Binding of Pseudomonas aeruginosa to neutral glycosphingolipids of rabbit corneal epithelium

Modulation of corneal and stromal matrix metalloproteinase by the mannose-induced Acanthamoeba cytolytic protein

The involvement of the mannose-induced Acanthamoeba cytopathic protein (MIP-133) in tissue injury and activation of metalloproteinase of corneal and stromal cells was examined in vitro. Activation of MMP-1, MMP-2, MMP-3, and MMP-9 induced by MIP-133 on human corneal epithelial and stromal cell cultures was examined by reverse transcriptase polymerase chain reaction (RT-PCR) and ELISA. MMP-1, MMP-2, MMP-3, and MMP-9 mRNA were expressed in both cultured human corneal epithelial and stromal cells. When the epithelial cells were exposed to MIP-133 protein, the mRNA expression for MMP-1 and MMP-9 was unchanged. However, the transcript for MMP-2 and MMP-3 was decreased by 2-fold. By contrast, the expression of MMP-2 and MMP-3 was significantly upregulated (2- to 4-fold) in the corneal stromal cells 1, 4, and 8h after MIP-133 stimulation. At the protein level, there was no significant difference in the level of MMPs between the corneal epithelial cells before and after stimulation with MIP-133. By contrast, the levels of MMP-2 and MMP-3 were significantly higher in the corneal stromal cells stimulated with MIP-133. The supernatants from corneal stromal cells stimulated with MIP-133 were incubated with PMSF and MIP-133 antibody and the level of MMP-2 was measured by ELISA. Activation of MMP-2 by MIP-133 was inhibited in the supernatants pretreated with the serine protease inhibitor, PMSF, and anti-MIP-133. Supernatants pretreated with the cysteine protease inhibitor E6 or control antibody produced the same amount of MMP-2 as the untreated supernatants. To verify possible homology between MMPs and Acanthamoeba castellanii proteases, the mRNA from A. castellanii was prepared and analyzed for the expression of MMP genes by PT-PCR. The results showed that A. castellanii did not express mRNA for MMP-1, MMP-2, MMP-3, or MMP-9. Thus, A. castellanii mRNA does not cross-react with human MMPs. Furthermore, ELISA was used to determine the cross-reactivity of MMP antibodies with the MIP-133 protein. Monoclonal antibodies against MMPs did not cross-react with either the MIP-133 protein or BSA (negative control antigen). The results indicate that the MIP-133 protein modulates MMP-2 and -3 expression differently in human corneal epithelial and stromal cells.

Environmental factors influence P. aeruginosa binding to the wounded mouse cornea

PURPOSE

To determine whether environmental factors or bacterial viability affect the binding of two strains of P. aeruginosa to mouse cornea.

METHODS

Scarified corneas were placed in organ culture and inoculated with P. aeruginosa cell suspensions containing either ATCC 19660 or PAO1 bacterial strains classed as cytotoxic or invasive, respectively. Eyes were incubated in vitro for 1 h after bacterial application at different pH or temperature conditions or in PBS containing various divalent cations. The adhesion of heat-killed or formalin-fixed bacteria was tested similarly. Scanning electron microscopy (scanning EM) was used to quantitate adherent bacteria.

RESULTS

P. aeruginosa ATCC 19660 showed an increase in binding at pH 8.0, favored higher temperatures and required both calcium and magnesium for optimum binding. Adherence of PAO1 was enhanced at pH 6.5 and decreased at pH 8.0. This strain favored binding at lower temperatures and did not require either divalent cation for optimum binding. In addition, the presence of magnesium ions resulted in reduced binding for this strain. Both strains exhibited less binding ability after formalin fixation or heat killing.

CONCLUSION

Environmental factors and bacterial viability are important factors which influence the ability of both cytotoxic and invasive strains of P. aeruginosa to bind to the scarified cornea.

