Published online before print October 6, 2006
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,
,1,
,
* Institute for Eye Research, Sydney, New South Wales Australia;
School of Optometry, University of New South Wales, Sydney, New South Wales, Australia;
Vision CRC, Sydney, New South Wales, Australia; and
Faculty of Veterinary Science, University of Sydney, New South Wales, Australia
1Correspondence: Institute for Eye Research, University of New South Wales, Sydney, NSW 2052, Australia. E-mail: e.hume{at}ier.org.au
ABSTRACT
Pseudomonas is one of the leading causes of contact lens-related microbial keratitis. Despite the use of antibiotics, the host inflammatory response continues to cause damage to the cornea, which may lead to blindness. CXCR2-binding chemokines have been implicated in the pathogenesis of Pseudomonas keratitis, and the exact role of this receptor remains to be elucidated. Corneas of CXCR2 knockout and wild-type mice (Cmkar 2–/– and Cmkar 2+/+) were scratched, and 2 x 106 CFU/mL Pseudomonas 6294 or 6206 was added to corneas. Twenty-four hours postinfection, mice were killed, and eyes were harvested for enumeration of bacteria, myeloperoxidase (MPO) levels, and inflammatory mediators. Cmkar 2–/– had 20- to 100-fold more bacteria than Cmkar 2+/+ mice. There were no differences in MPO levels between gene knockout and Cmkar 2+/+ mice. Histology revealed PMN were restricted to the limbal area. Levels of CXCR2 chemokines (keratinocyte-derived chemokine and MIP-2) were elevated significantly in gene knockout mice. A lack of CXCR2 leads to an inability to control bacterial numbers as a result of the inability of PMN to reach the site of infection in the avascular cornea. These results imply that CXCR2 is critical to the extravasation of neutrophils into the avascular cornea.
Key Words: cornea • gene knockout mice • microbial keratitis • CXCR2
INTRODUCTION
Pseudomonas aeruginosa accounts for 37% of all culture-proven cases of contact lens-related corneal infection but in some centers, may cause up to 60% of all cases [1 ]. Despite the use of antibiotics, the host inflammatory response continues to cause damage to the cornea, which may lead to vision loss or blindness. The CXC chemokines, MIP-2 and keratinocyte-derived chemokine (KC), have been implicated in the pathogenesis of this disease [2 , 3 ]. Although the timely recruitment of leukocytes into the cornea is necessary for the control of the bacterial infection, it has been found that down-regulation of MIP-2 results in improved corneal outcome [3 ]. This indicates that regulation of this pathway is pivotal to the host outcome. Lack of CXCR2 has been demonstrated to result in the restriction of neutrophil recruitment to the cornea during helminth corneal infection [4 ]. It is well known that PMN play a central role in the outcome of Pseudomonas infection of the cornea, and a greater understanding of the mechanisms controlling neutrophil recruitment is imperative.
CXCR2–/– mice (Cmkar 2–/–) and their wild-type control (Cmkar 2+/+) on a BALB/c background (Jackson Laboratories, Bar Harbor, ME) were infected with bacterial strains (6206 or 6294; n=22 per bacterial and mouse strain) with cytotoxic (6206) or invasive (6294) characteristics as described previously [5
]. One eye was scratched and infected, and the contralateral eye served as a noninfected scratch control. The animals were monitored during each experiment, and the Animal Care and Ethics Committees of the Universities of Sydney and New South Wales (Australia) approved all protocols for animal use. All experiments were repeated on a minimum of two occasions. Mice were clinically examined as described previously [6
]. Data were analyzed using a Kruskal-Wallis one-way ANOVA. Mice were killed at 1 day postchallenge. For histological examination, 5 µm sections of eyes were stained with Whitlock’s H&E. Homogenates were serially diluted and plated on nutrient agar to determine numbers of bacteria [7
]. Myeloperoxidase (MPO) levels, representing PMN numbers, were determined using an enzymatic assay using the substrate o-dianisidine dihydrochloride as described previously [7
]. Corneas for ELISA were homogenized and frozen immediately at –70°C. KC and MIP-2 ELISA (R&D Systems, Minneapolis, MN) and leukotriene B4 (LTB4) and IL-1ß (Becton Dickinson, San Jose, CA) were used according to the manufacturers’ instructions. Infected and control eyes (n=6) were collected at 1 day postchallenge from Cmkar 2+/+ and Cmkar 2–/– mice, and RNA was isolated and amplified as described previously [7
]. Changes in gene expression for the following inflammatory mediators were examined using semiquantitative RT-PCR as described previously [7
]: MIP-1
, ICAM-1, membrane-activated complex 1 (MAC-1), XCR1, lymphotactin, MCP-1, thymus and activation-regulated chemokine, IFN-inducible protein 10 (IP-10), T cell activation protein-3, eotaxin, MIG, RANTES, and G3PDH (housekeeping gene).
