Sympathetic Nerves Coordinate Corneal Epithelial Wound Healing by Controlling the Mobilization of Ly6C^hi Monocytes From the Spleen to the Injured Cornea

Abstract

Purpose: This study aims to investigate the potential involvement of spleen-derived monocytes in the repair process following corneal epithelial abrasion.

Methods: A corneal epithelial abrasion model was established in male C57BL/6J mice, and the dynamic changes of monocyte subpopulations in the injured cornea were analyzed using flow cytometry. The effects of Ly6C^hi monocyte depletion and local adoptive transfer of purified Ly6C^hi monocytes on wound closure and neutrophil recruitment to the injured cornea were observed. The effect of sympathetic nerves on the recruitment of spleen-derived Ly6C^hi monocytes to the injured cornea was also investigated using multiple methods. The emigration of fluorescence-labeled monocytes to the injured cornea was validated through intravital microscopy. Finally, differential genes between different groups were identified through high-throughput RNA sequencing and analyzed for functional enrichment, followed by verification by quantitative PCR.

Results: Ly6C^hi monocytes were present in large numbers in the injured cornea prior to neutrophil recruitment. Predepletion of Ly6C^hi monocytes significantly inhibited neutrophil recruitment to the injured cornea. Furthermore, surgical removal of the spleen significantly reduced the number of Ly6C^hi monocytes in the injured cornea. Further observations revealed that sympathetic blockade significantly reduced the number of Ly6C^hi monocytes recruited to the injured cornea. In contrast, administration of the β2-adrenergic receptor agonist significantly increased the number of Ly6C^hi monocytes recruited to the injured cornea in animals treated with sympathectomy and catecholamine synthesis inhibition.

Conclusions: Our results suggest that spleen-derived Ly6C^hi monocytes, under the control of the sympathetic nervous system, play a critical role in the inflammatory response following corneal injury.

The cornea is located in front of the eyeball and plays an important role in maintaining normal vision, structural integrity, and resisting damage caused by various external stimuli.¹ Its fine structure is a key feature of its ability to refract light and prevent infection. However, the location of the cornea makes it vulnerable to various injuries.² Rapid and complete healing of the corneal epithelium after external mechanical injury is extremely important to prevent microbial invasion and restore its structural integrity and ability to refract light on the retina.^3,4 Therefore, understanding the process and mechanism of corneal wound healing is critical for preventing infection and restoring normal visual acuity. 

Corneal wound healing is a complicated and overlapping cascade process, which mainly involves reepithelialization, inflammatory response, cell division, and extracellular matrix remodeling.² Reepithelialization can quickly cover the wound by flattening and expanding cells near the edge of the wound.⁵ Limbal stem cells with division ability proliferate and migrate to the wound site, thereby restoring its cellular composition.^6,7 In addition, corneal epithelial injury triggers an inflammatory response that promotes and supports these processes. Any imbalance or defect in this cascade can affect corneal wound repair and the perfect recovery of the corneal structure. Different types of inflammatory cells recruited to the injured cornea after corneal injury mediate wound healing through distinct mechanisms. As first-line inflammatory cells, neutrophils promote corneal wound healing, not only by means of phagocytosis and removing dead and necrotic cells but also by producing a variety of growth factors and cytokines.⁸ γδ T cells migrate to the injured cornea and recruit neutrophils to expand inflammation by releasing IL-17A.⁹ In addition, γδ T cells stimulate division and migration to the wound area via the release of IL-22.⁹ Resident CC motif chemokine receptor 2 (CCR2)⁺ and CCR2⁻ macrophages in the cornea coordinate the wound repair process by promoting or inhibiting the inflammatory response through the production of different cytokines, respectively.¹⁰ Furthermore, resident type 2 innate lymphoid cells in the corneal limbus promote corneal repair of wounds, mainly by producing amphiregulin, a ligand of the epidermal growth factor receptor, and triggering type 2 inflammatory responses.¹¹ However, little is known about the coordinated role of monocytes in corneal wound repair and their underlying cellular and molecular mechanisms. 

Monocytes are a heterogeneous population and the most important component of the “mononuclear phagocyte system.”^12,13 Based on their phenotypes, monocytes can be divided into classical, intermediary, and nonclassical subpopulations.¹⁴ Classical monocytes express the cell surface molecule Ly6C at high levels¹⁵ and are often considered “proinflammatory” cells.^16,17 In contrast, nonclassical monocytes express lower levels of Ly6C. Intermediate monocytes are a newly discovered subpopulation, which may be a transition state of nonclassical monocytes to classical monocytes¹⁶ that complement the function of nonclassical monocytes by preferential differentiation into dendritic cells in an inflammatory setting.¹⁸ A leading characteristic of classical monocytes is that they are rapidly mobilized in large numbers to inflamed sites throughout the body, where they serve as a particularly plastic “emergency squad” to provide proinflammatory or resolving activities.¹⁹ Classical monocytes are also involved in tissue injury, including those of the skin²⁰ and liver.²¹ However, it is not yet clear whether monocytes mobilize and migrate to the injured cornea to participate in the corneal repair process and inflammatory response. 

The spleen is the largest secondary lymphoid organ in the body and is involved in many immune responses via the storage of various immune cells.^22,23 After tissue damage, such as myocardial infarction^24,25 and ischemia–reperfusion injury,²⁶ spleen-derived monocytes quickly migrate to the injured tissue, to promote the healing process. In addition, the spleen is a major contributor to the exaggerated inflammatory response in sepsis, trauma, and burn injuries.²⁷ Recently, there has been increasing evidence of neural control of immunity and inflammation.^28–32 The sympathetic efferent nerves exclusively innervate the spleen^33–35 and directly contact the immune cells in the spleen.^35–40 Sympathetic nerves innervating the spleen can affect the migration and circulation of immune cells in the spleen by releasing the neurotransmitter norepinephrine (NE) or neuropeptide.^41,42 However, it is uncertain whether spleen-derived resident immune cells, particularly monocytes, are involved in the corneal wound repair process through sympathetic regulation after corneal injury. 

Recent studies have shown that the sympathetic nervous system (SNS) is involved in the inflammatory response after corneal wounding. First, circulating epinephrine (E) is rapidly elevated after corneal injury and regulates the inflammatory response after corneal abrasion.⁴³ Second, direct activation of the SNS by restraint experiments and smoke exposure can rapidly alter the normal course of the inflammatory response after corneal injury, which is mainly manifested as delayed repair and excessive inflammatory response.^43–45 Therefore, we hypothesized that besides directly influencing the function of different local immune cell populations, the SNS may also be involved in the regulatory process of corneal wound repair by regulating the recruitment of monocytes stored in the spleen to the injured cornea. 

To validate our hypothesis, we analyzed the phenotype and possible sources of monocytes in injured corneas of mice. Ly6C^hi monocytes were the main monocyte phenotype, and these were mainly derived from the spleen, not from the bone marrow. We also investigated the effects of surgical removal of the spleen and SNS interventions, including chemical sympathectomy, splenic denervation, and inhibition of catecholamine synthesis, on the recruitment of monocytes to the injured cornea. The data revealed that mobilization of Ly6C^hi monocytes from the spleen was controlled by the innervation of the sympathetic nerve. Furthermore, we found that the mobilization of these cells was controlled by the β2-adrenergic receptors (β2-ARs) predominantly expressed on Ly6C^hi monocytes. Overall, these observations highlight a previously unrecognized mechanism by which classical monocytes mobilized from the spleen regulate corneal wound repair and the inflammatory process, and they suggest that manipulation of sympathetic nerves in the spleen might represent a novel therapeutic strategy for impaired corneal healing conditions. 

Methods

Animals

Specific pathogen-free 10- to 12-week-old C57BL/6J male mice were purchased from the Medical Laboratory Animal Center (Guangdong, China). All animal protocols and procedures were approved by the Laboratory Animal Committee of Jinan University and were in accordance with the ARVO Statement for the Use of Animals in Ophthalmology and Vision Research. At the end of the experiment, all animals were anesthetized via inhalation of 2% isoflurane and sacrificed by an overdose of CO₂ and cervical dislocation. 

Corneal Wound-Healing Model

A corneal epithelial wound repair model was established as previously described.^9,46,47 Briefly, after general anesthesia of animals by means of intraperitoneal (IP) injection of sodium phenobarbital (80 mg/kg body weight; cat. 11715, Sigma-Aldrich, Burlington, MA, USA), a 2-mm-diameter central area of the cornea was marked using trephine under a dissecting microscope. The marked corneal epithelial area was scraped using a golf-like spatula (Accutome, Malvern, PA, USA). The wound was stained with sodium fluorescein to observe wound closure dynamics. The fluorescein-stained areas were photographed under a dissecting microscope, at 6-hour intervals, until the wound closed. ImageJ (National Institutes of Health, Maryland, MD, USA; https://imagej.nih.gov/ij/) was used to determine the dynamics of wound closure, by analyzing the pixels of the stained wounds. 

