Demonstration of biofilm in infectious crystalline keratopathy using ruthenium red and electron microscopy

Demonstration of Biofilm in Infectious Crystalline Keratopathy Using Ruthenium Red and Electron Microscopy Tim P. Fulcher, FRCOphth,1 John K. G. Dart, DM, FRCOphth,1 Louise McLaughlin-Borlace, MSc,2 Robin Howes,2 Melville Matheson, BSc,2 Ian Cree, MB, FRCPath2 Objective: Bacterial biofilm formation has been implicated in the pathogenesis of infectious crystalline keratopathy. Biofilm cannot be visualized by electron microscopy without the addition of a fixative that stabilizes the polysaccharide-rich bacterial extracellular matrix that surrounds the bacterial colonies in a biofilm. We used ruthenium red as a fixative to evaluate corneal biopsy specimens for the presence of bacterial biofilm in three cases of infectious crystalline keratopathy (ICK) and five cases of chronic microbial keratitis without crystalline changes. Design: Case series with clinicopathologic correlation. Participants: Eight patients underwent corneal biopsy or therapeutic keratoplasty as part of their management for chronic unresponsive microbial keratitis. Methods: The corneal specimens removed were trisected for microbiology, pathology, and transmission electron microscopy (TEM). The TEM specimens were fixed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer with 0.05% ruthenium red. Main Outcome Measures: Demonstration of bacterial biofilm with TEM. Results: TEM demonstrated organisms with a surrounding extracellular matrix consistent with a bacterial biofilm in the three cases of ICK but not in the five other cases of chronic microbial keratitis. Conclusions: The presence of biofilm in ICK can be demonstrated with TEM with appropriate fixation techniques that stabilize the bacterial extracellular matrix. Biofilm stains intensely with periodic acid–Schiff because of the polysaccharide-rich extracellular matrix and weakly with Gram stain because of the high proportion of nonviable organisms. Biofilm formation occurs in ICK but probably not in chronic bacterial keratitis without crystalline changes. Secretion of an extracellular matrix by bacteria to form a biofilm is a response to a nutrient-deprived environment in which growth and replication is depressed. The extracellular matrix of the biofilm may mask bacterial antigens, explaining the relative lack of inflammatory response in these infections. It may also be one of the mechanisms explaining the resistance to in vivo antimicrobial therapy when in vitro sensitivities have been proven. Ophthalmology 2001;108:1088 –1092 © 2001 by the American Academy of Ophthalmology. Infectious crystalline keratopathy (ICK) is an uncommon, chronic, progressive infection of the cornea. It was first described in 1983 by Govoroy et al1 as a noninflammatory bacterial infection in a corneal graft. It is characterized by branching intrastromal opacities associated with minimal inflammation. Predisposing factors include corneal surgery2– 4 (in particular penetrating keratoplasty) in conjunction with topical steroid use, herpes simplex keratitis, acanthamoeba keratitis, and local anesthetic abuse. It is relatively resistant to treatment, despite the fact that isolated organisms grow rapidly in the laboratory. A biofilm has been defined as a functional consortium of microorganisms organized within an extensive exopolymer Originally received: March 3, 1999. Accepted: January 24, 2001. Manuscript no. 99095. 1 Moorfields Eye Hospital, London, England. 2 Department of Pathology, Institute of Ophthalmology, London, England. Reprint requests to John K. G. Dart, Moorfields Eye Hospital, 162 City Road, London, EC1V 2PD, England.

1088

© 2001 by the American Academy of Ophthalmology Published by Elsevier Science Inc.

(extracellular polymer) matrix.5 Biofilm formation has been implicated in the pathogenesis of ICK6 – 8 and has been demonstrated in a single case report of ICK caused by candida.9 Biofilm cannot be visualized by electron microscopy unless fixation techniques are used that stabilize the bacterial extracellular matrix (glycocalyx) without which it collapses and is condensed during the dehydration process.10 Ruthenium red is a polyanionic stain that helps to maintain the structural integrity of the polysaccharide- and glycoprotein-rich bacterial extracellular matrix (glycocalyx), which is otherwise lost during the dehydration process.10 We used ruthenium red to fix corneal specimens in three cases of clinically diagnosed ICK and five additional cases of chronic antimicrobial-resistant keratitis to evaluate these for the presence of bacterial biofilm.

