Role of Pseudomonas aeruginosa ExsA in Penetration through Corneal Epithelium in a Novel In Vivo Model

Abstract

purpose. The scarified cornea keratitis model was modified to study Pseudomonas aeruginosa infection of healing corneal epithelium. The new model was then used to study the role of ExsA, a transcriptional activator of P. aeruginosa, in bacterial penetration through injured and healing corneal epithelia.

methods. Scratch-injured corneas of C57BL/6 mice were allowed to heal for 0, 6, 9, or 12 hours before inoculation with a cytotoxic (6206) or invasive (PAO1) P. aeruginosa strain. Disease progression was monitored for 14 days. The integrity of the healing epithelium was studied in uninfected eyes by fluorescein staining and by histologic examination. In other experiments, the effect of bacterial exsA mutation was studied after 0, 6, or 12 hours of healing. Three hours after infection, these eyes were used to quantify early bacterial colonization levels by viable counts, or they were sectioned to study bacterial penetration through the epithelium by microscopy.

results. Corneas remained susceptible to infection 6 but not 12 hours after scratch injury. By 6 hours, the previously exposed stroma was already completely covered by several layers of epithelial cells. Fluorescein staining unexpectedly occurred even after 12 hours of healing time, showing that resistance to infection preceded full restoration of epithelial barrier function. Mutation of exsA reduced both bacterial colonization levels and penetration through the epithelium 3 hours after bacterial inoculation, but only in the 6-hour healing situation, and only for the cytotoxic strain (PA103). Mutation of exsA in the invasive strain (PAO1) had no effect on 3-hour colonization or penetration levels under any circumstances.

conclusions. The 6-hour healing infection model showed a role for ExsA in early interactions with the corneal epithelium that was not detectable with the conventional (0-hour) scratch model. Comparison of the 6- and 12-hour healing models, which showed that factors additional to barrier function contribute to defense against infection, could be used to gain new insights into corneal defense mechanisms, and the methods used by bacteria to circumvent them.

The intact cornea is normally resistant to infection by Pseudomonas aeruginosa. The most common predisposing factors for keratitis caused by this opportunistic pathogen are extended wear of hydrogel contact lenses and corneal trauma. ^{1 2 3} Only some corneal diseases that cause breaks in the epithelium predispose to infection. For example, keratoconus frequently causes corneal abrasions, yet rarely leads to corneal infection. Similarly, epithelial injury after tonometry has not been associated with risk of infection. Factors that determine susceptibility to infection in patients who are at risk are not well understood. 

There are several animal models of Pseudomonas corneal infection that have clinical features similar to those of the disease in humans. Rabbits and mice are those most commonly used, ^{4 5 6} although rat and guinea pig models have also been described. ^{7 8} Each of these models requires some form of ocular surface manipulation to induce corneal infection. Simple inoculation of the intact ocular surface does not suffice. ⁹ To produce disease reliably, bacteria or their toxins are injected directly into the stroma, or they are applied topically after corneal scarification. Soft contact lens–associated infections have been studied in the rabbit, ^{10 11} but the method requires suturing of the eyelids, and the resultant infections are inconsistent in severity. 

The mouse scarification model has been used extensively for studying P. aeruginosa corneal infection. ^{4 12 13} After anesthesia, one cornea of each mouse is scratched with a sterile needle to produce full-thickness epithelial defects. A suspension of P. aeruginosa is then inoculated onto the damaged corneal surface, and the disease is allowed to progress for various periods before examination of infected corneas. With this model, most P. aeruginosa strains produce severe corneal disease that is detectable as early as 12 hours after inoculation. ^{13 14} Although this is a reliable infection model, contact lens wear, which is the most common underlying cause of P. aeruginosa keratitis, seldom causes epithelial injury severe enough to expose the stroma. It is more likely that P. aeruginosa infection in contact lens wearers is preceded by microtrauma of an otherwise intact epithelium or by subtle changes to the biology of individual epithelial cells that makes them more vulnerable. 

