Resolvin D1 attenuates the inflammatory process in mouse model of LPS‐induced keratitis

Abstract The aim of this study was to investigate the effects of the lipid mediator Resolvin D1 in experimental keratitis. C57BL/6J mice were injected with lipopolysaccharide (2 µg/eye), and after 24 hours, the corneal damage was assessed. Clinical score was quantified, and corneal inflammatory biomarkers were detected by immunohistochemistry. A robust accumulation of sub‐epithelial macrophages and polymorphonuclear leucocytes, chemokine (C‐X‐C motif) ligand 1 (also known as keratinocyte‐derived chemokine), interleukin‐10 and promoters of apoptosis was also observed in lipopolysaccharide‐treated mice. Formyl peptide receptor 2 corneal expression was also assessed. The corneal stroma treated with lipopolysaccharide was characterized by presence of macrophages of M1‐like subtype and immature fibroblastic cells, marked with Ki67, not fully differentiated in fibroblasts. Indeed, the staining of the cornea with anti‐vimentin antibodies, a marker of differentiated myofibroblasts, was very faint. Resolvin D1 attenuated all the inflammatory parameters assessed in the present study, except for IL‐10. In conclusion, the data presented here seem to be consistent with the hypothesis that Resolvin D1 protected the cornea from the lipopolysaccharide‐induced keratitis by acting on several inflammatory components of this damage, pivoted by Formyl peptide receptor 2 (FPR2) activation and macrophages‐leucocytes activity.


| INTRODUC TI ON
Keratitis is an ocular inflammatory disease characterized by corneal derangement, due to alterations of structural components such as tight junctions and infiltration of leucocytes into the cornea. These result in stromal keratitis, corneal thinning, and ultimately corneal perforation and scarring. Early neutrophil infiltration into the cornea causes local inflammation that greatly contributes to the host tissue destruction through the release of proteolytic enzymes and inflammatory mediators, ultimately leading to inflammation. 1 This latter accompanied by damage of cellular components including cell membrane, cytoplasmic organelles and the nucleus, as for example the mitochondria and the DNA.

| Induction of keratitis and RvD1 treatment schedule
Keratitis was induced in male C57BL/6J mice (8-10 weeks) by intrastromal injection of 2 μg LPS (Pseudomonas aeruginosa) (Sigma) dissolved in 2 μL of PBS (Sigma) in both the eyes. This LPS dose was previously reported as effective for the induction of keratitis in C57BL/6J mice. 22 The intrastromal injection of RvD1 (Cayman Chemical) was carried out as mentioned before, and based on previous studies. 5,6 This compound was injected 60 minutes after LPS treatment, through the same tunnel created earlier, at the doses of 10-, 100-and 1000 ng/eye/2 µL.
The mice were randomized in the following experimental group

| Clinical score
The development of keratitis was monitored 24 hours after LPS with a biomicroscope by an investigator unaware of treatments. Keratitis was scored as previously described, 1 with some modifications.
Briefly, the score was based upon the following criteria: 0 = normal, clear cornea, no inflammatory reaction; 1 = mild corneal haze with visible iris; 2 = moderate corneal flare with moderate corneal haze and superficial punctate keratitis; 3 = significant corneal opacity, infiltration of cells in the stroma; and 4 = damage and loss of corneal tissue. A score > 1 was assigned to positive keratitis.

| Immunohistochemistry
Following cornea isolation, immunohistochemical analysis were performed as previously described by Rossi et al. 5,6 Briefly, paraffin was removed from paraffin-embedded corneas by using a xylene substitute (Hemo-De; Fisher Scientific); then, ethanol gradient washes were used in order to rehydrated cornea sections. These were quenched sequentially in 3% hydrogen peroxide aqueous solution, then blocked at room temperature for 1 hour with PBS 6% non-fat dry milk (Bio-Rad).
The secondary antibodies used were biotin-conjugated goat anti-rabbit or antimouse IgG and avidin-biotin peroxidase complex (Biotin goat anti-rabbit IgG secondary antibody; dilution 1:1000; ab6720 Abcam and Biotin goat antimouse IgG secondary antibody; dilution 1:1000; ab6788 Abcam). A negative control was without primary antibody (data not shown). The immunohistochemical analysis was performed by an expert pathologist (intra-observer variability 6%) unaware of the procedure. Antigen expression was measured and calculated automatically with the image program LEICA IMM500 and with the statistics program LEICA QWIN. In each experimental group, six distinct preparations were analysed by observing 20 microscopic fields, for a total area of 4.2575e + 005µm 2 for 400× magnification.