Chemical Sympathectomy

Peripheral sympathetic nerves of mice were ablated using a treatment regimen that included the administration of freshly prepared 200 mg/kg 6-hydroxydopamine (6-OHDA; cat. H4381, Sigma-Aldrich) in phosphate-buffered saline (PBS) solution plus 10⁻⁷ M ascorbic acid on ice, followed by IP injection.^44,48 The control mice received injections of PBS plus 10⁻⁷ M ascorbic acid only. The efficacy of sympathectomy was validated by tyrosine hydroxylase (TH) expression level in the spleen and immunofluorescence labeling of TH in the whole-mount corneas after IP injection of 6-OHDA (Supplementary Fig. S1). Corneal injury was induced on day 2 after injection. 

Pharmacologic Administration

Salmeterol (cat. MB1325, Meilune, Dalian, China) was stored at −20°C and dissolved in PBS before oral gavage administration (10 mg/kg).⁴⁹ α-Methyl-L-tyrosine (α-MLT; cat. M107878, Aladdin, Shanghai, China) was formulated into a 3-mg/mL suspension in sterile PBS and then administered by means of oral gavage (300 mg/kg/d, 2.5-mL solution).⁵⁰ Sterile PBS was used as the vehicle. For Ly6C^hi monocyte depletion, mice were IP injected with 2 mg/kg of a CCR2 antagonist (cat. 227016, Millipore, Burlington, MA, USA).⁵¹ Mice pretreated with CCR2 antagonist for 24 hours were used to create a corneal abrasion model. For blocking β2-adrenergic receptors, butaxamine (cat. B1385, Sigma-Aldrich) was administered via intraperitoneal injection at a dose of 2.5 mg/kg.⁵² 

Immunofluorescence Analysis of Whole-Mounted Corneas

Whole mounting of the cornea was performed according to our previously described study.^9–11 Briefly, corneas were harvested at different time points after corneal abrasion, fixed with 2% paraformaldehyde at room temperature for 2 hours, and washed three times for 5 minutes each. The washed corneas were incubated with fluorescently labeled antibodies, at 4°C for 12 hours, and then washed three times for 5 minutes each. Finally, the corneas were radially cut into four flaps so that they were flattened on a glass slide, and the corneal tissues and coverslips were fixed with mounting medium (cat. F6182, Sigma-Aldrich) containing 1 µM 4′,6-diamidino-2-phenylindole (cat. D9524, Sigma-Aldrich) to detect nuclei. To label neutrophils, FITC-conjugated anti-mouse Ly-6G (clone 1A8, cat. 551460, BD Biosciences, San Jose, CA, USA) was used. To calculate the number of mitotic cells in the epithelial layer, the relative number of mitotic cells in the corneal epithelium in nine fields of view (from field 1 to 1′), along both the longitudinal and transverse directions of the cornea, was used as the average amount of mitosis in the corneal epithelium, as shown Supplementary Figure S2. Mitotic cells were defined as cells with the paired nuclei, as in our previously described study.⁴³ To capture the images, a DeltaVision Image System (General Electric, Boston, MA, USA) with 40× magnification was used. 

Flow Cytometric Analysis

Flow cytometric analysis of single-cell suspensions from the corneas,¹⁰ spleen,⁵³ bone marrow,⁵⁴ and peripheral blood⁵⁵ was performed as previously described. Single cells were blocked in flow cytometry staining buffer (cat. 00-4222, eBioscience, San Diego, CA, USA) containing anti-mouse CD16/32 antibody (cat. 14-0161-85, eBioscience) for 10 minutes at room temperature. The cells were then incubated for 30 minutes at room temperature with the following antibodies (diluted 1:100): APC-conjugated anti-mouse CD45 antibody (cat. 559864, BD Biosciences, San Jose, CA, USA), Brilliant Violet 421–conjugated anti-mouse CD64 (cat. 139309, BioLegend, San Diego, CA, USA), Brilliant Violet 421–conjugated anti-mouse Lineage Cocktail (cat. 139309, BioLegend), PE-conjugated anti-mouse F4/80 (cat. 565410, BD Biosciences), FITC-conjugated anti-mouse Ly-6G (cat. 551460, BD Biosciences), PerCP-conjugated anti-mouse CD11b (cat. 45-0112-82, Thermo, Waltham, MA, USA), and PE-CY7-conjugated anti-mouse Ly6C (cat. 25-5932-82, eBioscience). Cells labeled with fluorescent antibodies were run on a flow cytometer (BD Canto Plus, BD Bioscience, San Jose, CA, USA), and the resulting data were analyzed using FlowJo flow cytometry analysis software (version 10, Ashland, OR, USA). 

Monocyte Sorting Using Flow Cytometry

Corneal cell suspensions were digested with 0.2% collagenase type I (cat. C0130, Sigma-Aldrich) and stained with a mixture of antibodies (diluted 1:100), including an anti-mouse CD45 antibody conjugated with APC (cat. 559864, BD Biosciences), anti-mouse CD64 conjugated with Brilliant Violet 421 (cat. 139309, BioLegend), anti-mouse Lineage Cocktail conjugated with Brilliant Violet 421 (cat. 139309, BioLegend), anti-mouse F4/80 conjugated with PE (cat. 565410, BD Biosciences), anti-mouse Ly-6G conjugated with FITC (cat. 551460, BD Biosciences), anti-mouse CD11b conjugated with PerCP (cat. 45-0112-82, Thermo), and anti-mouse Ly6C conjugated with PE-CY7 (cat. 25-5932-82, eBioscience), at room temperature for 30 minutes. The stained corneal cells were sorted by means of flow cytometry, using a FACSAria instrument (BD Biosciences) to obtain CD45⁺CD11b⁺CD64⁻Lin^-Ly6G⁻F4/80⁺Ly6C^hi and CD45⁺CD11b⁺CD64⁻Lin^-Ly6G⁻F4/80⁺Ly6C^low monocytes. The cells labeled with antibodies were sorted using a flow cytometer (BD FACSAria II, BD Bioscience, San Jose, CA, USA). 

Subconjunctival Adoptive Transfer for Ly6C^hi Monocytes

Subconjunctival adoptive transfer of monocytes was performed as previously described.⁶⁹ Splenic Ly6C^hi monocytes sorted from normal mouse spleens by means of flow cytometry were stored in Dulbecco's modified Eagle medium containing 10% fetal bovine serum, at 37°C, in an atmosphere containing 5% CO₂, for 3 hours, and labeled with carboxyfluorescein diacetate (CFDA; cat. V12883, Thermo Fisher Scientific, Waltham, MA, USA). The supernatant and nonadherent cells were then aspirated. Six hours after corneal abrasion, approximately 1000 Ly6C^hi monocytes/2 µL were injected subconjunctivally. The existence and fate of spleen-derived Ly6C^hi monocytes in the injured corneas were verified at 12 hours after corneal abrasion, using flow cytometry. 

Single Monocyte Isolation and Transcriptional Analysis

Single monocytes were obtained by means of flow sorting, and the method was the same as the flow sorting of monocytes described earlier. Sorted monocytes were analyzed according to the protocol provided by the REPLI-gWTA Single Cell Kit (cat. 150063, Qiagen, Venlo, Netherlands). All the reagents used in the following steps of single-cell amplification are included in the reagent kit. Briefly, the sorted mononuclear cells were pipetted into a single-cell suspension and centrifuged at 400 × g for 5 minutes, following which the cells were collected. Next, 4 µL lysis buffer was added to the cells, mixed by vortexing, incubated at 24°C for 5  minutes and 93°C for 3 minutes, and cooled to 4°C. Following that, 2 µL gDNA Wipeout Buffer was added, mixed by vortexing, and incubated at 42°C for 10 minutes. Furthermore, 20 µL RT/polymerase buffer, 5 µL random primer, 5 µL oligo dT primer, and 5 µL Quantiscript RT Enzyme Mix were added to prepare the Quantiscript RT Mix. Next, 7 µL Quantiscript RT Mix was added to the incubated samples, incubated at 42°C for 60 minutes, and incubated at 95°C for 3 minutes to stop the reaction. Next, 10 µL Ligation Mix was added, incubated at 24°C for 30 minutes, and incubated at 95°C for 5 minutes to terminate the reaction. Finally, 30 µL REPLI-g SensiPhi amplification mix was added to the reaction system, incubated at 30°C for 2 hours, and incubated at 65°C for 5 minutes to terminate the reaction. The amplified cDNA obtained from the reaction was used for quantitative PCR (qPCR) analysis. 

Splenectomy

Splenectomy was performed as previously described.⁵⁶ Briefly, the mice were anesthetized with 1% sodium pentobarbital (80 mg/kg body weight, IP injection). After routine disinfection, the left abdominal wall was incised and the spleen was exposed. After ligation of the splenic artery and vein, the spleen was carefully excised and the abdominal wall was sutured. After spleen excision, the animals were allowed to recover for 2 weeks before formal experiments were performed. 

Splenic Denervation

Splenic denervation was performed as previously described.⁵⁷ Briefly, the mice were anesthetized by means of IP injection of 1% pentobarbital sodium (80 mg/kg body weight). The abdominal wall was incised along the left side of the abdomen to carefully expose the spleen and nerves. Under a dissecting microscope, the splenic artery was carefully separated from the surrounding fat and pancreatic tissues. A cotton-tipped applicator was soaked in 10% phenol–ethanol solution over the splenic artery for 30 seconds, until marked vasodilation was visible, while taking care to avoid contact with the spleen or surrounding tissue. After identifying the blood vessels and nerves, the nerves were carefully amputated under a dissecting microscope using corneal scissors. For sham-operated mice, only the left side of the abdomen was opened to expose the spleen and splenic artery, but other procedures were not performed. The animals were allowed to recover for 2 weeks. 