Soft contact lens wear has been shown to increase corneal epithelial permeability to fluorescein, suggesting alterations to epithelial barrier function. ¹⁵ Whether this involves microtrauma or changes in the regulation of tight junctions, it is likely that the initial interactions of bacterial pathogens with the cornea involve the epithelium rather than the stroma. Because expression of bacterial virulence determinants is dependent on environmental signals, those expressed in the corneal epithelium may differ from those induced by the stromal environment. Moreover, the nature of bacterial interactions with the corneal epithelium is likely to determine whether bacteria gain access to the stroma at all. 

P. aeruginosa isolates from corneal infections have been divided into two types based on their different effects on corneal epithelial cells in vitro. ^{16 17} Invasive and cytotoxic strains occur in approximately equal numbers in human corneal disease and both can cause keratitis in mice. ^{13 17} Invasive strains enter corneal epithelial cells and then replicate within their cytoplasm. ^{18 19} Although cytotoxic P. aeruginosa can also invade cells, ²⁰ these genotypically distinct strains can regulate their uptake by corneal epithelial cells by using active invasion inhibition ²¹ and can instead inject a toxin to kill the cell quickly from an extracellular location. ^{22 23} We have shown that some of this contrasting behavior relates to differences in four known ExsA-regulated proteins secreted via an ExsA-regulated type III system into the host corneal epithelial cell. ^{20 24} ExsA, a member of the AraC family of transcriptional activators, coordinately regulates a group of proteins that comprise a secretion apparatus and translocators to deliver effector molecules into target cells. Each of the effectors (ExoS, ExoT, ExoU, and ExoY) has the capacity to modify epithelial cell signal transduction events. ^{25 26 27 28}  

Although exsA mutation of cytotoxic strains impairs their ability to damage corneal epithelia and other cells, including macrophages and polymorphonuclear leukocytes (PMNs), ^{20 29 30} exsA mutation of invasive strains has almost no effect on their ability to invade cells. ²⁰ Accordingly, we have found that ExsA-regulated effectors of a cytotoxic, but not an invasive strain, contribute to P. aeruginosa keratitis pathogenesis in vivo in a murine scarification model. ³¹  

In this study, we set out to test the hypothesis that early epithelial interactions with bacteria can influence disease pathogenesis caused by P. aeruginosa. Thus, we developed a new model for P. aeruginosa keratitis that forces bacteria to interact with corneal epithelial cells before they can access the stroma. Our approach was to study scratched corneas during the healing process. Utilization of the resultant new model showed a role for ExsA in early epithelial cell interactions that was not detectable with the conventional scratch model. 

Materials and Methods

Bacterial Strains and Transformation with GFP

The following wild-type P. aeruginosa strains were used: cytotoxic strains 6206 and PA103 (serogroup O11) and the invasive strain PAO1 (serogroup O5). ¹⁶ Isogenic exsA mutants of PAO1 and PA103 were also used: PAO1exsA::Ω and PA103exsA::Ω. ³² Bacteria were grown overnight (18 hours) at 37°C on tryptic soy agar (TSA) plates. Inocula were prepared by resuspension of bacteria in buffered minimum Eagle’s medium (MEM; cat. no. M-4642; Sigma-Aldrich, St. Louis, MO) to a concentration of 10¹⁰ cfu/mL (approximately 10⁸ cfu in 5 μL). Bacterial concentrations were confirmed by viable count. 

To facilitate their visualization within the cornea, bacteria were transformed with a plasmid constitutively expressing green fluorescent protein (GFP). Strain PAO1 expressing GFP on the plasmid pSMC2 (PAO1-GFP) ³³ was kindly supplied by Gerald B. Pier (Harvard Medical School, Boston, MA). This plasmid was isolated from PAO1-GFP by alkaline lysis ³⁴ and transformed into competent PAO1exsA::Ω, PA103 and PA103exsA::Ω by heat pulse at 50°C. ³⁵ In control experiments, plasmid introduction did not affect the bacterial growth, cytotoxicity, or invasion capabilities of wild-type or mutant strains (data not shown). 