| Myeloperoxidase activity
The determination of myeloperoxidase activity was performed as described by D'Amico et al, 23 Briefly, a snap-frozen aliquot of the isolated cornea was homogenized in a buffer containing protease inhibitors (Sigma-Aldrich) and centrifuged for 30 minutes at 4000 g at 4°C. Twenty μL of the supernatant was collected and then added to a solution of tetra-methyl-benzidine (1.6 mmol/L) and 0.1 mmol/L H 2 O 2 . A spectrophotometer set to 620 nm was used to analyse the change in absorbance for each sample.

| Isolation and quantization of cornea protein content
Cornea tissues were homogenized in RIPA lysis buffer (Sigma; R0278) containing a protease inhibitor cocktail (Roche; 11873580001).
Then, they were centrifuged at 13 000 g for 10 minutes at 4°C, to isolate nucleic acids from the protein supernatants. Bio-Rad protein assay protocol (Bio-Rad Laboratories; 500-0006) was used to asses total protein concentration, used for ELISAs.

| FPR2, RvD1, Connexin 43, cytokines, LY6G and p53 quantization by ELISA
Fifty μL of cornea homogenate was used to assess the protein lev- and p53 (Myobiosource, MBS721665). These were determined by using specific ELISA kits according to the manufacturer's instructions.

| Statistical analysis
The results of each experiment were reported as the mean ± s.e.m. of n = 10 mice. One-way ANOVA followed by Bonferroni's test was used to assess statistical significance among the groups with GraphPad Prism 6 software. A probability of P < .05 was considered sufficient to reject the null hypothesis.

| Clinical score
All the mice showed signs of keratitis 24 hours after LPS (2 µg/eye)

| Haematoxylin/eosin stained cornea
Vehicle (PBS only) corneas showed normal epithelial cell and stromal morphologies. Corneas treated with LPS (2 μg/eye) had severe damage and showed a loss of cellularity of the basal and wing epithelial cells, flattening of the stromal lamellae and erosion of the epithelium compared to corneas of vehicle control mice (red arrows, Figure 2A).
Sub-epithelial inflammatory infiltrate, valuable by greater magnification, was observed with associated oedema and erythrocyte extravasation (black arrows, Figure 2A).

| RvD1 affected CXCL1/KC and IL-10 cytokines
As per mirror of the biomolecular changes occurring after leucocytes infiltration, specific ELISA kits revealed a marked increase of both the PMN-produced cytokine CXCL1/KC and the macrophage-produced IL-10 24 hours after the induction of keratitis (Figure 4). CXCL1/KC but not IL-10 was affected by RvD1, indicating a major role for PMN activation rather than for macrophages or a possible shift of this leucocyte line towards the anti-inflammatory phenotype (Figure 4).

| RvD1 reduced the ATM and p53 markers
Vehicle-treated mice did not show any sign of apoptosis within the epithelium, stoma and basal membrane of the cornea ( Figure 5). LPS significantly (+80%, P < .01 vs vehicle) increased the levels of the ATM serine/threonine kinase, sensor of DNA damage, and of the protein p53 (+73%, P < .01 vs vehicle) within the corneas of mice

| Ki67 and vimentin
Ki67 as marker of undifferentiated stromal fibroblasts was evident after 24 hours of LPS compared to vehicle-treated corneas and was reduced by 100 ng/eye RvD1 ( Figure 6). Vimentin staining as marker of differentiated fibroblasts was faint in LPS-corneas ( Figure 6) with no significant difference with respect to vehicle corneas and with the groups treated with RvD1 ( Figure 6).

