Inactivation of the miR-183/96/182 Cluster Decreases the Severity of Pseudomonas aeruginosa-Induced Keratitis

Mice

All animals were handled in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The miR-183C^GT/GT mice are on a 129S2/BL6-mixed background¹⁹ and were originally derived from a gene-trap ESC clone.^19,24,25 Male and female, young adult (8-week-old) miR-183C^GT/GT mice, their heterozygous (miR-183C^GT/+) and wild-type (wt) littermates (miR-183C^+/+) were used in this study.

Human Cornea, PMN, and Peripheral Blood Mononuclear Cells (PBMCs)

Human corneas were obtained from the Body Bequest Program, Department of Anatomy and Cell Biology, School of Medicine, Wayne State University. The entire human cornea (anterior to the limbus) was collected from cadaveric eyes within 12 hours postmortem and dissected into V-shaped slices from the periphery to the center of the cornea for RNA preparation. The donors were 79 (male), 41 (male), and 54 (female) years old without prior corneal or other eye diseases. Human PMNs and PBMCs were purchased from Zen-Bio, Inc. (Research Triangle Park, NC, USA).

Nerve Staining

Corneas were excised posterior to the limbal region and flattened by four evenly spaced cuts from the periphery toward the center. Corneas were fixed in 2% paraformaldehyde (PFA) in 0.1M phosphate buffer (PB), pH 7.4 overnight (O/N) at 4°C; then, permeabilized with 0.3% Triton X-100 in PBS for 1 hour at room temperature (RT) and blocked with PBS with 0.3% Triton X-100, 2.5% BSA, and 5% donkey serum (Covance, Battle Creek, MI, USA) for 2 hours. Then, corneas were incubated with rabbit anti-βIII-tubulin antibody (Covance) for 2 days at 4°C. After four rinses with PBS plus 0.3% Triton X-100 and 0.05% Tween-20, corneas were incubated with Alexafluor-488–conjugated donkey anti-rabbit secondary antibody (ThermoFisher, Grand Island, NY, USA) O/N at 4°C. Excess antibody was removed with rinses in PBS. Then, the corneas were flattened onto poly-L-lysine-coated slides and coverslipped using Vectashield mounting media (Vector Labs, Burlingame, CA, USA). Controls were treated similarly with omission of primary antibody or addition of nonspecific IgG.

Corneal Infection and Grading

P. aeruginosa infection of the cornea was performed as described before.⁵⁸ Briefly, mice were anesthetized with ethyl ether in a well-ventilated hood. The cornea of the left eye was wounded; 5.0 × 10⁶ colony forming units (CFU) of PA (strain 19660; ATCC, Manassas, VA, USA) in a 5 μL volume was topically delivered. Corneal disease was graded at 1, 3, and 5 dpi using an established scale⁵⁸: 0, clear or slight opacity, partially or fully covering the pupil; +1, slight opacity, covering the anterior segment; +2, dense opacity, partially or fully covering the pupil; +3, dense opacity, covering the entire anterior segment; and +4, corneal perforation. Photography with a slit lamp was used to illustrate disease.

RNA Preparation and Real-Time RT-PCR

Total RNA was prepared using the miRVana miRNA isolation kit (Life Technologies, Foster City, CA, USA) for miRNA studies, or the RNeasy (Qiagen, Frederick, MD, USA) and RNA STAT-60 kit (Tel-Test, Friendswood, TX, USA) for mRNA studies as described previously.^17,52,60 Quantitative (q)RT-PCR for miRNAs was performed using Taqman miRNA primers and RT-PCR kit (Life Technologies) on a CFX Connect Real-time System (Bio-Rad, Hercules, CA, USA) with snRNA U6 as an endogenous control as described before.^17,60 Quantitative RT-PCR assays of protein-coding genes were performed using QuantiFast SYBR Green RT-PCR kit and QuantiTect primers (Qiagen) with 18s rRNA and β-actin as endogenous controls.^17,19,60