Healing Model of Murine Corneal Infection In Vivo

After induction of anesthesia, three linear scratches were applied to one cornea of female C57BL/6 mice using a sterile 25-gauge needle, and then the scratches were allowed to heal for 0, 6, 9, or 12 hours. Animals were infected with 10⁸ cfu bacteria in 5 μL of buffered MEM. Six animals were assigned to each group for each experiment with the exception of those involving histologic evaluation of wound repair (Fig. 1) for which three animals were assigned to each group, and the evaluation of susceptibility of healing corneas to P. aeruginosa infection (Table 1) for which seven animals were assigned to each group. All experiments were repeated twice. 

At 1, 2, 4, 7, and 14 days after bacterial challenge, corneal disease was scored in a masked fashion with two grading systems. The overall severity of infections was scored as previously described ¹² : grade 0, eye macroscopically identical with the uninfected control eye; grade 1, faint opacity partially covering the pupil; grade 2, dense opacity covering the pupil; grade 3, dense opacity covering the entire anterior segment; and grade 4, perforation of the cornea and/or phthisis bulbi (shrinkage of the eyeball). An eye receiving an overall grade of 2 or greater was considered infected. Another 5-point grading system (grade 0, no infection to grade 4, severe infection) that assesses four different characteristics of the disease was also used. ³⁶ This involved scoring the area and the density of the central opacity, the density of the peripheral opacity, and the epithelial surface quality. The calculated sum of scores for these four characteristics ranges from a possible 0 (clear, normal) to a maximum of 16. 

Descriptive comments were also recorded and eyes were photodocumented using an three-chip cooled camera (Optronics, Goleta, CA) attached to a dissecting microscope (Stemi 2000-C; Carl Zeiss Meditec, Jena, Germany). All procedures were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the UC Berkeley Animal Care and Use Committee. 

Histologic Evaluation of Wound Repair

The healing of uninoculated scratched corneas was examined by microscopy. A drop of fixative (2% paraformaldehyde/0.1% glutaraldehyde in 0.1 M sodium cacodylate buffer/pH 7.2) was topically applied to the surface of unscratched corneas and scratched corneas that had healed for 0, 6, 9, or 12 hours. Eyes were immediately enucleated and fixed intact, dehydrated using a graded series of ethanol (35%, 50%, 70%, 80%, 95%, and 100%), and infiltrated with and embedded in resin (LR White; Ted Pella, Redding, CA), which was then polymerized at 60°C for 2 days. Serial sections (1 μm) were cut perpendicular to the orientation of the scratches with an ultramicrotome (Ultracut UCR; Leica, Deerfield, IL) to obtain cross sections of the scratches. Sections were stained with toluidine blue (0.1%) and examined by light microscopy (model IX-70; Olympus, Melville, NY). 

Evaluation of Corneal Epithelial Permeability in Healing Corneas

To qualitatively assess the patency of the epithelial barrier at the various time points during the healing process, scratched corneas that had healed for 0, 6, 9, or 12 hours and unscratched corneas were stained with 1 μL sodium fluorescein in vivo in anesthetized animals, and examined and photographed with the dissecting microscope (Stemi 2000-C; Carl Zeiss Meditec). 

Bacterial Localization and Colonization in Healing Corneas

To explore bacterial localization within infected tissue, corneas were inoculated with GFP-transformed bacteria at 0, 6, or 12 hours after scratch injury. After 3 hours, animals were killed, a drop of fixative was applied to the ocular surfaces, and the corneas were excised. Isolated corneas were cryoprotected in 10% and 30% sucrose, embedded in optimal cutting temperature (OCT; Tissue-Tek; Sakura, Torrance, CA), and 6-μm sections were cut perpendicular to the length of the scratch with a cryostat (CM1850; Leica Instruments GmbH, Nussloch, Germany). Serial sections were examined by light and fluorescence microscopy (IX-70; Olympus). In each section, the full thickness of the stratified epithelium and corneal stroma were visualized at 1200× magnification and examined for the presence of GFP bacteria. The number of bacteria visualized in the following corneal layers were counted and recorded: stroma, basal cell, wing cell, and superficial cell. The total number of bacteria in each layer, for the entire cornea, was then tabulated by adding together the bacterial counts from every section. 

To quantify bacterial colonization, mice inoculated at 0, 6, or 12 hours after scratch injury were killed 3 hours after bacterial challenge. The infected eyes were then enucleated and separately homogenized in 1 mL tryptic soy broth for enumeration of viable bacteria on TSA plates. 

