Critical Role of the CXCL10/C-X-C Chemokine Receptor 3 Axis in Promoting Leukocyte Recruitment and Neuronal Injury during Traumatic Optic Neuropathy Induced by Optic Nerve Crush

Abstract

Traumatic optic neuropathy (TON) is an acute injury of the optic nerve secondary to trauma. Loss of retinal ganglion cells (RGCs) is a key pathological process in TON, yet mechanisms responsible for RGC death remain unclear. In a mouse model of TON, real-time noninvasive imaging revealed a dramatic increase in leukocyte rolling and adhesion in veins near the optic nerve (ON) head at 9 hours after ON injury. Although RGC dysfunction and loss were not detected at 24 hours after injury, massive leukocyte infiltration was observed in the superficial retina. These cells were identified as T cells, microglia/monocytes, and neutrophils but not B cells. CXCL10 is a chemokine that recruits leukocytes after binding to its receptor C-X-C chemokine receptor (CXCR) 3. The levels of CXCL10 and CXCR3 were markedly elevated in TON, and up-regulation of CXCL10 was mediated by STAT1/3. Deleting CXCR3 in leukocytes significantly reduced leukocyte recruitment, and prevented RGC death at 7 days after ON injury. Treatment with CXCR3 antagonist attenuated TON-induced RGC dysfunction and cell loss. In vitro co-culture of primary RGCs with leukocytes resulted in increased RGC apoptosis, which was exaggerated in the presence of CXCL10. These results indicate that leukocyte recruitment in retinal vessels near the ON head is an early event in TON and the CXCL10/CXCR3 axis has a critical role in recruiting leukocytes and inducing RGC death.

Figure 1

Leukocyte recruitment is increased after traumatic optic neuropathy (TON). A–C: Wild-type (WT) mice received bone marrow transplants from green fluorescent protein (GFP) transgenic mice and 4 weeks later they were subjected to TON. A: Images of leukocyte distribution in the central retina were taken at 0, 3, 6, and 9 hours after TON using real-time scanning laser ophthalmoscopy (SLO) imaging. B: Bar graph represents the quantification of percentage leukocyte area relative to total image area at 0 hour [control (Con)] and 9 hours after TON. C: Representative SLO image sequences at 9 hours after TON. Arrows indicate an example of GFP-positive leukocyte rolling along the major retinal vein. D and E: Eight-week-old WT mice were subjected to TON, and 24 hours later, noninjured control or TON-performed eyes were collected. D: Immunofluorescence staining for leukocyte subtype markers CD45, Ly6G/Ly6C, Iba1, and CD3 (green) in retinal flat-mounts. E: Bar graphs represent the quantification of recruited cells that were specifically labeled with antibodies against CD45, Ly6G/Ly6C, Iba1, and CD3. n = 5 mice (B and C); n = 3 mice (D and E). ^∗P < 0.05, ^∗∗P < 0.01 versus control. Scale bars: 200 μm (A); 50 μm (C and D). T, time elapsed after initiation of imaging.

Figure 2

Analysis of retinal neuronal function and retinal ganglion cell loss. A–D: Wild-type mice were subjected to traumatic optic neuropathy (TON). Electroretinographic analysis for retinal neuronal function under scotopic conditions at 24 hours after TON. Graphs represent average amplitudes of positive scotopic threshold response (pSTR), negative STR (nSTR), and a- and b-waves over a range of stimulus strengths. E and F: Representative images of control (Con) and TON retinas labeled with Tuj1 antibody (green) at 24 hours (E) or 7 days (F) after TON. Graphs represent the number of Tuj1-positive cells per field. n = 3 mice (A–D); n = 4 mice (E and F). ^∗∗∗P < 0.001 versus control. Scale bars = 50 μm (E and F).

Figure 3

CXCL10/CXCR3 pathway is activated in TON. Wild-type mice were subjected to traumatic optic neuropathy (TON). A: Quantitative PCR analysis of CXCL10 mRNA expression in noninjured retinas [control (Con)] or injured retinas at 3, 6, 12, and 24 hours after TON. B: Normal and TON-performed eyes were collected at 6 hours after TON. CXCL10 mRNA localization was assessed in retinal frozen sections by fluorescence in situ hybridization with RNAscope Fluorescent Multiplex Kit. Green fluorescent signal reflects CXCL10 mRNA expression, and DAPI (blue) stains nuclei. Arrows indicate CXCL10-expressing retinal cells in the ganglion cell layer (GCL). C: Enzyme-linked immunosorbent assay analysis of CXCL10 protein in control or TON-performed retinas at 6 hours after TON. D: Quantitative PCR analysis of CXCR3 mRNA expression in control or injured retinas at 3, 6, 12, and 24 hours after TON. E: Representative images of CXCR3 immunostaining in retinal frozen sections from control and TON-performed eyes at 24 hours after TON. Fluorescent signal (red) reflects CXCR3 staining. n = 4 to 5 mice (E). ^∗P < 0.05, ^∗∗P < 0.01 versus control. Scale bars = 50 μm (B and E). INL, inner nuclear layer; ONL, outer nuclear layer.