In Situ Labeling of Splenic Monocytes With CFDA

Splenic monocytes were labeled in situ with CFDA, as previously described.⁵⁸ After the mice were anesthetized with an IP injection of sodium pentobarbital (80 mg/kg body weight), a 1.5-cm surgical incision was made in the skin and peritoneum on the left side of the abdomen. The spleen was exposed and partially bled by gently pulling fat. A premixed 20 µM CFDA solution was loaded into a 30-guage insulin syringe. The syringe needle was inserted into the lower pole of the spleen (approximately 0.3–0.4 mm), and the contents were slowly expelled as the end of the needle was pulled out. The spleen was gently pushed back into the abdominal cavity. The incision was intermittently sutured with 5-0 absorbable sutures in two layers (abdominal wall and skin). Postoperative analgesia was administered via buprenorphine (0.05 mg/kg, subcutaneous, twice a day) and meloxicam (1 mg/kg, subcutaneous, once a day). 

In Vivo Observation of the CFDA-Labeled Monocyte Emigration to the Injured Cornea Using Intravital Microscopy

Twelve hours after corneal abrasion, the mice were anesthetized by means of IP injection of pentobarbital sodium (80 mg/kg body weight). After complete anesthesia, the mice were placed on a custom-made plastic plate with the glasses facing up. The head was restrained with a head immobilizer (Narishige, Tokyo, Japan), and the limbus was moderately restrained by means of adhesive tape. Photographs were taken using a LSM780 two-photon laser scanning microscope (Carl Zeiss, Aalen, Germany), and all observations and imaging were performed with a 320 (numerical aperture ¼ 1.0) water-immersion objective, using a reflected light nonscanning detector with a filter window (LP490–SP485) to detect two-photon imaging signals. The two-photon excitation wavelength was 720 nm, and the power was 17% of the maximum power (3.5 W). The shooting software Zen2010B (Version 6.0, Carl Zeiss MicroImaging GmbH, Aalen, Germany) was used to set the pinhole to the maximum open setting and scan with a 3-μm z-axis step size. The scanning area was randomly selected from five fields of view—the central, paracentral, parawound, peripheral, and limbus—to observe and take pictures. 

Western Blot Analysis

The spleen tissue was harvested, cut using ophthalmic scissors, and ground in radioimmunoprecipitation lysis buffer, following which the obtained lysate was digested for 30 minutes and centrifuged at 12,000 rpm for 5 minutes. A bicinchonic acid kit (cat. PC0020, Solarbio, Beijing, China) was used to quantify the proteins, following which equal amounts of protein (20 mg/lane) were separated on a 6% Bis-Tris sodium dodecyl sulfate–polyacrylamide gel electrophoresis gel and transferred to polyvinylidene difluoride membranes (Millipore, San Diego, CA, USA). To detect TH or levodopa (l-DOPA), the membranes were blocked with 5% bovine serum albumin in PBS and washed three times, followed by incubation with an anti-TH antibody (cat. MAB7566, R&D Systems, Minneapolis, MN, USA) or anti–l-DOPA Ab (cat. NBP3-06881, Novus Biologicals, Englewood, CO, USA). The antibody was diluted in a ratio of 1:1000 (Cell Signaling Technology, Danvers, MA, USA) and left overnight at 4°C. The membranes were then incubated with an anti-rabbit horseradish peroxidase–linked secondary antibody (diluted in a ratio of 1:2000) for 1 hour at 25°C. Antibody-labeled proteins were developed using the ECL kit (cat. 34577, ThermoFisher) and imaged by the FluorChem Q imager (GE Healthcare, Chicago, IL, USA). 

RNA Extraction, RNA Sequencing, and Analysis of RNA Sequencing Data

The corneal tissue was cut into small pieces, placed in Buffer RZ (cat. RK145, Tiangen Biotechnology, Beijing, China), and minced well using a microtissue grinder (Qiagen). The minced tissue was left at room temperature for 5 minutes to facilitate the separation of nucleic acid–protein complexes. An RNAsimple Total RNA Kit (cat. DP419, Tiangen Biotechnology) was used to extract total RNA from the corneal tissue. A ReverTra Ace qPCR RT Kit (cat. FSQ˗101, Toyobo, Osaka, Japan) was used to obtain cDNA, and the expression of the target genes in the samples was analyzed using the THUNDERBIRD SYBR qPCR mix (cat. QPS˗201, Toyobo). 

RNA sequencing was performed as previously described.^59,60 The first step in the workflow involved the purification of poly A–containing mRNA molecules using poly T oligo–attached magnetic beads. Following purification, the mRNA was fragmented into small pieces using divalent cations at elevated temperatures. The cleaved RNA fragments were copied into first-strand cDNA using reverse transcriptase and random primers. This step was followed by the synthesis of second-strand cDNA using DNA Polymerase I and RNase H. These cDNA fragments were then subjected to addition of a single A base and subsequent ligation of the adapter. The products were purified and enriched using PCR amplification. Following that, the PCR yield was quantified with a Qubit 2.0 fluorometer (Thermo Fisher Scientific), and the samples were pooled together to create a single-stranded DNA circle, which produced the final library. DNA nanoballs were generated with the single-stranded DNA circle by rolling circle replication, to enlarge fluorescent signals in the sequencing process. The DNA nanoballs were loaded into the patterned nanoarrays, and single-end reads of 50 base pairs were read using the BGISEQ-500 platform (BGI, Shenzhen, China) for data analysis. For this step, the BGISEQ-500 platform combined DNA nanoball-based nanoarrays and stepwise sequencing using the combinational probe anchor synthesis sequencing method. 

The gene expression level was expressed as fragments per kilobase of transcript per million reads mapped. The R software DESeq2 (version 1.38.3, https://bioconductor.org/packages/release/bioc/html/DESeq2.html) package was used to analyze the differential expression between the different groups, and genes with a Q value of <0.05 and |log₂ (fold change)| of >1 were regarded as differentially expressed genes (DEGs).⁶¹ Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment and Gene Set Enrichment Analysis (GSEA) enrichment analyses were performed on the DEGs using the clusterProfiler (version 4.6.0, https://bioconductor.org/packages/ release/bioc/html/clusterProfiler.html) R software package,^62,63 and 0.05 was used as the significance threshold of the P value. 

qPCR

Corneal tissues were cut into pieces, placed in Buffer RZ (cat. R0424, Omega, Mogadore, OH, USA), and smashed using a TissueRuptor (cat. OSE-Y10, Tiangen). Total RNA was extracted from corneal tissues using the RNA Simple Total RNA Kit (cat. R6688R, Omega). cDNA was generated using a ReverTra Ace qPCR RT Kit (cat. FSQ-101, Toyobo). PCR primers used in this study are listed in the Table. 

Table.

 

PCR Primers Used in This Study

 

Statistical Analyses

SPSS for Windows (Version 16.0, SPSS, Inc., Chicago, IL, USA) was used for statistical analyses and generating graphs. Factorial design analysis of variance and unpaired Student's t-test were used to perform comparisons between groups. Multiple sets of quantitative data were analyzed using one-way analysis of variance, followed by pairwise comparisons using the Bonferroni method. The Pearson correlation coefficient, r, was computed to estimate the association between two continuous variables. Results are presented as mean ± SD. Statistical significance was set at P < 0.05. 

Results

Recruitment of Ly6C^hi Monocytes From the Spleen to the Injured Cornea After Corneal Abrasion

To address whether monocytes are recruited to the injured cornea and coordinate the wound healing, we mechanically scraped a 2-mm-diameter area of the central corneal epithelium in C57BL6/J mice and collected injured corneas at different time points (6, 12, 18, and 24 hours), for flow cytometric analysis, as demonstrated in our previous studies^10,45,64 (Supplementary Fig. S3A). As shown in Figures 1A and 1B, monocytes infiltrating injured corneas were classified into three subgroups, based on the level of Ly6C expression: Ly6C^hi, Ly6C^int, and Ly6C^low. Further dynamic analysis showed that the number of Ly6C^hi monocytes in the injured corneas started to increase at 6 hours after corneal abrasion, peaked at 12 hours, decreased at 18 hours, and returned to baseline levels at 24 hours (Fig. 1B). However, the numbers of both Ly6C^int and Ly6C^low monocytes in the injured cornea were relatively constant at different time points after corneal abrasion (Figs. 1C, 1D). Thus, these data indicated that Ly6C^hi monocytes, but not Ly6C^int or Ly6C^low monocytes, infiltrated the injured cornea over time after corneal abrasion. 

Figure 1.

 

Dynamic changes of Ly6C^hi monocytes in the cornea after corneal abrasion. (A) Representative flow cytometry plots for identification of three groups of monocytes in the injured cornea, at different time points after corneal abrasion. (B) Assessment of dynamic changes in the Ly6C^hi monocytes in the injured cornea using flow cytometry, at different time points after corneal abrasion. (C, D) Assessment of dynamic changes in the Ly6C^int and Ly6C^low monocytes in the injured cornea using flow cytometry, at different time points after corneal abrasion. For flow cytometry, each corneal sample was pooled from 10 corneas, with three independent replicates. Data are represented as mean ± SD. *P < 0.05, **P < 0.01.