Statistical Analysis

The statistical significance between the number of bacteria among three or more groups was determined nonparametrically using the Kruskal-Wallis test, whereas that between two groups was determined nonparametrically by the Mann-Whitney test. P < 0.05 was considered significant. 

Results

Effect of Epithelial Healing Time on P. aeruginosa Keratitis

All eyes inoculated with strain 6206 (cytotoxic) or strain PAO1 (invasive) after 6 hours of healing time became infected. In contrast, none of the eyes inoculated after 12 hours of healing time showed development of infection. Both strains caused infection in some eyes that had been allowed to heal for 9 hours (Table 1) . 

Development of P. aeruginosa–Induced Disease after 6 Hours of Healing Time

The time course and severity of infections that developed in eyes inoculated with 6206 or PAO1 after 6 hours’ healing time is presented in Table 2 . These results are similar to those that have been described with the conventional scratch model. ³¹ For both 6206 and PAO1, infections that developed in 9-hour healed eyes tended to be delayed in their development and milder in severity compared with infections that developed after 6 hours of healing time (data not shown). 

Examination of Healing Corneas 0, 6, 9, and 12 Hours after Scratch Injury

Scratch injury was found to produce full-thickness epithelial breaks that exposed the stroma (Fig. 1A) . After just 6 hours of healing time, several layers of epithelial cells filled in the scratched areas (Figs. 1B 1C) . However, those cells tended to stain more intensely with toluidine blue and they were often of irregular shape, compared with cells adjacent to the previously wounded area. In some parts of the cornea, the previously scratched area was also slightly indented at the epithelial surface, and many of the cells were flat with little visible detail (Fig. 1B) . In other areas, individual cellular detail was more apparent, but the overall stratified cellular architecture remained disordered (Fig. 1C) . After 9 and 12 hours of healing times, differences between normal and previously scratched epithelium were still apparent, but they were less obvious (Figs. 1D 1E) . Unexpectedly, significant fluorescein staining, indicating compromised epithelial barrier function, was noted at all time points including 12 hours (Fig. 2) . Although the intensity of staining along the actual scratches became less with longer healing times, diffuse staining of the entire corneal surface appeared after longer healing times (arrows). Histologic examination of the areas between scratches did not provide an obvious explanation for this diffuse staining pattern during healing. 

The Role of ExsA in Bacterial Colonization of Healing Corneas

The cytotoxic strain PA103 and the invasive strain PAO1 were chosen for the study of the role of ExsA in bacterial colonization at the various healing times because of the availability of well-defined exsA mutants. Control experiments confirmed that wild-type PA103, similar to wild-type PAO1 and wild-type 6206, reliably causes disease when inoculated onto 6-hour healed corneas (five of five mice, data not shown). 

Corneas that had healed for 0, 6, or 12 hours were inoculated with strain PA103 (cytotoxic), PAO1 (invasive), or their respective isogenic exsA mutants, and then the number of viable bacteria colonizing each cornea 3 hours after inoculation was quantified (Table 3) . For PA103, exsA mutation was found to reduce significantly the number of PA103 bacteria that colonized corneas after a 6-hour healing time (P = 0.0250), but not after 0 or 12 hours of healing time. 

For the invasive strain PAO1, exsA mutation did not reduce colonization at any of the three time points during healing. In fact, exsA mutation actually increased the ability of PAO1 to colonize corneas immediately after scratch injury (P = 0.0039). 

Other differences between the two strains were also noted while comparing the effect of healing time on their ability to colonize corneas. Both wild-type and exsA mutants of the cytotoxic strain PA103 had more difficulty colonizing corneas with increased healing time (comparison of 0, 6, and 12 hours of healing, P = 0.0148 and P = 0.0023, respectively, Kruskal-Wallis test). Opposite results were found for the invasive strain PAO1—that is, significantly more wild-type bacteria colonized corneas with increasing healing time (P = 0.0342), whereas exsA mutant colonization was unaffected (P = 0.1129). 