At 24 h postinfection, the Cmkar 2–/– mice infected with P. aeruginosa strain 6206 or 6294 showed edema extending to the periphery of the cornea and dense cellular infiltration at the limbus and periphery of the cornea, and pronounced hypopyon was also observed in both groups. Mice in both groups also showed extensive epithelial loss extending almost to the periphery of the cornea when viewed after instillation of fluorescein. Clinical scores were significantly more severe in Cmkar 2–/– mice compared with Cmkar 2+/+ mice for both strains of bacteria (Fig. 1A , P<0.05). However, the mice infected with 6294 again showed a more severe response in some variables measured (chemosis, hyperaemia, hypopyon, and edema) compared with the mice infected with strain 6206 (Fig. 1A) . These observations corresponded with our previous findings for corneal infection with these strains of bacteria (ref. [5 ] and Fig. 1A ). The progress of corneal infection at time-points beyond 24 h was not able to be made in the Cmkar 2–/– mice, as the infection was no longer localized to the cornea and was systemic.
 View larger version (12K): [in a new window] |
Figure 1. (A) Clinical scores of Cmkar 2–/– and Cmkar 2+/+ mice corneas infected with P. aeruginosa strain 6206 or 6294 24 h postchallenge. SC, Scratch control eyes. (B) Bacterial numbers in corneas of Cmkar 2–/– and Cmkar 2+/+ mice 24 h postchallenge with P. aeruginosa strain 6206 or 6294. The Cmkar 2–/– corneas had significantly higher bacterial loads compared with the Cmkar 2+/+ corneas (*, P<0.05).
|
 View larger version (80K): [in a new window] |
Figure 2. Histological examination of mouse corneas infected with P. aeruginosa 24 h postchallenge. Histological sections are stained with H&E. (A) Cmkar 2+/+ mouse challenged with P. aeruginosa strain 6206. (B) Cmkar 2–/– mouse challenged with P. aeruginosa strain 6206. (C) Cmkar 2+/+ mouse challenged with P. aeruginosa strain 6294. (D) Cmkar 2–/– mouse challenged with P. aeruginosa strain 6294. (E) Uninfected cornea of a Cmkar 2+/+ mouse. Arrows indicate corneal structures. S, Stroma; Epi, corneal epithelium; Endo, corneal endothelium. (F) Uninfected cornea of a Cmkar 2–/– mouse. (G) Cmkar 2–/– mouse challenged with P. aeruginosa strain 6294 showing the central cornea at high magnification. Arrows indicate bacteria, denoted by B, visible in the stroma, which is devoid of epithelium. Inset shows individual bacteria in the stroma stained with toluidine blue. (H) Cmkar 2–/– mouse challenged with P. aeruginosa strain 6206 showing the central cornea at high magnification. The stroma is devoid of epithelium, and neutrophils, denoted by N, can be seen in the anterior chamber (indicated by the arrow) but not in the stroma.
|
Estimation of numbers of neutrophils (PMN) in the whole eye using a MPO assay showed no significant difference in PMN recruitment to the eye between types of mice or infecting strain of bacteria. At 1 day postchallenge, estimates of PMN numbers in Cmkar 2–/– mice were 5 x 107 ± 6 x 106/eye in response to infection with strain 6206 and 5.0 x 107 ± 1 x 106/eye in response to infection with strain 6294 and in Cmkar 2+/+ mice, 3 x 107 ± 6 x 106/eye in response to infection with strain 6206 and 5 x 107 ± 8 x 106/eye in response to infection with strain 6294. Numbers of neutrophils in the scratch control eyes were below the limits of detection of this assay.
Levels of the inflammatory mediators IL-1ß, MIP-2, KC, IL-6, and LTB4 in the corneas of wild-type and Cmkar 2–/– mice were determined by ELISA, and the results are summarized in Table 1
. Levels of these inflammatory mediators were significantly higher in the Cmkar 2–/– mice compared with the Cmkar 2+/+ mice in response to both strains of infecting bacteria. MIP-2 was increased in the Cmkar 2–/– by 
1.5-fold compared with the Cmkar 2+/+ mice, and KC was increased in the Cmkar 2–/– by approximately 6-fold compared with the Cmkar 2+/+ mice. IFN-
levels were also determined by ELISA and were not different between strains of mice or strains of bacteria.
|
View this table: [in a new window] |
Table 1. Inflammatory Mediator Levels in the Corneas of Cmkar 2+/+ and Cmkar 2–/– Mice 24 h Postchallenge with P. aeruginosa Strains 6206 (cytotoxic) or 6294 (Invasive)
|
|
View this table: [in a new window] |
Table 2. Levels of mRNA Relative to G3PDH for Inflammatory Mediators in the Corneas of Cmkar 2+/+ and Cmkar 2–/– Mice 24 h Postchallenge with P. aeruginosa Strains 6206 (Cytotoxic) or 6294 (Invasive)
|
50% higher in response to strain 6294 compared with strain 6206 in the knockout mice. Levels of MAC-1 mRNA were down-regulated in Cmkar 2–/– mice compared with Cmkar 2+/+ mice. This study demonstrates that Cmkar 2–/– mice show a markedly more severe corneal response to infection with P. aeruginosa than Cmkar 2+/+ mice. At 24 h postinfection, neutrophils had failed to penetrate corneal tissue, despite the fact that the same number of neutrophils was present in the whole eye of wild-type and knockout mice. Histologically, the neutrophils were observed to be confined to the limbal area of the cornea and anterior chamber and appeared to be unable to cross the corneal endothelium and epithelium or penetrate the avascular stroma. These findings are consistent with other models of infection, where in the absence of CXCR2, neutrophils have been confined to the blood vessels [8 9 10 ]. Furthermore, when CXCR2 is absent, there is restriction of neutrophils but not other infiltrating leukocytes into the cornea during Onchocerca volvulus infection [4 ].