Figure 4

CXCL10/CXCR3 pathway is involved in leukocyte recruitment and RGC death in the retina. A and B: Wild-type (WT) mice received bone marrow transplants from WT-GFP⁺ or CXCR3^−/−-GFP⁺ mice (defined as WT-GFP⁺→WT or CXCR3^−/−-GFP⁺→WT) and 4 weeks later they were subjected to traumatic optic neuropathy (TON). A: Images of leukocyte distribution in the central retina were taken before or 9 hours after TON using real-time scanning laser ophthalmoscopy imaging. Bar graph represents the quantification of leukocyte area at 0 and 9 hours after TON. B: Representative images of leukocyte distribution in retinal flat mounts from WT mice receiving bone marrow (BM) from WT-GFP⁺ or CXCR3^−/−-GFP⁺ mice at 5 days after TON. High-magnification images are shown in insets. Bar graph represents the quantification of leukocyte number in the retina, which was normalized to WT-GFP⁺→WT mice after TON (control). C: WT mice received BM from WT or CXCR3^−/− (knockout) (defined as WT→WT or CXCR3^−/−→WT) and 4 weeks later they were subjected to TON. At 7 days after TON, retinas were collected and stained with Tuj1 antibody (green). Bar graph represents the number of Tuj1-positive cells per field. D: Primary RGCs were isolated, cultured alone, or co-cultured with bone marrow–derived or blood leukocytes, and treated with 10 or 100 ng/mL of CXCL10. At 24 hours after treatment, cells were subjected to terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay. The total TUNEL/Tuj1 double-positive cells and Tuj1-positive cells in each field were counted under a microscope. Bar graph represents the percentage of apoptotic RGCs, calculated as the ratio of total TUNEL/Tuj1 double-positive cells to total Tuj1-positive cells. Twenty fields were counted in each group. n = 6 mice (A); n = 3 (B, mice, and D, independent experiments). n = 5 to 7 mice (C). ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001. Scale bars: 200 μm (A); 500 μm (B); 100 μm (B, insets); 50 μm (C). Original magnification, ×200 (B).

Figure 5

AMG487, antagonist of CXCR3, preserves retinal function and prevents retinal ganglion cell loss after traumatic optic neuropathy (TON). Wild-type mice were i.p. injected with vehicle (Veh) or AMG487 (20 mg/kg) at 1 and 12 hours before TON procedure and continuously injected twice a day for 7 days after TON. A: Body weights of vehicle- and AMG487-treated TON mice. B: Graph represents average amplitudes of positive scotopic threshold response (pSTR) over a range of stimulus strengths. C: Representative pSTR patterns for stimuli of −4.0 log cd-second/m². D: Retinas were collected from noninjured mice [control (Con)] and TON mice treated with vehicle or AMG487. Representative images of retinal flat mounts labeled with Tuj1 antibody (green) were shown. Bar graph represents the number of Tuj1-positive cells per field. n = 5 to 7 mice (B); n = 8 to 9 mice (D). ^∗P < 0.05 versus noninjured control mice (Con); ^†P < 0.05 versus vehicle-treated TON mice; ^‡‡‡P < 0.001. Scale bars = 50 μm.

Figure 6

NF-κB signaling is not involved in traumatic optic neuropathy (TON)-induced CXCL10 expression. NF-κB inhibitor pyrrolidine dithiocarbamate (PDTC; i.p., 120 mg/kg) or vehicle (Veh) was injected into wild-type mice at 1 hour before TON. Retinas were collected at 6 hours after TON. The mRNA expressions of intercellular adhesion molecule (ICAM) 1 (A), inducible nitric oxide synthase (iNOS; B), and CXCL10 (C) were analyzed by quantitative PCR. n = 4 to 5 mice (A–C). ^∗P < 0.05, ^∗∗P < 0.01.

Figure 7

Traumatic optic neuropathy (TON)-induced CXCL10 up-regulation is modulated by STAT signaling. A: Wild-type (WT) mice were subjected to TON, and proteins were isolated from retinas at indicated time points after TON. The phosphorylation levels of STAT1, STAT2, STAT3, STAT5, and STAT6 were evaluated by Western blotting. α-Tubulin was used as loading control. B and C: STAT1/STAT3 inhibitor Stattic or vehicle (Veh) was injected intravitreally into WT mice at 1 hour after the induction of TON. Retinas were collected at 4 hours after TON. B: Phosphorylated and total STAT1 and STAT3 were assessed by Western blotting. C: The expression of CXCL10 mRNA was analyzed by quantitative PCR. n = 5 mice (A–C). ^∗P < 0.05, ^∗∗P < 0.01.