Characterization of Recruited Mononuclear Phagocytes following Corneal Chemical Injury

Mononuclear phagocytes (MP) have central importance in innate immunity, inflammation, and fibrosis. Recruited MPs, such as macrophages, are plastic cells and can switch from an inflammatory to a restorative phenotype during the healing process. However, the role of the MPs in corneal wound healing is not completely understood. The purpose of this study is to characterize the kinetics of recruited MPs and evaluate the role of macrophage metalloelastase (MMP12) in the healing process, using an in vivo corneal chemical injury model. Unwounded and wounded corneas of wild-type (WT) and Mmp12−/− mice were collected at 1, 3, and 6 days after chemical injury and processed for flow cytometry analysis. Corneal MP phenotype significantly changed over time with recruited Ly6Chigh (proinflammatory) cells being most abundant at 1 day post-injury. Ly6Cint cells were highly expressed at 3 days post-injury and Ly6Cneg (patrolling) cells became the predominant cell type at 6 days post-injury. CD11c+ dendritic cells were abundant in corneas from Mmp12−/− mice at 6 days post-injury. These findings show the temporal phenotypic plasticity of recruited MPs and provide valuable insight into the role of the MPs in the corneal repair response, which may help guide the future development of MP-targeted therapies.


Introduction
The mononuclear phagocyte (MP) system, also called the macrophage system, has central importance in the immune defense of organisms and is comprised of monocytes, macrophages, and dendritic cells [1,2]. Monocytes are the main mononuclear phagocytes in peripheral blood whereas macrophages and dendritic cells are mainly located in tissues. Inflammation is a normal defense response induced by infection or injury [3] and in some cases, it may also lead to tissue fibrosis. The cornea is the transparent tissue in the front part of the eye. The development of corneal fibrosis is associated with tissue opacification, which often leads to visual impairment and blindness. Dysregulated macrophage recruitment and clearance have been shown to be important factors in the development of fibrosis following tissue injury [4][5][6].
Macrophages are crucial for tissue repair and regeneration. In vivo populations of macrophages do not always fit into the traditional classifications of M1 (classically activated macrophages) and M2 (alternatively activated macrophages) defined by in vitro systems [3,7]. Recent studies have shown that macrophages are plastic cells whose phenotype and function can vary depending on the time after injury and that inflammatory monocytes can be converted in situ into macrophages exhibiting a reparative phenotype [8][9][10][11][12]. Thus, 2 of 13 the selective reprogramming of macrophages may be a promising therapeutic approach for preventing and treating tissue fibrosis [13].
Differential expression of the cell surface marker Ly6C has been used to identify functionally discrete monocyte/macrophage populations in various tissues. Ly6C + macrophages are thought to be derived from the peripheral circulation and have a pro-inflammatory function [14]. Ly6C − macrophages, on the other hand, are generally regarded as tissue-resident macrophages that play an important role in maintaining immune homeostasis and tissue regeneration [12,14]. Interestingly, in a model of liver injury and regeneration, a Ly6C lo restorative macrophage subset was identified and showed an increased expression of matrix metalloproteinases (MMPs), in particular macrophage metalloelastase (MMP12) [12]. We have previously shown that MMP12 similarly promotes tissue repair in a corneal chemical injury model where MMP12 expression decreased the expression of chemokine CCL2 and reduced the accumulation of macrophages into wounded corneas [15,16]. There are no prior studies assessing MP Ly6C expression over time using an epithelial/stromal chemical injury model of corneal healing. The purpose of this study is to characterize the phenotype and dynamics of recruited MPs in a corneal chemical injury model and evaluate the role of MMP12 in this process. We hypothesized that MP Ly6C expression after corneal injury is highly dynamic and helps to direct the immune and tissue repair response.

Results
Based on the level of F4/80 expression, three distinct leukocyte (CD45 + CD11b + ) subsets were observed in control (uninjured) and injured corneas from WT and Mmp12 −/− mice. A CD11b + F4/80 int population represented a recruited mononuclear phagocyte population (R1), and a CD11b + F4/80 high population represented a resident mononuclear phagocyte population (R2) (Figure 1). The CD45 + CD11b + F4/80 − subset (O) represents a population of other immune cell types that are not part of the MP system, and presumably comprise granulocytes, lymphocytes, and mast cells.