Next, to determine which compartment the infiltrated Ly6C^hi monocytes following corneal abrasion mainly come from, the number of Ly6C^hi monocytes in the peripheral blood, spleen, and bone marrow was analyzed using flow cytometry. The results revealed that the number of Ly6C^hi monocytes in the injured corneas increased significantly at 12 hours after corneal abrasion, as shown in Figure 1A. In parallel with this increase, the number of Ly6C^hi monocytes in the peripheral blood also increased (Figs. 2A, 2B). To evaluate whether the number of Ly6C^hi monocytes in the cornea is related to that in the peripheral blood, the ratio of Ly6C^hi monocytes in the two tissues, at 0, 6, 12, 18, and 24 hours, was analyzed using Pearson correlation analysis. There was a strong positive correlation between dynamic changes in Ly6C^hi monocytes in the injured corneas and peripheral blood (r = 0.9480, P < 0.001) (Fig. 2C). However, the number of Ly6C^hi monocytes in the spleen decreased significantly from 69.8% in the resting spleen to approximately 46% at 12 hours after corneal abrasion (Figs. 2D, 2E). Statistical analysis of correlations revealed a mildly negative correlation between the dynamics of Ly6C^hi monocytes in injured corneas and spleens (r = −0.5294, P = 0.0031) (Fig. 2F) but not in relation to the dynamics of Ly6C^hi monocytes in the bone marrow (r = 0.3290, P = 0.0814) (Figs. 2G–I). In conclusion, these data suggested that Ly6C^hi monocytes recruited to the cornea are predominantly derived from the spleen, rather than from the bone marrow. 

Figure 2.

 

Distribution and dynamics of Ly6C^hi monocytes in the peripheral blood and spleen, after corneal abrasion. (A) The change in the number of Ly6C^hi monocytes in the peripheral blood, before and 12 hours after corneal abrasion, as revealed using flow cytometry. (B) Representative flow cytometric plots showing the change in the number of Ly6C^hi monocytes in the peripheral blood, before and 12 hours after corneal abrasion. (C) Ly6C^hi monocytes in the cornea by that in the peripheral blood, at 0, 6, 12, 18, and 24 hours after corneal abrasion. The Pearson correlation coefficient (r) is shown. The gray band represents the 95% confidence interval. (D) The change in the number of Ly6C^hi monocytes in the spleen, before and 12 hours after corneal abrasion, as revealed using flow cytometry. (E) Representative flow cytometric plots showing the change in the number of Ly6C^hi monocytes in the spleen, before and 12 hours after corneal abrasion. (F) Ly6C^hi monocytes in the cornea by that in the spleen, at 0, 6, 12, 18, and 24 hours after corneal abrasion. The Pearson correlation coefficient (r) is shown. The gray band represents the 95% confidence interval. (G) The change in the number of Ly6C^hi monocytes in the bone marrow, before and 12 hours after corneal abrasion, as revealed using flow cytometry. (H) Representative flow cytometry plots showing the change in the number of Ly6C^hi monocytes in the bone marrow, before and 12 hours after corneal abrasion. (I) Ly6C^hi monocytes in the cornea by that in the bone marrow, at 0, 6, 12, 18, and 24 hours after corneal abrasion. The Pearson correlation coefficient (r) is shown. The gray band represents the 95% confidence interval. For flow cytometry, each corneal sample was pooled from 10 corneas, with three independent replicates; each peripheral blood sample was pooled from the combined peripheral blood of two mice, with three independent replicates; each spleen sample was obtained from one mouse, with six independent replicates; each bone marrow sample was obtained from two mouse, with six independent replicates. Data are represented as mean ± SD. *P < 0.05, **P < 0.01.

Ly6C^hi Monocytes Are Responsible for Neutrophil Extravasation and Wound Closure Following Corneal Abrasion

To determine the temporal relationship between the recruitment of neutrophils and Ly6C^hi monocytes to the cornea after corneal abrasion, we plotted the dynamic changes in the numbers of both cell types over wounding time, using flow cytometry. Neutrophils were defined as a cell subset characterized by the expression of CD45⁺CD11b⁺F4/80⁻/CD64⁻Lin⁻Ly6C⁻Ly6G⁺ (Supplementary Fig. S3B). Consistent with our previous data,⁶⁵ neutrophils peaked at 18 hours after corneal abrasion (Fig. 3A). However, the number of infiltrated Ly6C^hi monocytes peaked at 12 hours and 6 hours earlier than the neutrophils (Fig. 3A). This suggests that the early infiltration of Ly6C^hi monocytes may be responsible for the subsequent influx of neutrophils into the injured cornea. 

Figure 3.

 

Ly6C^hi monocytes participate in neutrophil recruitment after corneal abrasion. (A) Flow cytometry plots showing the dynamics of recruitment of Ly6C^hi monocytes and Ly6G⁺ neutrophils to the injured cornea at different time points after corneal abrasion. (B) Comparative analysis of Ly6C^hi monocytes counts in the resting spleen of vehicle- and CCR2 antagonist-treated mice for 24 hours by flow cytometry. Representative flow cytometric analysis of Ly6C^hi monocytes counts in the spleen of vehicle-treated mice (left) and CCR2 antagonist-treated mice (right). (C) Comparative analysis of Ly6C^hi monocytes counts in the peripheral blood of vehicle- and CCR2 antagonist-treated mice for 24 hours at 12 hours after corneal abrasion by flow cytometry. Representative flow cytometric analysis of Ly6C^hi monocytes counts in the peripheral blood of vehicle-treated mice (left) and CCR2 antagonist-treated mice (right). (D) Comparative analysis of Ly6C^hi monocytes counts in the cornea of vehicle- and CCR2 antagonist-treated mice for 24 hours at 12 hours after corneal abrasion by flow cytometry. Representative flow cytometric analysis of Ly6C^hi monocytes counts in the cornea of vehicle-treated mice (left) and CCR2 antagonist-treated mice (right). (E) Flow cytometry plots showing the effect of the administration of CCR2 antagonists on the recruitment of Ly6G⁺ neutrophils to the injured corneas at different times. (F) Left: a representative image acquired under a dissecting microscope showing the unclosed wound area 12 hours after corneal abrasion by 1% sodium fluorescein staining. Right: a representative image of FITC-conjugated Ly6G⁺ neutrophils at the wound area of the whole-mounted cornea under fluorescence microscopy (scale bar: 20 µm). (G) Functional KEGG enrichment for the differentially expressed genes between vehicle- and CCR2 antagonist-treated corneas, at 12 hours after corneal abrasion. (H) GSEA analysis showing that TNF, IL-17, NET formation, and chemokine signaling pathways were inhibited at 12 hours after corneal abrasion in the CCR2 antagonist-treated mice. (I) Effect of Ly6C^hi monocytes depletion with CCR2 antagonist on gene expression of proinflammatory molecules in the injured corneas, at 12 hours after corneal abrasion. For flow cytometry, each corneal sample was pooled from 10 corneas, with three independent replicates; each peripheral blood sample was pooled from the combined peripheral blood of two mice, with three independent replicates; each spleen sample was obtained from one mouse, with six independent replicates. Data are represented as mean ± SD. *P < 0.05, **P < 0.01. NET, neutrophil extracellular trap.

To test the proposed hypothesis, we first employed a CCR2 antagonist, a chemical compound (C₂₈H₃₄F₃N₅O₄S), to pretreat the animals and deplete monocytes. This compound efficiently competes with monocyte chemoattractant protein 1 (MCP-1) to bind to CCR2,⁵¹ thus reducing the number of Ly6C^hi monocytes. The efficacy of this approach was demonstrated by administering the CCR2 antagonist through intraperitoneal injection and measuring the depletion of Ly6C^hi monocytes in the resting murine spleens after 24 hours. The results showed a significant reduction in the number of Ly6C^hi monocytes in the CCR2 antagonist-treated spleens compared to the vehicle-treated spleens (4% [vol/vol] dimethyl sulfoxide in PBS) (Fig. 3B). To further determine whether Ly6C^hi monocytes in the peripheral blood and cornea of mice after corneal abrasion were derived from the spleen, we counted the number of Ly6C^hi monocytes in both the peripheral blood and cornea at 12 hours after corneal abrasion by flow cytometry in mice pretreated with CCR2 antagonist for 24 hours. The results indicated a significant reduction in the number of Ly6C^hi monocytes in the peripheral blood (Fig. 3C) and cornea (Fig. 3D) of mice subjected to corneal abrasion. In conclusion, these results confirm that administration of CCR2 antagonists effectively removes resident Ly6C^hi monocytes from the resting spleen and simultaneously reduces the number of recruitment through the peripheral blood to the injured cornea. 