X-Gal Staining

Staining with X-gal was performed as described before.^19,61,62 Briefly, corneas of young adult wt and heterozygous mice (∼8 weeks old) were excised posterior to the limbus. Corneas were fixed in 4% PFA in 0.1M PB, pH 7.4, for 30 minutes on ice, then rinsed in Buffer A (0.1 M PB, pH 7.4, 2 mM MgSO₄, 5 mM EGTA) at RT for 5 minutes and a second rinse for 25 minutes, and then soaked in Buffer B (0.1 M PB, pH 7.4, 2 mM MgSO₄, 0.01% sodium deoxycholate, 0.02% NP-40) at RT twice for 5 minutes. Corneas then were stained in Buffer C (0.1 M PB, pH 7.4, 2 mM MgSO₄, 0.01% sodium deoxycholate, 0.02% NP-40, 5 mM K₃Fe[CN]₆, 5 mM K₄Fe[CN]₆, 1 mg X-gal/mL) at 37°C O/N. After washing with PBS, corneas were observed and photographed under a dissection scope (Accu-scope, Inc., Comack, NY, USA). Then, for whole-mount microscopy, the cornea was flattened by four evenly spaced cuts from the periphery toward the center corneas and mounted on Superfrost plus slides (ThermoFisher); for cross-sections, the cornea was soaked in 20% sucrose in 0.1M PB pH 7.4 O/N and mounted in optimum cutting temperature (OCT; Fisher Healthcare, Houston, TX, USA), 12-μm cryoprotected cross-sections were made and examined using light microscopy (Leica DM4000B; Leica Microsystem, Inc., Buffalo Grove, IL, USA).

For TG X-staining, the whole head of a heterozygous mouse was fixed in 4% PFA in 0.1 M PB, pH 7.4 for 2 hours after the skull was opened. Then the head was sequentially soaked in Buffers A and B, stained in Buffer C and postfixed with 4% PFA as described above. Then, the stained head was carefully dissected under a dissection scope to expose the TG.

Enzyme-Linked Immunosorbent Assay

Protein levels for IL-1β and MIP-2 were tested using ELISA kits (R&D Systems, Minneapolis, MN, USA) as described before.^53,63 Briefly, corneas were homogenized in 1 mL PBS with 0.1% Tween-20 and protease inhibitors (Roche Diagnostics, Indianapolis, IN, USA). An aliquot of each supernatant was assayed in duplicate per the manufacturer's instructions. Sensitivities of the ELISA assays were less than 2.31 pg/mL for IL-1β, and less than 1.5 pg/mL for MIP-2.

Quantitation of Corneal PMN

A myeloperoxidase (MPO) assay was performed as described before.^64–67 Briefly, corneas were harvested, homogenized in 1 mL potassium phosphate buffer (50 mM, pH 6.0) containing 0.5% hexadecyltrimethylammonium bromide (Sigma-Aldrich Corp., St. Louis, MO, USA) using glass microtissue grinders. The samples were freeze-thawed four times and centrifuged at 21,000g for 10 minutes. To measure MPO, a 100-μL supernatant was added to fresh 2.9 mL o-dianisidine dihydrochloride substrate buffer (16.7 mg/mL; Sigma-Aldrich Corp.) with 0.0005% hydrogen peroxide. The change in absorbance at 460 nm was read every 30 seconds for 5 minutes on a Heλios α spectrophotometer (ThermoFisher). Afterward, units of MPO per cornea were calculated: 1 unit of MPO activity is equivalent to approximately 2 × 10⁵ PMN.⁶⁸

Quantitation of Viable Bacteria

Bacteria were quantitated as described before.^64–67 Briefly, each cornea was homogenized in 1.0 mL sterile saline containing 0.25% BSA; 0.1 mL corneal homogenate was serially diluted 1:10 in the same solution and selected dilutions were plated in triplicate on Pseudomonas isolation agar (Becton Dickinson, Sparks, MD, USA). Plates were incubated O/N at 37°C and the number of viable bacteria counted. Results are reported as log₁₀ number of CFU/cornea ± SEM.