Figure 1.
Representative gating strategy of flow cytometry used to identify recruited and resident leukocyte populations and kinetics in uninjured and injured corneas after corneal chemical injury. Corneas from wild type (WT) and Mmp12 −/− (KO) mice at 8-12 weeks old, were collected at 1, 3, and 6 days after an alkaline burn was created by applying filter paper soaked in 0.1 N NaOH. Four corneas from 2 mice were pooled per sample, cut into pieces, and digested with collagenase type I for 1.5 h at 37 °C. Cells were washed with FACS buffer and then incubated for 30 min in a FACS buffer solution containing CD16/32 Fc block. Cells were then washed with FACS buffer and stained for 30 min with the following antibodies: CD45-FITC, CD11B-Pacific Blue, F4/80-APC, Ly-6C-PerCP/Cy5, CD64-PE-Dazzle, CD11c-BrilliantViolet 605, Zombie-Acqua Viability Dye. Cells were washed and then fixed with 2% PFA in PBS overnight (at 4 °C). Cell profiles were acquired on an LSRII flow cytometer (Becton Dickinson, Franklin Lakes, NJ, USA), and were gated based on fluorescence Corneas from wild type (WT) and Mmp12 −/− (KO) mice at 8-12 weeks old, were collected at 1, 3, and 6 days after an alkaline burn was created by applying filter paper soaked in 0.1 N NaOH. Four corneas from 2 mice were pooled per sample, cut into pieces, and digested with collagenase type I for 1.5 h at 37 • C. Cells were washed with FACS buffer and then incubated for 30 min in a FACS buffer solution containing CD16/32 Fc block. Cells were then washed with FACS buffer and stained for 30 min with the following antibodies: CD45-FITC, CD11B-Pacific Blue, F4/80-APC, Ly-6C-PerCP/Cy5, CD64-PE-Dazzle, CD11c-BrilliantViolet 605, Zombie-Acqua Viability Dye. Cells were washed and then fixed with 2% PFA in PBS overnight (at 4 • C). Cell profiles were acquired on an LSRII flow cytometer (Becton Dickinson, Franklin Lakes, NJ, USA), and were gated based on fluorescence minus one control. The gating strategy is illustrated: Cells → Live → Singlets → CD45 high leukocytes (L). Fifty-six corneas were tested in each group (WT and KO); controls (n = 3 samples; 12 corneas); day 1 (n = 4 samples; 16 corneas); day 3 (n = 4 samples, 16 corneas); day 6 (n = 3 samples; 12 corneas). CD11b + F4/80 high resident mononuclear phagocytes (R2) can be seen in the uninjured samples. CD11b + F4/80 intermediate recruited mononuclear phagocytes (R1) and other CD11b + F4/80 − leukocytes (O) are seen mainly in day 1 after the chemical injury.
There was also a significant difference over time in the O subset (two-way ANOVA, p-value < 0.0001; WT uninjured vs. D1 adjusted p-value = 0.0012; KO uninjured vs. D1 adjusted p-value = 0.0019; Figure 2C).
The differences between WT and Mmp12 −/− were not statistically significant for R1, R2 or O subsets (p > 0.05).