Material and Methods We describe three case reports of patients who were clinically diagnosed as having ICK and who required surgical intervention in ISSN 0161-6420/01/$–see front matter PII S0161-6420(01)00561-9

Fulcher et al 䡠 Biofilm in Infectious Crystalline Keratopathy Table 1. Description of the Demography, Risk Factors, and History of the Episode of the Three Cases with Infectious Crystalline Keratopathy Together with the Clinicopathologic Correlates Demography, Risk Factors, and History of the Episode of Infectious Crystalline Keratopathy

Laboratory Findings on the Corneal Biopsies

Patient 1, female, aged 63, on topical steroids for left herpes zoster ophthalmicus keratouveitis for 8 mos. She had ICK develop (Fig 1A). Corneal scrapes showed gram-positive cocci, and cultures isolated viridans-type streptococci. Then treated with topical penicillin, 5000 units/ml hourly for 6 d, then 2 hourly for 2 wks. Subsequent cultures were negative, but she had a persistent epithelial defect develop for which she had a superficial keratectomy. Patient 2, female, aged 92, on topical steroids, mydriatics, and chloramphenicol ointment for the management of a painful blind right eye caused by rubeosis and bullous keratopathy after a central vein occlusion 2 yrs earlier. She had an asymptomatic corneal abscess develop, which included an area of ICK. She had an evisceration the day after presentation before starting antibiotic therapy.

Patient 3, aged 82, treated with topical chloramphenicol and a soft therapeutic lens for corneal decompensation and recurrent corneal ulceration secondary to pseudophakia and glaucoma. She had ICK develop. After an initial response to topical vancomycin 5%, and steroids, there was a central perforation requiring a keratoplasty.

Culture: no organisms were isolated. Histology: attenuated epithelium, Bowman’s layer, and superficial stromal lamellae. Within the lamellae there were PAS-positive and slightly basophilic deposits. Gram stain showed disruption of the corneal stroma, but no organisms. There was an absence of inflammatory cells. Transmission electron microscopy: bacteria encased in an extracellular matrix consistent with the appearance of a bacterial biofilm. Culture: no organisms were isolated. Histology: area of central ulceration with a perforation. Within the deep stroma, there was a densely cellular area containing acute and chronic inflammatory cells. Within the superficial stroma there was an area of basophilic deposits (Fig 1B), which contained gram-positive material thought to represent clumps of cocci (Fig 1C). A mild chronic inflammatory response was present in the superficial stroma peripherally. Transmission electron microscopy: bacteria encased in an extracellular matrix consistent with a bacterial biofilm (Fig 1D). Culture: no organisms were isolated. Histology: dissolution of the corneal stroma associated with nuclear debris. An intralamellar area of basophilic material was seen within the stroma. No active acute or chronic inflammatory cell infiltrate was present surrounding this region. The Gram stain was negative. Transmission electron microscopy: microorganisms surrounded by extracellular matrix consistent with a bacterial biofilm.

ICK ⫽ infectious crystalline keratopathy; PAS ⫽ periodic acid-Schiff stain.

the management of their disease. The surgically excised specimens were trisected for microbiology, histology, and electron microscopy. The microbiology specimen was placed in normal saline for transportation. In the laboratory, it was homogenized and cultured on solid and liquid phase media. Media used were chosen to ensure that fungal and bacterial pathogens were sought out. The histologic specimen was fixed in formol saline for routine histologic examination and stained using periodic acid–Schiff and Gram stain.11 The electron microscopy specimen was fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer with 0.05% ruthenium red for at least 4 hours. After a buffer wash, the material was postfixed in 2% osmium tetroxide in 0.1 M sodium cacodylate buffer and ruthenium red for another 4 hours. After a final buffer wash, the specimen was dehydrated through graded alcohols, propylene oxide, and embedded in araldite resin. Semithin (0.7 ␮m) and ultrathin (60 – 80 nm) sections were cut with a ReichertJung Ultracut E Ultramicrotome with glass and diamond knives (Reichert-Jung Ltd, Knowle Hill, Milton Keynes, England). Ultrathin sections were mounted on 150 hexagonal mesh copper grids and left either unstained or double stained with uranyl acetate and lead citrate. Sections were taken through the affected areas and viewed by transmission electron microscopy (TEM). TEM was carried out with the Hitachi H600 (Hitachi, Tokyo, Japan) and JEOL 1010 (Japan Electron Optics [UK] Ltd, Welwyn Garden City, Herts, England) electron microscopes. Five other patients with chronic keratitis also had corneal biopsy or penetrating keratoplasty specimens trisected for microbiology, histology, and electron microscopy in the same way to determine whether bacterial biofilm had developed. None of these cases showed the clinical appearances of ICK but were investi-

gated for the presence of ICK because of their chronic course and slow or failed response to medical therapy for microbial keratitis.