| Macrophage phenotypes were changed by intrastromal RvD1
Twenty-four hours after the induction of keratitis, the corneas of

| D ISCUSS I ON
In the present study, we showed that RvD1 ameliorated the patho-  [24][25][26][27][28] Once activated by RvD1, this receptor translates positive bioactive activities. FPR2 receptor is constitutively expressed in the cornea, as evidenced in the present study, and it is overexpressed in the cornea following LPS compared to vehicle cornea. In addition, this FPR2 overexpression is accompanied by low corneal endogenous RvD1 levels, data herein described for the first time. The current results reveal that these low levels are insufficient to fully activate the local overexpressed FPR2 receptors as their overexpression resulted in clear association with corneal damage in the LPS mice in our setting. Therefore, a strategic increase in RvD1 bioavailability and proper activation of FPR2 might counteract the corneal damage. In line with this contention, here it is demonstrated that an intrastromal administration of RvD1 increases corneal levels of RvD1 and was linked to reduction of LPS-inflammation of the cornea through real FPR2 activation.
Among the biological activities of FPR2 receptor, there is the capability to limit the trafficking of neutrophils in the inflammation site. 29,30 Here, it is shown that exogenous RvD1, and proper activation of FPR2, promotes the diminution of PMN (neutrophils) presence in the corneal stroma. Although it is not possible to exclude the participation of other cells expressing FPR2 receptor in the action F I G U R E 6 Ki67, marker of undifferentiated fibroblast into the cornea, was strongly reduced by RvD1. Vimentin (Vim), marker of differentiated fibroblasts, was low or absent in both LPS (2 μg/eye) and LPS + RvD1 (100 ng/eye) treated mice. Treatments as above. Graph showing the percentage of the total positive stained area per total area analysed at 40× magnification (scale bar = 100 μm). Values are mean ± s.e.m. of n = 10 observation for each group. *P < .05 and **P < .01 vs vehicle; °°P < .01 vs LPS of RvD1, PMN trafficking seems to be one of the first component affected by RvD1 in the present setting of keratitis.
The corneal damage is mediated by autocrine and paracrine interactions of cytokines, growth factors, chemokines and their receptors produced by epithelial, stromal, bone marrow-derived and neural cells that contribute to the cell damage. 31 We demonstrated an increase of the IL-8/CXCL8 homologue CCXL1/KC, probably generated by PMN leucocytes attracted into the corneal epithelial structures that may trigger the damaging phenomenon. 32 Usually, the first and earliest damaging events start few hours after an inflammatory insult to the cornea [33][34][35] and are followed by corneal cells necrosis mainly characterized by loss membrane integrity and random DNA degradation and is often associated with greater tissue injury and inflammatory response. 36,37 Here, we report that the marker of DNA breaking ATM serine/threonine kinase (ATM) is strongly augmented after LPS injection into the In contrast to this, RvD1 did not significantly affect the levels of the cytokine IL-10. It should be noted that RvD1 may not influence IL-10 levels as much as it would allow a macrophage shift from M1like to M2-like phenotype that would support the resolution of the deleterious migration of neutrophils and degradation factors in the cornea. Indeed, the levels of M2-like macrophages after RvD1 treatment were higher than the M1-like, consistent with the high levels of IL-10 within the cornea.

| CON CLUS IONS
The results seem to be consistent with the hypothesis that RvD1 protected the cornea from the LPS-induced keratitis by acting on several inflammatory components of this damage, pivoted by FPR2 activation and macrophages-leucocytes activity. A mention is due to the appearance of a phenomenon of bilateral ptosis observed in mice upon awakening from the use of RvD1 1000 ng/eye requiring deeper investigation.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest. Writing-original draft (lead).

DATA AVA I L A B I L I T Y S TAT E M E N T
The authors confirm that the data supporting the findings of this study are available within the article.