Discussion
In this study, we used flow cytometry to characterize the kinetics of corneal leukocytes, the impact of MMP12 gene knockout, and the phenotype of recruited MPs following corneal chemical injury. The MP system is the main first-line immune defense against pathogens and we found that this population of recruited cells responded acutely following corneal chemical injury, being detected at the highest levels on the first day post-injury and then decreasing steadily over the week following injury. Within the recruited MP population, cells expressing high levels of Ly6C were predominant in the early phase after injury (day 1) while cells expressing intermediate levels of Ly6C increased in frequency on day 3 and cells with low expression of Ly6C were the predominant cell type at 3 and 6 days after injury. These findings show the very dynamic changes in MP Ly6C expression acutely after corneal injury.
MP cells, in particular macrophages, have a key role in tissue repair [1,2]. Many studies have shown that inflammatory monocytes can be converted in situ into cells exhibiting a reparative phenotype [8,10,12]. Our findings show that following corneal chemical injury, Ly6C int restorative MP (CD11B hi F4/80 int ) are mostly highly expressed on day 3 post-injury. Differentiation of corneal MPs based on surface markers is not trivial, and although many studies have considered that all CD11b + F4/80 + represent macrophages, recent studies have suggested that CD64 surface marker can help differentiate corneal macrophages from monocytes [18]. Our subsequent analysis of the Ly6C expression gating for CD11b + F4/80 int CD11c − CD64 + recruited macrophages showed a similar frequency pattern with a gradual shift in frequency from Ly6C high to Ly6C int on day 3 and Ly6C neg on day 6.
Restorative Ly6C lo macrophages expressed during tissue repair and regeneration have been shown to highly express MMP12 [12,19]. Because we have previously shown MMP12 to promote corneal repair [15,16,20], we expanded our study to investigate the kinetics of MP cells not only in WT mice, but also in MMP12 knockout mice. MMP12 is a member of the matrix metalloproteinase (MMP) family and is secreted mainly by macrophages. MMP12 has been shown to be a key factor in regulating the development of fibrosis in several models of tissue injury, including lung and liver injuries [15,[21][22][23][24]. In a previous study, we observed an increased macrophage infiltration in response to chemical injury in the central corneas of the Mmp12 −/− mice compared to WT mice [15]. In the current study, we used flow cytometry to characterize the MP cells over the whole cornea (not only the central part), and despite initially observing a trend, the frequencies of MP subsets based on Ly6C expression did not reach statistical significance (p > 0.05) between MMP12 knockout and WT mice. Results from these two studies thus suggest MMP expression by monocytes as an important factor for their migration through the tissue matrix. In particular, MMP12 expression appears to inhibit monocyte migration through the cornea stroma. We previously showed that compared to WT mice, myeloid cells from Mmp12 −/− mice are less dynamic, show a decreased velocity, and have only a small increase in track lengths and displacement after corneal injury [15,16]. Because MP cells are not evenly distributed over the cornea [25,26], the role of different MP populations in corneal wound healing may depend on how far and where these cells migrate into the cornea.
The corneas from MMP12 knockout mice presented a statistically significant reduction in the frequency of CD64 + macrophages on day 6 compared to days 1 and 3 (p = 0.0196; p = 0.0438) and a statistically significant (p = 0.0061) increase in the frequency of dendritic cells on day 6 post chemical injury compared to day 1, while the trend of progressive increase in dendritic cells observed in wild-type corneas did not reach statistical significance (p = 0.1982). Dendritic cells are antigen-presenting cells considered to be part of the MP system. This result may be seen as complementary to the observation of a progressive increase in the frequency of dendritic cells, reaching statistical significance after 7 days, in a model of dry eye and ocular surface inflammation [27].
Using a model of corneal epithelial and anterior stroma scraping, Sahu et al. investigated the kinetics of inflammatory cell recruitment over the initial 12 h after injury [28]. They found that the frequency of mast cells progressively increased over the initial 6 h, before declining to baseline at 12 h after injury, while neutrophil infiltration of the cornea progressively increased during the 12 h follow-up after injury. In our study, we focused on later time points and on the kinetics of the mononuclear phagocytes, but we did notice the presence of a population of other immune cells (CD45 + CD11b + F4/80 − ) that includes neutrophils, that was not present in the uninjured corneas but was elevated at days 1 and 3 after injury. These findings complement the observations reported by Sahu et al. and may suggest that the frequency of neutrophils remains elevated for at least 3 days but not until day 6 post-corneal injury (Figure 2). Neutrophils are involved in the early phase of the immune response and can be mediators of collateral tissue damage [29].
The early immune response of the cornea has been also studied using flow cytometry by Liu et al. [18]. After producing a 2 mm corneal epithelial injury, they followed the mouse wounds for up to 36 h. Unlike the isolated epithelial injury model, the chemical corneal injury model used in this study is a well-established model for studying corneal fibrosis [30], a problem that can lead to visual impairment and blindness. The NaOH corneal injury model affects the epithelial and stromal layers and induces a robust increase in MMP expression, inflammatory cytokine release, and intense corneal immune response [31]. Using this model, we previously found that F4/80 + cells are heavily recruited during the first week after injury and that the loss of MMP12 resulted in a higher number of recruited F4/80 + cells at 6 days post-injury and increased corneal fibrosis at 3 weeks postinjury [15]. This study provides a more detailed analysis of the MP subsets in the acute response to chemical injury. The Ly6C high (proinflammatory) subset was predominant on day 1 post-injury for both WT and Mmp12 −/− mice, and trended higher in Mmp12 −/− mice compared with WT mice on days 1 and 6 post-injury ( Figure 3). Additionally, a significant increase in CD11c + dendritic cells was observed on day 6 post-injury in the Mmp12 −/− mice. This higher expression of Ly6C high cells and CD11c + dendritic cells during the first 6 days after injury in Mmp12 −/− mice compared with WT mice may contribute to the higher levels of corneal fibrosis we previously observed in the Mmp12 −/− mice compared with WT mice at 3 weeks post-injury [15]. Interestingly, in a mouse model of cardiac fibrosis, the depletion of bone marrow-derived CD11c + cells significantly reduced ventricular fibrosis [32] and suggests that therapies targeting these cells may have the potential for preventing tissue fibrosis.
Leukocytes that are not part of the MP system were not the focus of this study; therefore, specific markers for cells such as neutrophils, natural killers, and mast cells were not added to our antibody panel.
Our study builds on previous work by using well-known cell markers to characterize the phenotype and dynamic of MP cells over a period of 6 days after corneal injury. Our results show an early increase in the frequency of monocytes and Ly6C high cells, followed later by an increase in Ly6C int cells, Ly6C neg cells, and CD11c + dendritic cells. These findings provide an insight into the kinetics and functional roles of inflammatory and restorative MPs following corneal injury. Further analyses of these distinct MP populations using lineage tracing systems and single-cell transcriptomics may guide the development of MP-targeted therapies.