Case Reports Infectious Crystalline Keratopathy Patients Three cases, each demonstrating ICK, are described in Table 1. All cases had risk factors for the development of microbial keratitis, and all demonstrated the clinical appearances of ICK. The management of the episode of ICK is summarized together with the laboratory findings on the corneal biopsy specimens. The clinical and histologic features were similar for all three cases, and representative examples are shown in Figure 1A–D.

Cases of Chronic Microbial Keratitis without Crystalline Keratopathy The details of these patients are summarized in Table 2. All patients had multiple risk factors for microbial keratitis developing, and all had organisms cultured from the biopsy or keratoplasty specimens. Patient 6 had multiple positive cultures of Propionibacterium despite prolonged treatment with topical antibiotics to which the organisms were sensitive in vitro. Bacterial biofilm was not identified in any of these specimens using TEM.

1089

Ophthalmology Volume 108, Number 6, June 2001

Figure 1. A, Clinical appearance in patient 1 demonstrating the typical crystalline features of this infection as indicated by the arrow. Note the absence of surrounding stromal inflammation. B, Histologic appearance in patient 2 (stained with periodic acid-Schiff). This case demonstrates the basophilic deposits within the stromal lamellae very nicely (arrow). The infiltrate can be seen extending between the stromal lamellae, and the surrounding stroma is devoid of inflammatory cells. C, Gram stain in patient 2 showing large numbers of gram-positive bacteria within the corneal stroma (hollow arrow) (original magnification, ⫻400). D, Transmission electron micrograph in patient 2 showing typical bacterial biofilm within a degenerate cornea. Relatively normal stromal lamellae, without inflammatory cells, are present below, surrounding a mass of bacteria, many of which are probably nonviable (bar ⫽ 5 ␮m). The higher power inset shows bacteria within the biofilm (filled arrow) surrounded by the extracellular matrix (hollow arrow). Extracellular matrix can also be seen as fibrillar material bridging the gap between adjacent bacterial cells.

1090

Fulcher et al 䡠 Biofilm in Infectious Crystalline Keratopathy Table 2. Risk Factors and Organisms Cultured in Five Patients who had Chronic Keratitis Develop Without the Clinical Features of Infectious Crystalline Keratopathy and in Whom There Was No Histologic Evidence of Biofilm Formation Using the Same Technique as for the Infectious Crystalline Keratopathy Cases. The Organisms Were Cultured Either from Preoperative Corneal Scrapes or from the Surgical Specimen Rick Factors for Developing Chronic Keratitis

Biopsy Specimen(s) Examined

Culture Results

Patient 4: Herpes simplex keratitis leading to penetrating keratoplasty, cataract extraction, graft rejection and longterm use of topical steroids.

Tectonic keratoplasty

Fusarium

Patient 5: Contact lens wearer developing Acanthamoeba keratitis and managed with a penetrating keratoplasty followed by chronic use of topical steroids.

Penetrating keratoplasty

Streptococcus viridans

Patient 6: Atopic keratoconjunctivitis complicated by herpes simplex keratitis. Treated with a penetrating keratoplasty and cataract extraction followed by use of long-term topical steroids.

Multiple keratoplasties and a corneal biopsy for recurrent keratitis at graft interface

Propionibacterium acnes

Patient 7: Rubeotic glaucoma and bullous keratopathy.

Corneal biopsy

Staphylococcus epidermidis

Patient 8: Phacoemulsification cataract surgery followed by a postoperative corneoscleral tunnel, resulting in a corneal abscess with endophthalmitis and recurrent sclerokeratitis.