Mannose induces the release of cytopathic factors from Acanthamoeba castellanii

Acanthamoeba keratitis is a chronic inflammatory disease of the cornea which is highly resistant to many antimicrobial agents. The pathogenic mechanisms of this disease are poorly understood. However, it is believed that the initial phases in the pathogenesis of Acanthamoeba keratitis involve parasite binding and lysis of the corneal epithelium. These processes were examined in vitro, using Acanthamoeba castellanii trophozoites. Parasites readily adhered to Chinese hamster corneal epithelial cells in vitro; however, parasite binding was strongly inhibited by mannose but not by lactose. Although mannose prevented trophozoite binding, it did not affect cytolysis of corneal epithelial cells. Moreover, mannose treatment induced trophozoites to release cytolytic factors that lysed corneal epithelial cells in vitro. These factors were uniquely induced by mannose because supernatants collected from either untreated trophozoites or trophozoites treated with other sugars failed to lyse corneal cells. The soluble factors were size fractionated in centrifugal concentrators and found to be > or = 100 kDa. Treatment of the supernatants with the serine protease inhibitor phenylmethylsulfonyl fluoride inhibited most, but not all, of the cytopathic activity. These data suggest that the binding of Acanthamoeba to mannosylated proteins on the corneal epithelium may exacerbate the pathogenic cascade by initiating the release of cytolytic factors.

The inside story: why contact lens cases become contaminated

Contact lens case contamination has become an enigma, both because its role in the pathogenesis of lens-related keratitis has remained uncertain, and because current contact lens disinfection systems have been ineffective in eliminating it. This lecture reviews the evidence regarding the role of lens case contamination in the pathogenesis of keratitis and examines the reasons for the failure of disinfection systems to minimise lens case contamination.

Pathogenesis of corneal infection: binding of Pseudomonas aeruginosa to specific phospholipids

Clinical isolates of Pseudomonas aeruginosa were examined for binding interactions with phospholipids of corneal epithelium. Thin-layer chromatography (TLC) of lipids extracted from corneal epithelia followed by staining with an ammonium molybdate spray reagent revealed three phospholipid components, PL1, PL2, and PL3. The chromatographic mobility of PL1 was similar to that of the phospholipid standards phosphatidylinositol (PI) and phosphatidylserine (PS), which were not well resolved from one other; PL2 and PL3 comigrated with the standards phosphatidylcholine and phosphatidylethanolamine, respectively. By use of a TLC-bacterial overlay procedure, 35S-labeled P. aeruginosa organisms were shown to bind to PL1 but not to PL2 or PL3. P. aeruginosa binding to PL1 was concentration dependent. Alkaline methanolysis abolished the binding. PL1 was separated into two components, PL1-I and PL1-S, by chromatography on borate-treated TLC plates. Both PL1-I and PL1-S contained binding sites for P. aeruginosa. Mass spectral analysis identified PL1-I and PL1-S as PI and PS, respectively. Radiolabeled P. aeruginosa organisms were subsequently shown to bind to commercially available bovine PI and PS and synthetic dipalmitoyl-PS but not to other phospholipid standards, including bovine SM and PC or synthetic dioleoyl- and distearoyl-PC. A control Escherichia coli strain did not bind to either PS or PI. Tetramethylurea, a disrupter of hydrophobic associations, did not influence the binding of P. aeruginosa to PS or PI. P. aeruginosa bound to the monolayers of corneal epithelial cells. P. aeruginosa binding to the monolayer cultures as well as to rabbit corneas pretreated with exogenous PS and PI was significantly higher than that to those preincubated with PC or medium alone. The data suggest that phospholipids PS and PI present in mucus or on the cell surface may function as P. aeruginosa receptors and contribute to selective bacterium-host interactions responsible for initial colonization.

The effect of cleaning and disinfection of soft contact lenses on corneal infectivity in an animal model

PURPOSE

Bacterial contamination of previously worn soft contact lenses, especially at sites of lens deposits, might play a role in the pathogenesis of lens-associated bacterial keratitis. We studied the effects of three commercial contact lens cleaners and disinfectants in a rabbit model to determine whether cleaning and disinfection reduced infectivity.

METHODS

Duragel 75 soft contact lenses, designed to fit the eyes of rabbits, were worn by rabbits under tarsorrhaphies, then were removed and cleaned in one of three cleaner and disinfectant solutions according to the manufacturers' instructions. The lenses were contaminated by overnight incubation in a suspension of 10(8) Pseudomonas aeruginosa/ml and were placed under tarsorrhaphies on the eyes of fresh rabbits. The rabbits were observed for two weeks for signs of infection. Control rabbits wore new, uncleaned but contaminated lenses or worn, uncleaned but contaminated lenses.

RESULTS

The rates of infection with the three commercial cleaner and disinfectant solutions ranged from 18% (two of 11) to 31% (four of 13); these incidences were not significantly different from one another or from the 19% (three of 16) incidence with new, contaminated but uncleaned lenses. By contrast, when worn, uncleaned but contaminated lenses were placed in rabbits' eyes, seven of eight were infected, a rate that is significantly higher than that of the other four groups (P = .0003).