Next, to determine the effect of Ly6C^hi monocyte depletion on neutrophil recruitment to the injured cornea, we quantified the dynamics of Ly6G⁺ neutrophil recruitment to the injured corneas at different time points in the CCR2 antagonist- and vehicle-treated mice, using flow cytometry. To exclude the possibility that peripheral blood neutrophils were removed by CCR2 antagonists due to their expression of CCR2, we examined the expression of CCR2 in peripheral blood neutrophils of mice by RT-PCR. The results showed that peripheral blood neutrophils from mice undergoing corneal injury did not have any expression of CCR2 (Supplementary File). Notably, the number of neutrophils in the corneas treated with the CCR2 antagonist was significantly reduced at 6, 12, 18, and 24 hours postabrasion, as compared to that in the vehicle-treated mice (Figs. 3E, 3F). These data suggested that the early recruitment of monocytes to the injured cornea drives subsequent neutrophil recruitment after corneal abrasion. 

To understand the underlying molecular mechanism of monocyte-driven neutrophil recruitment to the injured cornea, we first compared DEGs (|log fold change|>1, P < 0.05) between the vehicle- and CCR2 antagonist-treated injured corneas (Supplementary File). A total of 3974 DEGs were identified between the two groups, including 1396 upregulated and 2578 downregulated genes. Functional KEGG enrichment analysis of these DEGs showed that the top 10 pathways were mainly associated with inflammatory responses, particularly the chemokine, IL-17, and nuclear factor κB signaling pathways (Fig. 3G). To further explore the activation statuses of the signaling pathways, GSEA was performed on these DEGs. Tumor necrosis factor (TNF), IL-17, neutrophil extracellular trap formation, and chemokine signaling pathways were significantly downregulated in the CCR2 antagonist-treated corneas, as compared with those in the vehicle-treated controls (Fig. 3H, Supplementary Fig. S4). In addition, to further verify this information, qPCR was performed to compare the expression of key cytokines between the two groups of wounded corneas. The expression of Mcp1p, Il17a, Tnf, Il1a, Arg1, and Retnlb was found to be significantly decreased in the corneas treated with a CCR2 antagonist, as compared to those in the vehicle-treated corneas (Fig. 3I). Thus, these results indicated that early recruitment of Ly6C^hi monocytes to the injured cornea drives further recruitment of subsequent neutrophils to the injured cornea, through higher-level production of proinflammatory cytokines. 

Finally, to determine the role of Ly6C^hi monocytes in reepithelialization after corneal abrasion, the process of corneal epithelial closure was recorded with fluorescein under a dissecting microscope, as previously described.⁶⁶ The data showed that wound closure after abrasion in the Ly6C^hi monocyte-depleted corneas was faster than that in the vehicle-treated animals (24 vs. 30 hours) (Figs. 4A, 4B). Meanwhile, the number of corneal epithelial mitotic cells in the Ly6C^hi monocyte-depleted animals peaked at 18 hours after corneal abrasion, while it peaked at 24 hours in the vehicle-treated animals (Figs. 4C, 4D). Therefore, these data suggested that Ly6C^hi monocytes may also be involved in corneal wound closure by affecting neutrophil infiltration into the injured cornea. 

Figure 4.

 

Ly6C^hi monocytes depletion promotes corneal wound repair. (A) Representative images of corneal wound closure in the Ly6C^hi monocyte-depleted and vehicle-treated mice. The staining of the wounded corneal area with fluorescein sodium, at each time point after epithelial abrasion. (B) Dynamic changes in corneal wound closure after corneal abrasion at 6 hours intervals (n = 6 corneas per group, at each time point). (C) Representative image of 4′,6-diamidino-2-phenylindole–labeled mitotic cells in the epithelial layer, which were acquired from zone 2 (peripheral region), 18 hours after corneal abrasion (40×; scale bar: 20 µm). (D) Plots showing the dynamic changes in mitotic cell numbers over time, after corneal abrasion in Ly6C^hi-depleted mice. n = 6 corneas for each time point. Data are represented as mean ± SD. *P < 0.05, **P < 0.01. n = 6 corneas.

Splenectomy Reduces the Recruitment of Ly6C^hi Monocytes to the Injured Cornea

To further confirm the splenic origin of Ly6C^hi monocytes in injured corneas, surgical splenectomies were performed in some animals. Two weeks after recovery from surgery, changes in the number of Ly6C^hi monocytes in the cornea, peripheral blood, and bone marrow of surgically splenectomized and sham-operated animals without corneal injury were examined using flow cytometry. The distribution of Ly6C^hi monocytes in the cornea, peripheral blood, and bone marrow was not altered in either the animals whose spleens were surgically removed or sham-operated animals (Figs. 5A–C). Similarly, we compared the distribution of Ly6C^hi monocytes in the cornea, peripheral blood, and bone marrow of splenectomized and sham-operated animals at 12 hours after corneal injury. There was a significant decrease in the number of Ly6C^hi monocytes in the cornea and peripheral blood after splenectomy (Figs. 5D, 5E) but no significant change in the number in the bone marrow (Fig. 5F). These data suggested that the Ly6C^hi monocytes recruited to the injured cornea may be primarily derived from the spleen rather than the bone marrow. 

Figure 5.

 

Splenectomy results with a change in the distribution of Ly6C^hi monocytes in the cornea, peripheral blood, and bone marrow after but not before corneal abrasion. (A) Distribution of Ly6C^hi monocytes in the resting cornea of animals whose spleen was surgically removed and those that were sham-operated (left). Representative plots of the flow cytometry analysis of the resting cornea (right). (B) Distribution of Ly6C^hi monocytes in the peripheral blood without corneal abrasion, in animals whose spleen was surgically removed and those that were sham-operated (left). Representative plots of the flow cytometry analysis of the peripheral blood (right). (C) Distribution of Ly6C^hi monocytes in the bone marrow without corneal abrasion, in animals whose spleen was surgically removed and those that were sham-operated (left). Representative plots of the flow cytometry analysis of the bone marrow (right). (D) Distribution of Ly6C^hi monocytes in the cornea, at 12 hours after corneal abrasion, in animals whose spleen was surgically removed and those that were sham-operated (left). Representative plots of flow cytometry analysis in the injured cornea (right). (E) Distribution of Ly6C^hi monocytes in the peripheral blood, at 12 hours after corneal abrasion, in animals whose spleen was surgically removed and those that were sham-operated (left). Representative plots of the flow cytometry analysis of the peripheral blood (right). (F) Distribution of Ly6C^hi monocytes in the bone marrow, at 12 hours after corneal abrasion, in animals whose spleen was surgically removed and those that were sham-operated (left). Representative plots of the flow cytometry analysis of the bone marrow (right). For flow cytometry, each corneal sample was pooled from 10 corneas, with three independent replicates. Each peripheral blood sample was pooled from the combined peripheral blood of two mice, with three independent replicates. Each spleen sample was obtained from one mouse, with six independent replicates. Each bone marrow sample was pooled from the combined peripheral blood of two mice, with six independent replicates. Data are represented as mean ± SD. NS, not significant, *P < 0.05, **P < 0.01.

In Situ–Labeled Spleen-Derived Monocytes Are Recruited to the Injured Cornea

To further determine whether the recruited Ly6C^hi monocytes in the injured cornea originate from the spleen, CFDA, a fixable cell-permeant fluorescein-based tracer for long-term cellular labeling, was administered by means of direct injection into the spleen, to locally label the spleen cells.^67,68 To exclude the possibility of injected dyes in the spleen leaking into the blood to label circulating immune cells, some mice were subjected to tail vein injection with CFDA, as a control. As predicted, few CFDA⁺ Ly6C^hi monocytes were detected in the corneas of mice injected with vehicle in their spleen or CFDA in their tail veins, at 12 hours after corneal injury, using both intravital microscopy (Figs. 6A–C) and flow cytometry (Figs. 6D–F). However, large numbers of CFDA⁺ Ly6C^hi monocytes were identified in the cornea of mice injected with CFDA in their spleen, at 12 hours after corneal abrasion (Fig. 6E). Further quantitative flow analysis showed that CFDA⁺ Ly6C^hi monocytes in the corneas of mice injected with CFDA in their spleen accounted for 88% of all Ly6C^hi monocytes in the injured corneas (Fig. 6G). Therefore, these data suggested that the spleen may be the primary source of Ly6C^hi monocyte recruitment to the injured cornea. 

Figure 6.

 

Recruitment of Ly6C^hi monocytes from CFDA-labeled spleens to the injured cornea. (A) A few cells carrying green fluorescence were found by means of intravital microscopy, in the corneas limbus of the mice injected with vehicle in the spleen, at 12 hours after corneal abrasion. (B) A large number of cells carrying green fluorescence were found by means of intravital microscopy, in the cornea limbus of mice injected with CFDA in the spleen, at 12 hours after corneal abrasion. (C) A few cells carrying green fluorescence were detected by means of intravital microscopy, in the cornea limbus of mice injected with CFDA in the tail vein, at 12 hours after corneal abrasion. (D) A few CFDA⁺ Ly6C^hi monocytes were detected by means of flow cytometry, in the corneas of mice injected with vehicle in the spleen, at 12 hours after corneal abrasion (Q2). (E) A large number of Ly6C^hi monocytes were detected by means of flow cytometry, in the corneas of mice injected with CFDA in the spleen, at 12 hours after corneal abrasion (Q2). (F) A few CFDA⁺ Ly6C^hi monocytes were detected by means of flow cytometry, in the corneas of mice injected with CFDA in the tail vein, at 12 hours after corneal abrasion (Q2). (G) Flow cytometric quantification of CFDA⁺ Ly6C^hi monocytes in the whole corneas of mice injected with vehicle in the spleen or CFDA in the tail vein or spleen, at 12 hours after corneal abrasion. Each corneal sample was pooled from 10 corneas, with three independent replicates. Data are represented as mean ± SD.

