In Vivo and In Vitro Collogenolytic Activity of ...

Investigative Ophthalmology & Visual Science, Vol. 31, No. 11, November 1990 Copyright © Association for Research in Vision and Ophthalmology

In Vivo and In Vitro Collogenolytic Activity of Acanthamoeba castellanii YuGuang He, Jerry Y. Niederkorn, James P. McCulley, George L. Stewart, Dale R. Meyer, Robert Silvany, and Joel Dougherty Axenic cultures of Acanthamoeba castellanii contained a collagenolytic enzyme that digested collagen shields and purified collagen in vitro. Specificity of biologic activity was determined by the addition of selected enzyme inhibitors to the assays and revealed that the parasite-conditioned medium contained both collagenase and lower concentrations of other proteolytic enzymes. However, most of the collagenolytic and pathogenic activity was directly attributable to specific collagenase. Intrastromal injection of sterile, Acanthamoeba-conditioned culture medium into naive Lewis rats produced cornea! lesions clinically similar to and closely resembling those found in biopsy specimens of human patients diagnosed with acanthamoebic keratitis. Histopathologic analysis revealed moderate-to-severe neutrophil infiltration, disruption of stromal lamellae, and edema. Identical pathologic sequelae were produced by intrastromal injection of purified collagenase (25 units/ml). The pathogenicity of the soluble parasite-derived product was removed by passage over affinity columns armed with antibody specific for collagenase. These results indicated that soluble parasite-derived factors were capable of producing lesions characteristic of acanthamoebic keratitis and that the pathogenicity of these factors was either directly or indirectly attributable to specific collagenase activity. Invest Ophthalmol Vis Sci 31:2235-2240,1990

Acanthamoebic keratitis is a serious sight-threatening disease caused by a ubiquitous, pathogenic, freeliving amoeba. Since its original description in 1975, the number of documented cases of acanthamoebic keratitis has increased dramatically and continues to rise.1'2 Although there have been major advances in the diagnosis and management of this ulcerative corneal disease,3 the pathogenesis of acanthamoebic keratitis remains poorly understood. Consistent clinical signs have been described and include a characteristic ringlike infiltrate, radial keratoneuritis, dendritiform epithelial lesions, stromal edema, and ulceration. Histopathologically, the necrotic stroma is infiltrated with neutrophils, but there is an interesting and conspicuous dearth of lymphocytes.4'5 The most severe stromal necrosis is typically associated with a heavy neutrophil infiltrate, and it has been speculated that proteases secreted by neutrophils contribute to extensive collagenolysis in infected corneas.5 However, we report results indicating that the parasite may directly contribute to stromal

degradation and the pathologic sequelae through the elaboration of collagenolytic enzymes. Materials and Methods Animals

Adult female Lewis strain rats were purchased from Harlan Sprague-Dawley (Indianapolis, IN) and used as experimental subjects when at least 8 weeks of age. Animals were handled in accordance with the ARVO Resolution on the Use of Animals in Research. Parasite Cultivation

Acanthamoeba castellanii (ATCC #30868), originally isolated from a human cornea, was obtained from American Type Culture Collection (Rockville, MD). Parasites were grown as axenic cultures in PYG medium containing 0.5 mM Ca++ as previously described6 at a concentration of 3 X 106 organisms/ml at 35°C for 4 weeks. The Acanthamoeba-conditioned culture medium (CCM) was then clarified by centrifugation (250 X g for 10 min), the supernatant filtered with a 0.22 ^M membrane filter (Millipore, Bedford, MA), and the filtrate stored at 4°C until used.

From the Department of Ophthalmology, University of Texas Southwestern Medical Center at Dallas, Dallas, and Center for Parasitology, University of Texas at Arlington, Arlington, Texas. Supported in part by NIH grants EY 07641 and EY 03650, an unrestricted grant from Research to Prevent Blindness, Inc., New York, and a grant from Bausch and Lomb, Inc. Reprint requests: Dr. J. Y. Niederkorn, Department of Ophthalmology, University of Texas Southwestern Medical Center at Dallas, Dallas, TX 75235.

