The Degradation of Type II Collagen in Rheumatoid Arthritis: An Immunoelectron Microscopic Study

Matrix Vol. 1111991, pp. 330~338 © 1991 by Gustav Fischer Verlag, Stuttgart

The Degradation of Type II Collagen in Rheumatoid Arthritis: An Immunoelectron Microscopic Study GEORGE R. DODGE, ISABELLE PIDOUX and A. ROBIN POOLE Joint Diseases Laboratory, Shriners Hospital for Crippled Children, Division of Surgical Research, Department of Surgery, McGill Uuiversity, 1529 Cedar Avenue, Montreal, Quebec, H3G 1A6, Canada.

Summary Rabbit antibodies were prepared that react only with denatured type II collagen a-chains and cleavage products. The epitopes that these antibodies recognize reside in cyanogen bromide peptides CB8 and CB1l. The antibodies do not react with triple helical collagen nor with any other collagen or protein present in hyaline cartilage (Dodge and Poole, J. Clin. Invest. 83: 647~661, 1989). These antibodies can therefore be used to detect denatured type II collagen produced, for example, by enzymatic cleavage. In this study they were used to determine, at the ultrastructural level, using immunogold staining, type II collagen fibril cleavage in articular cartilages remote from synovium and pannus of patients with rheumatoid arthritis. Comparisons were made with site- and age-matched healthy articular cartilages. Antibody binding was detected in the extracellular matrix, at the articular surface and in the deep zone, usually on visibly damaged collagen fibrils which exhibited a loss of the normal banding pattern of staining produced by lead citrate and uranyl acetate: binding was also observed in disrupted fibrils, sometimes at their ends. Binding was commonly associated with amorphous-looking material (and occasionally unstained sites) in the extracellular matrix which, because of the specificity of the antibody, can be identified as containing denatured or fragmented type II collagen, stained (and unstained) with heavy metals. In both rheumatoid and healthy articular cartilages, there was no antibody binding to intact well stained fibrils which exhibited a regular banding pattern. Little or no staining was detected at the ultrastructural level in normal cartilages. These studies provide ultrastructural evidence in support of the previously demonstrated specificity of antibodies to denatured type II collagen. Also they show that morphologically recognizable damage to type II collagen fibrils usually accompanies antibody binding. Damage to the articular surface may be caused by polymorphonuclear leukocytes and cytokines, such as interleukin-1, originating from the joint cavity. Damage to type II collagen in the deep zone may result from activation of chondrocytes by inflammatory cells in subchondral bone, where resorption is commonly observed. The usefulness of such specific antibodies is shown by studying cartilages from patients with rheumatoid arthritis where damage to type II collagen fibrils is demonstrated. Key words: cartilage, degradation, immunohistochemistry, type II collagen.

Introduction Rheumatoid arthritis is a chronic inflammatory joint disease which is mainly characterized by extensive erosive Present Address: Dept. of Pathology and Cell Biology, Thomas Jefferson University, 1020 Locust Street, Alumni Hall, Room 249, Philadelphia 19107, PA.

destruction of articular cartilage and intraarticular ligaments with associated bone loss (Kennedy and Lindsay, 1977; Parfitt, 1983). Evidence for increased bone turnover in periarticular sites has also been obtained (Deslauniers et al., 1974; Shimizu et al., 1985). Articular cartilage contains a fibrillar network of type II collagen. Type II collagen fibrils are composed of tropocollagen molecules in which

