- Email: [email protected]
Role and regulation of CXC-chemokines in acute experimental keratitis
Experimental Eye Research 76 (2003) 221–231 www.elsevier.com/locate/yexer
Role and regulation of CXC-chemokines in acute experimental keratitis M.L. Xuea,1, A. Thakura,1, M.D.P. Willcoxa,*, H. Zhua, A.R. Lloydb, D. Wakefieldb a
Cooperative Research Centre for Eye Research and Technology, Cornea and Contact Lens Research Unit, The University of New South Wales, Sydney, NSW 2052, Australia b School of Pathology, The University of New South Wales, Sydney, NSW 2052, Australia Received 29 January 2002; accepted in revised form 2 October 2002
Abstract The aim of this study was to elucidate the expression of chemokines, their role and regulation in bacterial corneal infection using three bacterial strains (Pseudomonas. aeruginosa- invasive, cytotoxic and contact lens induced acute red eye strains) which have been shown to produce three distinct patterns of corneal disease in the mouse. The predominant chemokine expressed in response to all three strains was MIP-2. Prolonged expression of high levels of MIP-2 was associated with increased severity of corneal inflammation. Significantly reduced disease severity upon administration of anti-MIP-2 antibodies suggested that MIP-2 may play an important role in the pathogenesis of Pseudomonas keratitis at least in part by being a major chemoattractant for polymorphonuclear leukocytes (PMN) recruitment. Interestingly, the numbers of bacteria in eyes with neutralized MIP-2 activity did not decrease even though the severity of the disease was decreased. This implies PMNs as the major destructive factor in microbial keratitis. Further, neutralization of IL-1b activity alone using monoclonal antibodies resulted in significant reduction of both MIP-2 and KC activity indicating that chemokine levels were regulated by IL-1b. These studies demonstrate that the regulation of MIP-2 activity may be beneficial in reducing corneal damage during microbial keratitis in rodents and perhaps that regulation of the human homologue of MIP-2, IL-8, may be useful for controlling keratitis in humans. q 2003 Published by Elsevier Science Ltd. Keywords: MIP-2; KC; Pseudomonas keratitis; corneal infection; inflammation
1. Introduction Bacterial corneal infection of the eye remains a significant health problem and presents a major therapeutic challenge. Among the bacterial pathogens, Pseudomonas aeruginosa is the most frequently isolated microorganism in cases of ulcerative keratitis associated with contact lens wear. Keratitis caused by P. aeruginosa is a highly destructive corneal infection (Kernacki and Berk, 1994). Both bacterial exotoxins and proteases (Kernacki et al., 1995), as well as the host’s own inflammatory response primarily derived from stimulated polymorphonuclear leukocytes (PMN) (Steuhl et al., 1989) and possibly inappropriate production of inflammatory cytokines and chemokines, may contribute to corneal damage. Although rapid infiltration of PMN is required to eliminate the infecting bacteria, the continued * Address correspondence to: M.D.P. Willcox, Cooperative Research Center for Eye Research and Technology, The University of New South Wales, Sydney NSW 2052, Australia. E-mail address: [email protected] (M.D.P. Willcox). 1 First and second authors have equal contribution in this paper. 0014-4835/03/$ - see front matter q 2003 Published by Elsevier Science Ltd. DOI:10.1016/S0014-4835(02)00270-1
presence of PMNs implies ongoing inflammation, which appears to promote scarring. The current concept of a multi-step process of leukocyte recruitment envisions chemotactic agonists or chemokines as one of the key effector molecules (Springer, 1994; Baggiolini et al., 1997; Rollins, 1997). The CXC-chemokines such as MIP-2 and KC are chemotactic for neutrophils, while in vitro studies on C –C chemokines (MCP-1 and MIP-1a) have revealed that they are predominantly chemotactic for monocytes and lymphocytes (Taub and Oppenheim, 1994). Understanding the molecular mechanisms of the host response, including induction of defenses required to influence microbial clearance, is important to devise better management strategies to control sight threatening ocular infection and inflammation. The cornea is known to produce chemokines specific for the recruitment and activation of PMNs. Human corneal tissue, including epithelium and stromal keratocytes, express IL-8 (murine homolog MIP-2) and GRO-a (murine homolog KC) mRNA and protein in vivo (Cubitt et al., 1993; Cubitt et al., 1997) and in vitro experiments (Rosenbaum et al.,
222
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
1995; Kernacki et al., 2000). Our previous studies have demonstrated the presence of chemotactic factors in tears collected during eye closure and keratitis (Thakur et al., 1998; Thakur and Willcox, 1998a,b). P. aeruginosa strains isolated from microbial keratitis are either invasive or cytotoxic in nature. In this study three phenotypes of P. aeruginosa (the invasive, cytotoxic and contact lens induced acute red eye (CLARE) strains) which have recently been shown to produce distinct patterns of corneal disease in mice were used (Cole et al., 1998). The invasive and cytotoxic strains were isolated from human microbial keratitis. The CLARE strain (Paer1; which is neither invasive nor cytotoxic) was isolated from the self-limiting condition known as CLARE (Cowell et al., 1998) and is not associated with tissue destruction in the cornea. The invasive or cytotoxic phenotypes of P. aeruginosa appear to be mutually exclusive (Fleiszig et al., 1997); this exclusivity has been shown to be dependent on possession of a number of genes. Invasive strains possess the genes exoS and exoT that encode the ADP-ribosylating toxins ExoS and ExoT (Fleiszig et al., 1996). Cytotoxic strains lack the gene exoS but possess the gene exoU (Fleiszig et al., 1996). The names ‘cytotoxic’ or ‘invasive’ refer to the interactions between epithelial cells and the bacteria. The cytotoxic strains cause a rapid cytotoxicity in the mammalian cells, whereas the invasive strains can remain inside cells for up to 24 hr before the cells show any overt cytopathic effect (Fleiszig et al., 1994). Expression of many of the virulence factors in P. aeruginosa appears to be highly controlled by the signal molecule-dependent cell – cell communication systems known as quorum sensing in a cell density-dependent fashion (Fuqua et al., 1996). Cell density-dependent signal molecules, N-acyl-L -homoserine lactones (AHLs), are fundamental regulatory agents of many bacterial processes. The AHLs produced by Gram-negative bacteria may not only be limited to induction of specific bacterial phenotypes, but may also act directly to control the host immune response through regulation of gene expression in host cells (Telford et al., 1998). The host response associated with each phenotype has not been clearly defined. The order of virulence of these strains was 6294 . 6206 . Paer1. We studied the hypothesis that the magnitude and temporal production of chemokines induced by each phenotype of P. aeruginosa would effect the outcome of corneal disease in a resistant mouse strain (BALB/c). Keeping this in view, we elucidated the expression of chemokines and the regulation of infiltrating cells in the corneas infected with the three different strains of P. aeruginosa and the contribution and regulation of chemokines in corneal pathology. In addition, bacterial virulence factors, which may induce chemokine production, were also investigated.
2. Materials and methods 2.1. Animal model BALB/c mice aged 8– 10 weeks were used in this study. All mice had undergone baseline measurements of corneal integrity, including slit lamp evaluation. Details of corneal transparency and vessel hyperemia were assessed and only those animals which were in the normal range for the above clinical variables were used in the study. ARVO guidelines on the use of animals in experimentation were adhered to and institutional ethics clearance was obtained. 2.2. Bacterial strains and growth conditions Three strains of P. aeruginosa, 6294 (invasive strain), 6206 (cytotoxic strain) and Paer1 (CLARE strain), were used in this study. Both the cytotoxic and invasive strains were isolated from human corneal ulcers and the Paer1 strain was isolated from a non-infectious condition called CLARE. Bacterial cells were grown in 10 ml of Tryptone Soy Broth (Oxoid Ltd., Sydney, Australia) overnight at 378C. Bacteria were harvested and washed three times in sterile PBS (pH 7·4) and re-suspended in PBS at 1 £ 109 cells ml21 (OD660 ¼ 1·0). 2.3. Animal infection Inbred 8– 10 weeks old BALB/c mice were challenged with P. aeruginosa. The central corneal surface of only the left eye was incised (two parallel incisions approximately 2 – 3 mm long and penetrating only the epithelium) with a sterile 27·5 gauge needle. Twelve animals were included at each time point (three eyes for bacterial enumeration, three for PMN quantitation, three for ELISA and three for RNA studies) for each bacterial strain. The mice were anaesthetized with Avertin (125 mg kg21, intraperitoneally) under a stereomicroscope and 5 ml of bacterial suspension (5 £ 106 cells) was applied topically onto the wounded cornea. The right eye of each animal, which served as a control, was scratched but not infected. 2.4. Administration of anti mouse-MIP-2, -KC and -IL-1b monoclonal antibodies Anti-mouse-MIP-2, KC and IL-1b monoclonal antibodies were purchased from R&D Systems (Minneapolis, MN, USA). Anesthetized animals were injected with 10 ml (1 mg ml21) of anti-mouse-MIP-2, KC or IL-1b monoclonal antibody sub-conjunctivally 2 hr before infection with the invasive strain of P. aeruginosa 6294. Control mice received an equal volume of isotype matched control antibody at the same time point and were also infected. Twelve animals were included at each time point (three eyes for bacterial enumeration, three for PMN quantitation, three
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
223
Table 1 PCR primers for cytokines and chemokines Size of PCR product
Primer sequences
MIP-2 (221 bp)
S: 50 GCC AGT GAG CTG CGC TGT CAG TGC-30 S: 50 GCC AAT GAG CTG CGC TGT CAG TGC-30 S: 50 GGC TCG CTC GGT GAC CCT AGT CTT T-30
KC (207 bp) BMG (300 bp)
A: 50 GTT AGC CTT GCC TTT GTT CAG TAT G-30 A: 50 CTT GGG GAC ACC TTT TAG CAT CTT-30 A: 50 TCT GCA GGC GTA TGT ATC AGT CTC A-30
S: sense primer, A: anti-sense primer.