Biopsy from anterior lamella of corneoscleral tunnel

Streptococcus pneumoniae

Discussion Infectious crystalline keratopathy is a microbial infection of the cornea characterized by a relative lack of inflammation, chronicity, difficulty in isolating the pathogen, and resistance to treatment. It is unclear why a small subgroup of patients have ICK develop as opposed to acute keratitis after microbial inoculation. Both patient predisposing risk factors and properties of the infecting organism seem to play a role. Common predisposing risk factors for ICK include prolonged topical steroid use, preceding corneal surgery (especially penetrating keratoplasty), herpetic keratitis2,12 and local anesthetic abuse,13 and neurotrophic keratopathy14 ICK is usually caused by low virulence organisms, most commonly viridans-type streptococci.2,3,6,8 Other reported organisms include Haemophilus,15 Staphylococcus epidermidis,3,13 Mycobacterium,16 Enterococcus,17 and Candida spp.9,18,19 Rarely, more virulent organisms have been described, including Streptococcus pneumoniae20 and Pseudomonas.16 Microbial proliferation in ICK typically causes lamellar separation within the corneal stroma.7,14 This may provide a surface on which a biofilm can be produced, analogous to the surfaces of a tooth or pipe on which bacterial biofilm formation occurs readily. Under suboptimal growth conditions, many species of bacteria switch from a planktonic phenotype to a sessile phenotype in which bacterial cells secrete a polysacchariderich extracellular polymer matrix (glycocalyx). Colonies of sessile bacteria within this extracellular matrix make up the biofilm. Biofilm-enclosed organisms elicit less of an immune response than their planktonic phenotype and are much more resistant to antibiotics. Viridans-type streptococci are well recognized as being glycocalyx producers.21 Streptococcus sanguis type 2 has been shown to produce exopolysaccharide in the presence of sucrose supplementation in 71% of cases when inoculated into rabbit corneas.8 This produced a clinical picture identical to that of ICK in humans. The bacterial extracellular matrix interferes with

host immune responses by inhibiting attachment to certain macrophages,22 phagocytosis,23 lymphocyte proliferative responses,24 and granulocyte myeloperoxidase release.23 Biofilm protects the organisms from antibiotics, possibly by binding large numbers of antibiotic molecules, thereby requiring much higher concentrations to achieve the same effect.25,26 At the same time, the reduced metabolic activity of organisms within biofilms also reduces their susceptibility to antibiotics. We have demonstrated biofilm formation in three cases of infectious crystalline keratopathy. The intense periodic acid–Schiff stain of the corneal lesions (Fig 1B) is striking and is probably due to the high concentration of mucopolysaccharides in the bacterial extracellular matrix, which are stained well by periodic acid–Schiff stain. The tissue Gram stains (Fig 1C) show the bacteria poorly, which has been previously recognized in infectious crystalline keratopathy.27 Gram stains do not stain bacteria in crystalline keratopathy, probably because many of the bacteria in the biofilm are nonviable and do not retain the Gram stain well during the differentiation step. Gram stainability is known to be a function of the cell wall,28 and our electron micrographs show that a high proportion of the organisms have damaged cell walls or are empty of chromatin, suggesting that they are nonviable, and thus are unlikely to take up the Gram stain well. Bacterial biofilm formation has previously been suggested as the possible pathogenic mechanism for ICK,6 – 8 although it has only been demonstrated in a single case report of ICK caused by Candida.9 This report confirms that typical biofilm formation, identical to that of environmental biofilms, does occur in this condition in some cases. We investigated five additional cases of chronic culture-positive bacterial keratitis, which did not have the clinical appearance of crystalline keratopathy, to see whether these too might show the presence of biofilm with identical techniques; biofilm was not demonstrated in these. This finding suggests that either the appearance or development of bio-

1091

Ophthalmology Volume 108, Number 6, June 2001 film is modified by the intensive antibiotic and anti-inflammatory therapy that these cases were exposed to or that the development of biofilms by the sessile phenotype of some organisms is not their only strategy accounting for resistance to both the host immune response and antibiotic treatment. On the basis of our findings, it seems probable that bacterial biofilm formation is restricted to cases showing clinical evidence of crystalline keratopathy and is not found in other cases of chronic antibiotic-resistant bacterial keratitis. Improved therapy for ICK might result from treatment to modify the biofilm-producing capacity of organisms as an adjunct to antibiotic therapy. Acknowledgments. The authors thank Mr. James Sheldrick for referring one of the cases and providing us with clinical pictures of this patient. We also thank Mr. B. Dowsett and Dr. J. Rogers for their advice on interpreting the electron micrographs, and Mrs. Isabel Moldon without whom the preparation of this manuscript would have been very difficult. We also thank Fight for Sight, London, for funding Mrs. Louise McLaughlin-Borlace.