Type II Collagen Degradation three-identical a-chains are wound together to form a triple helix. These tropocollagen molecules combine, through the involvement of intermolecular cross-links, to form microfibrils which form the mature fibril seen with the electron microscope (Eyre, 1980). These collagen fibrils contain type XI collagen (Mendler et aI., 1989) and type IX collagen which is covalently bound to type II collagen in the N- and C-terminal telopeptide regions (Eyre et aI., 1987; van der Rest and Mayne, 1988). These mixed fibrils can be demonstrated with immunoelectron microscopy (Mendler et aI., 1989). In adult human articular cartilage, tropocollagen molecules are mainly cross-linked to each other by nonreducible inter-molecular hydroxypridinium cross-links between non-helical and helical regions (Eyre et aI., 1984). These fibrils endow cartilage with its tensile stiffness and strength (Kempson et aI., 1973; Roth and Mow, 1980). Collagen can be cleaved by mammalian collagenase producing characteristic Y4 and V4 products, as a result of cleavage of each of the three a-chains of the triple helix (Woolley, 1984). Polymorph elastase and cathepsin G, as well the cysteine proteinase cathepsin B and the metalloproteinase stromelysin (Wu et aI., 1991), can cleave collagen fibrils in the non-helical domain (telopeptide region) of tropocollagen (Poole, 1990). The intermolecular crosslinks probably stabilize the collagen fibril, add strength and serve to limit disruption of the molecule. Extensive cleavage can result in the loss of the tensile properties of cartilage with little evidence of release of degraded collagen when measured as hydroxyproline (Bader et aI., 1981). This is probably in part a result of cross-links preventing the rapid break-up of the tight fibrillar structure leading to retention of cleavage products within the fibril. Recent immunological studies in this laboratory have demonstrated that it is possible to prepare a monospecific antiserum which recognizes only degraded type II collagen: it reacts with isolated a chains or fragments thereof (Dodge and Poole, 1989). Thus it can be used to detect cleavage of type II collagen since, at physiological temperatures, this results in unwinding of the triple helix with recognition of epitopes on unwound a-chains. The antiserum was used to demonstrate that cleavage of type II collagen can be detected in articular cartilage. In situ this was shown by using, with light microscopy, both immunohistochemical and immunochemical analyses of type II collagen a chain cleavage products. We showed that in rheumatoid arthritis cleavage of type II collagen is excessive in territorial regions around cells in the deep zone closest to subchondral bone. It was also enhanced in the superficial zone at the articular surface. Sometimes in advanced disease, extensive degradation was observed throughout the cartilage matrix. In contrast, in osteoarthritis, cleavage of type II collagen was first observed in the superficial and upper mid zones: there was never significant degradation in the deep zone, except in pericellular sites, at this stage. Much less evidence of collagen cleavage was seen in normal age-matched cartilage,

331

where the staining was mainly confined to pericellular sites (probably reflecting limited turnover of collagen molecules) and the articular surface. All these observations were made using light microscopy. In this paper we describe these changes at the ultrastructural level. We demonstrate that when the characteristic banded pattern of collagen fibrils is lost and fibril structure is visibly altered and disrupted, antibodies that react with denatured type II collagen and a-chains and fragments thereof, are commonly bound to visibly damaged collagen fibrils and to amorphous-looking material. This is in distinct contrast to those fibrils with a more normal banded appearance where no antibody binding is observed. We have studied both normal and rheumatoid articular cartilages. These investigations provide further evidence to indicate that in patients with rheumatoid arthritis extensive damage to type II collagen in articular cartilage occurs at and remote from the articular surface and remote from synovial and pannus tissue.

Materials and Methods Cartilage

As described previously (Dodge and Poole, 1989), cartilages from patients with classic rheumatoid arthritis which had been diagnosed by rheumatologists according to the classification of the American Rheumatism Association (Arnett et aI., 1988) were obtained at surgery for total joint arthroplasty. These patients were of ages 33, 59 and 63 years. The 33-year-old cartilage was from the distal radius, whereas the others were from femoral condyles. Control cartilages were obtained at autopsy from femoral condyles of non-arthritic persons of ages 33, 47 and 58 years. These autopsy specimens were obtained within 10 h of death and were isolated from healthly looking joints in which there was no macroscopic or microscopic evidence of cartilage fibrilation. Antibodies and immunohistochemical detection of type II collagen degradation