for ELISA and three for histology). These experiments were repeated at least three times to ensure reproducibility. 2.5. Bacterial enumeration Clearance of P. aeruginosa in the infected cornea was monitored by plate counts of viable bacteria in corneal homogenates. Three eyes at 4 hr, 8 hr, 16 hr, 1 d, 3 d, 5 d and 7 d time points were assessed for the dynamics of P. aeruginosa removal from the corneas. Small aliquots (20 ml in duplicate) of serial dilutions were plated onto the nutrient agar plates. Plates were incubated for 24 hr at 378C. Results were expressed as mean CFU/cornea ^ S.D. 2.6. Measurement of myeloperoxidase (MPO) activity Samples were assayed for MPO activity as previously described (Williams et al., 1983). Whole eyes enucleated at 1 hr, 4 hr, 8 hr, 16 hr, 1 d, 3 d, 5 d, 7 d were homogenized in 1 ml of hexadecyl trimethylammonium bromide (HTAB) buffer (0·5% HTAB in 50 mM phosphate buffer, pH 6·0) and sonicated for 10 sec in an ice bath. The samples were freezethawed three times and centrifuged at 14 000 rpm for 20 min. Supernatant (0·1 ml) was mixed with 2·9 ml of 50 mM phosphate buffer (pH 6·0), containing 0·167 mg ml21 O-dianisidine hydrochloride and 0·0005% hydrogen peroxide. The change in absorbance at 460 nm was continuously monitored for 5 min. Two eyes were used at each time point, and experiments were repeated at least two times. One unit of MPO activity was determined to be equivalent to approximately 2 £ 105 PMN cells ml21. 2.7. Reverse transcription-PCR for chemokine mRNA Total RNA was extracted from mouse eyes by using TRI reagent (Sigma, St Louis, USA). First-strand cDNA synthesis (First-strand cDNA synthesis kit from Promega, Madison, WI, USA) from 2 mg total RNA was accomplished by reverse transcription using oligo(dT)15 as primer. Specific primers and product lengths are listed in Table 1. Primers were based on published sequences (Huang et al., 1992; Yamamoto et al., 1995) or from gene sequences in Genbank (www.ncbi.nlm.nih.gov/Genbank). Polymerase chain reaction was performed as described
previously (Yamamoto et al., 1994, 1995). Briefly, 4 ml of cDNA product for MIP-2 was amplified in the presence of primer sets, dNTPs mixture, and Taq DNA polymerase in a final volume of 50 ml. The reaction mixture was overlaid with mineral oil, and PCR was performed in PC-960G Microplate Gradient Thermal Cycler (Corbett Research, Sydney, NSW, Australia) for 30 cycles. The first cycle consisted of a 5 minute denaturation at 94 8C, a 5 min annealing at 608C, and then 30 cycles each of 1·5 min at 728C, 45 sec at 948C, and 45 sec at 608C, with a final extension of 10 min at 728C. The amplification of KC gene fragment was performed with 2·5 U AmpliTaq polymerase in 25 mM /Tris/HCl (pH 8·3), 50 mM l21 KCl, 2·5 mM l21MgCl2, 0·01% gelatin, 5 mM /DTT, and 1 mmol l21 each primer, each cycle consisted of denaturation at 948C 1 min , annealing at 558C 2 min, and extension at 728C 3 min. PCR was performed with two primer sets (b2-microglobulin (BMG)-specific primers as an endogenous standard and cytokine-specific primers) in the same tube. RT-PCR products were separated by 3% agarose gel electrophoresis, detected by ethidium bromide staining and analyzed by Quantity One software (Bio-Rad, Sydney, NSW, Australia). The relative mRNA level was obtained by comparing the optical density of cytokine-specific bands to the internal standard. 2.8. Chemokine protein determination by ELISA Cytokine levels were measured in ocular homogenates of scratch control and challenged eyes at different time points using ELISA kits (R&D Systems, Minneapolis, MN, USA). Samples for ELISA were prepared by homogenizing mouse eyes in sterile PBS. Homogenates were centrifuged at 12,000 rpm for 20 min at 48C. The resulting supernatants were used to quantitate MIP-2 and KC proteins. Samples, diluted 1:5 in sample diluting buffer, were added in duplicate wells. Samples were analysed following the manufacturer’s instructions. 2.9. Immunostaining for MIP-2 and KC Whole mouse eyes were enucleated and covered with tissue freezing medium OCT (Pro SciTech, Qld, Australia), snap-frozen by immersing into liquid nitrogen and 5 mm
224
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
thick sections were cut. Sections were fixed in freshly prepared 4% paraformaldehyde for 8 min, rinsed in phosphate-buffered saline (PBS, 0·06 M , pH 7·2) and air dried. The sections were stained following the manufacturer’s instructions for staining (HRP – DAB System, R&D Systems, Minneapolis, MN, USA). The cytokine-specific biotin labelled monoclonal antibodies to MIP-2 and KC (R&D Systems, Minneapolis, MN, USA) were used at a concentration of 5 mg ml21 in PBS (0·01 M , pH 7·2) containing 1% BSA and 0·1% Saponin (Sigma), while control samples were treated in the same fashion using isotype matched control antibody. After staining with diaminobenzidine the sections were counter-stained with 0·1% (w/v) Toluene Blue a for few seconds, mounted with DPX (BDH Laboratory Supplies, Poole, UK) and observed using light microscopy by a masked observer. 2.10. Collection of corneas for histological examination Mouse eyes were enucleated at different time points (a suture was taken as an orientation mark) and fixed in 2·5% glutaradehyde (in PBS buffer, pH 7·4) for 2 d. After completely washing with PBS three times for 10 min each, eyes were kept in fresh PBS overnight (o/n) to remove all the fixative, then dehydrated in a graded ethanol series (30, 50, 70, 95%) for 15 min each. Following dehydration, eyes were infiltrated in 95% ethanol and pure resin (1:1) o/n at 48C, and transferred to pure resin o/n at 48C again. Finally, eyes were embedded at embedding mould, sectioned with a microtome (Leica RM 2155, Germany) in sections of 2 mm thickness. Specimens were stained with 0·1% (w/v) Toluene Blue for 7 min, mounted with DPX mountant (ProSciTech, Sydney, NSW, Australia) and observed under a microscope. At least two eyes were examined at each time point by a masked observer. 2.11. Bacterial virulence factors 2.11.1. Zymography for elastase and alkaline protease in P. aeruginosa Bacterial proteases were separated on the basis of molecular size by electrophoresis through a sodium dodecyl sulfate (SDS)-polyacrylamide gel within which 0·1% (w/v) gelatin was immobilized. Subsequently, the position of each enzyme in the gel was visualized by its ability to degrade the substrate, which was detected by staining the remaining gelatin. After overnight culture of P. aeruginosa, in Tryptone Soy Broth (Oxoid Ltd., Sydney, Australia) at 378C, culture supernatants were collected by centrifugation and used for quantitation of protein using the Bradford assay (Bio-Rad, Sydney, NSW, Australia). Approximately 10 mg of each sample was loaded per lane into the wells of each gel (10% polyacrylamide gel containing 0·1% gelatin). Electrophoresis was performed with a Tris –glycine running buffer at 125 V constant voltage for 1·5– 2 hr. The gel was removed and incubated for 1 hr at room temperature in 100 ml of
development buffer (50 mM Tris base, 40 mM HCl, 200 mM NaCl, 5 mM CaCl2 and 0·2% Brij35) at 378C for 14 –18 hr. Staining was performed with 100 ml of 0·5% Coomassie blue G-25 in 30% methanol and 10% acetic acid for 1 hr, and the gel was then de-stained with four changes of solutions. 2.11.2. Production of acyl homoserine lactones The level of acyl homoserine lactones (AHLs) in bacterial culture supernatants was quantified by using the reporter strain Agrobacterium tumefaciens A136 following the method previously described (Zhu, et al., 2000). A. tumefaciens traG < lacZ/traR reporter detects 3oxo-substituted AHL derivatives with acyl chain length from 4 – 12 carbons particularly OdDHL. The AHL reporter strain A. tumefaciens A136 was cultured on a supplement medium (Miller, 1976) at 308C. The levels of AHLs in the supernatants were examined as the ability to activate traR in a b-galactosidase reporter strain A. tumefaciens A136. 2.12. Statistical analysis Statistical analysis of data was performed using a one way ANOVA to test for differences in chemokine gene and protein expression in the corneas infected with the three different strains.
3. Results 3.1. Bacterial clearance Viable bacterial counts of infected mouse eyes were performed at various time points from 1 h to 7 d post challenge. Although bacterial numbers decreased transiently at 4 hr in the corneas infected with the invasive and cytotoxic strains, they increased again at 16 hr post challenge. All mice challenged with the invasive and cytotoxic bacterial strains showed a rapid increase in bacterial numbers which peaked 1d post challenge, dropped significantly ( p , 0·04) by day 3 and declined further by day 5 and 7 post challenge. Corneas inoculated with the CLARE strain showed a continuous decrease in bacterial numbers. Bacteria could not be recovered 16 hr post challenge from corneas inoculated with the CLARE strain. Results are presented in Fig. 1. 3.2. PMN infiltration We have used whole-eye homogenates for quantitation of PMNs, as recruitment of inflammatory cells are not only restricted to the cornea but other anterior segment tissues such as limbus and anterior chamber are also involved. Comparison of MPO activity at different time points in whole mouse eyes challenged with the invasive and cytotoxic P. aeruginosa revealed a significant infiltration of PMNs compared to scarified control corneas after 8 hr
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
225
Fig. 1. Bacterial numbers in whole eyes challenged with invasive, cytotoxic and CLARE bacterial strains and scratch control at 1 hr, 4 hr, 8 hr, 16 hr, 1 d, 3 d, 5 d, and 7 d post challenge. Results are presented as mean log10 bacterial number ^ S.E.M /eye (CLARE, CLARE strain; cytotoxic, cytotoxic strain; invasive, invasive strain).
post challenge ( p , 0·004). Corneas infected with the invasive and cytotoxic strains showed peak MPO activity at 1 d post challenge. MPO activity in response to the cytotoxic strain was significantly lower at all time points after 8 hr post challenge (16 hr p ¼ 0·003; p ¼ 0·01; 1 d p ¼ 0·005; 3 d p ¼ 0·005; 5 d p ¼ 0·001; 7 d p ¼ 0·002) compared to the corneas inoculated with the invasive strain. In response to the CLARE strain, peak activity was observed at 8 hr post infection and this remained constant up to 1 d post challenge. Results are presented in Fig. 2. 3.3. Kinetics of chemokine mRNA expression Corneas infected with the invasive or cytotoxic bacterial strains expressed MIP-2 and KC transcript as early as 1 hr
post challenge (Fig. 3). MIP-2 and KC mRNAs were significantly higher in the corneas infected with either the invasive (MIP-2 p ¼ 0·003; KC p ¼ 0·0001) or the cytotoxic (MIP-2 p ¼ 0·001; KC p ¼ 0·0001) strains, at 8 hr post challenge, compared to the CLARE strain. Corneas challenged with the invasive strain showed significantly higher ( p , 0·02) levels of MIP-2 mRNA compared to the corneas infected with the cytotoxic strain. Both MIP-2 and KC mRNA declined by 7d post challenge in response to both the invasive and cytotoxic strains (Fig. 3). In response to the CLARE strain, the transcripts of MIP-2 peaked at 16 hr post inoculation and declined by 1 d post challenge. In contrast, KC mRNA peaked at 8 hr post challenge and declined by 16 hr post challenge (Fig. 3). 3.4. Kinetics of chemokine protein expression
Fig. 2. MPO activity in whole eyes challenged with the invasive, cytotoxic or CLARE strains and scratch control at 1 hr, 4 hr, 8 hr, 16 hr, 1 d, 3 d, 5 d, and 7 d post challenge. Results are reported as mean log10 MPO activity ^ S.E.M. /eye. Note that the horizontal axis is a non-linear scale. (SC, scratch control; CLARE, CLARE strain; cytotoxic, cytotoxic strain; invasive ¼ invasive strain).
In the murine model of keratitis induced by the invasive bacterial strain, neutrophil chemoattractant MIP-2 was highly up-regulated and to a lesser extent so was KC. In response to the invasive strain, MIP-2 protein was produced as early as 4 hr post challenge, peaked at 16 hr and declined by 1 d post challenge. MIP-2 levels were significantly higher in response to the invasive strain at all time points, while KC levels were significantly higher at 16 hr and 1 d post challenge compared to the corneas inoculated with either the cytotoxic (MIP-2, p , 0·0001 KC, p , 0·0001) or CLARE (MIP-2, p , 0·0001, KC, p , 0·0001) strains. In response to the cytotoxic strain, the kinetics of chemokine appearance were different compared to the invasive strain. Both MIP-2 and KC levels peaked at 1 d post challenge (Fig. 3). In response to CLARE strain, levels of the chemokines peaked at 16 hr post challenge and were significantly lower compared to the invasive ( p , 0·0001) and cytotoxic ( p , 0·0001) strains (Fig. 3).
226
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
Fig. 3. Top panel. Time dependent expression of chemokine mRNA in mouse corneas at 1 hr, 4 hr, 8 hr, 16 hr, 1 d, 3 d, 5 d, and 7 d post challenge with invasive, cytotoxic or CLARE bacterial strains. Results are presented as mean relative intensity (^ S.E.M. ) of each chemokine band divided by intensity of the control and normalized by respective GAPDH band in the same lane. Bottom panel. Time dependent expression of chemokine protein in mouse corneas at 1 hr, 4 hr, 8 hr, 16 hr, 1 d, 3 d, 5 d, and 7 d post challenge with invasive, cytotoxic or CLARE bacterial strains. Results are presented as pg chemokine protein (^ S.E.M. )/eye. Note that horizontal axis is a non-linear scale. (SC, scratch control; CLARE, CLARE strain; cytotoxic, cytotoxic strain; invasive ¼ invasive strain).
3.5. MIP-2 and KC are produced by both resident corneal cells and infiltrating PMNs
3.6. Neutralization of MIP-2 or KC activities during infection with invasive strain of P. aeruginosa
Immunohistochemistry was used to evaluate the types of cells that produced each chemokine during the infection process. Immunoreactivity for MIP-2 and KC was initially detected in corneal epithelial and endothelial cells (4 hr). Later (24 hr) during the infection both MIP-2 and KC were largely immuno-localized in infiltrating cells (PMNs) in the corneal stroma (Fig. 4).
