A relevant enzyme in granulomatous reaction, active matrix metalloproteinase-9, found in bovine Echinococcus granulosus hydatid cyst wall and fluid
- First Online:
- Cite this article as:
- Marco, M., Baz, A., Fernandez, C. et al. Parasitol Res (2006) 100: 131. doi:10.1007/s00436-006-0237-5
- 78 Views
In addition to the ability of matrix metalloproteinases (MMP) to degrade components of the extracellular matrix and their involvement in pathology-related processes of tissue remodeling, they were recently reported to enhance inflammation by activation of proinflammatory cytokines, or their release from the cell surface. In the work reported here, proteolytic activity previously found for hydatid cysts was further characterized as MMP-9. Active host MMP-9 was found in walls and fluids of bovine hydatid cysts of Echinococcus granulosus in the environment of granulomatous reaction. Pooled walls and fluids of hydatid cysts obtained from infected cattle were processed. Strong proteolytic activity was detected by zymography. The proteolytic fraction was purified by anion exchange and gelatin-agarose affinity chromatography. Major proteinases of the purified fraction were subjected to mass spectrometry and their identities were further confirmed by Western blotting using commercial anti-human MMP-9 monoclonal antibodies. Two proteinases were characterized as latent and active forms of host MMP-9. Using the same antibody for immunoblot, activity was localized, in paraffin-embedded sections of the parasite and the local host environment, to epithelioid and giant multinucleated cells. It is proposed here that MMP-9 is secreted by specialized host cells of monocytic lineage (epithelioid/giant cells) as an effector, in an attempt to digest the persistent foreign body. In vivo activation of MMP-9 suggests its involvement in inflammatory reaction and in the chemotaxis of inflammatory cells to the cyst. However, E. granulosus can deal efficiently with MMP-9. Research is suggested into possible immune evasion mechanisms, including the secretion of an inhibitory molecule.
A universally distributed zoonosis, hydatid disease is caused by the cestode Echinococcus granulosus. The parasite life cycle involves two mammalian hosts, causing cystic (unilocular) hydatid disease affecting humans and a variety of livestock species. The disease is characterized by steady growth of fluid-filled cysts in the host’s internal organs, especially liver and lung. Cysts are bounded by a two-layer parasite-derived hydatid cyst wall (HCW) containing hydatid cyst fluid (HCF). The HCW is composed of an inner germinal layer, parasite tissue, and an outer acellular laminated layer produced by the germinal layer, rich in carbohydrate, elastic, and mechanically resistant (Eckert and Deplazes 2004). The biology of Echinococcus was previously described (Thompson 1995).
The host–parasite relationship is characterized by the coexistence of chronic infection with detectable antibodies and cellular response against the parasite. It was widely accepted that the physical barrier to host immune cells represented by the laminated layer plays a foremost role in parasite immune evasion. However, the laminated layer, or the HCW altogether, was reported to be permeable to host macromolecules (Varela-Díaz and Coltorti 1972, 1973; Coltorti and Varela-Díaz1972, 1974). Consistently, host albumin and Ig are found in HCF and HCW (Kagan and Norman 1963; Chordi and Kagan 1965; Shapiro et al. 1992; Díaz et al. 1997). Therefore, the live germinal layer is potentially exposed to humoral factors on both its external and internal surfaces.
The germinal layer thus constitutes the only potential barrier to the passage of macromolecules across the host–parasite interface. Extensive evidence shows that host macromolecules find their way into the HCF in vivo, and apparently do so in a nonspecific way. However, the concentrations of host IgG and albumin (used as markers) were found to be lower in cyst fluid than in host plasma (by three to four orders of magnitude) (Coltorti and Varela-Diaz 1972).
The production of proteolytic enzymes and their release as excretory–secretory products were reported to be essential to parasite survival in the host environment for various helminthes and/or in parasite evasion of the host’s immune system (Petralanda et al. 1986; Robertson et al. 1989; Uparanukraw et al. 2001; Wada et al. 1998; McKerrow 1989; McKerrow and Doenhoff 1988; McKerrow et al. 1990; Matthews 1982; Morris and Sakanari 1994; Trancart et al. 1992).
Proteolytic activity in hydatid cysts was studied in a previous work and was characterized as metalloproteinases on account of results of a battery of inhibition assays, resulting in positive inhibition for EDTA and o-phenantroline (Marco and Nieto 1991). Metalloproteolytic activity in E. granulosus metacestodes was found in HCF and HCW. Major fractions of apparent molecular weight of 92, 84, and 68 kDa were determined by gelatin zymography, showing enhanced activity in the presence of Ca+2 ions. No gelatinolytic activity were found in somatic and excretion/secretion preparations obtained from protoscoleces (PEs) of viability higher than 95% (Marco and Nieto 1991).