Purification of Conditioned Culture

Protein was extracted from the CCM by precipitation with ammonium sulfate (Sigma, St. Louis, MO) 2235

2236

INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / November 1990

at 65% saturation on ice for 30 min. The precipitated proteins were collected by centnfugation (5850 X g for 30 min) and the pellet resuspended to the original volume in fresh, serum-free PYG medium. The crude extract was subjected to affinity column purification using sheep anticollagenase antibody. Briefly, Econo-Pak columns packed with 2.0 ml of Affi-Gel protein A agarose (Bio-Rad, Richmond, CA) were each equilibrated with 10 ml of binding buffer (10 mM Na2HPO4 and 0.15 M NaCl at pH 8.2). One ml of sheep anticollagenase (microbial origin) immunoglobulin G (IgG; 4 mg/ml; Binding Sites, Birmingham, England) was applied to each column followed by 20 ml of binding buffer (flow rate, 1 ml/ min). These anticollagenase IgG-armed columns were used to remove suspected coUagenolytic activity from CCM by passage of 2 ml of CCM through each of the three columns (flow rate, 1 ml/min). Controls included 2.0-ml samples of CCM passed through protein A columns which had not been armed with anticollagenase antibody and 1.0-ml samples of a solution of commercially purified collagenase (25 units/ml; Worthington, Freehold, NJ) passed through columns armed with anticollagenase IgG. Removal of collagenase by this method was assessed in vivo and in vitro by the methods described. In Vitro Degradation of Collagen Shields

Collagen corneal shields (24 hr Collagen Shield; Chiron, Irvine, CA) were cut into quarters, and each quarter was placed into individual wells of a 24-well tissue culture plates (Costar, Cambridge, MA) along with 0.5 ml of CCM with and without 3 X 106/ml viable parasites. Control wells contained fresh PYG culture medium and collagen shields. Cultures were then incubated for 24 hr at 35°C and examined with an inverted compound microscope at various times throughout the incubation period. The specificity of the coUagenolytic activity was determined by adding either N-ethylmaleimide and phenymethylsulfonyl fluoride and/or sodium ethylenediaminetetraacetate (EDTA-Na) to companion cultures to inhibit nonspecific proteases and collagenase, respectively.7'8 In Vitro Collagenolysis Assays Degradation of purified collagen was assessed by the Coomassie blue assay and by enzyme-linked immunosorbent assay (ELISA). Both procedures are routinely used for the rapid screening and quantification of coUagenolytic activity in biologic samples.910 For Coomassie blue assays, purified type I collagen (Virogen 100; Collagen, Palo Alto, CA) was suspended in phosphate-buffered saline (PBS, pH, 7.2) at a concentration of 30 Mg/ml. Purified collagen was

Vol. 31

dispensed into individual wells (100 ^I/well) of 96well microtiter plates (Costar, Cambridge, MA) and held for 7 days at 4°C or 35°C for 48 hr. Unbound collagen solution was discarded and the plates thoroughly washed with deionized water, and air dried. Aliquots of undiluted and serial dilutions of CCM and purified collagenase (Worthington) were prepared in a buffer consisting of 0.05 M Tris HC1, 0.15 M NaCl, and 5 mM CaCl2 and were added to each well (100 /il/well). Plates were then incubated at 35°C in a humidified chamber for 48 hr, the supernate discarded, and the wells washed thoroughly with "assay buffer" (50 mM Tris HC1, 100 mM NaCl, 10 mM CaCl2, and 0.02% NaN3; pH 7.5). The wells were air dried, and Coomassie blue stain G250 was then added to each well. Coomassie blue staining was evaluated by inspecting individual microtiter wells under a dissecting microscope (8X). Scoring was reported as: 0, no trace of blue stain; 1+, faint blue stain discernible only under 8X magnification; 2+, residual blue staining with a faint but discernible circle of blue stain; 3+, circle of blue stain evident but not completely circumscribing the bottom of the microtiter well; and 4+, complete, deeply stained blue circle evident in well and detectable without use of microscope. An ELISA procedure was used to confirm results from the Coomassie blue assays. Collagen-coated microtiter plates were prepared as described, and serial dilutions of CCM were added to each well (100 ^1/ well). After the samples were added, the plates were incubated for 4-5 hr at 35 °C. Reactions were arrested by the addition of 100 fi\ of 50 mM EDTA. Plates were then washed thoroughly with PBS-Tween 20 solution, and 100 n\ of a rabbit anticollagen type I antisera (diluted 1:100; Chemicon, Temecula, CA) was added to each well. The plates were reincubated for 45 min at room temperature, washed thoroughly with PBS-Tween 20, and 100 ^1 of goat antirabbit IgG peroxidase conjugated antiserum (Accurate Chemical & Scientific, Westbury, NY) diluted 1:1000 in PBS-Tween 20 containing 1% bovine serum albumin was added to each well. Plates were reincubated for 45 min at room temperature, washed with PBSTween 20 solution, and 100 /x\ of the enzyme substrate (8 mM O-phylenediamine: 0.3% H2O2; 1:9) was added to each well. Plates were then incubated an additional 15 min at room temperature and the reaction arrested by the addition of 100 ^1 of a 4 N sulfuric acid solution. Substrate degradation was evaluated colorimetrically using a spectrophotometer (MR700; Dynatech, Baton Rouge, LA) with an absorbance band of 450 nm. In this experiment, a value above 0.2 was considered positive. Controls consisted of normal rabbit serum instead of anticollagen and