Antibodies (Antiserum R181) to degraded type II collagen were prepared and characterized and light microscopic visualization of collagen degradation was performed as described previously (Dodge and Poole, 1989). For immunoelectron microscopic visualization, 6-flm thick frozen sections on microscope slides were fixed for 20 min in 4% formaldehyde (Dodge and Poole, 1989) at room temperature. They were then washed in phosphate-buffered saline (PBS) (Dodge and Poole, 1989) for 30 min and treated with 5% non-immune donkey serum for 30 min to block free aldehyde groups. After rinsing in PBS, they were treated for 1 h at 37 °C with 0.2 % hyaluronidase (Sigma, type Is, bovine testicular) in PBS containing proteinase

332

G. R. Dodge et al.

Fig. 1. Electron micrograph of the articular surface of femoral cartilage of a 47-year-old non-arthritic person reacted with rabbit antibody F(ab')z to type II collagen peptides. There is little evidence of immunogold staining. Occasional gold particles (arrowheads) are indicated. Scale = 200 nm. inhibitors (Dodge and Poole, 1989). After washing in PBS containing 0.1 % bovine serum albumin (BSA) for 10 min at room temperature, sections were treated with antibody F(ab'h (R181 to type II collagen peptides) at 0.1 mglml or with non-immune F(ab'h at 0.1 mglml. 1001-11 of reagent was applied to each section in a humidified chamber overnight at 4°C. This was followed by washing for 90 min at room temperature in PBS with 0.1 % BSA. A biotin-streptavidin gold detection system (Amersham Corp., Arlington Heights, IL) was used to detect bound F(ab'h. Thus 1001-11 of biotinylated donkey F(ab')z antirabbit IgG (diluted II 100 with PBS and 1 % BSA) was applied for 4.5 h at room temperature. After washing in PBS and 0.1 % BSA for 1.5 h, sections were incubated overnight at 4°C in 100-1-11 streptavidin-gold (15 nm diameter), diluted 1110 in PBS and 1 %

BSA. Sections were finally washed in PBS and 0.1 % BSA for 2 h at room temperature. Tissues were dehydrated and flatembedded on glass slides in Spurr resin. Blocks were detached from slides on a hot plate by applying lateral pressure to the block. Ultrathin sections were cut and counterstained with lead citrate and uranyl acetate and examined in a Phillips 400 electron microscope at 80 KV. To determine the density and distribution of immunostaining, gold particles were counted with the assistance of a 2-cm square graticule on micrographs printed at a final magnification of 78 000.

Fig. 2. Electron micrographs of femoral condylar cartilage of a 59-year-old patient with rheumatoid arthritis treated with antibody F(ab')z to type II collagen pep tides to show immunogold staining at the articular surface in (a) non-pericellular and (b) pericellular sites. In (a) the arrowheads point to staining of collagen fibrils in regions that lack a well defined banded pattern. In (b) they identify staining at the "ends" of fibrils. The pericellular region ofthe same cartilage treated with non-immune rabbit F(ab')z is shown in (c): no immunogold staining is visible. Scale = 200 nm. An example of these differences in staining is provided by the following analyses. Pericellular fields, each 25 [.lm x 25 [.lm, were counted for gold particles. Sections treated with the immune F(ab'h contained 63.0 ± 25.7 (mean ± S.D., n = 7) particles. In contrast, non-immune F(ab'h treated specimens contained 2.2 ± 2.3 (mean ± S.D., n = 7) gold particles.

~

Type II Collagen Degradation

333

334

G. R. Dodge et al.

••

e.

,.

..

•

•

e e

• ••

Fig. 3. Electron micrographs of the territorial region of the deep zone of the patient described in Figure 2 treated (a) with F(ab'h to type II collagen peptides or (b) non-immune F(ab'h. In (a) small arrowheads pointto amorphous-looking material oflow electron density. Large arrowheads point to staining at the" ends" of fibrils. Scale = 200 nm.

