TEXT

Top Abstract Text References

Infectious endophthalmitis is a sight-threatening clinical crisis that occurs as a complication of ocular surgery (postoperative endophthalmitis) or penetrating ocular injury (posttraumatic endophthalmitis). The severity of vision loss in endophthalmitis is related to the pathogenic potential of the infecting organism (7, 19, 25-29, 33). Coagulase-negative staphylococci are generally associated with final visual acuities of 20/40 or better, whereas in endophthalmitis caused by more virulent organisms such as Staphylococcus aureus, Enterococcus faecalis, or Bacillus cereus, visual outcomes ranging from 20/100 to enucleation occur in approximately 50 to 90% of cases (1, 5, 7). Despite a general association between visual outcome and the infectious agent, relatively little is known about the species-specific factors that account for the characteristic severity of each disease.

We previously focused on determining the role of secreted bacterial toxins in the pathogenesis of endophthalmitis caused by S. aureus and E. faecalis by using well-characterized laboratory strains in animal models of disease. These studies showed that cytolytic E. faecalis not only causes more fulminant disease but also renders the infection unresponsive to therapeutic intervention (17, 18). The production of most secreted and cell surface proteins in S. aureus is coordinately controlled by chromosomal regulatory loci termed accessory gene regulator (agr) and staphylococcal accessory regulator (sar) (6, 16, 20). Mutant strains of S. aureus with insertional mutations in the sar and agr loci are attenuated in virulence in experimental endophthalmitis compared with parental strains (3, 4). Since sar and agr affect the expression of 12 or more unrelated genes (6, 20), the staphylococcal toxin(s) that contributes most significantly to the severity of disease has not yet been identified.

To further analyze the bacterial factors that contribute to the pathogenesis of endophthalmitis, we performed a genomic DNA fingerprint analysis on 39 S. aureus and 28 E. faecalis strains isolated from the vitreous or aqueous humor of endophthalmitis patients treated at multiple clinical centers. The purpose of this investigation was to assess whether common traits that may be related to ocular colonization and/or the severity of disease outcome exist among isolates of a particular species.

The S. aureus and E. faecalis isolates analyzed in this study were collected from patients with endophthalmitis between 1984 and 1995 at Cullen Eye Institute, Houston, Tex. (CE), Dean A. McGee Eye Institute, Oklahoma City, Okla. (DM), University of Pittsburgh School of Medicine, Pittsburgh, Pa. (UP), King Fahd Hospital, Al Hasa, Saudi Arabia (KF) (a kind gift from LouAnn Bartholomew), and Bascom Palmer Eye Institute, Miami, Fla. (BP). S. aureus strains were collected from DM (7 isolates), UP (7 isolates), and BP (25 isolates), while E. faecalis strains were collected from CE (10 isolates), DM (3 isolates), UP (4 isolates), KF (2 isolates), and BP (9 isolates). Twenty-nine additional S. aureus clinical isolates of extraocular origin were a kind gift from Mark Huycke, Veterans Administration Medical Center, Oklahoma City, Okla. Twenty-one S. aureus keratitis isolates were obtained from the Alcon Microbiology Culture Collection (Fort Worth, Tex.).