3.6.1. Histopathologic changes in mouse corneas injected with anti-MIP-2 or -KC before P. aeruginosa infection Histopathological examination of corneas injected with anti-MIP-2 monoclonal antibodies before infection with invasive bacterial strain showed markedly reduced infiltrates in the corneal stroma but large number of PMNs were present in the anterior chamber at 1 d post challenge
Fig. 4. Immuno-localization of MIP-2 and KC in the corneas at 1 d post challenge with invasive bacterial strain. Frozen sections stained for MIP-2 protein (25 £ ) in the corneal epithelium and infiltrating cells both intra-cellularly and extra-cellularly (A) and KC protein (25 £ ) largely stained extracellularly in the corneal epithelium and stroma around the infiltrating cells (B). When frozen sections were incubated with isotype control monoclonal antibody no signals were observed (C). (Ep, corneal epithelium; S, corneal stroma.
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
227
Fig. 5. Histological examination of mouse corneas at 1 d post challenge. (A) Corneas treated with anti-MIP-2 monoclonal antibodies before infection with invasive strain showed markedly reduced infiltrates (arrowhead) in the corneal stroma but large numbers were in the anterior chamber (arrow) at 1d post challenge (10 £ ). The original scratch site had healed and there was no epithelial defect present at 1 d post challenge. (B) Corneas treated with anti-KC monoclonal antibodies before infection with the invasive strain showed reduced infiltrates (arrow) in the corneal stroma compared to control corneas (arrows) at 1 d post challenge (10 £ ). The original scratch site had somewhat healed at 1d post challenge. (C) Isotype control antibody injected corneas before infection showed prominent cellular infiltration (arrows), with marked edema (5 £ ). The epithelium was completely destroyed in the central cornea (open arrows) at 1 d post challenge and large number of infiltrates are also present in the anterior chamber (Ep, epithelium; S, stroma; E, endothelium; AC, anterior chamber).
(Fig. 5(A)) compared to control antibody treated corneas. There was no epithelial defect present. On day seven corneas appeared almost normal. Corneas injected with anti-KC monoclonal antibodies showed reduced infiltration in the corneal stroma compared to control corneas (Fig. 5(B)). Corneas injected with isotype control antibody sub-conjunctivally before infection showed prominent cellular infiltration with complete denudation of corneal stroma in the central cornea at 1d post challenge. The corneal stroma was markedly edematous and a large number of infiltrates were present in the anterior chamber (Fig. 5(C)). On day 7, in both antiKC monoclonal antibody or isotype control antibody treated animals, the corneas had recovered noticeably compared to 1d, although there was still large number of PMNs evident in the corneal stroma and anterior chamber. The epithelium had healed.
isotype control antibody treated animals at any time point. Mice injected with anti-KC antibodies showed a slight increase in bacterial number at later time points, however, the difference was not statistically significant (Fig. 6). 3.6.3. PMN infiltration PMN numbers quantified by measuring MPO activity showed significantly lower ( p , 0·001) numbers of PMNs in anti-MIP-2 treated mice compared to both control antibody treated ( p , 0·0001) and anti-KC treated ( p , 0·0002) mice at all time points (Fig. 7). Neutralization of KC protein also showed reduced MPO activity at 1 and 3 d post challenge. However, MPO activity increased significantly on day 7 ( p , 0·02) post challenge compared to control antibody treated animals (Fig. 7).
3.6.2. Bacterial enumeration Administration of anti-MIP-2 antibodies did not result in alteration in bacterial numbers. There was no significant difference found between anti-MIP-2 antibody treated or
3.6.4. Chemokine proteins The MIP-2 protein levels were reduced significantly in anti-MIP-2 treated mice as expected compared to control antibody treated animals at all time points. Mice treated with anti-KC antibodies did not show any significant difference in MIP-2 levels at 1 and 3 d post challenge, but
Fig. 6. Bacterial enumeration in the whole eye of anti-MIP-2, anti-KC and isotype control monoclonal antibody treated animals challenged with invasive strain at 1–7 d post challenge. Results are presented as mean log10 bacterial number ^ S.D. /eye.
Fig. 7. MPO activity in whole eyes of anti-MIP-2, anti-KC and isotype control monoclonal antibody treated animals challenged with invasive strain at 1–7 d post challenge. Results are reported as mean log10 MPO activity ^ S.D. /eye.
228
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
Fig. 8. Kinetics and levels of chemokine proteins in the corneas of anti-MIP-2, anti-KC and isotype control monoclonal antibody treated animals challenged with invasive strain at 1, 3, 5, and 7 d post challenge.
levels rose significantly at 5 d post challenge compared to controls (Fig. 8). KC protein levels were significantly reduced in anti-KC treated animals at 1 d but levels rose at 3 and 5 d post challenge compared to control mice. Interestingly, KC levels were significantly up-regulated at 1 d post challenge in anti-MIP-2 antibody treated mice compared to both controls and anti-KC antibody treated mice (Fig. 8). 3.6.5. MIP-2 is regulated by the pro-inflammatory cytokine IL-1b MIP-2 protein levels were significantly reduced in antiIL-1b treated mice. There was approximately 18-fold ( p ¼ 0·002) less MIP-2 at 1 d, 22-fold ( p ¼ 0·002) less MIP-2 at 3 d and about 25-fold ( p ¼ 0·001) less MIP-2 at 5 d in the IL-1b mAb treated corneas compared to isotype control antibody treated corneas (Fig. 9(A)). Levels of KC were also reduced in anti-IL-1b treated mice by approximately 5-fold ( p , 0·04) at 1 d post challenge. However, there was no significant difference found at other time points (Fig. 9(B)).
3.7. Bacterial virulence factors 3.7.1. Production of proteases In order to determine whether the invasive strain could produce proteases capable of destroying corneal stroma, the culture supernatants were examined by zymography. Zymography of culture supernatants of P. aeruginosa showed presence of considerable amounts of proteases especially elastase (145 kDa) and alkaline protease (53 kDa) in the invasive strain compared to cytotoxic strain (Fig. 10(A)). 3.7.2. Production of N-acyl-L -homoserine lactones The invasive strain produced significantly higher levels of AHLs compared to cytotoxic strain as detected by its ability to activate traR in a b-galactosidase reporter strain A. tumefaciens A136 (Fig. 10(B))
4. Discussion This study demonstrates for the first time that MIP-2 is the predominant chemokine expressed during corneal
Fig. 9. Kinetics and levels of MIP-2 (a) and KC (b) proteins in the corneas of anti-IL-1b and isotype control monoclonal antibody treated animals challenged with invasive strain at 1 d post challenge.
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
Fig. 10. (A) Demonstration of protease activity in the culture supernatant of the invasive and cytotoxic strains of P. aeruginosa. Culture supernatant showing elastase at 145 kDa band and alkaline protease at 53 kDa band. (B) Demonstration of AHL production by measuring beta-galactosidase activity in the supernatants of the invasive and cytotoxic strains.
infection by three distinct types of P. aeruginosa that produce distinct pathologies in the cornea (Cole et al., 1998) and have been shown to be genetically or phenotypically distinct (Fleiszig et al., 1994, 1996, 1997). Furthermore, this study has confirmed that subsequent to MIP-2 expression, the recruitment and activation of PMNs is the major cause of corneal destruction. Neutralization of MIP-2 (during corneal infection induced by the invasive P. aeruginosa strain) and the consequent reduction in PMN recruitment into the cornea resulted in almost clear corneas even though bacterial numbers were high. This demonstrates that an invasive P. aeruginosa strain which was shown to produce appreciable amounts of both elastase and alkaline protease in vitro was not responsible for the major part of the corneal destruction during microbial keratitis in the mouse model. MIP-2 is a neutrophil chemoattractant (Baggiolini et al., 1994; Taub and Oppenheim, 1994; Rollins, 1997) and primarily involved in PMN recruitment during acute corneal inflammation. Peak levels of MIP-2 were detected in severe infection induced by the invasive strain, and to a lesser extent in corneas infected by the cytotoxic strain. In contrast, in response to the CLARE strain which induced only a mild
229
inflammatory response, significantly lower levels of MIP-2 were induced. Further we examined cause and effect relationships between MIP-2/KC and corneal disease. Previous studies have shown that antibody neutralization of endogenous MIP-2 can alter the host response in animal models of bacterial infection (Greenberger et al., 1996; Yan et al., 1998; Dumont et al., 2000). In vivo neutralization of MIP-2, using anti-MIP-2 monoclonal antibodies, substantially reduced the severity of corneal disease induced by the invasive bacterial strain (which produced high levels of MIP-2 and severe corneal disease) and resulted in reduced PMN infiltration as shown by MPO activity and histopathological examination. Collectively these results indicate that both the magnitude and kinetics of appearance of MIP-2 may effect the severity of corneal disease, primarily by regulating the infiltration of PMNs and their activation. A recent study by Kernacki et al. (2000) has also shown that MIP-2 is the major chemokine responsible for the initial recruitment and persistence of neutrophils in Pseudomonas keratitis. The major differences between our study of MIP-2/KC neutralization and that by Kernacki et al. (2000) are that they used a susceptible mouse strain (C57BL/6, cornea perforates), the cytotoxic P. aeruginosa strain-19660 and a systemic intra-peritoneal injection of anti-MIP-2 polyclonal antibodies, whereas in the present study we used a resistant mouse strain (BALB/c, cornea heals), an invasive strain and local subconjunctival injection of monoclonal antibodies. Importantly, since both the cytotoxic and invasive strains are isolated from the Pseudomonas keratitis cases, it was critical to examine whether neutralization of MIP-2 after corneal infection with an invasive strain lead to reduced corneal disease. The results of our study indicate the potential for antiCXC chemokine therapies as an adjunct in cases of Pseudomonas keratitis caused by either of the lineage of P. aeruginosa that have been isolated from microbial keratitis. Further work would need to be performed in this area to demonstrate effectiveness of anti-chemokine therapy once infection had been initiated. In vivo neutralization of KC using anti-KC antibodies showed significantly reduced MPO activity on day 1 and 3, although histological examination revealed large number of infiltrating cells (similar to untreated controls) in the central cornea. One explanation for this could be that PMNs at the site of infection have undergone necrosis/apoptosis, and KC may have some role in delaying necrosis/apoptosis. Later during the infection when KC levels returned back to untreated control levels on day 5 and 7, MPO activity also increased (equivalent to untreated controls) but by this time point the number of bacteria had increased significantly compared to untreated controls. Interestingly, KC levels were significantly higher in anti-MIP-2 treated mice but this did not seem to compensate for MIP-2 as significantly reduced number of PMNs (as judged by MPO activity and histology) were present in these animals. This further
230
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
indicates that KC may not involved in PMN recruitment and that MIP-2 is the dominant chemokine involved in the PMN recruitment in bacterial corneal infection. MIP-2 has been shown to be a major chemokine in HSV-1 infected corneas and neutrophil chemoattractant KC did not compensate when MIP-2 was neutralized, even though the concentration of KC was more than two times higher than MIP-2 (Yan et al., 1998). These authors also reported that KC was 3– 4 times less potent compared to MIP-2 in promoting PMN migration in vitro. Tessier et al. also observed that KC was a less potent chemoattractant than MIP-2 in a murine subcutaneous air pocket model of inflammation (Tessier et al., 1997). Our in vitro studies have shown that MIP-2 is regulated by IL-1b at both the transcriptional and translational levels in a human corneal epithelial cell line (Xue et al., 2000, 2001). Thus we sought to determine whether MIP-2 expression was regulated by IL-1b in this keratitis model. IL-1b protein was neutralized using anti-IL-1b monoclonal antibodies before corneal infection and as a consequence MIP-2 levels were significantly reduced as was the infiltration of PMNs in the corneal stroma. This suggests that MIP-2 expression is largely regulated by IL-1b in the corneas infected by an invasive strain of P. aeruginosa. These results are in accordance with previous study which showed that IL-1b regulates MIP-2 in susceptible mouse (C57BL/6J) strain for P. aeruginosa corneal infection with a cytotoxic strain (Rudner et al., 2000). The virulence factors in P. aeruginosa are highly regulated by the signal molecule-dependent cell – cell communication systems known as quorum sensing in a cell density-dependent fashion (Fuqua et al., 1996). Cell density-dependent signal molecules, AHLs, are fundamental regulatory agents of many bacterial processes. Interestingly, the invasive strain also produced extremely high levels of AHLs compared to cytotoxic strain. Recently there have been reports that AHLs can cause IL-8 production by respiratory epithelial cells (Telford et al., 1998; Saleh et al., 1999). Corneal damage is also likely to be attributed to the ability of each Pseudomonas strain to produce proteolytic enzymes i.e. LasB elastase, LasA protease, alkaline protease and protease IV (Heck et al., 1986a,b). This study showed that the invasive strain produced larger amounts of proteases compared to the cytotoxic strain. Previous studies in Pseudomonas strains have demonstrated that cytokine production in vivo is correlated to a large extent with the virulence of the pathogen (Hirakata et al., 1999; Epelman et al., 2000). Consistent with these studies, the three strains used in this study differed significantly in the kinetics and magnitude of cytokine induction. A recent study has shown that exoenzyme S (present in the invasive strain, a highly virulent strain) can rapidly induce various cytokines (IL-1a, IL-1b and IL-6) in host cells (Epelman et al., 2000). The cytotoxic strain used in the present study lacks the exoS gene (Fleiszig et al., 1996). It appears likely that the greater magnitude of the cytokine responses produced by
the invasive strain may be related to the capability of key virulence factors in the microorganism to induce cytokines production in the host. It is noteworthy that attenuation of MIP-2 activity only resulted in reduced PMN numbers but did not effect the bacterial numbers. There are several possible mechanisms that could explain the apparent discrepancy between low numbers of neutrophils but reductions in bacterial numbers. Our current hypothesis is that the neutrophils that are recruited into the cornea of anti-chemokine-treated mice retain their ability to phagocytose bacteria and are sufficient to reduce the bacterial numbers. The large numbers of neutrophils that are recruited into normal corneas (or corneas treated with irrelevant antibodies) are not only involved in bacterial clearance but also in destruction of the cornea. Based on these findings we speculate that continued up-regulation of MIP-2 is the major factor contributing to severe corneal disease induced by invasive and cytotoxic Pseudomonas strains. Whereas the CLARE strain, which induced mild inflammation, showed lower levels of MIP-2 expression. In conclusion, our data suggests that the severity of infection (bacterial clearance/PMN infiltration) is associated with the magnitude of expression of chemokine MIP-2 but not KC. These findings point to the therapeutic potential of using both anti-MIP-2 and anti-IL-1b antibodies as possible treatment strategies in bacterial keratitis.
