Pseudomonas aeruginosa Keratitis in Knockout Mice Deficient in Intercellular Adhesion Molecule 1

Abstract

In this study, the role of intercellular adhesion molecule 1 (ICAM-1) in the pathogenesis of Pseudomonas aeruginosa keratitis was examined by using inbred ICAM-1-deficient knockout mice. These mice had significantly less (P ≤ 0.02) ocular disease than wild-type mice, suggesting that ICAM-1 contributes to a more severe disease response following P. aeruginosa infection.

Pseudomonas aeruginosa is capable of causing devastating corneal infections. Several studies suggest that bacterial proteases (2, 6, 15, 16) contribute to the pathogenesis of P. aeruginosa keratitis. The host inflammatory response to P. aeruginosa corneal infection consists primarily of an influx of polymorphonuclear leukocytes (PMN) which also destroy corneal tissue (11, 18), despite their being essential to resolving the infection (7, 9).

PMN migrate from the tear film and from the limbal and iridial vasculature into the avascular cornea. Intercellular adhesion molecule 1 (ICAM-1) is a key molecule for PMN recruitment into infected tissue. Up-regulation of ICAM-1 is correlated with the ability of outbred mice to restore corneal clarity after infection with P. aeruginosa (12). However, immunostaining of inflamed human ocular tissue (1, 4) suggests that ICAM-1 is a key mediator of acute ocular inflammation. The purpose of this study was to clarify the role of ICAM-1 in ocular P. aeruginosa infection.

Breeding pairs of homozygous 129/Sv wild-type (WT) mice and knockout (KO) mice deficient in ICAM-1 expression (21) were supplied by Jose-Carlos Gutierrez-Ramos of the Center for Blood Research, Harvard Medical School (Boston, Mass.). Eight-week-old ICAM-1 KO mice and age- and sex-matched WT mice were used. Mice in this study were treated humanely, in accordance with the ARVO Resolution on the Use of Animals in Research.

The preparation of the bacterial inoculum used in this study to initiate pseudomonal keratitis in mice has been described (13). Briefly, overnight broth cultures of P. aeruginosa ATCC 19660 (American Type Culture Collection, Manassas, Va.) were pelleted by centrifugation, washed once with sterile saline (pH 7.2), and resuspended in saline to a concentration of approximately 2.0 × 10¹⁰ CFU/ml. The mouse scratch wound model of P. aeruginosa keratitis also has been previously described (13). Mice were anesthetized, and the corneal surface of the left eye of each animal was wounded with a sterile 26-gauge needle. A 5.0-μl bacterial suspension containing 1.0 × 10⁷ CFU was then delivered topically onto the surface of the wounded cornea.

Mice were macroscopically evaluated in a masked fashion and disease graded with a scale of 0 to +4 (8), with 0 signifying no disease and +4 signifying corneal perforation. The ocular pathology observed at each time point is expressed as the mean clinical score ± standard error of the mean (SEM). Means represent the sum of scores at each time point divided by the number of corneas scored (n ≥ 10).

Interleukin 1β (IL-1β) and tumor necrosis factor alpha (TNF-α) enhance ICAM-1 expression 5- to 10-fold above constitutive levels on receptive cells in vitro (3). ICAM-1 expression peaks 24 h after induction with these cytokines and remains elevated for another 48 h (5). At 6, 12, 24, and 48 h postinfection (p.i.), ICAM-1-deficient KO mice and WT mice were sacrificed and corneas were processed for quantitation of IL-1β and TNF-α by enzyme-linked immunosorbent assay (ELISA), as described previously (13). Uninfected corneas from five naive KO mice and five WT mice were similarly processed as controls. Cytokine concentrations are expressed as picograms of cytokine per milligram of corneal tissue. The sensitivities of the ELISAs for IL-1β and TNF-α are 10 and 15 pg/ml of sample, respectively. Each cytokine assay was repeated at least twice.

At 6, 12, 24, and 48 h p.i., six ICAM-1-deficient KO mice and six WT mice were sacrificed, and corneas were harvested as described previously (13) for quantitation of viable bacteria (by direct plate count) and of infiltrating PMN (by assay of myeloperoxidase activity [14]). The numbers of viable bacteria in corneas are expressed as mean log CFU of bacteria ± SEM. Infiltrating PMN in corneal tissue are expressed as mean log PMN ± SEM per milligram of cornea.

