary containing PMN with necrotic and disrupted endothelium and expanded subendothelial space. Images representative of 3 mice per condition. Thick scale bars = 5 μm, thin scale bars = 1 μm. Original magnification x15 000.

Figure 6

Focal Endothelial Necrosis Requires Retention of Ly6C^low Monocytes, which Requires Expression of TLR7 on the Kidney (A and B) Endothelial thickness measurements and microscopic features from the TEM results described in Figure 4G. At least 30 capillaries were examined per condition. Mean ± 95% confidence interval (A) or SEM (B).^∗, p < 0.05. (C) Representative FACS dot plots from the blood of bone marrow chimera. Mean ± SEM, n = 5–8 mice per condition. (D) Intravascular mononuclear cells (left) and polymorphonuclear cells (right) quantified by TEM in bone marrow chimera. Cell counts are from individual mice. Mean and SEM, n = 2–5 mice. (E) Intravascular monocytes (left) and granulocytes (right) quantified by intravital microscopy in the kidney cortex of bone marrow chimera following R848 painting. Granulocytes (×10²) were quantified at 5 hr immediately after i.v. injection of the Ly6G antibody. Mean ± SEM, n = 3 or 4 mice per condition. Also see Figure S3.

Figure 7

Neutrophils Kill Endothelial Cells (A) Endothelial cell microscopic features of chimeric mice described in Figure 6D expressed as the percentage of mononuclear cells (mono) or PMN-containing fields that present with the indicated lesions. (B) Representative FACS dot plots of peripheral blood cells of mice treated 8 hr earlier with Ly6G-depleting Ab (1A8) or isotype control (2A3). The arrow in the FSC/SSC panel indicates granulocytic cells, the percentage of Lin⁻ CD115⁻ granulocytes are indicated in red. n = 3 mice per group, mean ± SEM. (C) Presence of intravascular mononuclear (left) and polymorphonuclear cells (right) as quantified by TEM in mice treated with 1A8 or 2A3 8 hr before kidney painting with R848. n = 3 mice per group, mean ± SEM. (D) Endothelial cell microscopic features of granulocyte-depleted and control mice. n = 3 mice per group, mean ± SEM. (E) Representative peritubular capillary containing a monocyte from a 1A8-treated mouse. (F) Proinflammatory cytokine production in vitro by sorted Ly6C^low and Ly6C⁺ monocytes after 24 hr stimulation with medium alone or R848 (top) in the absence or presence of a MEK inhibitor (PD) or for medium alone or LPS (bottom) in the absence or presence of the MEK inhibitor (PD) (bottom). Multiplexed ELISA, n = 3 mice per condition. (G) Schematic representation of the molecular and cellular features of the interaction of Ly6C^low monocytes with the endothelium in a steady state and TLR7-mediated endothelial “safe disposal.” Also see Figure S3.