CONCLUSIONS

These data indicate that the three commercial lens cleaner and disinfectant solutions were of similar efficacy in reducing the infectivity of contaminated contact lenses to a level similar to that of new, unworn lenses.(ABSTRACT TRUNCATED AT 250 WORDS)

Pseudomonas aeruginosa infection of the cornea and asialo GM1

Extensive immunohistochemical and thin-layer chromatogram-immunostain analyses were carried out to establish whether asialo GM1, a glycolipid which contains binding sites for Pseudomonas aeruginosa, is present in corneal epithelium. The data suggest that rabbit corneal epithelium does not contain detectable levels of asialo GM1 even after corneas are scarified and incubated with trypsin, P. aeruginosa, or P. aeruginosa exoproducts to expose potential cryptic sites. Preliminary immunohistochemical analyses indicated that asialo GM1 is also not found in human corneas.

Protection of human respiratory epithelium from Pseudomonas aeruginosa adherence by phosphatidylglycerol liposomes

The ability of phosphatidylglycerol (DSPG) liposomes to prevent adherence of Pseudomonas aeruginosa to primary cultures of non-cystic fibrosis (CF) and delta F508 homozygous CF human respiratory epithelium was studied. The culture model was characterized by the simultaneous presence of various cellular phenotypes: well-differentiated respiratory epithelial cells, ciliated and nonciliated cells, and migrating cells which can be assimilated into a regenerating epithelium after injury. DSPG liposomes significantly decreased the binding of P. aeruginosa to migrating cells of both non-CF and delta F508 homozygous CF cultures compared with control cultures (35.5 x 10(-3) +/- 8.1 x 10(-3) bacteria per micron 2 versus 23.9 x 10(-3) +/- 2.5 x 10(-3); P < 0.01 for non-CF cultures and 88.8 x 10(-3) +/- 17.2 x 10(-3) bacteria per micron 2 versus 29.1 x 10(-3) +/- 0.6 x 10(-3), P < 0.001 for CF cultures). After treatment with DSPG liposomes, the size of P. aeruginosa aggregates bound to migrating cells in both non-CF cultures and delta F508 homozygous CF cultures was significantly decreased (14.4 +/- 3 bacteria per aggregate versus 11.9 +/- 2.5 bacteria per aggregate [P < 0.05] and 29.9 +/- 8.4 bacteria per aggregate versus 17.3 +/- 2.3 bacteria per aggregate [P < 0.01], respectively). Moreover, the control cultures were characterized by a differential P. aeruginosa adherence according to both the cellular phenotype and the mutation. The migrating cells bound more bacteria than the stationary cells of both non-CF and delta F508 homozygous CF cultures. The CF migrating cells bound significantly more bacteria than the non-CF migrating cells (88.8 x 10(-3) +/- 17.2 x 10(-3) bacteria per microns 2 versus 35.5 x 10(-3) +/- 8.1 x 10(-3) bacteria per micron 2, P < 0.001). These results suggest that DSPG liposomes are able to decrease P. aeruginosa adherence to CF and non-CF respiratory epithelium, particularly to migrating cells, which mimic a regenerating epithelium after injury. DSPG liposomes could also represent a hydrophobic barrier limiting the deleterious action of P. aeruginosa exoproducts.

Evidence for asialo GM1 as a corneal glycolipid receptor for Pseudomonas aeruginosa adhesion

Anti-gangliotetraosylceramide (anti-asialo GM1) and antiparagloboside monoclonal antibodies (MAbs) were used in immunofluorescence, immunoelectron-microscopic, and in vitro binding inhibition assays to determine whether either of the glycolipids was detectable in the normal cornea, whether levels changed following corneal scarification and either trypsin treatment or incubation in vitro with Pseudomonas aeruginosa, and whether either of the MAbs could competitively inhibit P. aeruginosa binding to cornea. No immunostaining above background for either glycolipid was observed in frozen, unfixed sections or in lightly fixed, K4M-embedded antibody-gold-labeled thin sections of normal cornea. In frozen sections of organ-cultured scarified cornea, no increased immunostaining for anti-asialo GM1 or antiparagloboside reactivity was noted immediately or 60 min after corneal scarification. However, at 60 min after scarification and in vitro incubation of the eye with either trypsin or P. aeruginosa, enhanced immunostaining for both glycolipids was associated with cells within or immediately adjacent to the wound site. Trypsin increased immunoreactivity in the wound site more markedly compared with incubation with P. aeruginosa, but immunostaining was similarly localized with either treatment. No staining above background was seen in control sections. Similarly, with immunoelectron microscopy, increased immunogold-MAb staining for both glycolipids was seen on the plasma membranes of the wound-site cells of eyes incubated with either trypsin or P. aeruginosa compared with controls that were similarly immunostained but with the primary antibody either omitted or substituted with a nonspecific MAb. Competitive binding inhibition assays, in which the bacterial inoculum or the eye in organ culture was incubated with anti-asialo GM1 MAb prior to topical ocular application of the bacteria, showed significantly decreased P. aeruginosa adhesion compared with preparations similarly treated with phosphate-buffered saline or antiparagloboside MAb. These data provide evidence to support the hypothesis that asialo GM1, not paragloboside, serves as a receptor for P. aeruginosa binding to the scarified cornea of the adult mouse and spatially localizes both glycolipids in the wound site.