Locally Adoptive Transfer of Ly6C^hi Monocytes Restores the Defective Inflammatory Response Caused by Ly6C^hi Monocyte Depletion

To further verify Ly6C^hi monocytes originating from the spleen play an important role in triggering an inflammatory response after corneal abrasion, we purified Ly6C^hi monocytes from the spleen using a flow-sorting protocol. Purified Ly6C^hi monocytes were administered to mice that underwent splenectomy via subconjunctival injection as described previously.⁶⁹ A corneal abrasion model was established 6 hours after subconjunctival injection of CFDA-labeled Ly6C^hi monocytes, and monocytes were detected in the injured corneas at 12 hours after abrasion, using flow cytometry. Subconjunctival injection of CFDA-labeled Ly6C^hi monocytes effectively complemented the reduced Ly6C^hi monocytes recruited to the injured corneas due to spleen removal, as compared to that observed in case of Ly6C^hi monocytes in injured corneas from sham-operated animals (Fig. 7A). Next, we compared the effect of subconjunctival adoptive transfer of Ly6C^hi monocytes on postwound neutrophil recruitment to the injured corneas. As splenic removal caused a decrease in the recruitment of Ly6C^hi monocytes to the injured corneas, there was a significant decrease in the number of neutrophils at the same time. In contrast, after compensating for the loss of Ly6C^hi monocytes in the injured corneas of splenectomized animals by subconjunctival replenishment of Ly6C^hi monocytes, the neutrophils in the injured cornea were effectively replenished (Fig. 7B). 

Figure 7.

 

Subconjunctival adoptive transfer of Ly6C^hi monocytes ameliorates abnormal corneal wound repair caused by spleen removal. (A) Comparison of Ly6C^hi monocytes in the injured cornea, at 12 hours after corneal abrasion, between the sham-operated, splenectomized, and splenectomized mice with subconjunctival adoptive transfer of Ly6C^hi monocytes (left). The gating strategy of flow cytometry for analyzing Ly6C^hi monocytes in three different kinds of injured corneas (right). (B) Dynamics of subconjunctival adoptive transfer of Ly6C^hi monocytes on neutrophil recruitment to the wounded corneas, at different time points after corneal abrasion (left); the gating strategy for flow cytometry analysis of Ly6G⁺ neutrophils(Q2) in three different kinds of injured corneas (right). (C) Observation of wound closure in the cornea from sham-operated + subconjunctival injection of PBS, splenectomized + subconjunctival injection of PBS, and splenectomized + subconjunctival injection of Ly6C^hi monocytes in mice. The staining of the wounded corneal area with fluorescein sodium, at each time point after epithelial abrasion. (D) Dynamic changes in corneal wound closure after corneal abrasion at 6 hours intervals (n = 6 corneas per group, at each time point). (E) Comparison of the number of dividing epithelial cells in the corneas of mice subjected to subconjunctival injection of PBS as a normal control, splenectomized + subconjunctival injection of PBS, and splenectomized + subconjunctival injection of Ly6C^hi monocytes, at each time point after abrasion (n = 6 corneas per group, at each time point). (F) Comparison of the number of dividing epithelial cells at all time points in the corneas of mice subjected to subconjunctival injection of PBS as a normal control, splenectomized + subconjunctival injection of PBS, and splenectomized + subconjunctival injection of Ly6C^hi mice (n = 6 corneas per group, at each time point). (G) Functional KEGG enrichment for DEGs between splenectomy- and splenectomy + Ly6C^hi monocytes adoptive transfer-treated corneas, at 12 hours after corneal abrasion. (H) GSEA analysis showing that IL-17 and TNF signaling pathways were activated in splenectomy + Ly6C^hi monocyte adoptive transfer-treated corneas, as compared to splenectomy only-treated corneas, at 12 hours after corneal abrasion. (I) Effect of splenectomy, as well as, splenectomy + subconjunctival adoptive transfer of Ly6C^hi monocytes on the expression of Mcp1p, Il17a, Il1a, Retnlb, and Tnf in the injured corneas, at 12 hours after corneal abrasion. Each corneal sample was pooled from 10 corneas, with three independent replicates. Data are presented as mean ± SD. *P < 0.05, **P < 0.01. *Splenectomy versus splenectomy + Ly6C^hi monocytes. ^&&Splenectomy vs. control, P < 0.01. ^##Splenectomy + Ly6C^hi monocytes versus control, P < 0.01.

Moreover, we compared the differences in wound closure and epithelial cell division after corneal abrasion between spleen-removed animals and spleen-removed animals with locally adoptive transfer of Ly6C^hi monocytes. Results show that epithelialization of the injured corneas of the spleen-removed mice was significantly faster than that of spleen-removed animals with locally adoptive transfer of Ly6C^hi monocytes for 12 hours (Figs. 7C, 7D). However, in the spleen-removed animals with locally adoptive transfer of Ly6C^hi monocytes, reepithelialization was delayed for 6 hours compared with that in sham-treated animals. Notably, the number of mitotic cells in the corneal epithelium peaked earlier (at 18 hours) in only the spleen-removed animals and peaked at 30 hours in the spleen-removed animals with locally adoptive transfer of Ly6C^hi monocytes compared with the peaking at 24 hours in the sham-treated animals (Fig. 7E). Furthermore, we compared the cumulative number of mitotic cells in the corneal epithelium at different times after corneal abrasion among the three treatments. The results showed that the three different treatments did not change the cumulative number of mitotic cells from 0 to 36 hours postwounding, although they changed the peak timing of mitotic cells, as described before (Fig. 7F). Overall, these data confirmed that the subconjunctival adoptive transfer of Ly6C^hi monocytes significantly compensated for the abnormal inflammatory response but not wound closure after corneal abrasion due to spleen removal. 

Finally, we compared the expression levels of proinflammatory molecules between the splenectomized animals and spleen-removed animals in which local Ly6C^hi monocytes were adopted (Supplementary File). A total of 2586 DEGs (absolute value of fold change >2, Q value <0.01), including 1185 upregulated and 1401 downregulated genes, were identified in the abrated corneas from the splenectomized group alone, as compared to those from the splenectomized group with subconjunctival adoptive transfer of Ly6C^hi monocytes. Functional enrichment analysis of these DEGs revealed the activation of several inflammatory pathways, including TNF, IL-17, nuclear factor κB, and transforming growth factor β signaling pathways (Fig. 7G). Further, GSEA analysis showed that the IL-17 and TNF signaling pathways were activated in the Ly6C^hi monocyte-adoptive transfer-treated corneas (Fig. 7H, Supplementary Fig. S5). In addition, qPCR performed to verify the expression levels of proinflammatory cytokines in the corneas at 12 hours postwounding showed that the expression of proinflammatory cytokines, including Mcp1p, Tnf, Il1a, Retnlb, and Il17a, was significantly increased in the injured corneas of the splenectomized mice treated with additional Ly6C^hi monocyte adoptive transfer, as compared to those of only splenectomized mice (Fig. 7I). In conclusion, these data suggested that adoptive transfer of Ly6C^hi monocytes improves the local inflammatory response in injured corneas following splenic removal. 

Chemical Sympathectomy Blocks the Recruitment of Ly6C^hi Monocytes to the Injured Cornea

To ascertain sympathetic nerves regulate the recruitment of Ly6C^hi monocytes from the spleen to the injured cornea, we used 6-OHDA, a commonly used sympathetic-specific neurotoxin that chemically ablates adrenergic neurons by generating reactive oxygen species.^44,70 As expected, 6-OHDA–treated animals had significantly lower numbers of Ly6C^hi monocytes in the injured cornea after corneal abrasion than vehicle-treated animals (Fig. 8A). Consistent with the changes in the number of Ly6C^hi monocytes in the cornea, the number of Ly6C^hi monocytes in the peripheral blood of the 6-OHDA–treated mice was also significantly reduced, as compared to that in the vehicle-treated animals (Fig. 8B). However, in the vehicle-treated group, the number of Ly6C^hi monocytes in the spleen decreased with the increase of Ly6C^hi monocytes in the corneal tissue and peripheral blood. However, in 6-OHDA–treated mice, the number of Ly6C^hi monocytes in the spleen was significantly higher than that in the vehicle-treated group (Fig. 8C). Collectively, these data suggested that sympathetic nerves control the recruitment of Ly6C^hi monocytes from the spleen to the injured cornea. 

Figure 8.

 

Chemical sympathectomy alters the recruitment of Ly6C^hi monocytes to the injured cornea and peripheral blood after corneal abrasion. (A) 6-OHDA treatment reduced the recruitment of Ly6C^hi monocytes to the injured cornea 12 hours after corneal abrasion, as assessed using flow cytometry. (B) 6-OHDA treatment reduced the recruitment of Ly6C^hi monocytes to the peripheral blood in the corneal injury host 12 hours after corneal abrasion, as assessed using flow cytometry. (C) 6-OHDA treatment did not reduce the emigration of Ly6C^hi monocytes from the spleen 12 hours after corneal abrasion, as assessed using flow cytometry. Each corneal sample was pooled from 10 corneas, with three independent replicates. Peripheral blood samples were obtained from the combined peripheral blood of two mice, with three independent replicates. Each spleen sample was obtained from one mouse, with six independent replicates. Data are represented as mean ± SD. *P < 0.05, **P < 0.01.