No. 11

2237

COLLAGENOLYTIC ACTIVITY OF ACANTHAMOEDA / He er ol

fresh PYG medium instead of CCM. Purified collagenase (Worthington), at a concentration of 25 units/ ml, served as the positive control standard.

Table 2. Degradation of purified collagen by Acanthamoeba-conditioned culture medium (CCM) Medium

Intrastromal Injection of Parasite Supernates and Purified Collagenase

The pathogenic properties of parasite products were assayed by the intrastromal injection of CCM diluted in PYG medium. Panels of Lewis rats (4-6 animals/group) were anesthetized with a combination of ketamine (100 mg/kg; Parke-Davis, Morris Plains, NJ) and promazine (60 mg/kg; Wyeth, Philadelphia, PA) given intramuscularly. Purified collagenase (between 5-100 units/ml), affinity purified CCM, crude CCM, or fresh PYG culture medium supplemented with bovine serum albumin (200 /zg protein/ml; Sigma) was injected directly into the corneal stroma using a microinjection technique described previously." Additional experiments evaluated the effect of the collagenase inhibitor, EDTANa, on the pathogenic potential of CCM and purified collagenase. The EDTA-Na was added to undiluted, sterile CCM and purified collagenase (25 units/ml) to create a final concentration of 20 mM. The enzyme/ EDTA-Na preparations were incubated for 30 min at room temperature and injected intrastromally into panels of naive Lewis rats (4 rats/group). Eyes were observed daily using a slit lamp. Animals were killed on days 1, 2, 3, and 7 postinjection, and the eyes were processed for histopathologic analysis. Results Collagen shields incubated in CCM were partially disintegrated after 4 hr and were completed degraded by 8 hr (Table 1). Similar results were found with shields incubated in purified collagenase solution (25

CCM CCM CCM CCM

Fresh PYG Collagenase Collagenase Collagenase

Coomassie blue reaction

Inhibitor None NEM + PMSF EDTA-Na NEM + PMSF + EDTA-Na None None NEM + PMSF EDTA-Na

1+ 1+ 2+ 4+ 4+ 0

0 4+

Collagenase was used at a concentration of 25 U/ml. Inhibitors were added at the following concentrations: NEM = 10 nM, PMSF = I mM, and EDTA-Na = 20 mM. There were seven specimens in each test group. Coomassie blue staining intensity was evaluated by inspecting individual microtiter wells under a dissecting microscope. (X8).

units/ml). By contrast, collagen shields incubated in fresh PYG culture medium maintained their structural integrity and could be readily identified after 24 hr of incubation. The specificity of the collagenolytic activity was ascertained by adding specific enzyme inhibitors to companion cultures. A combination of nonspecific protease inhibitors and EDTA-Na completely blocked the degradation of collagen shields; individual inhibitors were less effective. Thus, the parasitederived CCM probably contained considerable amounts of collagenase and lesser concentrations of other proteolytic enzymes (Table 1). Specific collagenolytic activity of the parasite-derived product was confirmed by incubating CCM with purified collagen under the same conditions as demonstrated by the Coomasie blue G25O assay (Table 2), later substantiated by ELISA (Figs. 1, 2). Collectively, the results 1.4-00 j