Acanthamoebae bind to glycolipids of rabbit corneal epithelium

By use of a thin-layer chromatogram (TLC) overlay procedure, 35S-labeled acanthamoebae were shown to bind to seven glycolipids of rabbit corneal epithelium. Corneal epithelial cells were grown in culture and were subjected to Folch extraction to isolate a chloroform-rich lower phase containing neutral glycosphingolipids (NGSL) and an aqueous upper phase containing gangliosides, i.e., sialic acid-containing glycolipids. Thin-layer chromatography of the upper phase revealed the presence of 10 ganglioside components. Acanthamoebae were shown to bind to four of these components, referred to as 2, 3, 6, and 7. On TLC plates, ganglioside components 2 and 3 migrated slightly ahead of the glycolipid standard GD1a, component 7 comigrated with standard GM3, and component 6 migrated a little more slowly than GM3. Likewise, of the 10 NGSL known to be present in the lower phase, acanthamoebae bound to components 1, 5, and 6. NGSL components 1, 5, and 6 migrated on TLC plates with relative mobilities similar to those of standards asialo GM1, asialo GM2, and ceramidetrihexoside, respectively. We propose that one or more of the Acanthamoeba-reactive glycolipids of corneal epithelium identified in this study may play a role in the pathogenesis of Acanthamoeba keratitis by mediating the adherence of the parasites to the cornea.

Binding of nonmucoid Pseudomonas aeruginosa to normal human intestinal mucin and respiratory mucin from patients with cystic fibrosis

Lung infections due to Pseudomonas aeruginosa and Pseudomonas cepacia are common in patients with cystic fibrosis. Initial colonization is due to nonmucoid P. aeruginosa, while later mucoid variants emerge and are associated with chronic infection. P. cepacia colonization tends to be more prevalent in older patients. The present study was conducted to discover whether highly purified mucins (from cystic fibrosis sputum and control intestinal secretions) exhibited specific binding of nonmucoid P. aeruginosa. In vitro solid phase microtiter binding assays (with or without a blocking agent) as well as solution phase assays were conducted. Bacteria bound to both mucins via bacterial pili, but no differences in binding capacity were noted between the mucins. Unlike P. cepacia (described in the accompanying manuscript) there was also no preferential binding of P. aeruginosa to mucins versus bovine serum albumin, casein, gelatin, or a host of structurally unrelated proteins and glycoproteins. Carbohydrate hapten inhibition studies did not suggest the existence of specific mucin carbohydrate receptors for P. aeruginosa. In solid phase assays a low concentration (0.05 M) of tetramethylurea abolished P. aeruginosa bacterial binding to both mucins as well as to BSA, whereas in solution phase assays mucin binding to bacteria was not completely disrupted by tetramethylurea. Specific monoclonal antipilus antibodies did not inhibit binding to a greater extent than did Fab fragments of normal mouse IgG. Binding of strains PAO1 and PAK (and isolated PAK pili) to buccal epithelial cells was not influenced by the presence of mucin in binding assay mixtures. Our findings do not support the widely held notion that specific mucin receptors are responsible for the attachment of P. aeruginosa pili, nor do they support the idea that there is a competitive interference by mucins of bacterial binding to respiratory cells. In patients with cystic fibrosis, it would seem unlikely therefore that initial colonization of the lungs by P. aeruginosa is due to a 'selective tropism' of these bacteria for respiratory mucin.