Our findings indicate that concomitant with the lack of neutrophil recruitment into the cornea is an increase in bacterial numbers. This result would be expected, as the infection proceeds unchecked in the absence of phagocytes and is consistent with findings during Staphylococcus aureus infection of the CXCR2–/– mouse brain [8 ]. Consistent with previous findings, infection with strain 6294 resulted in a more severe clinical appearance than infection with strain 6206 [3 ], although there was no significant difference in bacterial numbers or PMN in this model. However, the two strains have been reported to result in differing immunological responses [3 , 7 , 11 ]. These differing responses were observed in the Cmkar 2+/+ but not the Cmkar 2–/– mice (Tables 1 and 2) . Therefore, this receptor may have a role in determining the pathogenic differences, which are dependent on cytotoxic or invasive characteristics of bacterial strains.
The absence of the CXCR2 results in an increased production of KC and MIP-2, which bind preferentially to this receptor, and MIP-2 has a tenfold greater affinity for the receptor than KC [12 ], which was up-regulated five times more than MIP-2 in the Cmkar 2–/– mice. The relative increase in KC production compared with MIP-2 has been reported previously [4 , 10 ]. In the mouse, KC binds only to CXCR2; the high levels in the gene knockout mice suggest that there is a presence of a negative feedback loop during infection to limit neutrophil chemotaxis, which may be responsible for limiting tissue damage. Kernacki et al. [2 ] showed that KC does not compensate for the loss of MIP-2 in neutrophil recruitment into the cornea. In this study, the restriction of neutrophils to the limbal area may suggest a role for KC in corneal tissue penetration for neutrophils as opposed to recruitment into the cornea, but this remains to be elucidated.
Other chemokines, MCP-1 and IP-10, which bind alternative receptors, were found to be up-regulated in Cmkar 2–/– mice. These changes may result from direct induction by bacteria or downstream alteration of chemokine modulation as a result of a failure of CXC chemokines to bind to their receptor. Cytokines IL-1ß and IL-6 were up-regulated in Cmkar 2–/– mice, which is most likely a response to the uncontrolled bacterial replication. Consistent with our findings, IL-6 has previously been found to be up-regulated 100-fold in HSV-infected corneas of Cmkar 2–/– mice [13 ]; however, the mechanisms and consequences of this remain to be determined. The chemotactic agent LTB4, which does not act via the CXCR2, was increased significantly in Cmkar 2–/– mice but failed to compensate for the recruitment of neutrophils into the cornea, suggesting CXCR2 is essential for neutrophil recruitment and penetration into the cornea. The intergrin MAC-1 was down-regulated in Cmkar 2–/– mice, which suggests that binding to the CXCR2 may be important in the regulation of intergrins involved in neutrophil infiltration of the avascular cornea; however, no changes in ICAM-1 mRNA levels were observed in this model.
CXCR2 is essential for neutrophil recruitment into the avascular cornea and bacterial clearance. Other mechanisms by which neutrophils are recruited do not compensate for this loss. Consequently, abrogation of CXC chemokines early in the infectious process is not appropriate but may be a useful adjunct therapy after bacterial clearance to control PMN-mediated destruction of the cornea.
ACKNOWLEDGEMENTS
This work was supported by a National Health and Medical Research Council project grant.
Received May 23, 2006; revised August 30, 2006; accepted September 7, 2006.
REFERENCES
This article has been cited by other articles:
 |
|
 |
|
  M. Lin, P. Jackson, A. M. Tester, E. Diaconu, C. M. Overall, J. E. Blalock, and E. Pearlman Matrix Metalloproteinase-8 Facilitates Neutrophil Migration through the Corneal Stromal Matrix by Collagen Degradation and Production of the Chemotactic Peptide Pro-Gly-Pro Am. J. Pathol., July 1, 2008; 173(1): 144 - 153. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Chiavolini, J. Alroy, C. A. King, P. Jorth, S. Weir, G. Madico, J. R. Murphy, and L. M. Wetzler Identification of Immunologic and Pathologic Parameters of Death versus Survival in Respiratory Tularemia Infect. Immun., February 1, 2008; 76(2): 486 - 496. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. C. Carlson, M. Lin, C.-Y. Liu, W. W-Y. Kao, V. L. Perez, and E. Pearlman Keratocan and Lumican Regulate Neutrophil Infiltration and Corneal Clarity in Lipopolysaccharide-induced Keratitis by Direct Interaction with CXCL1 J. Biol. Chem., December 7, 2007; 282(49): 35502 - 35509. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