Distribution of P. aeruginosa within Infected Corneas

Corneas that had healed for 0, 6, or 12 hours were infected with bacteria transformed with GFP to allow study of bacterial penetration into the corneal epithelium. Results of control experiments showed that transformation with GFP did not affect virulence (data not shown). Three hours after inoculation, the distribution of bacteria within the corneal epithelium was determined by microscopy (Fig. 3) . 

Mutation of exsA was associated with reduced bacterial penetration to the stroma of cytotoxic strain PA103 when corneas were inoculated 6 hours after scratch injury (P = 0.0374), but not when they were inoculated immediately after injury (P = 0.2002). Mutation of exsA did not affect bacterial penetration through the epithelium for the invasive strain PAO1 under either circumstance (for 0 hours, P = 0.8728; for 6 hours, P = 0.6310). 

The length of healing time had a significant effect on the number of bacteria that reached the stroma for both strains, with or without mutation of exsA (PA103, P = 0.0008; PA103exsA, P = 0.0010; PAO1, P = 0.0021; PAO1exsA, P = 0.0021: simultaneous comparison of all healing times by the Kruskal-Wallis test). Few bacteria were found in the stroma when bacteria were added after corneas had healed for 12 hours, which explain those corneas’ resistance to infection. Even so, the invasive strain (PAO1) tended to penetrate further into the epithelium of 12-hour healed eyes than did the cytotoxic strain PA103. Most of the PAO1 bacteria had accessed the wing cell layer within 3 hours, whereas most if the PA103 remained in the superficial cell layer. 

Discussion

Human bacterial keratitis involving P. aeruginosa may not always be preceded by full-thickness epithelial defects. In those situations, bacterial interactions with the corneal epithelium early in the disease could be critical to pathogenesis. In this study, we developed a murine infection model that forces bacteria to interact with the corneal epithelium before they can access the stroma. Our approach to developing this model involved studying healing corneal epithelium, which was found to remain susceptible to P. aeruginosa infection 6 hours after corneas were scratched and then to regain its resistance after 12 hours. We then used the healing model to validate findings in our prior in vitro epithelial cell infection studies. 

Histologic evaluation revealed that several layers of epithelial cells had already covered the wounded area by the 6-hour healing time point, although cells occupying the wound site were irregular in shape and stained more densely than cells in adjacent areas. By the 12-hour healing time point when the corneas were no longer susceptible to infection, these differences remained, although they were less obvious. Thus, histology showed no clearly observable differences between the 6- and 12-hour healed eyes, despite their very different responses to bacterial inoculation. 

Similarly, fluorescein-staining patterns did not predict susceptibility—that is, significant staining was observed even after 12 hours of healing time. In a normal healthy cornea, fluorescein staining does not occur, because the dye is excluded from entering the cornea by the tight junctions between superficial cells of the corneal epithelium. These junctions normally exclude debris and fluid, and they are also essential for the maintenance of epithelial cell polarity. It has generally been assumed that increased epithelial permeability is associated with increased risk of infection, because it could allow bacteria to penetrate between cells and also because it would expose the more susceptible basolateral cell surfaces to bacteria. ³⁷ The results showing resistance to infection at 12 hours despite significant fluorescein staining suggest that other defense factors can compensate for loss of barrier function. Those could include antimicrobial peptides, mucins, cytokines, or PMNs, all of which are part of innate corneal defenses against infection. ^{38 39 40 41 42 43} Defensins, cytokines, and PMNs have already been shown to be upregulated during wound healing. ^{44 45 46}  

Alternatively, resistance to infection after 12 hours of healing time may mean that tight junctions do not need to be completely formed to exclude bacteria. Indeed, fluorescein staining along the scratches was more subtle with increasing healing time, suggesting that more cells occupied the scratched areas, that more tight junctions had reestablished between cells, or both. Studies of tight junction reformation after large scrape wounds in rabbits in vivo have shown that tight junctions become progressively more complex and mature over a period of at least a day. ⁴⁷ The significance of fluorescein staining away from the previously scratched area during wound healing is yet to be determined. 

To explore the usefulness of the 6-hour healing infection model for examining P. aeruginosa interactions with the corneal epithelium, the effect of exsA mutation on P. aeruginosa colonization and penetration of the regenerating corneal epithelium was investigated. 