Table 1. Degradation of collagen shields by Acanthamoeba-conditwned culture medium Medium CCM CCM CCM CCM

Fresh PYG Collagenase Collagenase Collagenase Collagenase

1.200*£ c

1.000 • •

Degradation time* (hr)

g

0.800 ••

Inhibitors

«S

0.600-

None NEM + PMSF EDTA-Na NEM + PMSF + EDTA-Na None None NEM + PMSF EDTA-Na NEM ± PMSF + EDTA-Na

8 8 16 >24 >24 4 4 >24 >24

§

0.400 - •

Collagenase was used at a concentration of 25 U/ml. Inhibitors added at the following concentrations: NEM = 10 mM, PMSF = 1 mM, EDTA-Na = 20 mM. Degradation time is defined as the interval between beginning of assay and time when absolutely no remnants of collagen shields could be detected by inspection under a dissection microscope (X8). There were four specimens in each experimental group.

CZ) •• BS3 ES

FRESH MEDIUM CCM w / o PARASITES CCM with PARASITES PURE COLLAGENASE

0.200 •• 0.000EXPERIMENTAL GROUP

Fig. 1. Degradation of purified collagen by parasite-conditioned culture medium (CCM). Wells of microtiter plates were coated with purified collagen and incubated for 48 hr with sterile CCM alone, CCM plus 3 X 10/ml viable Acanthamoeba castallanii, or purified collagenase (25 U/ml). Residual collagen was quantified spectrophotometrically by ELISA using rabbit anti-collagen Type I. Nondigested collagen is represented by the optical density (OD) of the supernate in the microtiter wells evaluated by ELISA.

2208

INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / November 1990

Vol. 01

1.400-•

q 6

0.600-f 0.400-

DILUTION OF CCM (Log 2 )

Fig. 2. Titration of collagenolytic activity of parasite-conditioned culture medium. Twofold serial dilutions of filtered CCM were evaluated for collagenolytic activity by ELISA. The presence of residual, undigested collagen was detected by ELISA and represented here as the optical density (OD) of the supernate following the immunoiogic indicator response measured at 450 nm.

confirmed the suspicion that collagenase was the most important enzyme responsible for the observed proteolytic activity. The rapid and complete digestion of purified collagen and collagen shields by the parasite-derived products suggested that the same degradative processes might occur in situ in the infected cornea. This was examined by directly injecting sterile CCM into the corneal stroma of naive Lewis rats. Significant changes were apparent within 24 hr of injection. Corneas became opaque, edematous, inflamed, and developed ringlike infiltrates (Figs. 3, 4) reminiscent of those found in patients diagnosed with acanthamoebic keratitis. A more severe acute reaction, with very prominent corneal edema is shown in Figure 4. Pathologic sequelae intensified over the next 48-72

Fig. 3. Production of a characteristic ringlike infiltrate in a Lewis rat cornea 24 hr after an intrastromal injection of 1.0 fil of undiluted, sterile Acanthamoeba-condftioned culture medium. Note ring-like infiltrate (bracket). Injection site indicated by arrow.

Fig. 4. Acute pathologic effects of sterile parasite-conditioned culture medium injected inlrastromally in Lewis rats. Approximately 1 ^1 of sterile, undiluted CCM was injected intrastromally. Cornea is severely edematous, opaque, and displays ringlike infiltrate (bracket) 24 hr after injection of CCM. Injection site is indicated by arrow.

hr in some cases culminated in severe stromal ulceration and frank necrosis by day 7 (Fig. 5). The role of parasite-derived collagenase in producing ulcerative keratitis was evaluated further in experiments in which corneas of naive rats were injected with either purified collagenase (5-100 units/ ml) or sterile CCM. Corneas injected with 100 units/ml collagenase developed intense necrosis that bore little resemblance to the keratitis observed in corneas injected with CCM. By contrast, injection of collagenase at a concentration of 25 units/ml produced pathologic responses that closely mimicked the keratitis produced by CCM. Histologic examination of corneas injected with CCM confirmed the clinical impressions and revealed moderate-to-severe neutrophil infiltration, corneal edema, and extensive disruption of stromal lamellae (Fig. 6). These patho-

Fig. 5. Stromal ulceration in a Lewis rat cornea injected intrastromally with 1.0 n\ of undiluted, sterile CCM 7 days earlier.