Acknowledgements This research was supported by the National Health and Medical Research Council. The authors would like to thank Reg Wong for excellent statistical analysis, Wen Wang for technical assistance, Denise Lawler and Robyn Lawler for animal handling.
References Baggiolini, M., Dewald, B., Moser, B., 1994. Interleukin-8 and related chemotactic cytokines: CXC and CC chemokines. Adv. Immunol. 55, 97 –179. Baggiolini, M., Dewald, B., Moser, B., 1997. Human chemokines: an update. Ann. Rev. Immunol. 15, 675–705. Cole, N., Willcox, M.D.P., Fleiszig, S.M.J., Stapleton, F., Bao, S., Tout, S., Husband, A., 1998. Different strains of Pseudomonas aeruginosa isolated from ocular infections or inflammation display distinct corneal pathologies in an animal model. Curr. Eye Res. 17, 730 –735. Cowell, B.A., Willcox, M.D.P., Hobden, J.A., Schneider, R.P., Tout, S., Hazlett, L.D., 1998. An ocular strain of Pseudomonas aeruginosa is inflammatory but not virulent in the scarified mouse model. Exp. Eye Res. 67, 347 –351. Cubitt, C.L., Lausch, R.N., Oakes, J.E., 1997. Differential induction of GROa (gene expression in human corneal epithelial cells and keratocytes exposed to proinflammatory cytokines. Invest. Ophthalmol. Vis. Sci 38, 1149– 1158. Cubitt, C.L., Tang, Q., Monteiro, C.A., Lausch, R.N., Oakes, J.E., 1993. IL-8 gene expression in cultures of human corneal epithelial cells and keratocytes. Invest. Ophthalmol. Vis. Sci. 34, 3199–3206.
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231 Dumont, R.A., Car, B.D., Voitenok, N.N., Junker, U., Moser, B., Zak, O., O’Reilly, T., 2000. Systemic neutralization of interleukin-8 markedly reduces neutrophilic pleocytosis during experimental lipopolysaccharide-induced meningitis in rabbits. Infect. Immun. 68, 5756–5763. Epelman, S., Bruno, T.F., Neely, G.G., Woods, D.E., Mody, C.H., 2000. Pseudomonas aeruginosa exoenzyme S induces transcriptional expression of proinflammatory cytokines and chemokines. Infect. Immun. 68, 4811–4814. Fleiszig, S.M., Zaidi, T.S., Fletcher, E.L., Preston, M.J., Pier, G.B., 1994. Pseudomonas aeruginosa invades corneal epithelial cells during experimental infection. Infect. Immun. 62, 3485–3493. Fleiszig, S.M., Zaidi, T.S., Preston, M.J., Grout, M., Evans, D.J., Pier, G.B., 1996. Relationship between cytotoxicity and corneal epithelial cell invasion by clinical isolates of Pseudomonas aeruginosa. Infect. Immun. 64, 2288–2294. Fleiszig, S.M.J., Wiener-Kronish, J.P., Miyazaki, V., Vallas, V., Mastov, K.E., Kanada, D., Sawa, T., Yen, T.S.B., Frank, D.W., 1997. Pseudomonas aeruginosa-mediated cytotoxicity and invasion correlate with distinct genotype at the loci encoding exoenzyme S. Infect. Immun. 65, 579 –585. Fuqua, C., Winans, S.C., Greenberg, E.P., 1996. Census and consensus in bacterial ecosystems: the LuxR-LuxI family of quorum-sensing transcriptional regulators. Ann. Rev. Microbiol. 50, 727 –751. Greenberger, M.J., Strieter, R.M., Kunkel, S.L., et al., 1996. Neutralization of MIP-2 attenuates neutrophil recruitment and bacterial clearance in murine Klebsiella pneumonia. J. Infect. Dis. 173, 159–165. Heck, L.W., Morihara, K., Abrahamson, D.R., 1986a. Degradation of soluble laminin and depletion of tissue-associated basement membrane laminin by Pseudomonas aeruginosa elastase and alkaline protease. Infect. Immun. 54, 149 –153. Heck, L.W., Morihara, K., McRae, W.B., Miller, E.J., 1986b. Specific cleavage of human type III and IV collagens by Pseudomonas aeruginosa elastase. Infect. Immun. 51, 115 –118. Hirakata, Y., Kirikae, T., Kirikae, F., Yamaguchi, T., Izumikawa, K., Takemura, H., Maesaki, S., Tomono, K., Yamada, Y., Kamihira, S., Nakano, M., Kitamura, S., 1999. Effect of Pseudomonas aeruginosa exotoxin A on endotoxin-induced tumour necrosis factor production in murine lung. J. Med. Microbiol. 48, 471–477. Huang, S., Paulaukis, J.D., Godleski, J.J., Kobzik, L., 1992. Expression of macrophage inflammatory protein-2 and KC mRNA in pulmonary inflammation. Am. J. Pathol. 141, 981 –988. Kernacki, K.A., Barrett, R.P., Hobden, J.A., Hazlett, L.D., 2000. Macrophage inflammatory protein-2 is a mediator of polymorphonuclear neutrophil influx in ocular bacterial infection. J Immunol. 164, 1037–1045. Kernacki, K.A., Berk, R.S., 1994. Characterization of the inflammatory response induced by corneal infection induced by Pseudomonas aeruginosa. J. Ocul. Pharmacol. 10, 281–288. Kernacki, K.A., Hobden, J.A., Hazlett, L.D., Fridman, R., Berk, R.S., 1995. In vivo bacterial protease production during Pseudomonas aeruginosa corneal infection. Invest. Ophthalmol. Vis. Sci. 36, 1371–1377. Miller, J., 1976. Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, New York. Rollins, B.J., 1997. Chemokines. Blood. 90, 909–928. Rosenbaum, J.T., Planck, S.R., Huang, X.-N., Rich, L., Ansel, J.C., 1995. Detection of mRNA for the cytokines IL-1a and IL-8 in corneas from patients with pseudophakic bullous keratopathy. Invest. Ophthalmol. Vis. Sci. 36, 2151–2155. Rudner, X.L., Kernacki, K.A., Barrett, R.P., Hazlett, L.D., 2000. Prolonged elevation of IL-1 in Pseudomonas aeruginosa ocular infection regulates
231
macrophage-inflammatory protein-2 production, polymorphonuclear neutrophil persistence, and corneal perforation. J. Immunol. 164, 6576–6582. Saleh, A., Figarella, C., Kammouni, W., Marchand-Pinatel, S., Lazdunski, A., Tubul, A., Brun, P., Merten, M.D., 1999. Pseudomonas aeruginosa quorum-sensing signal molecule N-(3-oxododecanoyl)-L -homoserine lactone inhibits expression of P2Y receptors in cystic fibrosis tracheal gland cells. Infect. Immun. 67, 5076–5082. Springer, T.A., 1994. Traffic signal for lymphocyte recirculation and leukocyte emigration: the mutistep paradigm. Cell 76, 301–314. Steuhl, K.P., Doring, G., Theil, H.J., 1989. The significance of bacterial and host factors in corneal infections caused by Pseudomonas aeruginosa. Fortschr. Ophthalmol. 86, 283–289. Taub, D.D., Oppenheim, J.J., 1994. Chemokines, inflammation and immune system. Ther. Immunol. 1, 229–246. Telford, G., Wheeler, D., Williams, P., Tomkins, P.T., Appleby, P., Sewell, H., Stewart, G.S., Bycroft, B.W., Pritchard, D.I., 1998. The Pseudomonas aeruginosa quorum-sensing signal molecule N-(3-oxododecanoyl)-L-homoserine lactone has immunomodulatory activity. Infect. Immun. 66, 36–42. Tessier, P.A., Naccache, P.H., Clark-Lewis, I., Gladue, R.P., Kuldeep, S.N., McColl, S.R., 1997. Chemokine networks in vivo. J. Immunol. 159, 3595–3602. Thakur, A., Willcox, M.D.P., 1998a. Chemotactic activity in tears and bacteria isolated during contact lens induced adverse events. Exp. Eye Res. 66, 129– 137. Thakur, A., Willcox, M.D.P., 1998b. Cytokines and lipid inflammatory mediator profile of human tears of contact lens associated inflammatory diseases. Exp. Eye Res. 67, 9–19. Thakur, A., Willcox, M.D.P., Stapleton, F., 1998. The proinflammatory cytokines and arachidonic acid metabolites in human tears: homeostatic mechanism. J. Clin. Immunol. 18, 61 –70. Williams, R.N., Patterson, C.A., Eakins, K.E., 1983. Quantitation of ocular inflammation: evaluation of polymorphonuclear leukocyte inflammation by measuring myeloperoxidase activity. Curr. Eye Res. 2, 465– 470. Xue, M.L., Thakur, A., Lutze-Mann, L., Willcox, M.D., 2000. Proinflammatory cytokine/chemokine gene expression in human corneal epithelial cells colonized by Pseudomonas aeruginosa. Clin. Exp. Ophthalmol. 28, 197 –200. Xue, M.L., Willcox, M.D.P., Lloyd, A., Wakefield, D., Thakur, A., 2001. Regulatory role of IL-1beta in the expression of IL-6 and IL-8 in human corneal epithelial cells during Pseudomonas aeruginosa colonization. Clin. Exp. Ophthalmol. 29, 171–174. Yamamoto, Y., Okubo, T., Klein, W., Onozaki, K., Saito, T., Friedman, H., 1994. Binding of Legionella pneumophila to macrophages increases cellular cytokine mRNA. Infect. Immun. 62, 3947–3956. Yamamoto, Y., Retzlaff, C., Klein, T., Friedman, H., 1995. Quantiative reverse transcription-PCR analysis of Legionella pneumophila-induced cytokine mRNA in different macrophage populations by highperformance liquid chromatography. Clin. Diagn. Lab. Immunol. 2, 18– 24. Yan, X., Tumpey, T.M., Kunkel, S.L., Oakes, J.E., Lausch, R.N., 1998. Role of MIP-2 in neutrophil migration and tissue injury in the herpes simplex virus-1 infected cornea. Invest. Ophthalmol. Vis. Sci. 39, 1854–1862. Zhu, H., Thuruthyil, S.J., Willcox, M.P.D., 2000. Invasive strains of Pseudomonas aeruginosa are able to cause epithelial cell cytotoxicity that is dependent on bacterial cell density. Clin. Exp. Ophthalmol. 28, 201– 204.