Our research has shown that the transcriptional activator ExsA regulates various genes that influence interactions between P. aeruginosa and corneal epithelial cells in vitro, but only in cytotoxic strains. ²⁰ These factors have since been shown to have effects on other mammalian cell types in vitro, including macrophages and PMNs. ^{29 30} We have also demonstrated that ExsA affects the 48-hour disease outcome in the conventional scratch model for a cytotoxic, but not an invasive, strain. ³¹ The involvement of ExsA in that model at 48 hours does not directly imply effects on epithelial cells, nor does it provide information about when it exerts its effect. 

In this study, results obtained with the 6-hour healed corneas showed the effect of ExsA early in the infectious process for a cytotoxic strain that were not apparent with the conventional scratch model (0 hours of healing). This included involvement in bacterial penetration through the epithelium to access the stroma and in overall colonization efficacy. ExsA was not involved in either of these events for the invasive strain. These results correlated well with our previously published findings in corneal epithelial cells in vitro that the exsA mutation profoundly influences cytotoxic strain interactions with epithelial cells, although it has little effect on the invasive strain’s behavior. ²⁰  

Factors regulated by ExsA of cytotoxic strains include ExoU, a potent cell toxin, and ExoT, which can disrupt the cytoskeleton of epithelial cells. The role of ExsA in vivo may involve ExoU-mediated cytotoxicity, which would allow bacteria to penetrate the epithelial cell layer more easily, while releasing nutrients from cells and exposing potential bacterial binding sites. ⁴⁸ ExoT may also help bacteria to penetrate between cells by altering cell shape or by inhibiting epithelial cell migration and healing, as demonstrated in vitro. ⁴⁹ Either ExoU or ExoT may also compromise epithelial cell innate immune defenses through effects on epithelial cell health and function. 

Comparison of the 6- and 12-hour healing models with the scratch model (0 hours of healing) showed that regeneration of the corneal epithelium gradually reduced overall colonization levels by the cytotoxic, but not by the invasive, strain. Regeneration of the corneal epithelium actually enhanced colonization of the cornea by the invasive strain. The significance of this unexpected finding may or may not relate to the interesting finding that the invasive bacteria penetrated more quickly and deeply into the 12-hour healed epithelium than did the cytotoxic bacteria. Fortunately, however, the number of bacteria actually reaching the stroma progressively decreased with increased healing time in both the invasive and cytotoxic strains. By the 12-hour healing time point, when the cornea was no longer susceptible to infection, few bacteria penetrated to the stroma of either strain type. Whether the critical factor distinguishing between resistance and susceptibility in these models relates to the ability of bacteria to penetrate into the stroma is to be determined. 

The pathogenesis of P. aeruginosa keratitis is a multifactorial process requiring a combination of different bacterial and host cell factors, the relative contributions of which may vary at different times during the infection process. The results of this study show that healing models are useful for studying P. aeruginosa keratitis, because they have the potential to identify bacterial virulence factors with activity against epithelial cells in vivo that is not detected by traditional scratch models. Using a 6-hour healing model, ExsA was found to be essential for corneal epithelial colonization and penetration for a cytotoxic, but not an invasive, P. aeruginosa strain, showing at least one difference in ExsA-regulated pathogenic mechanisms between these two strains. Further studies are needed to determine whether this difference in colonization and penetration applies to other cytotoxic and invasive strains. 

The use of fluorescein staining and histologic methods to study epithelial health did not clearly indicate why corneas were susceptible to infection at 6 but not at 12 hours after scratch injury. However, the data suggested that bacteria may not have to be deposited directly onto the stroma for them to cause disease and that disruption of the epithelial barrier function does not always leave the cornea susceptible to infection. Studies to elucidate differences at the ocular surface between these two healing time points could yield interesting information about corneal defense mechanisms. 

 Supported by National Eye Institute Grant EY11221, a UC Berkeley Faculty Research Grant, and a Hellman Family Faculty Award (SMJF). EJL was supported by an Ezell Fellowship awarded by the American Optometric Foundation.

 Submitted for publication March 5, 2003; revised July 10, 2003; accepted July 28, 2003.