Animals
Breeding colonies of mice homozygous for the null allele of the Mmp12 [33] and wild-type mice on an FVB/n background were maintained under pathogen-free conditions in the UCSF barrier facility. Experiments were performed with littermate 8-to 12-weekold male and female mice. All animal protocols were approved by the UCSF Institutional Animal Care and Use Committee and all procedures were performed in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research.

Corneal Chemical Injury
Corneal alkali burn injuries were performed on mice as previously described [34]. Briefly, mice were anesthetized by isoflurane inhalation (Baxter Pharmaceutical, Deerfield, IL) and by topical application of 0.5% Proparacaine (Akorn Inc., Buffalo Grove, IL, USA) placed on the cornea before and after the chemical burn. An alkaline burn was created by applying filter paper 2.5 mm in diameter soaked in 0.1 N NaOH (Sigma, St. Louis, MO, USA) for 30 s to the central cornea followed by rinsing with 1 mL of phosphate-buffered saline (PBS).

Flow Cytometry Analysis
Unwounded (control) and wounded whole globes were collected at 1, 3, and 6 days after injury. Corneas were removed and 4 corneas from 2 mice were combined into a single sample. The pooled corneas were cut into small pieces and digested with collagenase Post-acquisition analysis was performed using FCS Express 7 software (De Novo Software, Glendale, CA, USA), gating was based on unstained and fluorescence minus one controls (Supplementary Figure S1).

Statistical Analysis
The average percentages of each cell population were obtained using FCS Express 7 software. All data are expressed as the mean ± SEM. The effects of time and gene knockout on the cell population's phenotype were evaluated using two-way ANOVA followed by Tukey post hoc test for multiple comparisons. Statistical comparisons were made using Prism 8.3 statistical software (GraphPad, San Diego, CA, USA). An alpha level of P less than or equal to 0.05 was chosen as the criterion of significance.

Data Availability Statement:
The data presented in this study are available from the corresponding author upon reasonable request.