Role and regulation of CXC-chemokines in acute experimental keratitis M.L. Xuea,1, A. Thakura,1, M.D.P. Willcoxa,*, H. Zhua, A.R. Lloydb, D. Wakefieldb a
Cooperative Research Centre for Eye Research and Technology, Cornea and Contact Lens Research Unit, The University of New South Wales, Sydney, NSW 2052, Australia b School of Pathology, The University of New South Wales, Sydney, NSW 2052, Australia Received 29 January 2002; accepted in revised form 2 October 2002
Abstract The aim of this study was to elucidate the expression of chemokines, their role and regulation in bacterial corneal infection using three bacterial strains (Pseudomonas. aeruginosa- invasive, cytotoxic and contact lens induced acute red eye strains) which have been shown to produce three distinct patterns of corneal disease in the mouse. The predominant chemokine expressed in response to all three strains was MIP-2. Prolonged expression of high levels of MIP-2 was associated with increased severity of corneal inflammation. Significantly reduced disease severity upon administration of anti-MIP-2 antibodies suggested that MIP-2 may play an important role in the pathogenesis of Pseudomonas keratitis at least in part by being a major chemoattractant for polymorphonuclear leukocytes (PMN) recruitment. Interestingly, the numbers of bacteria in eyes with neutralized MIP-2 activity did not decrease even though the severity of the disease was decreased. This implies PMNs as the major destructive factor in microbial keratitis. Further, neutralization of IL-1b activity alone using monoclonal antibodies resulted in significant reduction of both MIP-2 and KC activity indicating that chemokine levels were regulated by IL-1b. These studies demonstrate that the regulation of MIP-2 activity may be beneficial in reducing corneal damage during microbial keratitis in rodents and perhaps that regulation of the human homologue of MIP-2, IL-8, may be useful for controlling keratitis in humans. q 2003 Published by Elsevier Science Ltd. Keywords: MIP-2; KC; Pseudomonas keratitis; corneal infection; inflammation
1. Introduction Bacterial corneal infection of the eye remains a significant health problem and presents a major therapeutic challenge. Among the bacterial pathogens, Pseudomonas aeruginosa is the most frequently isolated microorganism in cases of ulcerative keratitis associated with contact lens wear. Keratitis caused by P. aeruginosa is a highly destructive corneal infection (Kernacki and Berk, 1994). Both bacterial exotoxins and proteases (Kernacki et al., 1995), as well as the host’s own inflammatory response primarily derived from stimulated polymorphonuclear leukocytes (PMN) (Steuhl et al., 1989) and possibly inappropriate production of inflammatory cytokines and chemokines, may contribute to corneal damage. Although rapid infiltration of PMN is required to eliminate the infecting bacteria, the continued * Address correspondence to: M.D.P. Willcox, Cooperative Research Center for Eye Research and Technology, The University of New South Wales, Sydney NSW 2052, Australia. E-mail address: [email protected] (M.D.P. Willcox). 1 First and second authors have equal contribution in this paper. 0014-4835/03/$ - see front matter q 2003 Published by Elsevier Science Ltd. DOI:10.1016/S0014-4835(02)00270-1
presence of PMNs implies ongoing inflammation, which appears to promote scarring. The current concept of a multi-step process of leukocyte recruitment envisions chemotactic agonists or chemokines as one of the key effector molecules (Springer, 1994; Baggiolini et al., 1997; Rollins, 1997). The CXC-chemokines such as MIP-2 and KC are chemotactic for neutrophils, while in vitro studies on C –C chemokines (MCP-1 and MIP-1a) have revealed that they are predominantly chemotactic for monocytes and lymphocytes (Taub and Oppenheim, 1994). Understanding the molecular mechanisms of the host response, including induction of defenses required to influence microbial clearance, is important to devise better management strategies to control sight threatening ocular infection and inflammation. The cornea is known to produce chemokines specific for the recruitment and activation of PMNs. Human corneal tissue, including epithelium and stromal keratocytes, express IL-8 (murine homolog MIP-2) and GRO-a (murine homolog KC) mRNA and protein in vivo (Cubitt et al., 1993; Cubitt et al., 1997) and in vitro experiments (Rosenbaum et al.,
222
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
1995; Kernacki et al., 2000). Our previous studies have demonstrated the presence of chemotactic factors in tears collected during eye closure and keratitis (Thakur et al., 1998; Thakur and Willcox, 1998a,b). P. aeruginosa strains isolated from microbial keratitis are either invasive or cytotoxic in nature. In this study three phenotypes of P. aeruginosa (the invasive, cytotoxic and contact lens induced acute red eye (CLARE) strains) which have recently been shown to produce distinct patterns of corneal disease in mice were used (Cole et al., 1998). The invasive and cytotoxic strains were isolated from human microbial keratitis. The CLARE strain (Paer1; which is neither invasive nor cytotoxic) was isolated from the self-limiting condition known as CLARE (Cowell et al., 1998) and is not associated with tissue destruction in the cornea. The invasive or cytotoxic phenotypes of P. aeruginosa appear to be mutually exclusive (Fleiszig et al., 1997); this exclusivity has been shown to be dependent on possession of a number of genes. Invasive strains possess the genes exoS and exoT that encode the ADP-ribosylating toxins ExoS and ExoT (Fleiszig et al., 1996). Cytotoxic strains lack the gene exoS but possess the gene exoU (Fleiszig et al., 1996). The names ‘cytotoxic’ or ‘invasive’ refer to the interactions between epithelial cells and the bacteria. The cytotoxic strains cause a rapid cytotoxicity in the mammalian cells, whereas the invasive strains can remain inside cells for up to 24 hr before the cells show any overt cytopathic effect (Fleiszig et al., 1994). Expression of many of the virulence factors in P. aeruginosa appears to be highly controlled by the signal molecule-dependent cell – cell communication systems known as quorum sensing in a cell density-dependent fashion (Fuqua et al., 1996). Cell density-dependent signal molecules, N-acyl-L -homoserine lactones (AHLs), are fundamental regulatory agents of many bacterial processes. The AHLs produced by Gram-negative bacteria may not only be limited to induction of specific bacterial phenotypes, but may also act directly to control the host immune response through regulation of gene expression in host cells (Telford et al., 1998). The host response associated with each phenotype has not been clearly defined. The order of virulence of these strains was 6294 . 6206 . Paer1. We studied the hypothesis that the magnitude and temporal production of chemokines induced by each phenotype of P. aeruginosa would effect the outcome of corneal disease in a resistant mouse strain (BALB/c). Keeping this in view, we elucidated the expression of chemokines and the regulation of infiltrating cells in the corneas infected with the three different strains of P. aeruginosa and the contribution and regulation of chemokines in corneal pathology. In addition, bacterial virulence factors, which may induce chemokine production, were also investigated.
2. Materials and methods 2.1. Animal model BALB/c mice aged 8– 10 weeks were used in this study. All mice had undergone baseline measurements of corneal integrity, including slit lamp evaluation. Details of corneal transparency and vessel hyperemia were assessed and only those animals which were in the normal range for the above clinical variables were used in the study. ARVO guidelines on the use of animals in experimentation were adhered to and institutional ethics clearance was obtained. 2.2. Bacterial strains and growth conditions Three strains of P. aeruginosa, 6294 (invasive strain), 6206 (cytotoxic strain) and Paer1 (CLARE strain), were used in this study. Both the cytotoxic and invasive strains were isolated from human corneal ulcers and the Paer1 strain was isolated from a non-infectious condition called CLARE. Bacterial cells were grown in 10 ml of Tryptone Soy Broth (Oxoid Ltd., Sydney, Australia) overnight at 378C. Bacteria were harvested and washed three times in sterile PBS (pH 7·4) and re-suspended in PBS at 1 £ 109 cells ml21 (OD660 ¼ 1·0). 2.3. Animal infection Inbred 8– 10 weeks old BALB/c mice were challenged with P. aeruginosa. The central corneal surface of only the left eye was incised (two parallel incisions approximately 2 – 3 mm long and penetrating only the epithelium) with a sterile 27·5 gauge needle. Twelve animals were included at each time point (three eyes for bacterial enumeration, three for PMN quantitation, three for ELISA and three for RNA studies) for each bacterial strain. The mice were anaesthetized with Avertin (125 mg kg21, intraperitoneally) under a stereomicroscope and 5 ml of bacterial suspension (5 £ 106 cells) was applied topically onto the wounded cornea. The right eye of each animal, which served as a control, was scratched but not infected. 2.4. Administration of anti mouse-MIP-2, -KC and -IL-1b monoclonal antibodies Anti-mouse-MIP-2, KC and IL-1b monoclonal antibodies were purchased from R&D Systems (Minneapolis, MN, USA). Anesthetized animals were injected with 10 ml (1 mg ml21) of anti-mouse-MIP-2, KC or IL-1b monoclonal antibody sub-conjunctivally 2 hr before infection with the invasive strain of P. aeruginosa 6294. Control mice received an equal volume of isotype matched control antibody at the same time point and were also infected. Twelve animals were included at each time point (three eyes for bacterial enumeration, three for PMN quantitation, three
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
223
Table 1 PCR primers for cytokines and chemokines Size of PCR product
Primer sequences
MIP-2 (221 bp)
S: 50 GCC AGT GAG CTG CGC TGT CAG TGC-30 S: 50 GCC AAT GAG CTG CGC TGT CAG TGC-30 S: 50 GGC TCG CTC GGT GAC CCT AGT CTT T-30
KC (207 bp) BMG (300 bp)
A: 50 GTT AGC CTT GCC TTT GTT CAG TAT G-30 A: 50 CTT GGG GAC ACC TTT TAG CAT CTT-30 A: 50 TCT GCA GGC GTA TGT ATC AGT CTC A-30
S: sense primer, A: anti-sense primer.