 Disclosure: E.J. Lee, None; D.J. Evans, None; S.M.J. Fleiszig, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Histology of corneas after scratch injury. Representative photographs (from three animals per group) of toluidine blue–stained sections showing scratches after 0 (A), 6 (B, C), 9 (D), or 12 (E) hours of healing time (arrows).

Table 1.

 

Percentage of Eyes That Developed Infections by Day 4 after Inoculation with 10⁸ cfu P. aeruginosa at 6, 9, or 12 Hours after Scratch Injury

Healing Time	6206	PAO1
6h	100.0	100.0
9h	14.3	28.6
12h	0.0	0.0

Table 2.

 

Course of Infections in Eyes Challenged with 10⁸ cfu P. aeruginosa Strain 6206 (Cytotoxic) or Strain PAO1 (Invasive) 6 Hours after Scratch Injury

Strain	Day 1	Day 2	Day 4	Day 7	Day 14
6206
Overall*	3 (3:3)	3 (3:3)	3 (3:3)	4 (4:4)	4 (4:4)
Total, †	10 (9:12)	13 (12:14)	15 (13:15)	16 (16:16)	16 (16:16)
PAO1
Overall*	3 (2:3)	3 (3:3)	3 (3:3)	3 (3:3)	3 (2:3)
Total, †	8 (5:9)	8 (7:11)	9 (8:9)	9 (6:9)	6 (4:7)

 Data are expressed as the median score, with lower and upper quartiles in parentheses.

*  Overall scores assigned on a scale of 0 to 4, as described by Beisel et al. ¹²

†  Total scores range from 0 to 16 and represent the sum of scores given to each of four disease characteristics, as described in Cowell et al. ³⁶

Figure 2.

 

Representative photographs (from six animals per group) of fluorescein staining after 0 (A), 6 (B), 9 (C), or 12 (D) hours of healing time. (C, D, arrows) areas of diffuse staining, distant from the wound site, that were more pronounced with longer healing time.

Table 3.

 

Comparison of P. aeruginosa Colonization of Scratched and Healing Corneas for PA103 vs. its exsA Mutant and PAO1 vs. its exsA Mutant 3 Hours after Inoculation with 10⁸ cfu Bacteria

Healing Time	Strain	Bacterial Colonization (CFU)	Percentage Colonization*
PA103 vs. exsA mutant
0 h	PA103	46,750 (37,000:69,250)	0.1201 (0.0496:0.2059)
	PA103 exsA	16,625 (11,400:49,488)	0.0327 (0.0162:0.1153)
6 h	PA103	5,200 (1,963:9,150)	0.0105 (0.0062:0.0150), †
	PA103 exsA	730 (190:1,420)	0.0009 (0.0004:0.0012), † , ‡
12 h	PA103	5,025 (4,063:5,500)	0.0125 (0.0039:0.0226), †
	PA103 exsA	3,775 (2,375:5,438)	0.0093 (0.0026:0.0163)
PAO1 vs. exsA mutant
0 h	PAO1	16,125 (4,475:37,000)	0.0460 (0.0219:0.0660)
	PAO1 exsA	362,500 (247,500:372,500)	0.4224 (0.2169:0.6545), †
6 h	PAO1	123,325 (5,975:251,250)	0.1539 (0.1138:0.1896)
	PAO1 exsA	69,500 (26,263:84,125)	0.0454 (0.0372:0.1923)
12 h	PAO1	77,250 (59,250:84,000)	0.5025 (0.1646:0.1619)
	PAO1 exsA	135,250 (50,875:146,125)	0.2453 (0.0297:0.4618)

 Data are expressed as the median, with lower and upper quartiles in parentheses.

*  Percentage of original inoculum that colonized the eye.

†   P < 0.05, Mann-Whitney test, compared to 0 hours healing time.

‡   P < 0.05, Mann-Whitney test, compared to wild-type.

Figure 3.

 

Distribution of colonizing bacteria within the cornea. The number of bacteria counted in each corneal layer at 3 hours after inoculation with 10⁸ cfu bacteria is expressed as the median and quartiles for strains PA103 (A) and PAO1 (B).

The authors thank Myra N. Mendoza and Tan N. Truong for excellent technical assistance.