In Vitro Transfection and Phagocytosis and Intracellular Killing Assays

A total of 4 × 10⁵ RAW264.7 Mϕ (a cell line derived from the mouse; ATCC) were plated in a 48-well plate and transfected with anti-miR-183/anti-miR-96/anti-miR-182 (10 nM for each) or negative control anti-miR oligonucleotides with scrambled sequences (30 nM; Exiqon, Vedbaek, Denmark) and miRNA mimics of miR-183, miR-96, and miR-182 (10 nM for each) or negative control mimics (30 nM; ThermoFisher) using RNAiMax Lipofectamine (ThermoFisher) as described before.⁶⁹ All experiments were conducted with four replicate wells. Forty-eight hours after transfection, 1 × 10⁷ CFU PA (strain 19660)/well was added. One hour after coculture at 37°C, gentamicin (300 μg/mL) was added to kill extracellular bacteria. Thirty minutes after gentamicin treatment, cells were washed with 250 μL Hanks' balanced salt solution (HBSS) twice to remove gentamicin and dead extracellular bacteria. For the phagocytosis assay, cells were lysed in 250 μL ice-cold sterile water. The cell lysate was plated on PA isolation agar plates (Becton Dickinson) and incubated at 37°C for 24 hours as described before.⁶⁴ Colonies were manually counted; CFU/mL of cell lysate was calculated based on dilution factors. For the intracellular killing assay, fresh media was added to the wells after washing with HBSS to remove gentamicin and dead extracellular bacteria, and then cells were incubated for an additional hour before being washed, lysed, and plated. Killing efficiency was calculated as ([CFU/mL at 1 hour] − [CFU/mL at 2 hour])/(CFU/mL at 1 hour).

Mouse Peritoneal PMN Isolation and Ex Vivo Functional Assay

To harvest peritoneal PMN, mice were injected intraperitoneally with 1 mL 9% casein solution 27 and 3 hours before harvesting; a standard peritoneal PMN isolation protocol was followed as described previously.^64,67–70 A total of 4 × 10⁵ PMNs were plated on a 48-well plate and infected with 1 × 10⁷ CFU PA (strain 19660). Two hours after infection, gentamicin (300 μg/mL) was added to kill extracellular bacteria. For the phagocytosis assay, 30 minutes after gentamicin treatment, cells were washed, lysed, and plated on PA isolation agar plates as described above. For the intracellular killing assay, cells were incubated for an additional 2 hours after washing; then, cells were lysed and plated. Killing efficiency was calculated as ([CFU/mL at 2 hour] − [CFU/mL at 4 hour])/(CFU/mL at 2 hour).

Statistical Analysis

The difference in clinical score between groups was tested by the Mann-Whitney U test. One-way ANOVA with Bonferroni test (GraphPad Prism; La Jolla, CA, USA) was used to analyze qRT-PCR data of proinflammatory neuropeptides and cytokines. An unpaired, two-tailed Student's t-test was used to determine the significance of qRT-PCR of miRNAs, ELISA, MPO, and plate count data^17,19,60; P < 0.05 was considered significant. Each experiment was repeated at least twice to ensure reproducibility and data from a representative experiment are shown. Quantitative data are expressed as the mean ± SEM.

MicroRNA-183/96/182 is Expressed in the Cornea and TG and Is Inactivated in miR-183C^GT/GT Mice

Quantitative RT-PCR analysis showed that all three members of miR-183/96/182 are expressed in the cornea of wt mice (Fig. 1A). Among them, miR-182 and miR-183 showed higher expression than miR-96 (∼32- and 4-fold higher, respectively) (Fig. 1A), suggesting that miR-182 and miR-183 play major roles in the cornea. Consistently, miR-183/96/182 is also expressed in human cornea with a similar pattern as in the mouse (Fig. 1B).

In miR-183C^GT/GT mice, miR-183/96/182 is inactivated in the cornea. In heterozygous mice, expression of both miR-182 and miR-183 were significantly decreased to approximately one-half of that in wt mice, suggesting a dosage effect because heterozygous mice carry only one wt allele of the miR-183/96/182 gene (Fig. 1A).