for ELISA and three for histology). These experiments were repeated at least three times to ensure reproducibility. 2.5. Bacterial enumeration Clearance of P. aeruginosa in the infected cornea was monitored by plate counts of viable bacteria in corneal homogenates. Three eyes at 4 hr, 8 hr, 16 hr, 1 d, 3 d, 5 d and 7 d time points were assessed for the dynamics of P. aeruginosa removal from the corneas. Small aliquots (20 ml in duplicate) of serial dilutions were plated onto the nutrient agar plates. Plates were incubated for 24 hr at 378C. Results were expressed as mean CFU/cornea ^ S.D. 2.6. Measurement of myeloperoxidase (MPO) activity Samples were assayed for MPO activity as previously described (Williams et al., 1983). Whole eyes enucleated at 1 hr, 4 hr, 8 hr, 16 hr, 1 d, 3 d, 5 d, 7 d were homogenized in 1 ml of hexadecyl trimethylammonium bromide (HTAB) buffer (0·5% HTAB in 50 mM phosphate buffer, pH 6·0) and sonicated for 10 sec in an ice bath. The samples were freezethawed three times and centrifuged at 14 000 rpm for 20 min. Supernatant (0·1 ml) was mixed with 2·9 ml of 50 mM phosphate buffer (pH 6·0), containing 0·167 mg ml21 O-dianisidine hydrochloride and 0·0005% hydrogen peroxide. The change in absorbance at 460 nm was continuously monitored for 5 min. Two eyes were used at each time point, and experiments were repeated at least two times. One unit of MPO activity was determined to be equivalent to approximately 2 £ 105 PMN cells ml21. 2.7. Reverse transcription-PCR for chemokine mRNA Total RNA was extracted from mouse eyes by using TRI reagent (Sigma, St Louis, USA). First-strand cDNA synthesis (First-strand cDNA synthesis kit from Promega, Madison, WI, USA) from 2 mg total RNA was accomplished by reverse transcription using oligo(dT)15 as primer. Specific primers and product lengths are listed in Table 1. Primers were based on published sequences (Huang et al., 1992; Yamamoto et al., 1995) or from gene sequences in Genbank (www.ncbi.nlm.nih.gov/Genbank). Polymerase chain reaction was performed as described
previously (Yamamoto et al., 1994, 1995). Briefly, 4 ml of cDNA product for MIP-2 was amplified in the presence of primer sets, dNTPs mixture, and Taq DNA polymerase in a final volume of 50 ml. The reaction mixture was overlaid with mineral oil, and PCR was performed in PC-960G Microplate Gradient Thermal Cycler (Corbett Research, Sydney, NSW, Australia) for 30 cycles. The first cycle consisted of a 5 minute denaturation at 94 8C, a 5 min annealing at 608C, and then 30 cycles each of 1·5 min at 728C, 45 sec at 948C, and 45 sec at 608C, with a final extension of 10 min at 728C. The amplification of KC gene fragment was performed with 2·5 U AmpliTaq polymerase in 25 mM /Tris/HCl (pH 8·3), 50 mM l21 KCl, 2·5 mM l21MgCl2, 0·01% gelatin, 5 mM /DTT, and 1 mmol l21 each primer, each cycle consisted of denaturation at 948C 1 min , annealing at 558C 2 min, and extension at 728C 3 min. PCR was performed with two primer sets (b2-microglobulin (BMG)-specific primers as an endogenous standard and cytokine-specific primers) in the same tube. RT-PCR products were separated by 3% agarose gel electrophoresis, detected by ethidium bromide staining and analyzed by Quantity One software (Bio-Rad, Sydney, NSW, Australia). The relative mRNA level was obtained by comparing the optical density of cytokine-specific bands to the internal standard. 2.8. Chemokine protein determination by ELISA Cytokine levels were measured in ocular homogenates of scratch control and challenged eyes at different time points using ELISA kits (R&D Systems, Minneapolis, MN, USA). Samples for ELISA were prepared by homogenizing mouse eyes in sterile PBS. Homogenates were centrifuged at 12,000 rpm for 20 min at 48C. The resulting supernatants were used to quantitate MIP-2 and KC proteins. Samples, diluted 1:5 in sample diluting buffer, were added in duplicate wells. Samples were analysed following the manufacturer’s instructions. 2.9. Immunostaining for MIP-2 and KC Whole mouse eyes were enucleated and covered with tissue freezing medium OCT (Pro SciTech, Qld, Australia), snap-frozen by immersing into liquid nitrogen and 5 mm
224
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
thick sections were cut. Sections were fixed in freshly prepared 4% paraformaldehyde for 8 min, rinsed in phosphate-buffered saline (PBS, 0·06 M , pH 7·2) and air dried. The sections were stained following the manufacturer’s instructions for staining (HRP – DAB System, R&D Systems, Minneapolis, MN, USA). The cytokine-specific biotin labelled monoclonal antibodies to MIP-2 and KC (R&D Systems, Minneapolis, MN, USA) were used at a concentration of 5 mg ml21 in PBS (0·01 M , pH 7·2) containing 1% BSA and 0·1% Saponin (Sigma), while control samples were treated in the same fashion using isotype matched control antibody. After staining with diaminobenzidine the sections were counter-stained with 0·1% (w/v) Toluene Blue a for few seconds, mounted with DPX (BDH Laboratory Supplies, Poole, UK) and observed using light microscopy by a masked observer. 2.10. Collection of corneas for histological examination Mouse eyes were enucleated at different time points (a suture was taken as an orientation mark) and fixed in 2·5% glutaradehyde (in PBS buffer, pH 7·4) for 2 d. After completely washing with PBS three times for 10 min each, eyes were kept in fresh PBS overnight (o/n) to remove all the fixative, then dehydrated in a graded ethanol series (30, 50, 70, 95%) for 15 min each. Following dehydration, eyes were infiltrated in 95% ethanol and pure resin (1:1) o/n at 48C, and transferred to pure resin o/n at 48C again. Finally, eyes were embedded at embedding mould, sectioned with a microtome (Leica RM 2155, Germany) in sections of 2 mm thickness. Specimens were stained with 0·1% (w/v) Toluene Blue for 7 min, mounted with DPX mountant (ProSciTech, Sydney, NSW, Australia) and observed under a microscope. At least two eyes were examined at each time point by a masked observer. 2.11. Bacterial virulence factors 2.11.1. Zymography for elastase and alkaline protease in P. aeruginosa Bacterial proteases were separated on the basis of molecular size by electrophoresis through a sodium dodecyl sulfate (SDS)-polyacrylamide gel within which 0·1% (w/v) gelatin was immobilized. Subsequently, the position of each enzyme in the gel was visualized by its ability to degrade the substrate, which was detected by staining the remaining gelatin. After overnight culture of P. aeruginosa, in Tryptone Soy Broth (Oxoid Ltd., Sydney, Australia) at 378C, culture supernatants were collected by centrifugation and used for quantitation of protein using the Bradford assay (Bio-Rad, Sydney, NSW, Australia). Approximately 10 mg of each sample was loaded per lane into the wells of each gel (10% polyacrylamide gel containing 0·1% gelatin). Electrophoresis was performed with a Tris –glycine running buffer at 125 V constant voltage for 1·5– 2 hr. The gel was removed and incubated for 1 hr at room temperature in 100 ml of
development buffer (50 mM Tris base, 40 mM HCl, 200 mM NaCl, 5 mM CaCl2 and 0·2% Brij35) at 378C for 14 –18 hr. Staining was performed with 100 ml of 0·5% Coomassie blue G-25 in 30% methanol and 10% acetic acid for 1 hr, and the gel was then de-stained with four changes of solutions. 2.11.2. Production of acyl homoserine lactones The level of acyl homoserine lactones (AHLs) in bacterial culture supernatants was quantified by using the reporter strain Agrobacterium tumefaciens A136 following the method previously described (Zhu, et al., 2000). A. tumefaciens traG < lacZ/traR reporter detects 3oxo-substituted AHL derivatives with acyl chain length from 4 – 12 carbons particularly OdDHL. The AHL reporter strain A. tumefaciens A136 was cultured on a supplement medium (Miller, 1976) at 308C. The levels of AHLs in the supernatants were examined as the ability to activate traR in a b-galactosidase reporter strain A. tumefaciens A136. 2.12. Statistical analysis Statistical analysis of data was performed using a one way ANOVA to test for differences in chemokine gene and protein expression in the corneas infected with the three different strains.
3. Results 3.1. Bacterial clearance Viable bacterial counts of infected mouse eyes were performed at various time points from 1 h to 7 d post challenge. Although bacterial numbers decreased transiently at 4 hr in the corneas infected with the invasive and cytotoxic strains, they increased again at 16 hr post challenge. All mice challenged with the invasive and cytotoxic bacterial strains showed a rapid increase in bacterial numbers which peaked 1d post challenge, dropped significantly ( p , 0·04) by day 3 and declined further by day 5 and 7 post challenge. Corneas inoculated with the CLARE strain showed a continuous decrease in bacterial numbers. Bacteria could not be recovered 16 hr post challenge from corneas inoculated with the CLARE strain. Results are presented in Fig. 1. 3.2. PMN infiltration We have used whole-eye homogenates for quantitation of PMNs, as recruitment of inflammatory cells are not only restricted to the cornea but other anterior segment tissues such as limbus and anterior chamber are also involved. Comparison of MPO activity at different time points in whole mouse eyes challenged with the invasive and cytotoxic P. aeruginosa revealed a significant infiltration of PMNs compared to scarified control corneas after 8 hr
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
225
Fig. 1. Bacterial numbers in whole eyes challenged with invasive, cytotoxic and CLARE bacterial strains and scratch control at 1 hr, 4 hr, 8 hr, 16 hr, 1 d, 3 d, 5 d, and 7 d post challenge. Results are presented as mean log10 bacterial number ^ S.E.M /eye (CLARE, CLARE strain; cytotoxic, cytotoxic strain; invasive, invasive strain).
post challenge ( p , 0·004). Corneas infected with the invasive and cytotoxic strains showed peak MPO activity at 1 d post challenge. MPO activity in response to the cytotoxic strain was significantly lower at all time points after 8 hr post challenge (16 hr p ¼ 0·003; p ¼ 0·01; 1 d p ¼ 0·005; 3 d p ¼ 0·005; 5 d p ¼ 0·001; 7 d p ¼ 0·002) compared to the corneas inoculated with the invasive strain. In response to the CLARE strain, peak activity was observed at 8 hr post infection and this remained constant up to 1 d post challenge. Results are presented in Fig. 2. 3.3. Kinetics of chemokine mRNA expression Corneas infected with the invasive or cytotoxic bacterial strains expressed MIP-2 and KC transcript as early as 1 hr
post challenge (Fig. 3). MIP-2 and KC mRNAs were significantly higher in the corneas infected with either the invasive (MIP-2 p ¼ 0·003; KC p ¼ 0·0001) or the cytotoxic (MIP-2 p ¼ 0·001; KC p ¼ 0·0001) strains, at 8 hr post challenge, compared to the CLARE strain. Corneas challenged with the invasive strain showed significantly higher ( p , 0·02) levels of MIP-2 mRNA compared to the corneas infected with the cytotoxic strain. Both MIP-2 and KC mRNA declined by 7d post challenge in response to both the invasive and cytotoxic strains (Fig. 3). In response to the CLARE strain, the transcripts of MIP-2 peaked at 16 hr post inoculation and declined by 1 d post challenge. In contrast, KC mRNA peaked at 8 hr post challenge and declined by 16 hr post challenge (Fig. 3). 3.4. Kinetics of chemokine protein expression
Fig. 2. MPO activity in whole eyes challenged with the invasive, cytotoxic or CLARE strains and scratch control at 1 hr, 4 hr, 8 hr, 16 hr, 1 d, 3 d, 5 d, and 7 d post challenge. Results are reported as mean log10 MPO activity ^ S.E.M. /eye. Note that the horizontal axis is a non-linear scale. (SC, scratch control; CLARE, CLARE strain; cytotoxic, cytotoxic strain; invasive ¼ invasive strain).
In the murine model of keratitis induced by the invasive bacterial strain, neutrophil chemoattractant MIP-2 was highly up-regulated and to a lesser extent so was KC. In response to the invasive strain, MIP-2 protein was produced as early as 4 hr post challenge, peaked at 16 hr and declined by 1 d post challenge. MIP-2 levels were significantly higher in response to the invasive strain at all time points, while KC levels were significantly higher at 16 hr and 1 d post challenge compared to the corneas inoculated with either the cytotoxic (MIP-2, p , 0·0001 KC, p , 0·0001) or CLARE (MIP-2, p , 0·0001, KC, p , 0·0001) strains. In response to the cytotoxic strain, the kinetics of chemokine appearance were different compared to the invasive strain. Both MIP-2 and KC levels peaked at 1 d post challenge (Fig. 3). In response to CLARE strain, levels of the chemokines peaked at 16 hr post challenge and were significantly lower compared to the invasive ( p , 0·0001) and cytotoxic ( p , 0·0001) strains (Fig. 3).
226
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
Fig. 3. Top panel. Time dependent expression of chemokine mRNA in mouse corneas at 1 hr, 4 hr, 8 hr, 16 hr, 1 d, 3 d, 5 d, and 7 d post challenge with invasive, cytotoxic or CLARE bacterial strains. Results are presented as mean relative intensity (^ S.E.M. ) of each chemokine band divided by intensity of the control and normalized by respective GAPDH band in the same lane. Bottom panel. Time dependent expression of chemokine protein in mouse corneas at 1 hr, 4 hr, 8 hr, 16 hr, 1 d, 3 d, 5 d, and 7 d post challenge with invasive, cytotoxic or CLARE bacterial strains. Results are presented as pg chemokine protein (^ S.E.M. )/eye. Note that horizontal axis is a non-linear scale. (SC, scratch control; CLARE, CLARE strain; cytotoxic, cytotoxic strain; invasive ¼ invasive strain).
3.5. MIP-2 and KC are produced by both resident corneal cells and infiltrating PMNs
3.6. Neutralization of MIP-2 or KC activities during infection with invasive strain of P. aeruginosa
Immunohistochemistry was used to evaluate the types of cells that produced each chemokine during the infection process. Immunoreactivity for MIP-2 and KC was initially detected in corneal epithelial and endothelial cells (4 hr). Later (24 hr) during the infection both MIP-2 and KC were largely immuno-localized in infiltrating cells (PMNs) in the corneal stroma (Fig. 4).