The above characteristics are similar to those of matrix metalloproteinases (MMPs), a large family (more than 20 members) of Zn+2-and Ca+2-dependent endopeptidases involved in processes of tissue remodeling and chronic inflammation (Egeblad and Werb 2002; Bjorklund and Koivunen 2005; Van den Steen et al. 2002; Opdenakker et al. 2001, 2003; Okada et al. 1992; McCawley and Matrisian 2001). MMPs possess broad and overlapping specificities and have the capacity to collectively degrade all components of the extracellular matrix (ECM). MMPs are produced by many types of cell, including lymphocytes and granulocytes, and particularly by activated macrophages. MMPs are secreted as proenzymes, which are activated by proteolytic cleavage and regulated by a family of inhibitors called tissue inhibitors of matrix metalloproteinases (TIMPs), constitutively produced by a variety of cells. Changes in MMP activity are thus dependent on the balance between production and activation of MMPs and local levels of TIMPs. In addition to a proteolytic activity for degrading ECM proteins, other functions of MMPs were reported in the past few years. MMPs can increase inflammation by activating proinflammatory cytokines or releasing them from the cell surface. The generation of chemotactic fragments from ECM proteins may also contribute to the recruitment of inflammatory cells (Fridman et al. 1995, 2003; Nagase 1997, Nagase and Woesner 1999).
Isolation, purification, and characterization of different hydatid cyst pools led to the determination of a pattern of proteolytic activity by gelatin zymography, revealing latent and activated forms of MMP-9 secreted from the chronic reaction surrounding the cysts upon induction by the parasite—the active form binding strongly to HCW.
Materials and methods
Cyst fluid samples and cyst tissue extracts
E. granulosus cysts were obtained from sheep and cattle livers and lungs supplied by Uruguayan abattoirs. Human cysts were obtained upon surgery. HCFs from different hosts were prepared according to Marco and Nieto 1991.
Hydatid cysts were punctured aseptically. HCF containing PEs (fertile material) was allowed to settle for 20 min at room temperature before careful removal of the cyst fluid. EDTA and sodium azide were immediately added to the fluid to a final concentration of 5 mM and 0.02%, respectively.
The extraction procedure used on the HCW comprised initial washing with phosphate-buffered saline (PBS), followed by detergent extractions [PBS-Tween 20 (0.05%)] aimed at solubilizing the cellular layer, and by pulverization under liquid nitrogen, aimed at disorganizing the resilient laminated layer. Fractions of the insoluble pellet obtained after pulverization were further extracted with a battery of different solvents, including 2 M of NaCl, Triton X-100, and guanidine. A cocktail of protease inhibitors consisting of iodoacetamide at 0.2 mM, EDTA at 5 mM, and pepstatin A (Sigma, St. Louis, MO, USA) at 2 μg/ml was added at all steps during extraction (Díaz et al. 1997).
Preparation of control human MMP-9
Frozen 1-ml aliquots of human fibrosarcoma HT1080 cells (supplied by ATCC) (1–2 per 106 cells/ml) were brought to 37°C and washed in 10 ml of RPMI (Gibco, UK) enriched with 10% fetal bovine serum. Cells were resuspended in 20 ml of the same medium and seeded into a 75-cm3 flask (Falcon, Becton-Dickinson, UK). The medium was replaced every 48 h and the cell cultures grew to confluence in approximately 10 days. Cells were later washed in situ with Hanks balanced salt solution and incubated for a further 24 h in serum-free RPMI added with dexamethasone (1 μM) to obtain latent MMP-9 (Rasheed et al. 1974).
Latent MMP-9 was purified from the supernatants obtained by gelatin-agarose affinity chromatography. Activation was accomplished using the following in vitro methodology for MMP activation using p-aminophenylmercuric acid (APMA) (Sigma): add APMA to latent MMP-9 solution to a final concentration of 1 mM, incubate sample at room temperature, monitor activity vs activation time using fluorogenic substrate, stop reaction when it plateaus by inserting the tube in ice, and dialyze vs standard metalloproteinase buffer at 4°C.