Differential adhesion of Pseudomonas aeruginosa to human respiratory epithelial cells in primary culture

Human nasal polyps in outgrowth culture were used to study the Pseudomonas aeruginosa adhesion to respiratory cells. By scanning electron microscopy, P. aeruginosa were seen associated with ciliated cells, but by transmission electron microscopy, bacteria were never seen at the interciliary spaces or attached along cilia, but were identified trapped at the extremities of cilia, usually as bacterial aggregates. A fibronectin-containing fibrillar material was seen associated with aggregated bacteria. By time-lapse video microscopy, bacteria were seen to aggregate in the culture medium following their addition to the culture wells. Progressively, these aggregates were trapped by cilia or attached to migrating cells of a lower cell layer that protruded beneath the upper layer cells, at the outgrowth periphery. P. aeruginosa adhesion to these lower cell layer migrating cells was significantly higher than to ciliated or nonciliated cells of the upper cell layer. Migrating cells were intensely labeled by the complexes Con A and arachis hypogea agglutinin (PNA)-FITC, in contrast to the other cells. The percentage of PNA-labeled cells with attached bacteria was significantly higher than that without bacteria. These results suggest that changes of cell surface glycoconjugates related with cell migration may favor P. aeruginosa adhesion to respiratory cells.

Analysis of adhesion, piliation, protease production and ocular infectivity of several P. aeruginosa strains

The role of bacterial piliation and protease production in Pseudomonas aeruginosa adhesion to the injured corneal epithelial surface and subsequent infectivity was examined using several bacterial strains, including three that were hyperpiliated. To initiate this study, bacteria were examined by transmission EM to confirm their piliation characteristics. The PAK strain, like pseudomonas ATCC 19660, possessed about 1-4 polar pili. The mutant PAK/PR11 lacked pili while PAK/PR1, DB2, a mutant of PAO1, and PA1244, a wild-type clinical isolate, were hyperpiliated. Ocular infectivity of these bacterial strains and mutants was examined macroscopically and histopathologically in mice and these data compared to the well-characterized ocular disease response of a murine model of infection with pseudomonas ATCC 19660. The PAK strain was infective, but less virulent than strain 19660 by both macroscopic grading and histopathological analysis of infected eyes. Infectivity of the PR11 mutant was similar to the PAK parent strain, while PR1, DB2 and 1244, all hyperpiliated, were not infective. To explore the hypothesis that hyperpiliated bacteria bound less well to cornea and thus failed to induce corneal disease, in vitro quantitative studies of bacterial adhesion were done using an ocular organ culture model. The PR1 hyperpiliated mutant bound significantly less well to cornea than the PAK parent strain, PR11 mutant or pseudomonas 19660, while DB2 and 1244 binding did not differ significantly from 19660 or PAK. Examination of protease production, another factor which may influence adhesion, revealed that only 19660 and DB2 produced detectable protease. This study provides evidence that non-piliated, non-protease producing strains such as PAK/PR11 possess alternate virulence mechanisms to facilitate binding to and infectivity of corneal tissue.

Characterization of Pseudomonas aeruginosa adherence to mouse corneas in organ culture

The present study was designed to obtain further information on the nature of the corneal macromolecule(s) to which Pseudomonas aeruginosa adheres and how adherence might be prevented. Scarified adult mouse corneas in organ culture were treated with trypsin or lipase to determine whether the receptor molecule(s) was protein or lipid in nature. Trypsin (20 micrograms/ml) treatment of the cornea for 5 min had no significant effect on bacterial adherence, and longer periods of enzyme exposure resulted in extensive surface cell lysis. In contrast, lipase treatment (50,000 U/ml) for 1 h caused little visible cell lysis and significantly reduced bacterial adherence. To test further the lipid nature of the receptor, a highly purified monosialoganglioside (GM1) preparation (500 micrograms/ml) was used to preincubate (1 h) the cornea prior to bacterial application, and this also inhibited bacterial adherence. Similar corneal treatment with gangliotetraosylceramide (asialo GM1) (500 micrograms/ml) had little effect on ocular bacterial binding. Premixing of the bacterial inoculum with GM1 prior to corneal application had no significant effect on inhibiting bacterial binding, but similarly premixing the bacterial inoculum with asialo GM1 transiently decreased adherence. Lastly, premixing of the bacterial inoculum or preincubation of corneas with fibronectin (500 micrograms/ml for 1 h) both decreased bacterial adherence. These findings provide evidence that the receptor-adhesin interactions of P. aeruginosa at the ocular surface in organ culture are complex, involve a glycolipid moiety, and may be blocked by a ganglioside containing at least one sialosyl residue or by fibronectin, which may bind to membrane-associated gangliosides.