Type II Collagen Degradation

Results The regions of articular cartilage which were examined in this study are those described by us previously (Dodge and Poole, 1989). These studies were made of the superficial zone (including the articular surface), the mid-zone and the deep zone (which is adjacent to calcified cartilage and subchondral bone). The pericellular, territorial and interterritorial regions were examined. These zones and regions have been defined by us previously (Poole et aI., 1982). In brief, the pericellular region extends to approximately 2 flm from the chondrocyte surface. The territorial region is that matrix which surrounds the pericellular region of all cells in the superficial and mid-zones . In the deep zone, a separate interterritorial matrix is recognizable, which surrounds the territorial region of each chondrocyte and is the matrix most remote from the chondrocyte. Examination of the articular surfaces of normal looking cartilages from healthy joints revealed the characteristic presence of relatively thin fibrils of type II collagen generally arranged in parallel to each other (Fig. 1). There was little or no immunostaining. In contrast, cartilage at the articular surfaces of rheumatoid joints exhibited less organized fibrils and significant staining with antibodies to degraded type II collagen (Fig. 2 a). Staining was re-

335

presented by single or clustered gold particles which were often associated with collagen fibrils which contained little or no banded pattern (arrowhead). Other staining sites often corresponded to regions where there was little or no electron density. In pericellular sites in rheumatoid cartilage, staining was more clearly associated with fibrils of even smaller diameter, sometimes at the "ends" of fibrils (arrowheads) (Fig. 2 b). When sections of rheumatoid articular cartilage were first treated with non-immune antibody subunit, little or no staining was observed, including these pericellular sites (Fig. 2 c). An example of quantitation of these staining differences is provided in the legend to Fig. 2. No staining was seen with non-immune antibody F(ab')z in the deep zone in interterritorial sites (Fig. 3 b). But when the same specimen was stained with the antiserum F(ab'h, immunogold staining was commonly seen, which was mainly associated with amorphous-looking material of moderate or low electron density (small arrowheads), with fibrils lacking a banding pattern, or at the ends of such fibrils (arrowheads) (Fig. 3 a). Staining was also observed in sites where no electron density was detectable. In some patients, well defined unbanded fibrils of irregular organization were occasionally seen, often stained with antibody (Fig. 4). When the deep zones of control cartilages were

Fig. 4. Electron micrograph of the femoral cartilage of the 63-year-old patient to show territorial region of the deep zone treated with antibody F(ab'h to type II collagen pep tides. Scale = 200 nm.

336

C. R. Dodge et al.

Fig. 5. Electron micrographs of the territorial regions (a) of the 59-year-old patient and (b) the 58-year-old non-arthritic cartilage both treated with antibody F(ab'h to type II collagen peptides. Scale = 200 nm.

Type II Collagen Degradation

337

Table I. Distribution of immunogold staining in femoral condylar cartilage of a 59-year-old rheumatoid patient. % Fibril Staining

% Non-Fibril Staining Amorphous electron dense No electron dense background background

Banded

Non-banded

Superficial zone Territorial region n = 645

0

39

7

35

19

Deep zone Territorial region n = 274

0

30

13

41

16

Ends

Colloidal gold particles (n) were counted using a graticule on electron micrographs (see text). In those instances indicated, gold particles were detected in sites where there was no electron density or in sites where amorphous electron dense material of no recognizable form was observed. examined, only occasional staining was observed (Fig. 5 b) in contrast to the same sites in rheumatoid cartilages (Fig.5a). Table I provides a summary of the typical distribution of immunogold staining for degraded type II collagen seen in rheumatoid patients. At both the articular surface and in the deep zone, the majority of staining was mainly found in association with both non-banded fibrils or non-fibrillar amorphous electron dense material. Some staining was also observed at the ends of fibrils or in association with material with no detectable electron density. Such staining was rarely observed in specimens treated with non-immune immunoglobulin. Banded collagen fibrils never stained with the antibodies.