Because the β-gal cassette in the miR-183C^GT allele follows the endogenous expression pattern of miR-183/96/182,¹⁹ to test its spatial expression pattern, we performed X-gal staining in the cornea of miR-183C^GT/+ mice. Our results showed that miR-183/96/182 is mainly expressed in the corneal epithelium (Figs. 1C–F). A punctate staining pattern in the corneal epithelium (Fig. 1F) suggested that the expression of miR-183/96/182 may be due to staining of corneal nerve endings.

Inactivation of miR-183/96/182 Resulted in Decreased Nerve Density and Decreased Expression of Proinflammatory Neuropeptides in the Cornea

Sensory innervation of the cornea is from the TG.^38–40 To test miR-183/96/182 expression in SN of the TG, we performed X-gal staining on miR-183C^GT/+ mouse brain, and showed that miR-183/96/182 is highly expressed in the TG (Figs. 2A, 2B). Quantitative RT-PCR further confirmed its expression in TG (Fig. 2C). Similar to the cornea, miR-182 and miR-183 are predominately expressed in the TG, compared with miR-96 (∼16- and 8-fold higher, respectively) (Fig. 2C). MicroRNA-183/96/182 is inactivated in the TG of miR-183C^GT/GT mice. In the TG of heterozygous miR-183C^GT/+ mice, miR-183 is significantly decreased to approximately one-half of that in wt mice; miR-182 expression is also reduced, although did not reach significance (Fig. 2C). Similarly as in the cornea, these data suggest that miR-183/96/182 expression in TG also has a dosage effect based on numbers of wt alleles (Fig. 2C).

To test whether inactivation of miR-183/96/182 has significant impact on corneal innervation, we performed immunofluorescence analysis using an anti-β-III tubulin antibody, a marker for neurons. Our results showed that, in ko animals, corneal nerve density is drastically decreased; the normal whorl-like pattern of the subbasal nerve plexus^39,71–77 is disrupted (Fig. 2D). Local axon enlargements or “beads,” which contain high densities of mitochondria and share similarities to nociceptor terminals,^38,39 are decreased in ko compared with wt cornea (Fig. 2D), suggesting that inactivation of miR-183/96/182 has significant consequences on corneal sensory innervation. Consistent with this hypothesis, qRT-PCR using corneal RNA of ko and wt mice showed decreased expression of nociceptor transient receptor potential vanilloid 1 (TRPV1) and SN-released proinflammatory neuropeptides Tac1, the precursor gene for substance P (sP), calcitonin gene-related peptide (CGRP), and chemokine (C-X3-C motif) ligand 1 (Cx3cl1) (Figs. 2E–H).

Inactivation of miR-183/96/182 Resulted in Significant Decrease of Corneal Inflammatory Response to PA Infection and Severity of PA-Induced Keratitis

Because inactivation of miR-183/96/182 resulted in decreased nerve density and proinflammatory neuropeptides, we hypothesized that this disrupts neuroimmune cross-talk and modulates immune/inflammatory responses in the cornea. To test this hypothesis, the cornea of ko mice and their wt littermates were infected with PA. Our results showed that miR-183/96/182 ko mice had significantly decreased keratitis at 1, 3, and 5 dpi (Fig. 3A). Most of the ko mice had only slight corneal opacity; none of the corneas perforated (Figs. 3A, 3B). Quantitative RT-PCR analysis of the cornea at 5 dpi showed that many proinflammatory factors and their receptors, which have been shown to play important roles in PA-induced keratitis, were decreased in the infected cornea of ko mice (Figs. 3C–J). In wt animals, pattern-recognition receptor Toll-like receptor 4 (TLR4), IL-1β, IL-6, IL-12p40, chemokine (C-C motif) ligand 2 (CCL2) and its receptor CCR2, and intercellular adhesion molecule 1 (ICAM1) and vascular cell adhesion molecule 1 (VCAM1) were increased in the infected eyes compared with noninfected contralateral eyes (Figs. 3C–J); whereas in ko mice, increased expression of these proinflammatory factors in infected eyes was mitigated (Figs. 3C–J). Expression levels of most of these inflammatory factors were decreased in the infected eyes of ko mice when compared with their wt littermates (Figs. 3C–J). Enzyme-linked immunosorbent assays demonstrated that, at the protein level, IL-1β was significantly decreased at 1 dpi, whereas macrophage inflammatory protein-2 (MIP-2) was decreased at both 1 and 5 dpi (Figs. 4A, 4B). These results suggest that inactivation of miR-183/96/182 resulted in a decreased inflammatory response in infected corneas and contributed to decreased severity of PA-induced keratitis.