Purification of proteinase activity from HCF
Two liters of pooled bovine HCF were centrifuged at 10,000 rpm for 30 min at 4°C and applied to a Q-Sepharose column (2.5 per 10 cm) previously equilibrated in 20 mM phosphate buffer at pH 7.2. After washing with the phosphate buffer, 0.2 M of NaCl at pH 7.2, the proteolytic activity was eluted, increasing the NaCl concentration up to 0.4 M. The fraction obtained by chromatography in the anionic exchanger was purified using gelatin-agarose affinity chromatography. A 1-ml gelatin-agarose (6 mg gelatin/ml; Sigma, UK) affinity column was packed. The column was washed with several column volumes of equilibration buffer (0.05 M of Tris/HCl at pH 7.6 containing 0.5 M of NaCl, 0.005 M of CaCl2, 0.05% v/v Brij 35, and 0.02% w/v NaN3) and was stored in 2 ml of this solution at 4°C until used.
The fraction containing proteolytic activity was treated in a gelatin-agarose column at a flow rate of approximately 200 ml/h in a cool bath and the nonbinding fraction was collected.
The column was washed with 2.5 ml of a modified equilibration buffer (0.05M Tris/HCl, pH 7.6, containing 1 M NaCl, 0.005 M CaCl2, 0.05% v/v Brij 35, 0.02% w/v NaN3). The binding material was eluted with 2.5 ml of the same buffer containing 10% dimethyl sulfoxide.
The eluate was dialyzed overnight at 4°C against a standard metalloproteinase assay buffer (0.05M Tris/HCl, pH 7.6 containing 0.005 M CaCl2, 0.05% v/v Brij 35, 0.02% w/v NaN3). The fraction was then subjected to both gelatin zymography (as described below) and standard sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (7.5% gels in reducing and nonreducing conditions). The standard polyacrylamide gels were stained with Coomassie blue following routine laboratory protocol. Molecular masses of the detected proteins were determined by referring to molecular weight standards.
The above fraction was subjected to gelatin zymography. Gelatin (cell culture grade, Sigma) was copolymerized into 1-mm-thick, 7.5% polyacrylamide gel at a final concentration of 2.5 mg/ml.
The sample was diluted 1:1.25 in a 50-mM Tris/HCl sample buffer containing 2.5% SDS at pH 7.6 and incubated at 37°C for 1 h before electrophoresis. The sample was loaded onto a gel and electrophoresed at 100 V for approximately 2 h using a Miniprotean II system (BioRad, UK). The gel was washed at room temperature for 1 h in 2.5% Triton X-100 (Sigma, UK), rinsed in distilled water (3×30 s), and incubated for 18 h at 37°C in a 50-mM Tris/HCl buffer at pH 7.6, containing 50 mM of CaCl2 (Sigma), 20 mM of NaCl, and 0. 005% Brij 35 (Sigma). After incubation the gel was stained with 1% Coomassie brilliant blue R-200 (Sigma) (in 30% methanol and 7% glacial acetic acid), destained, and dried. Gelatin degradation was manifest as zones of clearing on a blue-stained gel. Calculation of apparent molecular masses of the gelatinolytic bands was made by reference to standard prestained molecular mass markers (Sigma, UK).
Samples were separated on 10% SDS-PAGE at 100 V for 2 h and electrophoretically transferred to a nitrocellulose membrane (Schleider and Shull) as described in a previous work (Towbin et al. 1979). One percent Tween in PBS was used to block nonspecific binding sites on the nitrocellulose membrane and the membrane was subsequently incubated with anti-MMP-9 monoclonal antibody [1:1,000 dilution in Tris–PBS (T–PBS), Oncogene Science, Cambridge, MA, USA] for 10 h. After 3×15 min of washing with T-PBS, the membrane was incubated for 1 h with alkaline phosphatase-conjugated goat anti-mouse IgG antibody (Nordic) at 1:1,000 dilution in T-PBS.
After a 1×15-min wash with T-PBS, immunoblots were developed by the addition of chemiluminescent reagents (Amersham) in a darkroom and exposed for 1 min on Polaroid films. All incubations were carried out at room temperature (Gómez et al. 1999).
Molecular masses of the tryptic fragments derived from the 92- and 84-kDa bands were determined by MALDI MS using Bruker REFLEX TOF equipment furnished with reflection system and N2 laser (337 nm). The sample (1 μl, containing 5–25 pmol μl−1) was mixed with an equal volume of sinapinic acid (10 mg ml−1 in 10% ethanol), spotted onto the stainless steel tip, and dried at room temperature. Spectra were recorded under an acceleration voltage of 30 kV.
Protein identification analysis was performed by comparison of the peptide maps with the database of National Center for Biotechnology Information (NCBI).
HCW and the surrounding host tissue was removed from infected bovine lungs and immediately fixed in phosphate-buffered 10% formalin embedded in paraffin, using standard procedures. The tissue was sectioned at a thickness of 5 μm and sections were stained with hematoxylin and eosin or antibodies.