The procedure for embedding eyes in resin and sectioning for light microscopy has been described before (10). Eyes (n = 3) at each time point were enucleated at 6 and 12 h and at 1, 2, 3, 5, 7, 14, and 21 days p.i., fixed in a 1:1:1 solution of osmium tetroxide, glutaraldehyde, and Sorenson’s phosphate buffer (pH 7.4), dehydrated in graded ethanols and propylene oxide, infiltrated with epon-araldite, and embedded. Thick (1.75-μm) sections were cut with an ultramicrotome and stained with Richardson’s stain (17) for histopathology.

Eyes (n = 3) from uninfected WT and KO mice were enucleated for determination of constitutive expression of ICAM-1. Eyes (n = 3 [each]) from WT and KO mice infected with P. aeruginosa were enucleated at 48 h p.i. to test for up-regulation of ICAM-1 expression. ICAM-1 immunostaining was performed as described previously (12) except that a Metal Enhanced DAB substrate kit (Pierce, Rockford, Ill.) was used for development of the sections.

Statistical analyses were performed with StatView SE + Graphics, version 4.5 (Abacus Concepts, Inc., Berkeley, Calif.). An unpaired Student’s t test using a two-tailed hypothesis was used to test for statistical significance in differences in the numbers of PMN per milligram of cornea, the numbers of viable bacteria per cornea, and the concentrations of cytokines per milligram of cornea. A nonparametric Mann-Whitney U test was used to compare mean clinical scores of infected ICAM-1-deficient KO mice and WT mice. Statistical significance was established at the P ≤ 0.05 confidence interval.

P. aeruginosa ocular disease in ICAM-1 KO mice and WT mice.

The corneal disease responses of ICAM-1-deficient KO mice and WT mice, expressed as mean clinical scores (± SEM), are shown in Fig. 1. At least 10 eyes from each group were examined per time point. At 2 days p.i., ocular pathology was more severe in WT than in ICAM-1-deficient KO mice (P ≤ 0.02). By 3 days p.i., the difference in disease response between WT and ICAM-1-deficient KO mice was more pronounced (P ≤ 0.0001). Ocular disease continued to be more severe in WT than in ICAM-1-deficient KO mice through 15 days p.i. (P ≤ 0.02). By 21 days p.i., eyes of both groups of mice were similar in appearance (P ≥ 0.6). Uninfected corneas of both groups of mice were similar upon microscopic examination (data not shown). By 6 h p.i., inflammatory cells had localized in conjunctival and iridial blood vessels in both ICAM-1-deficient KO mice and WT mice and were beginning to emigrate into the peripheral cornea and the anterior chamber (Fig. 2A), respectively. With the exception of the epithelial wounds created 6 h earlier and the presence of a few inflammatory cells in the peripheral cornea and the anterior chamber, the corneas of ICAM-1-deficient KO mice and WT mice were morphologically similar to uninfected corneas (Fig. 2A and B).

FIG. 1.

Mean clinical scores (± SEM) of ocular disease responses in ICAM-1-deficient KO mice and WT mice. Means represent the sum of scores at each time point divided by the number of corneas scored (n ≥ 10). At 1 day p.i., P was ≥0.17; at 2 to 15 days p.i., P was ≤0.02; and at 20 and 21 days p.i., P was ≥0.56.

FIG. 2.

Histopathology of the central region of corneas from WT mice at 6 (A), 24 (C), and 48 (E) h p.i. and from ICAM-1-deficient KO mice at 6 (B), 24 (D) and 48 (F) h p.i. EP, epithelium; ST, stroma; AC, anterior chamber. Small arrowheads indicate inflammatory cell infiltrate in anterior chamber; large arrowheads indicate inflammatory cell infiltrate in corneal stroma. Richardson’s stain was used. Magnification (A and B), ×200; magnification (C to F), ×100.

At 12 and 24 h p.i., inflammatory cells had infiltrated the corneas of both ICAM-1-deficient KO mice and WT mice. Although the numbers of inflammatory cells in the corneal stroma and the peripheral corneal epithelium of both groups of mice appeared to be similar, other pathological differences were evident. For example, the corneal epithelium of WT mice was denuded over the central cornea, which was markedly swollen (Fig. 2C). In contrast, the central corneal epithelium of ICAM-1 KO mice was partially intact and the corneal thickness was normal (Fig. 2D). In contrast to ICAM-1-deficient KO mice, there was a significant number of inflammatory cells (too numerous to count) in the anterior chamber of WT mice.