Polymorphonuclear neutrophils are the predominant infiltrating immune cells in bacterial keratitis and play important roles in bacterial clearance.^53–57 To begin to understand the pathological processes involved, we first performed MPO assays to quantify PMN in the cornea. Our results showed that PMN in the infected cornea of ko animals were significantly decreased at 1 dpi and slightly more decreased at 5 dpi. This result is consistent with the observation of decreased expression of chemoattractant neuropeptides (e.g., sP) (Fig. 2E) and cytokines (e.g., CCL2 and MIP-2), and leukocyte adhesion molecules, ICAM1 and VCAM1 (Figs. 3G, 3I, 3J, 4B).

To gain further insight into the progression of pathological processes, we quantified viable bacteria in the infected cornea. Intriguingly, our results showed that bacterial load was increased in the infected cornea of ko compared with wt mice at 1 dpi. This observation is consistent with decreased PMN infiltration in ko animals at this time. However, at 5 dpi, bacterial load in the corneas of wt controls was further increased when compared with 1 dpi; whereas the bacterial load in ko animals was decreased at 5 dpi compared with 1 dpi, and was significantly lower than wt controls. These results suggest that, although they had a higher bacterial burden at 1 dpi, bacteria continued to be cleared more efficiently in ko mice than in wt controls.

MicroRNA-183/96/182 is Expressed in Mϕ and PMN and Plays an Important Role in Modulation of Innate Immune Cell Function

Quantitative RT-PCR results showed that miR-183/96/182 is expressed in both PMNs and Mϕ of wt mice, while inactivated in miR-183C^GT/GT animals (Figs. 5A, 5B; Table). In Mϕ, miR-183 and miR-182 are predominantly expressed, whereas miR-96 is expressed at a significantly lower level, approximately one-sixth of the level of miR-183 and miR-182 (Fig. 5A). In PMN, only miR-182 and miR-183 are expressed, with miR-182 predominant (∼3.3-fold over miR-183); whereas miR-96 is barely detectable (Fig. 5B). Similarly, miR-183/96/182 is also expressed in human PMNs and PBMCs (Figs. 5C, 5D; Table), suggesting that miR-183/96/182 modulates the functions of PMNs and Mϕ in both mice and humans.

In vitro phagocytosis and intracellular killing assay using a mouse Mϕ cell line, RAW264.7, showed that simultaneous overexpression of miR-182, -183, and -96 in Mϕ significantly decreased their phagocytosis of PA by approximately 55.1% (Fig. 6A) and abolished their killing capacity (Fig. 6B); whereas knockdown of miR-182, -183, and -96 resulted in a significant increase of phagocytosis (by ∼67.3%) (Fig. 6C) and intracellular killing efficiency (4.2%–18.8%) (Fig. 6D). To test the role of miR-183/96/182 in PMNs, we isolated peritoneal PMNs from ko and wt mice for phagocytosis and intracellular killing assays. Our results showed that, consistent with the knockdown experiment in Mϕ, inactivation of miR-183/96/182 in PMNs resulted in a significant increased phagocytosis capacity (∼66.5%) (Fig. 6E) and intracellular killing efficiency (∼37%; range, 41.0%–78.0%) (Fig. 6F).