Immunohistochemical staining was performed on formalin-fixed paraffin sections using the avidin-biotin immunoperoxidase technique. Unstained sections mounted on polylysine-coated slides were deparaffinized with xylene and rehydrated with decreasing concentrations of ethanol. Nonspecific binding was blocked with 2% normal goat serum. Sections were incubated for 30 min at room temperature with mouse anti-human MMP-9 monoclonal antibodies diluted in PBS added with normal goat serum (as used for inmunoblot) (5 μg/ml).
Tissue sections were washed and incubated with a secondary biotinylated anti-mouse IgG (free of cross-reaction with the host) (Nordic). Endogenous peroxidase activity was blocked using 0.3% H2O2 (Vector Laboratories). Peroxidase activity was visualized using diaminobenzidine (Vector Laboratories). Slides were rinsed in water and lightly counterstained with hematoxylin. Before the blocking procedure, samples incubated with anti-MMP-9 monoclonal antibodies were preincubated with 0.1% trypsin in PBS for 12 min at 37°C, following the manufacturer’s instructions.
Immunohistochemical control procedures
Negative control immunohistochemical procedures included omission of the primary antibody from the described staining protocol and its replacement with PBS added with normal goat serum. An irrelevant monoclonal IgG1 (Sigma, USA) was also tested to rule out the possibility of artifacts leading to erroneous results.
Characterization of gelatinolytic activities by zymography
The electrophoretic mobility pattern of the purified fraction in gelatin gels was identical with that of human gelatinase MMP-9. A highly specific method, zymography enabled the detection of 1–2 ng of protein as an intense band of digestion, ensuring the absence of any gelatinase other than MMP-9 in the samples. A detection order of more than ten times higher than suggested in the procedure was used to ensure the absence of MMP-2 even at extremely low concentrations. Other proteases not digesting gelatin as a substrate were not identified by this method.
Gelatinolytic activity of about 84 kDa (presumably corresponding to an active form of the enzyme) was found in cyst wall in all preparations. Both the 84- and 68-kDa forms of the enzyme were found in fluids, the former consistently occurring in a greater proportion at a ratio varying according to batch (data not shown). The 92-kDa fraction (latent MMP-9) was detected in all tested samples except in preparations with proteins strongly bound to the cyst wall. In different preparations obtained from bovine HCW, gelatinolytic activity was predominantly due to active MMP-9 (84 and 68 kDa). No difference was found (data not shown) among proteolytic activity patterns obtained using HCF of different origin (liver or lung). Major forms of 92- and 84-kDa were found at varying ratios among different bovine fluid samples (data not shown).
Characterization by SDS-PAGE and immunoblot analysis of the purified fraction
Immunoblot analysis using an anti-human MMP-9 monoclonal antibody demonstrated the presence of immunoreactive protein bands of 92 and 84 kDa consistent with gelatinolytic activity of the same MW determined by gelatin zymography (Fig. 4b), but did not recognize the digested active form of 68 kDa, presumably on account of an extremely low content. Controls using secondary antibodies alone (goat anti-mouse antibodies) did not react with the bands detected using anti-MMP-9 monoclonal antibodies in Western blotting (data not shown).
Mass spectrometry characterization of gelatinolytic activity
Measured mass (Da)
Error with computed Mass (Da)
Latent MMP-9 (92 kDa)
Active MMP-9 (84 kDa
Parasite survival in vivo depends on efficient host evasion mechanisms starting to operate as the parasite develops toward the hydatid cyst stage.
As a long-lasting nonself-structure lodged in the parenchyma of the host’s internal organs, the hydatid cyst elicits a local inflammatory response, which may resolve leaving behind a collagenous capsule. In the bovine host, however, resolution does not generally take place and granuloma is formed (like Mycobacterium tuberculosus) (Taggart et al. 2005). HCW used in this work was obtained from infected cattle, thus closely apposed in vivo to host inflammatory cells. The laminated layer of the HCW, constituting a bulky yet permeable extracellular structure, appears to bind considerable amounts of host protein. Biochemical research on HCW extracts provided relevant findings regarding host secretory activities around the cyst. Host factor H, annexin II, and cathepsin K (Díaz et al. 1997, 2000a,b) were reported in HCW studies, the latter two being associated with host inflammatory reaction.
In human infections, the cyst is surrounded by a fibrotic layer (adventitial) resulting from the resolution of an initial inflammatory response. The resolution of the local inflammatory reaction appears to be associated with parasite viability for various hosts and situations, while the persistence of local granulomatous reaction correlates with cyst degeneration.