No. 11

2239

COLLAGENOLYTIC ACTIVITY OF ACANTHAMOEBA / He er ol

logic sequelae were not due to the host's nonspecific response to a foreign protein introduced into the stroma since an intrastromal injection of fresh culture medium, supplemented with bovine serum albumin (200 jug/ml) to produce a concentration equivalent to the protein concentration of undiluted, Acanthamoeba-conditioned culture medium, did not elicit either inflammatory responses, necrosis, or disruption of the stromal lamellae that could be detected either clinically or histologically (data not shown). By contrast, corneas from rats injected with purified collagenase (25 units/ml) had histopathologic features that were essentially identical to those observed in rats injected with CCM (data not shown), thereby demonstrating the specificity of collagenase in the pathologic process. Additional experiments verified the contribution of collagenase in the pathogenic effect produced by the parasite culture medium. Collagenase was removed from CCM by passage over protein A affinity columns activated with anticollagenase antibody. Affinity-purified CCM specimens were tested for resid-

?>•

—o

50

75

100

* OF MEDIUM CONTAINING AFFINITY PURIFIED CCM

Fig. 7. Removal of collagenolylic activity of CCM by anti-collagenase affinity column purification. Collagenase activity of CCM was removed by affinity columns activated with sheep anti-collagenase antibody. Collagenase-depleted CCM was tested undiluted or in various ratios with untreated CCM. Untreated CCM contained cither 0%, 20%, 50%, or 90% collagenase-depleted CCM. Collagenolytic activity was evaluated by measuring residual undigested collagen by the Coomassie Blue G25O assay.

ual collagenolytic activity by the previously described Coomassie blue assay. As shown in Figure 7, affinitycolumn purification removed virtually all of the collagenolytic activity of the parasite culture medium. Intrastromal injection of the affinity-column eluate produced only minimal keratitis and was insignificantly different from that produced by the intrastromal injection of fresh culture medium supplemented with bovine serum albumin (data not shown). By contrast, passage of CCM over protein A columns that were not armed with anticollagenase antibody did not reduce either the collagenolytic activity or the severe keratitis produced by intrastromal injection of CCM. Moreover, the severity and persistence of the keratitis in this group was indistinguishable from the original CCM. The pathogenic activity of CCM was almost entirely eliminated by treatment with the collagenase inhibitor, EDTA-Na, before intrastromal injection. A mild, transient keratitis was observed in eyes injected with EDTA-Na-treated CCM. However, the affected eyes never developed ringlike infiltrates, ulceration, or vascularization. Corneas became slightly edematous, but they eventually returned to near normal clarity within 5 days. We attribute the inability of EDTA-Na to remove the pathogenic effect of CCM completely to the limited ability of EDTA-Na to chelate calcium ions present in the corneal tissues. Discussion

Fig. 6. Histopathology of parasite-produced collagenolytic factor. Sterile, undiluted CCM was injected (1.0 jtl) intrastromally 24 hr earlier. Note disruption of collagen lamellae, edema, and intense neutrophil infiltrate. (H&E stain; original magnification X62.5).

Although the diagnosis of acanthamoebic keratitis has increased sharply in recent years, surprisingly little is known about the pathophysiology of this dis-