Discussion We have previously shown that it is possible to detect with light microscopy sites of type II collagen degradation using polyclonal antibodies that only react with "unwound" a-chains of type II collagen (Dodge and Poole, 1989). Use of the antibodies in immunogold staining has revealed significant damage to type II collagen in patients with rheumatoid arthritis as well as osteoarthritis. In rheumatoid arthritis, we have seen that this staining is most pronounced at the articular surface and in territorial sites of the deep zone, as well as in pericellular sites throughout rheumatoid cartilage. Staining throughout cartilage matrix is sometimes observed. This contrasted to sites of predominantly pericellular staining in normal age- and site-matched cartilages which was observed in specimens obtained at autopsy. Staining was unchanged at different times after death up to 18 hours. In the present study we have examined this immunogold staining at the electron microscopic level. We have used the same localization technique employing immunogold but without the silver enhancement step required for light microscopy. The specificity of this method and antibody-binding controls have already been described previously in detail (Dodge and Poole,

1989). This ultrastructural approach has enabled us to determine the relationship of staining to the morphology of collagen fibrils. Type II collagen fibrils normally exhibit a characteristic banding pattern as shown in control cartilages. In this study, as reported previously (Mitchell and Shepard, 1978) articular cartilages of rheumatoid patients, remote from synovium and pannus, frequently display a loss of this banding pattern, with disruption of fibrillar organization. We found that both unbanded fibrils, disorganized fibrils and amorphous-looking material of moderate, low and no electron density stained with the R181 antibodies. Banded collagen never stained. In view of the demonstrated specificity of the antibody we now know that this material identified with the electron microscope represents, in part at least, denatured or fragmented type II collagen. Thus the inferred association of antibody staining with denatured fibrils based on immunochemical analyses (Dodge and Poole, 1989) is supported by these ultrastructural observations. The accessibility of epitopes to antibody is clearly dependant on the exposure of epitopes on collagen a-chains which occurs when the triple helix unwinds, such as may result from proteolytic cleavage. The immunogold staining in areas of no electron density must represent, by definition of the antibody specificity proven previously (Dodge and Poole, 1989), denatured type II collagen a-chains or fragments thereof. Previously, collagen degradation in articular cartilages of patients with rheumatoid arthritis has been demonstrated morphologically at the interface between pannus and cartilage (Harris et aI., 1970; Kobayashi and Ziff, 1975). In other cases, degradation of cartilage collagen fibrils has been recognized morphologically in rheumatoid arthritis, in pericellular sites and remote from chondrocytes (Mitchell and Shepard, 1973), as well as at the articular surface (Kimura et aI., 1979). In all these studies electron dense material of an amorphous nature was described. In view of our present findings, some of this probably represents denatured unwound or fragmented type II collagen. The association of staining with a loss of the banded pattern and