3.6.1. Histopathologic changes in mouse corneas injected with anti-MIP-2 or -KC before P. aeruginosa infection Histopathological examination of corneas injected with anti-MIP-2 monoclonal antibodies before infection with invasive bacterial strain showed markedly reduced infiltrates in the corneal stroma but large number of PMNs were present in the anterior chamber at 1 d post challenge
Fig. 4. Immuno-localization of MIP-2 and KC in the corneas at 1 d post challenge with invasive bacterial strain. Frozen sections stained for MIP-2 protein (25 £ ) in the corneal epithelium and infiltrating cells both intra-cellularly and extra-cellularly (A) and KC protein (25 £ ) largely stained extracellularly in the corneal epithelium and stroma around the infiltrating cells (B). When frozen sections were incubated with isotype control monoclonal antibody no signals were observed (C). (Ep, corneal epithelium; S, corneal stroma.
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
227
Fig. 5. Histological examination of mouse corneas at 1 d post challenge. (A) Corneas treated with anti-MIP-2 monoclonal antibodies before infection with invasive strain showed markedly reduced infiltrates (arrowhead) in the corneal stroma but large numbers were in the anterior chamber (arrow) at 1d post challenge (10 £ ). The original scratch site had healed and there was no epithelial defect present at 1 d post challenge. (B) Corneas treated with anti-KC monoclonal antibodies before infection with the invasive strain showed reduced infiltrates (arrow) in the corneal stroma compared to control corneas (arrows) at 1 d post challenge (10 £ ). The original scratch site had somewhat healed at 1d post challenge. (C) Isotype control antibody injected corneas before infection showed prominent cellular infiltration (arrows), with marked edema (5 £ ). The epithelium was completely destroyed in the central cornea (open arrows) at 1 d post challenge and large number of infiltrates are also present in the anterior chamber (Ep, epithelium; S, stroma; E, endothelium; AC, anterior chamber).
(Fig. 5(A)) compared to control antibody treated corneas. There was no epithelial defect present. On day seven corneas appeared almost normal. Corneas injected with anti-KC monoclonal antibodies showed reduced infiltration in the corneal stroma compared to control corneas (Fig. 5(B)). Corneas injected with isotype control antibody sub-conjunctivally before infection showed prominent cellular infiltration with complete denudation of corneal stroma in the central cornea at 1d post challenge. The corneal stroma was markedly edematous and a large number of infiltrates were present in the anterior chamber (Fig. 5(C)). On day 7, in both antiKC monoclonal antibody or isotype control antibody treated animals, the corneas had recovered noticeably compared to 1d, although there was still large number of PMNs evident in the corneal stroma and anterior chamber. The epithelium had healed.
isotype control antibody treated animals at any time point. Mice injected with anti-KC antibodies showed a slight increase in bacterial number at later time points, however, the difference was not statistically significant (Fig. 6). 3.6.3. PMN infiltration PMN numbers quantified by measuring MPO activity showed significantly lower ( p , 0·001) numbers of PMNs in anti-MIP-2 treated mice compared to both control antibody treated ( p , 0·0001) and anti-KC treated ( p , 0·0002) mice at all time points (Fig. 7). Neutralization of KC protein also showed reduced MPO activity at 1 and 3 d post challenge. However, MPO activity increased significantly on day 7 ( p , 0·02) post challenge compared to control antibody treated animals (Fig. 7).
3.6.2. Bacterial enumeration Administration of anti-MIP-2 antibodies did not result in alteration in bacterial numbers. There was no significant difference found between anti-MIP-2 antibody treated or
3.6.4. Chemokine proteins The MIP-2 protein levels were reduced significantly in anti-MIP-2 treated mice as expected compared to control antibody treated animals at all time points. Mice treated with anti-KC antibodies did not show any significant difference in MIP-2 levels at 1 and 3 d post challenge, but
Fig. 6. Bacterial enumeration in the whole eye of anti-MIP-2, anti-KC and isotype control monoclonal antibody treated animals challenged with invasive strain at 1–7 d post challenge. Results are presented as mean log10 bacterial number ^ S.D. /eye.
Fig. 7. MPO activity in whole eyes of anti-MIP-2, anti-KC and isotype control monoclonal antibody treated animals challenged with invasive strain at 1–7 d post challenge. Results are reported as mean log10 MPO activity ^ S.D. /eye.
228
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
Fig. 8. Kinetics and levels of chemokine proteins in the corneas of anti-MIP-2, anti-KC and isotype control monoclonal antibody treated animals challenged with invasive strain at 1, 3, 5, and 7 d post challenge.
levels rose significantly at 5 d post challenge compared to controls (Fig. 8). KC protein levels were significantly reduced in anti-KC treated animals at 1 d but levels rose at 3 and 5 d post challenge compared to control mice. Interestingly, KC levels were significantly up-regulated at 1 d post challenge in anti-MIP-2 antibody treated mice compared to both controls and anti-KC antibody treated mice (Fig. 8). 3.6.5. MIP-2 is regulated by the pro-inflammatory cytokine IL-1b MIP-2 protein levels were significantly reduced in antiIL-1b treated mice. There was approximately 18-fold ( p ¼ 0·002) less MIP-2 at 1 d, 22-fold ( p ¼ 0·002) less MIP-2 at 3 d and about 25-fold ( p ¼ 0·001) less MIP-2 at 5 d in the IL-1b mAb treated corneas compared to isotype control antibody treated corneas (Fig. 9(A)). Levels of KC were also reduced in anti-IL-1b treated mice by approximately 5-fold ( p , 0·04) at 1 d post challenge. However, there was no significant difference found at other time points (Fig. 9(B)).
3.7. Bacterial virulence factors 3.7.1. Production of proteases In order to determine whether the invasive strain could produce proteases capable of destroying corneal stroma, the culture supernatants were examined by zymography. Zymography of culture supernatants of P. aeruginosa showed presence of considerable amounts of proteases especially elastase (145 kDa) and alkaline protease (53 kDa) in the invasive strain compared to cytotoxic strain (Fig. 10(A)). 3.7.2. Production of N-acyl-L -homoserine lactones The invasive strain produced significantly higher levels of AHLs compared to cytotoxic strain as detected by its ability to activate traR in a b-galactosidase reporter strain A. tumefaciens A136 (Fig. 10(B))
4. Discussion This study demonstrates for the first time that MIP-2 is the predominant chemokine expressed during corneal
Fig. 9. Kinetics and levels of MIP-2 (a) and KC (b) proteins in the corneas of anti-IL-1b and isotype control monoclonal antibody treated animals challenged with invasive strain at 1 d post challenge.
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
Fig. 10. (A) Demonstration of protease activity in the culture supernatant of the invasive and cytotoxic strains of P. aeruginosa. Culture supernatant showing elastase at 145 kDa band and alkaline protease at 53 kDa band. (B) Demonstration of AHL production by measuring beta-galactosidase activity in the supernatants of the invasive and cytotoxic strains.
infection by three distinct types of P. aeruginosa that produce distinct pathologies in the cornea (Cole et al., 1998) and have been shown to be genetically or phenotypically distinct (Fleiszig et al., 1994, 1996, 1997). Furthermore, this study has confirmed that subsequent to MIP-2 expression, the recruitment and activation of PMNs is the major cause of corneal destruction. Neutralization of MIP-2 (during corneal infection induced by the invasive P. aeruginosa strain) and the consequent reduction in PMN recruitment into the cornea resulted in almost clear corneas even though bacterial numbers were high. This demonstrates that an invasive P. aeruginosa strain which was shown to produce appreciable amounts of both elastase and alkaline protease in vitro was not responsible for the major part of the corneal destruction during microbial keratitis in the mouse model. MIP-2 is a neutrophil chemoattractant (Baggiolini et al., 1994; Taub and Oppenheim, 1994; Rollins, 1997) and primarily involved in PMN recruitment during acute corneal inflammation. Peak levels of MIP-2 were detected in severe infection induced by the invasive strain, and to a lesser extent in corneas infected by the cytotoxic strain. In contrast, in response to the CLARE strain which induced only a mild
229
inflammatory response, significantly lower levels of MIP-2 were induced. Further we examined cause and effect relationships between MIP-2/KC and corneal disease. Previous studies have shown that antibody neutralization of endogenous MIP-2 can alter the host response in animal models of bacterial infection (Greenberger et al., 1996; Yan et al., 1998; Dumont et al., 2000). In vivo neutralization of MIP-2, using anti-MIP-2 monoclonal antibodies, substantially reduced the severity of corneal disease induced by the invasive bacterial strain (which produced high levels of MIP-2 and severe corneal disease) and resulted in reduced PMN infiltration as shown by MPO activity and histopathological examination. Collectively these results indicate that both the magnitude and kinetics of appearance of MIP-2 may effect the severity of corneal disease, primarily by regulating the infiltration of PMNs and their activation. A recent study by Kernacki et al. (2000) has also shown that MIP-2 is the major chemokine responsible for the initial recruitment and persistence of neutrophils in Pseudomonas keratitis. The major differences between our study of MIP-2/KC neutralization and that by Kernacki et al. (2000) are that they used a susceptible mouse strain (C57BL/6, cornea perforates), the cytotoxic P. aeruginosa strain-19660 and a systemic intra-peritoneal injection of anti-MIP-2 polyclonal antibodies, whereas in the present study we used a resistant mouse strain (BALB/c, cornea heals), an invasive strain and local subconjunctival injection of monoclonal antibodies. Importantly, since both the cytotoxic and invasive strains are isolated from the Pseudomonas keratitis cases, it was critical to examine whether neutralization of MIP-2 after corneal infection with an invasive strain lead to reduced corneal disease. The results of our study indicate the potential for antiCXC chemokine therapies as an adjunct in cases of Pseudomonas keratitis caused by either of the lineage of P. aeruginosa that have been isolated from microbial keratitis. Further work would need to be performed in this area to demonstrate effectiveness of anti-chemokine therapy once infection had been initiated. In vivo neutralization of KC using anti-KC antibodies showed significantly reduced MPO activity on day 1 and 3, although histological examination revealed large number of infiltrating cells (similar to untreated controls) in the central cornea. One explanation for this could be that PMNs at the site of infection have undergone necrosis/apoptosis, and KC may have some role in delaying necrosis/apoptosis. Later during the infection when KC levels returned back to untreated control levels on day 5 and 7, MPO activity also increased (equivalent to untreated controls) but by this time point the number of bacteria had increased significantly compared to untreated controls. Interestingly, KC levels were significantly higher in anti-MIP-2 treated mice but this did not seem to compensate for MIP-2 as significantly reduced number of PMNs (as judged by MPO activity and histology) were present in these animals. This further
230
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231
indicates that KC may not involved in PMN recruitment and that MIP-2 is the dominant chemokine involved in the PMN recruitment in bacterial corneal infection. MIP-2 has been shown to be a major chemokine in HSV-1 infected corneas and neutrophil chemoattractant KC did not compensate when MIP-2 was neutralized, even though the concentration of KC was more than two times higher than MIP-2 (Yan et al., 1998). These authors also reported that KC was 3– 4 times less potent compared to MIP-2 in promoting PMN migration in vitro. Tessier et al. also observed that KC was a less potent chemoattractant than MIP-2 in a murine subcutaneous air pocket model of inflammation (Tessier et al., 1997). Our in vitro studies have shown that MIP-2 is regulated by IL-1b at both the transcriptional and translational levels in a human corneal epithelial cell line (Xue et al., 2000, 2001). Thus we sought to determine whether MIP-2 expression was regulated by IL-1b in this keratitis model. IL-1b protein was neutralized using anti-IL-1b monoclonal antibodies before corneal infection and as a consequence MIP-2 levels were significantly reduced as was the infiltration of PMNs in the corneal stroma. This suggests that MIP-2 expression is largely regulated by IL-1b in the corneas infected by an invasive strain of P. aeruginosa. These results are in accordance with previous study which showed that IL-1b regulates MIP-2 in susceptible mouse (C57BL/6J) strain for P. aeruginosa corneal infection with a cytotoxic strain (Rudner et al., 2000). The virulence factors in P. aeruginosa are highly regulated by the signal molecule-dependent cell – cell communication systems known as quorum sensing in a cell density-dependent fashion (Fuqua et al., 1996). Cell density-dependent signal molecules, AHLs, are fundamental regulatory agents of many bacterial processes. Interestingly, the invasive strain also produced extremely high levels of AHLs compared to cytotoxic strain. Recently there have been reports that AHLs can cause IL-8 production by respiratory epithelial cells (Telford et al., 1998; Saleh et al., 1999). Corneal damage is also likely to be attributed to the ability of each Pseudomonas strain to produce proteolytic enzymes i.e. LasB elastase, LasA protease, alkaline protease and protease IV (Heck et al., 1986a,b). This study showed that the invasive strain produced larger amounts of proteases compared to the cytotoxic strain. Previous studies in Pseudomonas strains have demonstrated that cytokine production in vivo is correlated to a large extent with the virulence of the pathogen (Hirakata et al., 1999; Epelman et al., 2000). Consistent with these studies, the three strains used in this study differed significantly in the kinetics and magnitude of cytokine induction. A recent study has shown that exoenzyme S (present in the invasive strain, a highly virulent strain) can rapidly induce various cytokines (IL-1a, IL-1b and IL-6) in host cells (Epelman et al., 2000). The cytotoxic strain used in the present study lacks the exoS gene (Fleiszig et al., 1996). It appears likely that the greater magnitude of the cytokine responses produced by
the invasive strain may be related to the capability of key virulence factors in the microorganism to induce cytokines production in the host. It is noteworthy that attenuation of MIP-2 activity only resulted in reduced PMN numbers but did not effect the bacterial numbers. There are several possible mechanisms that could explain the apparent discrepancy between low numbers of neutrophils but reductions in bacterial numbers. Our current hypothesis is that the neutrophils that are recruited into the cornea of anti-chemokine-treated mice retain their ability to phagocytose bacteria and are sufficient to reduce the bacterial numbers. The large numbers of neutrophils that are recruited into normal corneas (or corneas treated with irrelevant antibodies) are not only involved in bacterial clearance but also in destruction of the cornea. Based on these findings we speculate that continued up-regulation of MIP-2 is the major factor contributing to severe corneal disease induced by invasive and cytotoxic Pseudomonas strains. Whereas the CLARE strain, which induced mild inflammation, showed lower levels of MIP-2 expression. In conclusion, our data suggests that the severity of infection (bacterial clearance/PMN infiltration) is associated with the magnitude of expression of chemokine MIP-2 but not KC. These findings point to the therapeutic potential of using both anti-MIP-2 and anti-IL-1b antibodies as possible treatment strategies in bacterial keratitis.