2240

INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / November 1990

ease. The characteristics of the parasite would predict a high incidence of corneal infection. The cysts of Acanthamoeba are ubiquitous and can survive under a wide range of seemingly hostile environmental conditions including: under the ice in freshwater lakes, in hot tubs, in chlorinated swimming pools, and under conditions of low pH.12 Moreover, the major predisposing condition for infection—soft contact lens wear—occurs in over 20,000,000 individuals in this country alone. These conditions would lead one to expect a high incidence of acanthamoebic keratitis, yet the frequency of disease is remarkably low. Our results provide important clues for understanding the pathophysiology of acanthamoebic keratitis. In vitro findings revealed that the parasite produces a collagenolytic enzyme that is capable of digesting purified collagen. More importantly, in vivo studies showed the pathogenic features of this parasite product. Intrastromal injection of undiluted, sterile, Acanthamoeba-conditioned medium produced lesions that resembled acanthamoebic keratitis in several key areas including: edema, corneal opacity, neutrophilic infiltration, ring infiltrate, stromal disintegration, and keratocyte necrosis. Moreover, the pathologic sequelae induced by the parasite-derived product could be mimicked by the injection of purified collagenase (25 units/ml). Affinity-column removal of collagenase eliminated the pathogenicity of the soluble parasite product and thereby confirmed the role of collagenase in producing corneal lesions. It has been suggested that the characteristic stromal necrosis of acanthamoebic keratitis is the result of collagenolytic enzymes released from lysosomes of neutrophils that infiltrate infected corneas.5 However, our findings raise the possibility that the parasites might also directly contribute to collagenolysis and stromal thinning. These results further demonstrate that parasite-derived collagenase could either directly or indirectly stimulate the migration of neutrophils. It is conceivable that the infiltrating neutrophils exacerbate collagen disintegration by releasing collagenolytic enzymes as suggested by others.5 It should be emphasized, however, that the pathologic sequelae of stromal lysis, edema, opacity, ring infiltrate formation, and neutrophilic inflammation occurred in the absence of the infectious parasite and were initiated entirely by soluble parasite-derived factors. It is not surprising that Acanthamoeba cultures have collagenolytic activity since elastase activity has been demonstrated in Naegleria and Acanthamoeba species.13

Vol. 31

It is tempting to speculate that in the initial infection, Acanthamoeba-derived collagenase acts as a priming agent to produce collagenolysis, edema, and infiltration of neutrophils. The infiltrating, activated neutrophils in turn might exacerbate collagen lysis and keratocyte necrosis through the release of various nonspecific lysosomal enzymes and oxygen metabolites. The validity of this hypothesis, however, awaits confirmation. Key words: cornea, Acanthamoeba, keratitis, collagenase, stroma

Acknowledgments The authors thank Jessamee Mellon for excellent technical assistance.

References 1. Moore MB: Acanthamoeba keratitis. Arch Ophthalmol 106:1181, 1988. 2. Jones DB: Acanthamoeba: The ultimate opportunist? Am J Ophthalmol 102:527, 1986. 3. Moore MB and McCulley JP: Acanthamoeba keratitis associated with contact lenses: Six consecutive cases of successful management. Br J Ophthalmol 73:271, 1989. 4. Mathers W, Stevens Jr GS, Rodrigues M, Chan CC, Gold J, Visvesvara GS, Lemp MA, and Zimmerman LE: Immunopathology and electron microscopy of Acanthamoeba keratitis. Am J Ophthalmol 103:626, 1987. 5. Garner A: Pathology of corneal acanthamoebic infection. In The Cornea: Transactions of the World Congress on the Cornea III, Cavanagh HD, editor. New York, Raven Press, 1988, pp. 535-539. 6. Silvany RE, Dougherty JM, McCulley JP, Wood TS, Bowman RW, and Moore MB: The effect of currently available contact lens disinfection systems on Acanthamoeba casiellanii and Acanthamoeba polyphaga. Ophthalmology 97:286, 1990. 7. Shaw E: Chemical modification of active-site directed reagents. In The Enzyme: I. Structure and Control, Boyer PO, editor. New York, Academic Press, 1970, pp. 94-96. 8. Nagai Y, Lapiere CM, and Gross J: Tadpole collagenase: Preparation and purification. Biochemistry 5:3123, 1966. 9. Nethery A, Lyons JG, and O'Grady RL: A spectrophotometric collagenase assay. Anal Biochem 159:390, 1986. 10. Yoshioka H, Oyamada I, and Usuku G: An assay of collagenase activity using enzyme-linked immunosorbent assay for mammalian collagenase. Anal Biochem 166:172, 1987. 11. Niederkorn JY, Peeler JS, and Mellon J: Phagocytosis of particulate antigens by conreal epithelial cells stimulates interleukin-1 secretion and migration of Langerhans cells into the central cornea. Regional Immunology 2:83, 1989. 12. Auran JD, Starr MB, and Jakobiec FA: Acanthamoeba keratitis. Cornea 6:26, 1987. 13. Ferrante A and Bates EJ: Elastase in the pathogenic free-living amoebae Naegleria and Acanthamoeba spp. Infect Immun 56:3320, 1988.