Splenic Denervation Inhibits the Recruitment of Ly6C^hi Monocytes to the Injured Cornea and Interferes With the Repair Process of the Injured Cornea

To minimize the influence of the administration of 6-OHDA to systematically remove peripheral sympathetic nerve termini in the bone marrow and other immune organs, we further observed the impact of selective surgical removal of splenic nerves on Ly6C^hi monocyte recruitment to the injured cornea (Supplementary Fig. S6). The splenic-denervated animals displayed a significantly reduced number of Ly6C^hi monocytes in the injured cornea, at 12 hours after abrasion, as compared to the sham-operated animals (Fig. 9A). Similarly, the splenic-denervated animals also displayed a significantly reduced number of Ly6C^hi monocytes in the peripheral blood after corneal abrasion than the sham-operated animals (Fig. 9B). In the sham-treated group, the number of Ly6C^hi monocytes in the spleen was decreased, whereas the number of Ly6C^hi monocytes in the corneal tissue and peripheral blood was increased. However, in the splenic denervation–treated group, the number of Ly6C^hi monocytes in the spleen was significantly higher than in the sham-treated group (Fig. 9C). These results suggested that splenic innervation controls the recruitment of splenic Ly6C^hi monocytes to the injured cornea. 

Figure 9.

 

Splenic denervation blocks the recruitment of Ly6C^hi monocytes to the injured cornea. (A) The number of Ly6C^hi monocytes in the cornea of mice that underwent splenic denervation decreased at 12 hours after corneal abrasion, as assessed using flow cytometry. The right side represents typical flow cytometry data. (B) Splenic denervation reduced the number of Ly6C^hi monocytes in the peripheral blood, at 12 hours after corneal abrasion. (C) Splenic denervation prevented the decrease the number of Ly6C^hi monocytes in the spleen, at 12 hours after corneal abrasion. (D) Representative images of corneal wound closure in splenic denervation and sham-operated mice with fluorescein sodium staining after corneal abrasion. (E) Dynamic changes in corneal wound closure after corneal abrasion at 6 hours intervals (n = 6 corneas per group, at each time point). (F) Comparison of the number of dividing epithelial cells in the corneas of splenic denervation and sham-operated mice at each time point after epithelial abrasion (n = 6 corneas per group, at each time point). (G) Splenic denervation reduced proinflammatory gene expression in the injured corneas, at 12 hours after corneal abrasion. Each corneal sample was pooled from 10 corneas, with three independent groups. Peripheral blood samples were obtained from the combined peripheral blood of two mice, with three independent replicates. Each spleen sample was obtained from one mouse, with six independent replicates. For corneal wound healing, data were obtained from six corneal samples. Data are represented as mean ± SD. *P < 0.05, **P < 0.01.

Next, we investigated the effect of splenic denervation on wound closure and epithelial cell mitosis following corneal abrasion. The reepithelialization process was significantly accelerated in the splenic denervation group, as compared to that in the sham-operated group (Figs. 9D, 9E). Similarly, the number of mitotic divisions in corneal epithelial cells peaked 6 hours earlier in the splenic denervation group than in the sham-operated group, and the number of mitotic cells was significantly increased at 6, 12, and 18 hours postwounding (Fig. 9F). In addition, we examined the expression levels of proinflammatory cytokines in the corneas at 12 hours after injury by qPCR. The results showed that the expression of proinflammatory cytokines, including Mcp1p, Il17a, Il1a, Tnf, and Retnlb, was significantly decreased in the injured corneas of splenic neurectomy mice compared with sham-operated mice (Fig. 9G). In conclusion, these data suggested that splenic denervation accelerates wound repair in corneal wounds. 

Inhibition of NE Synthesis Blocks the Recruitment of Ly6C^hi Monocytes to the Injured Cornea

To determine the role of the SNS in the recruitment of Ly6C^hi monocytes to the injured corneas, we pretreated animals with an IP injection of α-MLT. To determine the effect of α-MLT on the production of l-DOPA in the spleen, Western blot analysis was used to compare the differences between α-MLT– and vehicle-treated animals. α-MLT treatment was found to significantly inhibit l-DOPA synthesis in the spleen (Fig. 10A). Second, we compared the difference in the number of Ly6C^hi monocytes recruited to the injured cornea, at 12 hours postinjury, between α-MLT– and vehicle-treated animals, using flow cytometry. The number of Ly6C^hi monocytes decreased from 70.8% in the vehicle-treated animals to 51.3% in the α-MLT–treated animals (P < 0.05) (Fig. 10B). Similarly, the percentage of Ly6C^hi monocyte subpopulations was significantly lower in the peripheral blood of the α-MLT–treated animals than that of the vehicle-treated control animals (Fig. 10C). However, the number of Ly6C^hi monocytes was considerably higher in the spleens of the α-MLT–treated animals than in the spleens of vehicle-treated control animals (P < 0.05) (Fig. 10D). Overall, these data suggested that inhibition of NE synthesis significantly inhibits recruitment of Ly6C^hi monocytes to the injured cornea. 

Figure 10.

 

α-MLT treatment reduces the recruitment of splenic Ly6C^hi monocytes to the injured cornea by inhibiting the synthesis of NE. (A) Western blot analysis showed that the administration of α-MLT by means of IP injection inhibited l-DOPA expression in the spleen. (B) Flow cytometry analysis showed that IP injection of α-MLT decreased the number of Ly6C^hi monocytes in the cornea, at 12 hours after corneal abrasion. (C) Flow cytometry analysis showed that α-MLT treatment reduced the increase in the number of Ly6C^hi monocytes in the peripheral blood of mice after corneal abrasion. (D) Flow cytometry analysis showed that α-MLT treatment prevented the decrease in the number of Ly6C^hi monocytes in the spleens of the corneal-injured mice. Each corneal sample was pooled from 10 corneas, with three independent replicates. Peripheral blood samples were obtained from the combined peripheral blood of two mice, with three independent replicates. Each spleen sample was obtained from one mouse, with six independent replicates. Data are represented as mean ± SD. *P < 0.05, **P < 0.01.

The Recruitment of Ly6C^hi Monocytes to the Injured Cornea Depends on the β2-ARs on the Ly6C^hi Monocytes

To further determine the molecular mechanism by which the SNS regulates the recruitment of splenic Ly6C^hi monocytes to the injured cornea, we determined whether and what ARs are expressed by splenic Ly6C^hi monocytes, using a highly sensitive qPCR technique described in our previous study,¹⁰ the REPLI-gWTA Single Cell Kit for cDNA amplification and qPCR. We first obtained purified Ly6C^hi monocytes from the cornea, spleen, and peripheral blood of normal mice and mice that received corneal abrasion, using flow cytometric sorting. We then examined the transcriptional profiles of the ARs from these different sources of monocytes. Notably, the Ly6C^hi monocytes from the above sources predominantly expressed β2-ARs after corneal abrasion (Figs. 11A–C). 

Figure 11.

 

AR expression on Ly6C^hi monocytes in the injured corneas, peripheral blood, and spleen and the effect of pharmacologic intervention of β2-ARs on recruitment and distribution of Ly6C^hi monocytes following corneal injury. (A–C) Transcriptional profile of ARs on the Ly6C^hi monocytes in the cornea, peripheral blood, and spleen after corneal abrasion. Data are represented as mean ± SD, n = 6. (D) Effect of IP injection of the β2-AR–specific blocker butylamine on the recruitment of Ly6C^hi monocytes to the injured cornea. (E) Effect of IP injection of the β2-AR–specific blocker butylamine on the Ly6C^hi monocytes in the peripheral blood after corneal abrasion. (F) Effect of IP injection of the β2-AR–specific blocker butylamine on the Ly6C^hi monocytes in the spleen after corneal abrasion. (G) Effect of the β2-AR receptor agonist salmeterol on the recruitment of Ly6C^hi monocytes to the injured cornea, upon administration of 6-OHDA. (H) Effect of the β2-AR receptor agonist salmeterol on the recruitment of Ly6C^hi monocytes to the injured cornea, upon administration of α-MLT. Each corneal sample was pooled from 10 corneas, with three independent replicates. Peripheral blood samples were obtained from the combined peripheral blood of two mice, with three independent replicates. Each spleen sample was obtained from one mouse, with six independent replicates. Data are represented as mean ± SD. NS, not significant, *P < 0.05, **P < 0.01.

Next, to verify the importance of β2-ARs in the recruitment of spleen-derived Ly6C^hi monocytes to the injured cornea, we observed the effect of IP injection of the β2-AR–specific blocker, butaxamine, on the recruitment of Ly6C^hi monocytes to the injured cornea. As predicted, the number of Ly6C^hi monocytes significantly decreased in the injured corneas of the butaxamine-treated mice, as compared to those in the vehicle-treated injured corneas (Figs. 11D, 11E). However, the number of Ly6C^hi monocytes in the spleen of animals with corneal injury was significantly higher than that in the vehicle-treated animals (Fig. 11F). Therefore, these results indicated that the β2-AR blocker prevents the recruitment of spleen-derived Ly6C^hi monocytes to the injured cornea. 