At 48 and 72 h p.i., inflammatory cells continued to infiltrate corneas of ICAM-1-deficient KO mice and WT mice, appearing more dense in the central cornea (Fig. 2E and F). The number of inflammatory cells in the anterior chamber of ICAM-1-deficient KO mice had visibly decreased, whereas inflammatory cells filled the anterior chamber of WT mice. By 5 to 7 days p.i., the number of inflammatory cells in the anterior chamber of ICAM-1-deficient KO mice was diminished while the anterior chamber of WT mice remained infiltrated with inflammatory cells. At 14 days p.i., the cell infiltrate was decreased in both KO and WT mice; however, corneas of WT mice remained swollen. By 21 days p.i., corneas of both KO and WT mice demonstrated some stromal scarring and neovascularization, but no inflammatory cells were seen in the anterior chamber.

Corneal inflammation and infection.

No ICAM-1 immunostaining was observed in uninfected or infected eyes of ICAM-1-deficient KO mice (data not shown). In WT mice, ICAM-1 was constitutively expressed (weakly) on stromal keratocytes and corneal epithelium (data not shown). The intensity of immunostaining increased by 48 h after infection with P. aeruginosa in WT mice.

IL-1β and TNF-α were not constitutively expressed in the uninfected corneas of either KO or WT mice. Concentrations of IL-1β and TNF-α in infected corneas are shown in Fig. 3. There were no significant differences (P ≥ 0.10) in cytokine concentrations in KO and WT infected eyes.

FIG. 3.

Concentrations of TNF-α and IL-1β in corneas of ICAM-1-deficient KO mice and WT mice expressed as picograms of TNF-α or IL-1β per milligram of corneal tissue. No significant differences were observed between the two groups of mice (P ≥ 0.10).

There were no significant differences in the mean log base 10 numbers of viable bacteria recovered from corneas of KO or WT mice at any time point (P ≥ 0.29). At 6 h p.i., no myeloperoxidase activity was found in homogenates of corneas from ICAM-1-deficient KO mice (≤3.0 log PMN). In contrast, approximately 4.8 log PMN/mg of tissue infiltrated the corneas of WT control mice. At 12, 24, and 48 h p.i., there was no significant difference (P ≥ 0.27) in the numbers of infiltrating PMN per milligram of cornea between KO and WT mice.

ICAM-1 and P. aeruginosa keratitis.

The inbred mouse strain 129/Sv is resistant to ocular infection with P. aeruginosa; i.e., mice are able to resolve the infection and restore ocular integrity within 21 days p.i. Mutant mice deficient in ICAM-1 expression are phenotypically similar to WT mice with the exception of the ICAM-1 deficiency and a moderate elevation in the number of circulating PMN, probably resulting from the lack of cell trafficking brought about by the deleted adhesion molecule (21). Cytokine production in ICAM-1-deficient 129/Sv KO mice is unaffected by a lack of ICAM-1 expression. Inflammation-induced serum concentrations of cytokines such as IL-1α, IL-6, gamma interferon, and TNF-α in ICAM-1 deficient KO mice were shown to be comparable to those of WT animals (19, 21). In the present study, we similarly found no significant differences in corneal cytokine concentrations in WT versus ICAM-1-deficient KO mice after infection with P. aeruginosa.

The adhesion molecule ICAM-1, when up-regulated by proinflammatory cytokines, appears to play a significant role in development of the severe pathology characteristic of P. aeruginosa eye infections, presumably by the recruitment of an overwhelming number of inflammatory cells (PMN) into the anterior chamber from the iris/ciliary body. However, the presence of ICAM-1 does not appear to be critical in recruiting a sufficient number of PMN into the cornea to eliminate the bacteria and resolve the infection. It is also likely that other, as yet undefined, mechanisms are present in ICAM-1-deficient KO mice which compensate for the absence of the adhesion molecule. One likely candidate is intercellular adhesion molecule 2 (ICAM-2). ICAM-2 is constitutively expressed in the ICAM-1 KO mouse on the 129/Sv background (21). The basal expression of ICAM-2 is higher than that of ICAM-1, but, unlike ICAM-1, it is not cytokine inducible (20). Future studies will seek to determine if ICAM-2 or other adhesion molecules compensate for ICAM-1 in P. aeruginosa keratitis.