Pulsed-field gel electrophoretic analysis of endophthalmitis isolates. Bacterial genomic DNA was prepared as previously described (24), except that lysostaphin (50 µg/ml) was added to the lysis solution for the preparation of S. aureus chromosomal DNA. Isolates with similar banding patterns and no more than three band differences were considered clonally related (32). Isolates with banding patterns similar to clonally related strains but with no more than four band differences were considered subtypes of the clonal group. Once isolates were recognized as having identical or similar banding patterns, a second gel containing all isolates from the same group was run to verify clonal relationships. Twenty-one distinct fingerprint patterns were identified among the S. aureus isolates. Of these, five clonotypes were present more than once and accounted for 58.9% (23 of 39) of the total number of isolates. The clonotype represented most frequently was designated SA1 and accounted for 25.6% (10 of 39) of the isolates tested (Fig. 1). Isolates in this group were derived from each of the clinical centers from which S. aureus isolates were obtained (DM, UP, and BP). Clonotypes SA2 (n = 4) and SA3 (n = 2) were also derived from multiple clinical centers (DM and BP). All isolates comprising clonotypes SA4 (n = 3) and SA5 (n = 4) were derived from the same clinical center (BP) (Fig. 1). The remaining 16 isolates (41%) were present only once (data not shown) and were derived from all three clinical centers. To ensure that the general clonality observed among the S. aureus endophthalmitis isolates was not attributable to a methicillin-resistant S. aureus (MRSA) genotype (21), strains comprising each of the five S. aureus clonotypes were analyzed for the presence of the mecA antibiotic resistance determinant (8). Briefly, bacteria from a 0.5-ml suspension of bacterial cells in phosphate-buffered saline were lysed by boiling in a sealed tube for 10 min, followed by centrifugation (10,000 × g for 1 min) to remove cell debris. PCR was performed on cell lysates with previously published mecA-specific primers (8). Only clonotype SA4 (three isolates), was found to be mecA positive; all other clonotypes were mecA negative.

View larger version (47K): [in this window] [in a new window]   FIG. 1.   PFGE of SmaI-digested chromosomal DNA of endophthalmitis-derived S. aureus clinical isolates collected from three clinical centers, BP, DM, and UP. Separate panels show clonally related isolates. Where present, subtypes are designated with the suffix 1 or 2. Molecular size standards are the New England Biolabs lambda ladder.

View larger version (43K): [in this window] [in a new window]   FIG. 2.   PFGE of SmaI-digested chromosomal DNA of selected endophthalmitis-derived E. faecalis clinical isolates collected from five clinical centers, BP, DM, UP, KF, and CE. Clonally related strains are identified below the panels. Molecular size standards are the New England Biolabs lambda ladder.

Comparison of S. aureus endophthalmitis clonotypes with S. aureus isolated from various sources. Since it was determined that the general clonality observed among the endophthalmitis-derived S. aureus isolates was not due to an MRSA genotype, it was considered that the clonotypes identified might represent species subsets uniquely associated with ocular infection. To test this hypothesis, we examined chromosomal DNA fingerprints for 21 S. aureus keratitis isolates and 29 S. aureus strains isolated from extraocular infections, such as soft-tissue, catheter-associated, and surgical-wound infections. The frequency of occurrence of clonotypes SA1 to SA5 within these populations of isolates is shown in Table 1. Clonotypes SA1, SA3, SA4, and SA5 were found among the S. aureus keratitis isolates and accounted for 47.6% (10 of 21) of the total number of isolates analyzed; the remaining isolates in this group all showed unique genomic DNA fingerprints. Interestingly, as in the case of endophthalmitis-derived clonotypes SA1, SA2 and SA3, keratitis-derived clonotypes SA1, SA3, SA4, and SA5 were collected from geographically diverse clinical centers (6a). All five endophthalmitis-derived S. aureus clonotypes occurred at least once among the extraocular-infection isolates. One new clonotype consisting of three strains (SA6), which was not represented among either the endophthalmitis- or keratitis-derived isolates, was identified among the soft-tissue-wound isolates. The frequency of occurrence of clonotype SA1 among the extraocular-infection isolates was approximately 2.5-fold lower than that observed for the endophthalmitis isolates (10% versus 25.6%); however, the difference was not statistically significant (P = 0.136 [Fisher's exact test]). When analyzed in combination, the frequency of occurrence of clonotypes SA1 to SA5 was not significantly different between the groups (P = 0.340 [² test]). These results indicate that although clonotypes SA1 to SA5 are isolated frequently from ocular infections, these isolates are not unique to this site of infection.