Acknowledgements This research was supported by the National Health and Medical Research Council. The authors would like to thank Reg Wong for excellent statistical analysis, Wen Wang for technical assistance, Denise Lawler and Robyn Lawler for animal handling.
References Baggiolini, M., Dewald, B., Moser, B., 1994. Interleukin-8 and related chemotactic cytokines: CXC and CC chemokines. Adv. Immunol. 55, 97 –179. Baggiolini, M., Dewald, B., Moser, B., 1997. Human chemokines: an update. Ann. Rev. Immunol. 15, 675–705. Cole, N., Willcox, M.D.P., Fleiszig, S.M.J., Stapleton, F., Bao, S., Tout, S., Husband, A., 1998. Different strains of Pseudomonas aeruginosa isolated from ocular infections or inflammation display distinct corneal pathologies in an animal model. Curr. Eye Res. 17, 730 –735. Cowell, B.A., Willcox, M.D.P., Hobden, J.A., Schneider, R.P., Tout, S., Hazlett, L.D., 1998. An ocular strain of Pseudomonas aeruginosa is inflammatory but not virulent in the scarified mouse model. Exp. Eye Res. 67, 347 –351. Cubitt, C.L., Lausch, R.N., Oakes, J.E., 1997. Differential induction of GROa (gene expression in human corneal epithelial cells and keratocytes exposed to proinflammatory cytokines. Invest. Ophthalmol. Vis. Sci 38, 1149– 1158. Cubitt, C.L., Tang, Q., Monteiro, C.A., Lausch, R.N., Oakes, J.E., 1993. IL-8 gene expression in cultures of human corneal epithelial cells and keratocytes. Invest. Ophthalmol. Vis. Sci. 34, 3199–3206.
M.L. Xue et al. / Experimental Eye Research 76 (2003) 221–231 Dumont, R.A., Car, B.D., Voitenok, N.N., Junker, U., Moser, B., Zak, O., O’Reilly, T., 2000. Systemic neutralization of interleukin-8 markedly reduces neutrophilic pleocytosis during experimental lipopolysaccharide-induced meningitis in rabbits. Infect. Immun. 68, 5756–5763. Epelman, S., Bruno, T.F., Neely, G.G., Woods, D.E., Mody, C.H., 2000. Pseudomonas aeruginosa exoenzyme S induces transcriptional expression of proinflammatory cytokines and chemokines. Infect. Immun. 68, 4811–4814. Fleiszig, S.M., Zaidi, T.S., Fletcher, E.L., Preston, M.J., Pier, G.B., 1994. Pseudomonas aeruginosa invades corneal epithelial cells during experimental infection. Infect. Immun. 62, 3485–3493. Fleiszig, S.M., Zaidi, T.S., Preston, M.J., Grout, M., Evans, D.J., Pier, G.B., 1996. Relationship between cytotoxicity and corneal epithelial cell invasion by clinical isolates of Pseudomonas aeruginosa. Infect. Immun. 64, 2288–2294. Fleiszig, S.M.J., Wiener-Kronish, J.P., Miyazaki, V., Vallas, V., Mastov, K.E., Kanada, D., Sawa, T., Yen, T.S.B., Frank, D.W., 1997. Pseudomonas aeruginosa-mediated cytotoxicity and invasion correlate with distinct genotype at the loci encoding exoenzyme S. Infect. Immun. 65, 579 –585. Fuqua, C., Winans, S.C., Greenberg, E.P., 1996. Census and consensus in bacterial ecosystems: the LuxR-LuxI family of quorum-sensing transcriptional regulators. Ann. Rev. Microbiol. 50, 727 –751. Greenberger, M.J., Strieter, R.M., Kunkel, S.L., et al., 1996. Neutralization of MIP-2 attenuates neutrophil recruitment and bacterial clearance in murine Klebsiella pneumonia. J. Infect. Dis. 173, 159–165. Heck, L.W., Morihara, K., Abrahamson, D.R., 1986a. Degradation of soluble laminin and depletion of tissue-associated basement membrane laminin by Pseudomonas aeruginosa elastase and alkaline protease. Infect. Immun. 54, 149 –153. Heck, L.W., Morihara, K., McRae, W.B., Miller, E.J., 1986b. Specific cleavage of human type III and IV collagens by Pseudomonas aeruginosa elastase. Infect. Immun. 51, 115 –118. Hirakata, Y., Kirikae, T., Kirikae, F., Yamaguchi, T., Izumikawa, K., Takemura, H., Maesaki, S., Tomono, K., Yamada, Y., Kamihira, S., Nakano, M., Kitamura, S., 1999. Effect of Pseudomonas aeruginosa exotoxin A on endotoxin-induced tumour necrosis factor production in murine lung. J. Med. Microbiol. 48, 471–477. Huang, S., Paulaukis, J.D., Godleski, J.J., Kobzik, L., 1992. Expression of macrophage inflammatory protein-2 and KC mRNA in pulmonary inflammation. Am. J. Pathol. 141, 981 –988. Kernacki, K.A., Barrett, R.P., Hobden, J.A., Hazlett, L.D., 2000. Macrophage inflammatory protein-2 is a mediator of polymorphonuclear neutrophil influx in ocular bacterial infection. J Immunol. 164, 1037–1045. Kernacki, K.A., Berk, R.S., 1994. Characterization of the inflammatory response induced by corneal infection induced by Pseudomonas aeruginosa. J. Ocul. Pharmacol. 10, 281–288. Kernacki, K.A., Hobden, J.A., Hazlett, L.D., Fridman, R., Berk, R.S., 1995. In vivo bacterial protease production during Pseudomonas aeruginosa corneal infection. Invest. Ophthalmol. Vis. Sci. 36, 1371–1377. Miller, J., 1976. Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, New York. Rollins, B.J., 1997. Chemokines. Blood. 90, 909–928. Rosenbaum, J.T., Planck, S.R., Huang, X.-N., Rich, L., Ansel, J.C., 1995. Detection of mRNA for the cytokines IL-1a and IL-8 in corneas from patients with pseudophakic bullous keratopathy. Invest. Ophthalmol. Vis. Sci. 36, 2151–2155. Rudner, X.L., Kernacki, K.A., Barrett, R.P., Hazlett, L.D., 2000. Prolonged elevation of IL-1 in Pseudomonas aeruginosa ocular infection regulates
231
macrophage-inflammatory protein-2 production, polymorphonuclear neutrophil persistence, and corneal perforation. J. Immunol. 164, 6576–6582. Saleh, A., Figarella, C., Kammouni, W., Marchand-Pinatel, S., Lazdunski, A., Tubul, A., Brun, P., Merten, M.D., 1999. Pseudomonas aeruginosa quorum-sensing signal molecule N-(3-oxododecanoyl)-L -homoserine lactone inhibits expression of P2Y receptors in cystic fibrosis tracheal gland cells. Infect. Immun. 67, 5076–5082. Springer, T.A., 1994. Traffic signal for lymphocyte recirculation and leukocyte emigration: the mutistep paradigm. Cell 76, 301–314. Steuhl, K.P., Doring, G., Theil, H.J., 1989. The significance of bacterial and host factors in corneal infections caused by Pseudomonas aeruginosa. Fortschr. Ophthalmol. 86, 283–289. Taub, D.D., Oppenheim, J.J., 1994. Chemokines, inflammation and immune system. Ther. Immunol. 1, 229–246. Telford, G., Wheeler, D., Williams, P., Tomkins, P.T., Appleby, P., Sewell, H., Stewart, G.S., Bycroft, B.W., Pritchard, D.I., 1998. The Pseudomonas aeruginosa quorum-sensing signal molecule N-(3-oxododecanoyl)-L-homoserine lactone has immunomodulatory activity. Infect. Immun. 66, 36–42. Tessier, P.A., Naccache, P.H., Clark-Lewis, I., Gladue, R.P., Kuldeep, S.N., McColl, S.R., 1997. Chemokine networks in vivo. J. Immunol. 159, 3595–3602. Thakur, A., Willcox, M.D.P., 1998a. Chemotactic activity in tears and bacteria isolated during contact lens induced adverse events. Exp. Eye Res. 66, 129– 137. Thakur, A., Willcox, M.D.P., 1998b. Cytokines and lipid inflammatory mediator profile of human tears of contact lens associated inflammatory diseases. Exp. Eye Res. 67, 9–19. Thakur, A., Willcox, M.D.P., Stapleton, F., 1998. The proinflammatory cytokines and arachidonic acid metabolites in human tears: homeostatic mechanism. J. Clin. Immunol. 18, 61 –70. Williams, R.N., Patterson, C.A., Eakins, K.E., 1983. Quantitation of ocular inflammation: evaluation of polymorphonuclear leukocyte inflammation by measuring myeloperoxidase activity. Curr. Eye Res. 2, 465– 470. Xue, M.L., Thakur, A., Lutze-Mann, L., Willcox, M.D., 2000. Proinflammatory cytokine/chemokine gene expression in human corneal epithelial cells colonized by Pseudomonas aeruginosa. Clin. Exp. Ophthalmol. 28, 197 –200. Xue, M.L., Willcox, M.D.P., Lloyd, A., Wakefield, D., Thakur, A., 2001. Regulatory role of IL-1beta in the expression of IL-6 and IL-8 in human corneal epithelial cells during Pseudomonas aeruginosa colonization. Clin. Exp. Ophthalmol. 29, 171–174. Yamamoto, Y., Okubo, T., Klein, W., Onozaki, K., Saito, T., Friedman, H., 1994. Binding of Legionella pneumophila to macrophages increases cellular cytokine mRNA. Infect. Immun. 62, 3947–3956. Yamamoto, Y., Retzlaff, C., Klein, T., Friedman, H., 1995. Quantiative reverse transcription-PCR analysis of Legionella pneumophila-induced cytokine mRNA in different macrophage populations by highperformance liquid chromatography. Clin. Diagn. Lab. Immunol. 2, 18– 24. Yan, X., Tumpey, T.M., Kunkel, S.L., Oakes, J.E., Lausch, R.N., 1998. Role of MIP-2 in neutrophil migration and tissue injury in the herpes simplex virus-1 infected cornea. Invest. Ophthalmol. Vis. Sci. 39, 1854–1862. Zhu, H., Thuruthyil, S.J., Willcox, M.P.D., 2000. Invasive strains of Pseudomonas aeruginosa are able to cause epithelial cell cytotoxicity that is dependent on bacterial cell density. Clin. Exp. Ophthalmol. 28, 201– 204.