In the established cyst phase, the parasite is far less susceptible to the host immune system. In addition to the protection provided by the laminated layer, other various evasion mechanisms may be involved. The established cyst is nonetheless vulnerable to host response, the destruction of well-developed cysts being a common occurrence. Morphologically, cyst destruction is expressed as the degeneration of the germinal layer, and occasionally the collapse of the cyst wall into the cyst cavity, or calcification of the lesion. Parasite structures, the laminated layer in particular, appear not to be completely resorbed by the host. Immunologically, the killing of the established cyst was not traced to any particular immune status, thriving and degenerated cysts often coexisting in an individual host, suggesting that the phenomenon may be local and immunologically nonspecific. There is strong circumstantial evidence that the killing of the established cyst is associated with the lack of resolution of the local host inflammatory response. There is also circumstantial evidence of the live parasite downregulating the local inflammatory response. This stems from the observations that (1) E. granulosus elicits a relatively weak local host response in comparison, for example, with Schistosoma ovum, another structure of helminth origin lodged in organ parenchyma; and (2) local response is enhanced upon parasite death.
The above findings arise from studies made in human infection. However, analysis of the literature reveals that the local reaction to the cyst shows, among different hosts, organs, and cyst sizes and shapes, a continuous spectrum of morphologies ranging from a fairly intense granulomatous-type response to a noninfiltrated collagenous capsule (Thompson 1995). The latter case arises from the resolution of inflammation. Resolution nearly always takes place in humans, less generally in sheep and pigs, and least often in cattle, in agreement with results obtained in our laboratory. Across the spectrum of host response, cyst fertility correlates with resolution of the inflammation, while an intense response is associated with infertility and, more extremely, degeneration of the cyst. It was proposed that the situation where inflammation resolves reflects a balanced host–parasite relationship (Thompson 1995).
The inflammatory response is typically tri-layered (Turk 1989; Williams and Williams 1983; Thomas and Kothare 1975). The innermost layer is composed of epithelioid and giant multinucleated cells in radial arrangement with respect to the cyst. These cells are in close contact with the laminated layer, but an area of necrosis is sometimes found. The second outer layer, commonly referred to as “microcellular infiltrate,” is composed of lymphocytes, a small number of eosinophils, and possibly monocytes. On the outside is the collagenous layer in which fibroblasts lay down what may, upon resolution of inflammation, become the collagenous capsule. The two outer layers are not always sharply separated. Outside of the inflammation area, the host organ parenchyma is generally minimally disrupted.
MMP-9 was reported to be secreted by activated macrophages and granulomas and to play an important role in the pathology of the granulomatous reaction (Parks et al. 2004; Quiding-Jarbrink et al. 2001), yet it does not have the capacity of self-activation but depends on another enzyme. Not present in HCW, MMP-2 is not among the possible activators, independent of host and localization.
Active and latent forms of MMP-9 (an effector of the host immune system) were found in HCFs in proportions varying according to pool (data not shown). Latent MMP-9 was extracted by washing from HCW-occluded material; while the active form of the enzyme was obtained from walls using the strongest extractions (Fig. 3), suggesting that the activation process might take place in the hydatid wall.
Analysis of Fig. 2 shows a high fertility index and complete absence of MMP-9 in pools of cyst fluid obtained from high adaptation scenarios, such as those of human, ovine, and porcine host; while hosts where adaptation is scarce show a high concentration of MMP-9 within the cyst, particularly for nonfertile ones. It may be concluded from the results reported here that a higher activity correlates with the lack of parasite adaptation to host.
MMP-like activity was recently reported for Caenorhabditis elegans (Wada et al. 1998) and other nematodes such as Heterodera glycines and Globodera rostochiensis. (Kovaleva et al. 2004). Mass spectrometry was here used to further characterize major fraction components (92- and 84-kDa bands). In the light of the high degree of coincidence, both bands were confirmed to correspond to latent and active forms of bovine MMP-9, ruling out the possibility of activity being generated by the parasite.
Further research is required into the function of active MMP-9 in the interface and inside of cysts, as well as the mechanisms for its activation. MMP-9 was reported to be involved in the recruitment of neutrophils in inflammation, and the inhibition of its activity may lead to the formation of granuloma (Izzo et al. 2004). It remains to be determined whether a parasite-derived inhibitor may be involved in MMP-9 modulation, and whether MMP-9 may have a direct role over the parasite. Work in this field is currently being developed in our laboratory.
This study was supported by SAREC, CONICYT, CSIC (University of Uruguay), and CEE. The text was revised by Eduardo Speranza.