Frequency of cytolysin expression among E. faecalis endophthalmitis isolates. The frequency of the cytolytic genotype among the E. faecalis isolates was determined by performing PCR on bacterial cell lysates with primers specific for cylA, the proteolytic activator gene of the E. faecalis cytolysin operon. The following oligonucleotide primers were selected from published sequences: 5' AAT GGA TAA TAT TTC AGA ATT TGA AGT 3' (cylA1) and 5' TTC CCA CGA AAA TTT TAT AAA CCC 3' (cylA2) (9). Briefly, a suspension of each isolate was prepared by removing bacterial colonies from an overnight plate culture with a moistened sterile swab and resuspending them in 1 ml of sterile 10 mM Tris, pH 7.5. Bacteria from 0.5 ml of the suspension were lysed with a Mini-Beadbeater (Biospec Products, Bartlesville, Okla.) according to the manufacturers instructions. One hundred fifty microliters of the lysate was removed to a clean tube and centrifuged (10,000 × g for 1 min) to remove cell debris. PCR was performed in 10-µl reaction mixtures containing 1 µl of cell lysate, 1 µl of 3 mM MgSO₄, 1 µl of cylA1, 1 µl of cylA2 (10 µM each in sterile H₂O), 1 µl of diluted Taq polymerase (diluted 1:12.5), 1 µl of 2 mM deoxynucleotide triphosphates, and 4 µl of H₂O in a Rapidcycler PCR machine (Idaho Technologies, Idaho Falls, Idaho). Following an initial hold step (94°C for 30 s), the PCR mixtures were cycled 30 times as follows: denaturation, 94°C for 0 s; annealing, 50°C for 0 s; and elongation, 72° for 35 s. An additional hold step of 72°C for 2 min was included at the end of the 30 cycles. The cytolytic phenotype was confirmed by observing zones of hemolysis on brain heart infusion agar plates containing 5% rabbit blood incubated for 2 days at 37°C. E. faecalis FA2-2 (pAM714) and plasmid-free E. faecalis FA2-2 were used as positive and negative controls, respectively, for the detection of both cytolytic phenotype and genotype (14). Of the 28 E. faecalis endophthalmitis isolates collected for this study, 13 (46.4%) possessed the cylA gene, and all cylA-positive strains were phenotypically positive for cytolysin expression as indicated by zones of hemolysis on brain heart infusion agar. This represents an enrichment for the cytolytic phenotype among endophthalmitis isolates compared with its occurrence among isolates from the gastrointestinal tracts of healthy subjects (0 to 17%; P < 0.028 [² test]) (12, 15). All isolates comprising both EF1 and EF2 clonotypes were cytolytic.

Relationship between clonality of endophthalmitis isolates and final visual outcome. In the present study, a number of S. aureus clonotypes which are known to have caused endophthalmitis in multiple, apparently unrelated cases were identified. Therefore, it was of interest to determine whether a correlation between strains of S. aureus and disease outcome exists. The severity-of-outcome measure used in this study was final best-corrected visual acuity achieved following treatment for endophthalmitis and was ascertained following a retrospective review of patient records. The range of final visual acuities observed was in agreement with that found in a previous series for S. aureus endophthalmitis: 20/40 or better, 30.7%; 20/100 or better, 48.7%; 5/200 or better, 64% (7). The relationship between final best-corrected visual acuity and the clonality of the endophthalmitis-derived isolates was analyzed by Pearson's chi square (²) test. Due to the wide range of visual acuities recorded in patient charts, only two levels of severity were analyzed: better than 20/200 and 20/200 or worse. Clonotypes were analyzed either individually (when adequate numbers of isolates made up the group) or in combination. All isolates that occurred more than once were designated "clonal" and all those occurring only once were designated "nonclonal." Table 2 shows the distribution of isolates comprising each clonotype for each level of severity. Visual acuities of 20/200 or worse were found in 70%, 50%, 50%, 66%, and 75% of cases infected with clonotypes SA1, SA2, SA3, SA4, and SA5, respectively, compared with 31% of the nonclonal isolates (for SA1 versus nonclonal isolates, P was 0.053 [² test]). When clonal isolates were combined (SA1 to SA5; n = 23) and compared for severity of outcome with isolates occurring only once (nonclonal isolates; n = 16), it was found that a statistically significant relationship existed between clonality and visual outcomes of 20/200 or worse (P = 0.036 [² test]). These results suggest that clonotypes SA1 to SA5 not only possess traits that enhance ocular colonization, thereby favoring their occurrence at this site, but also possess traits that contribute to poor visual outcomes following intraocular infection.