Finally, to further validate the importance of β2-ARs in the recruitment of spleen-derived Ly6C^hi monocytes to the injured cornea, we pretreated animals by chemically removing peripheral sympathetic nerve endings using 6-OHDA or inhibiting l-DOPA synthesis with α-MLT, to reduce NE release from sympathetic nerve endings. Sympathetic activation was then simulated by means of administration of the β2-AR agonist salmeterol prior to corneal abrasion. We found that the number of Ly6C^hi monocytes in the cornea was significantly increased in both the 6-OHDA + β2-AR agonist-treated and α-MLT + β2-AR agonist-treated groups, at 12 hours postinjury, as compared to that in the injured mice treated with vehicle alone (Figs. 11G, 11H). In conclusion, these results confirmed that the NE released from sympathetic nerve endings plays a key role in the recruitment of Ly6C^hi monocytes from the spleen to the injured cornea after corneal abrasion. 

Discussion

To the best of our knowledge, this study is the first to show that after corneal abrasion, spleen-derived Ly6C^hi monocytes are recruited to the injured cornea earlier than neutrophils recruited from peripheral blood, and they orchestrate inflammation through the production of proinflammatory cytokines. More important, the mobilization of Ly6C^hi monocytes from the spleen to the injured cornea is primarily controlled by the SNS through β2-ARs expressed on the Ly6C^hi monocytes. These observations suggest that manipulation of Ly6C^hi monocyte recruitment will likely provide a novel target for improving corneal wound repair. Our results are depicted in Figure 12. 

Figure 12.

 

Diagram depicting how sympathetic nerves coordinate recruitment of spleen-derived Ly6C^hi monocytes to the injured cornea and the inflammatory process after corneal abrasion. Classical Ly6C^hi monocytes are released from the spleen after corneal epithelial abrasion, in a sympathetic nerve–dependent manner. After recruitment to the cornea, they produce chemokines and cytokines to promote inflammation. Blocking of sympathetic nerve activities results in the alteration of the wound repair process, which promotes reepithelialization but inhibits neutrophil extravasation.

Early studies have shown that neutrophils are the first-line inflammatory cells recruited to the wound area after corneal injury.⁸ Neutrophils clean the wound area primarily through the release of proteases and phagocytosis, thereby providing a favorable background for subsequent wound closure.^71,72 However, this study found that Ly6C^hi monocytes appeared before neutrophils to the injured corneas. When Ly6C^hi monocytes were predepleted, subsequent neutrophil recruitment was significantly inhibited. Comparative transcriptomic analysis of normal and Ly6C^hi monocyte-depleted corneas revealed that the expression of major inflammatory cytokines was significantly reduced in Ly6C^hi monocyte-depleted corneas. This suggests that monocyte recruitment following corneal injury drives neutrophil recruitment to the injured corneas. This result seems to be consistent with the inflammatory process in pulmonary ischemia–reperfusion injury.²⁶ In a pulmonary injury model, Ly6C^hi monocytes promoted neutrophil recruitment through MyD88-dependent IL-1β production. Similarly, in a model of paracetamol-induced liver injury, monocytes were essential for subsequent neutrophil recruitment.¹⁷ In addition, in a mouse model of inflammatory bowel disease, Ly6C^hi monocytes, together with neutrophils, constitute the major local inflammatory cells. The removal of Ly6C^hi monocytes significantly reduces the local inflammatory response in the intestine.⁷³ In conclusion, these results suggested that Ly6C^hi monocytes recruited early in the injured cornea play a unique role in triggering corneal wound repair, by modulating the inflammatory response. Therefore, further understanding of the origin of Ly6C^hi monocytes and trafficking power is extremely valuable. 

The classical view is that adult monocytes differentiate from precursor cells in the bone marrow and are continuously released into circulation. They then migrate to peripheral tissues in response to homeostatic or inflammatory cues.^14,17 However, several recent studies have found that the spleen contains precursor cells of monocytes and can serve as a reservior for monocytes.²³ In the settings of peripheral tissue infection, ischemic injuries of the heart and lung,²⁵ muscle trauma,⁷⁴ hyperglycemia,⁷⁵ and tumor-induced systemic changes,^76,77 these monocyte precursors can differentiate, expand, and mobilize to the tissues and organs affected. Consistent with these results, we found a positive correlation between the rapid increase in the number of Ly6C^hi monocytes in the peripheral blood and injured corneas after corneal abrasion. However, the number of Ly6C^hi monocytes in the injured cornea and peripheral blood after corneal injury was negatively correlated with the number of Ly6C^hi monocytes in the spleen but not in the bone marrow. This suggests that these monocytes may be of splenic origin. Consistent with this speculation, a decrease in the number of Ly6C^hi monocytes in the peripheral blood and cornea following corneal abrasion was observed when the spleen was preexcised prior to corneal injury. Notably, we demonstrated that Ly6C^hi monocytes in the cornea were predominantly of splenic origin, using in situ immunofluorescent cell tracer labeling of the spleen. Thus, this evidence suggests that spleen-derived Ly6C^hi monocytes are the primary source of corneal Ly6C^hi monocytes after corneal injury. Therefore, it is essential to further explore the mechanisms of spleen-derived Ly6C^hi monocyte recruitment to injured corneas. 

It is well known that all immune organs of the body, including the bone marrow, thymus, and lymph nodes, are directly and indirectly under sympathetic innervation.⁷⁸ In addition, sympathetic nerves play a crucial regulatory role in the inflammatory response to infection and posttraumatic conditions in many tissues and organs.^79,80 Similarly, the spleen is densely innervated by sympathetic nerves and regulates immune cell migration.^40,81,82 Previous studies by our team have shown that sympathetic nerves also play an important regulatory role in coordinating the inflammatory response after corneal injury.^43–45 Therefore, we hypothesized that sympathetic nerves also play an important role in regulating the recruitment of spleen-derived monocytes to injured corneas. 

To test this hypothesis, we first confirmed that the Ly6C^hi monocytes recruited to the injured cornea and peripheral blood after corneal abrasion predominantly expressed β2-ARs, using single-cell transcriptional techniques. Second, the systemic administration of β2-AR blockers significantly inhibited the recruitment of spleen-derived monocytes to the injured cornea and peripheral blood. In contrast, treating injured animals with the β2-AR agonist salmeterol significantly increased the number of Ly6C^hi monocytes in the injured corneas. Further studies found that the spleen expressed the catecholamine (including dopamine, E, and NE)–synthesizing enzyme TH. Therefore, to further verify the importance of adrenergic neurotransmitters, we administered the TH competitive inhibitor, α-MLT, intraperitoneally to mice. The results showed that, similar to the results obtained following β2-AR blockage, α-MLT intervention prevented Ly6C^hi monocyte trapping in the spleen from migrating to the injured cornea. In conclusion, these data suggested that adrenergic neurotransmitters play a key role in the recruitment of spleen-derived monocytes to the injured cornea. 

To further test the “hardwire” aspect of this hypothesis, we first examined the effect of splenic neurectomy on the recruitment of Ly6C^hi monocytes to the injured cornea. As predicted, splenic neurectomy significantly inhibited the recruitment of Ly6C^hi monocytes from the spleen to the injured cornea. However, splenic neurectomy indiscriminately transects all nerves. Therefore, it is difficult to exclude the possibility of modulation by other nerve fibers, such as sensory neural networks. 6-OHDA, a specific toxic hydroxylated derivative of NE, was used to chemically remove peripheral sympathetic terminals. Similar to splenic neurectomy, splenic Ly6C^hi monocytes could not be recruited to the circulation and injured cornea after chemical sympathectomy. In addition, we demonstrated that direct administration of β-AR agonists to chemically sympathectomized or l-DOPA-synthesizing blocked mice significantly improved recruitment of splenic Ly6C^hi monocytes to the injured cornea. Taken together, these data suggest that sympathetic nerves direct the migration and recruitment of Ly6C^hi monocytes from the spleen to the injured cornea after corneal abrasion. 

This study has several limitations. First, other immune cells in the cornea that are regulated by the sympathetic nervous system were not examined in this study.^43–45 Second, the role of nonclassical monocytes in corneal wound repair and their expression of adrenergic receptors remains to be investigated.^83,84 Third, the fate evolution of monocytes during wound repair and their different functions at each stage should be considered in future studies.^{15,19,85–88} Finally, only male mice were used, and data from female mice would provide more sex-specific information. 

In summary, our findings demonstrate a novel regulatory pathway wherein corneal epithelial abrasion activates the SNS, leading to the mobilization of classic Ly6C^hi monocytes from the spleen to the injured cornea via β2-ARs. These infiltrated monocytes trigger neutrophil influx to the injured cornea by producing proinflammatory cytokines, thereby influencing the wound-healing process. Our study provides important insights into the mechanisms of corneal wound repair following injury and suggests that manipulation of spleen-derived monocytes and their mediators could be a promising therapeutic target for efficient intervention in corneal epithelial wound healing. 

Acknowledgments

Supported by the National Natural Science Foundation of China (81700810, SH; 82171014, 81770962, ZL). This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. 

Disclosure: S. He, None; J. Liu, None; Y. Xue, None; T. Fu, None; Z. Li, None