References 1. Govoroy MS, Stern GA, Hood CI, Allen C. Intrastromal noninflammatory bacterial colonization of a corneal graft [case report]. Arch Ophthalmol 1983;101:1749 –52. 2. Meisler DM, Langston RHS, Naab TJ, et al. Infectious crystalline keratopathy [case report]. Am J Ophthalmol 1984;97: 337– 43. 3. Reiss GR, Campbell RJ, Bourne WM. Infectious crystalline keratopathy. Surv Ophthalmol 1986;31:69 –72. 4. Lubniewski AJ, Houchin KW, Holland EJ, et al. Posterior infectious crystalline keratopathy with Staphylococcus epidermidis. Ophthalmology 1990;97:1454 –9. 5. Costerton JW, Cheng KJ, Geesey GG, et al. Bacterial biofilms in nature and disease. Annu Rev Microbiol 1987;41:435– 64. 6. Ormerod LD, Ruoff KL, Meisler DM, et al. Infectious crystalline keratopathy. Role of nutritionally variant Streptococci and other bacterial factors [review]. Ophthalmology 1991;98: 159 – 69. 7. Elder MJ, Stapleton F, Evans E, Dart JKG. Biofilm-related infections in ophthalmology [review]. Eye 1995;9:102–9. 8. Hunts JH, Matoba AY, Osato MS, Font RL. Infectious crystalline keratopathy. The role of bacterial exopolysaccharide. Arch Ophthalmol 1993;111:528 –30. 9. Elder MJ, Matheson M, Stapleton F, Dart JKG. Biofilm formation in infectious crystalline keratopathy due to Candida albicans [case report]. Cornea 1996;15:301– 4. 10. Barber PM, Fletcher JM, Newman HN. Demonstration of the glycocalyces associated with three oral gram-negative bacterial species using a modern acrylic resin technique. J Struct Biol 1993;110:180 –7.

1092

11. Bancroft JD, Stevens A, eds. Theory and Practice of Histological Techniques. New York: Churchill Livingstone, 1977; 210 –1. 12. Zabel RW, Mintsioulis G, MacDonald I, Tuft S. Infectious crystalline keratopathy. Can J Ophthalmol 1988;23:311– 4. 13. Kintner JC, Grossniklaus HE, Lass JH, Jacobs G. Infectious crystalline keratopathy associated with topical anesthetic abuse [case report]. Cornea 1990;9:77– 80. 14. Matsumoto A, Sano Y, Nishida K, et al. A case of infectious crystalline keratopathy occurring long after penetrating keratoplasty. Cornea 1998;17:119 –22. 15. Groden LR, Pascucci SE, Brinser JH. Haemophilus aphrophilus as a cause of crystalline keratopathy. Am J Ophthalmol 1987;104:89 –90. 16. Hu FR. Infectious crystalline keratopathy caused by Mycobacterium fortuitum and Pseudomonas aeruginosa. Am J Ophthalmol 1990;109:738 –9. 17. Lam S, Meisler DM, Krachmer JH. Enterococcal infectious crystalline keratopathy [case report]. Cornea 1993;12:273– 6. 18. Wilhelmus KR, Robinson NM. Infectious crystalline keratopathy caused by Candida albicans [case report]. Am J Ophthalmol 1991;112:322–5. 19. Rhem MN, Wilhelmus KR, Font RL. Infectious crystalline keratopathy caused by Candida parapsilosis [case report]. Cornea 1996;15:543–5. 20. Matoba AY, O’Brien TP, Wilhelmus KR, Jones DB. Infectious crystalline keratopathy due to Streptococcus pneumoniae. Possible association with serotype. Ophthalmology 1994;101:1000 – 4. 21. Mills J, Pulliam L, Dall L, et al. Exopolysaccharide production by viridans streptococci in experimental endocarditis. Infect Immun 1984;43:359 – 67. 22. Whitnack E, Bisno AL, Beachey EH. Hyaluronate capsule prevents attachment of group A streptococci to mouse peritoneal macrophages. Infect Immun 1981;31:985–91. 23. Schwarzmann S, Boring JR III. Antiphagocytic effect of slime from a mucoid strain of Pseudomonas aeruginosa. Infect Immun 1971;3:762–7. 24. Brown EJ, Joiner KA, Cole RM, Berger M. Localization of complement component 3 on Streptococcus pneumoniae: anticapsular antibody causes complement deposition on the pneumococcal capsule. Infect Immun 1983;39:403–9. 25. Costerton JW, Cheng KJ, Geesey GG, et al. Bacterial biofilms in nature and disease. Annu Rev Microbiol 1987;41:433– 64. 26. Anwar H, Strap JL, Costerton JW. Establishment of aging biofilms: possible mechanism of bacterial resistance to antimicrobial therapy [review]. Antimicrob Agents Chemother 1992;36:1347–51. 27. Lee WR. Ophthalmic Histopathology. New York: SpringerVerlag, 1993;302. 28. Popescu A, Doyle RJ. The Gram stain after more than a century. Biotech Histochem 1996;71:145–51.