338

G. R. Dodge et al.

disruption of fibril structure also demonstrates in situ that recognizable disruption of type II fibrils is often associated with an unwinding of the triple helix of type II collagen in fibrils thereof. The present approach permits a more direct and specific demonstration of type II collagen cleavage in situ by using antibody staining. Further, it is apparent that major degenerative changes in the fibrillar architecture of cartilage collagen accompany rheumatoid arthritis. These changes are often most pronounced in the deep zone (Dodge and Poole, 1989), a region previously thought to be less affected in arthritis. It is well recognized that there is net loss of bone and remodelling in subchondral sites close and adjacent to articular cartilages (Kennedy and Lindsay, 1977; Parfitt, 1983). This is reflected by increased technetium polyphosphonate incorporation (Deslauniers et aI., 1974) and increased resorption sites and osteoid surfaces (Shimizu et aI., 1985). This may be caused by subchondral inflammation in bone leading to release of cytokines, such as interleukin-1, which can not only induce local bone resorption (Heath et aI., 1985; Thomson et aI., 1986) and, indirectly, bone remodelling, but can also induce type II collagen degradation in adjacent cartilage (Dodge and Poole, 1989). Therefore, chondrocytes may be activated to degrade their extracellular matrix by inflammatory cells in subchondral bone. Acknowledgements This study was funded by the Shriners of North America. We thank Michele Burman Turner for processing this manuscript. We thank the Department of Pathology, Royal Victoria Hospital, McGill University and Dr. Nelson Mitchell, Division of Orthopaedics, Department of Surgery, McGill University for providing tissues. References Arnett, F.e., Edworthy, S.M., Bloch, D.A., McShane, D.J., Fries, J.F., Cooper, N.S., Healey, L.A., Kaplan, S.R., Liang, M.H., Luthra, H.S., Medsger, T.A., Jr., Mitchell, D.M., Neustadt, D. H., Pinals, R. S., Schaller,]. G., Sharp, J. T., Wilder, R. L. and Hunder, G. G.: The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthr. Rheum. 31: 315-324, 1988. Bader, D.e., Kempson, G.E., Barrett, A.J. and Webb, W.: The effects of leucocyte elastase on the mechanical properties of adult human cartilage in tension. Biochim. Biophys. Acta 677: 103-108,1981. Deslauniers, M., Fuks, A., Hawkins, D., Lacourciere, Y. and Rosenthal, L.: Radiotechnetium polyphosphonate joint imaging.]. Nucl. Med. 15: 417-423, 1974. Dodge, G. R. and Poole, A. R.: Immunohistochemical detection and immunochemical analysis of type II collagen degradation in human normal, rheumatoid and osteoarthritic articular cartilages and in explants of bovine articular cartilage cultured with interleukin 1.]. Clin. Invest. 83: 647-661, 1989. Eyre, D. R.: Collagen: molecular diversity in the body's protein scaffold. Science 207: 1315 -1322, 1980. Eyre, D. R., Apone, S., Wu, ].-J., Ericsson, L. H. and Walsh, K. A.:

Collagen type IX: evidence for covalent linkages to type II collagen in cartilage. FEBS Lett. 220: 337 - 341, 1987. Eyre, D. R., Pax, M. A. and Gallop, P. M.: Cross-linking in collagen and elastin. Ann. Rev. Biochem. 53: 717-748, 1984. Harris, E.D., Jr., DiBona, D.R. and Krane, S.M.: A mechanism for cartilage destruction in rheumatoid arthritis. Trans. Assoc. Am. Physicians 83: 267-276, 1970. Heath, ].K., Saklatvala, J., Meikle, M.e., Atkinson, S.]. and Reynolds, ].J.: Pig interleukin-1 (catabolin) is a potent stimulator of bone resorption in vitro. Calc. Tiss. Int. 37: 95-97,1985. Kempson, G.E., Muir, H., Pollard, e. and Tuke, M.: The tensile properties of the cartilage of human femoral condyles related to the content of collagen and glycosaminoglycans. Biochim. Biophys. Acta 297: 456-472,1973. Kennedy, A. e. and Lindsay, R.: Bone involvement in rheumatoid arthritis. Clin. Rheum. Dis. 3: 403-420, 1977. Kimura, H., Tateishi, H. and Ziff, M.: Surface ultrastructure of rheumatoid articular cartilage. Arthr. Rheum. 20: 1085 -1098, 1977. Kobyashi, 1. and Ziff, M.: Electron microscopic studies of the cartilage-pannus junction in rheumatoid arthritis. Arthr. Rheum. 18: 475-483,1975. Mendler, M., Eich-Bender, S. G., Vaughan, L., Winterhalter, K. H. and Bruckner, P.: Cartilage contains mixed fibrils of collagen types II, IX and XL]. Cel/Biol.108: 191-197,1989. Mitchell, N. S. and Shepard, N.: Changes in proteoglycan and collagen in cartilage in rheumatoid arthritis. A study by light and electron microscopy.]. Bone Jt. Surg. 60-A: 349-354, 1978. Parfitt, A. M.: Stereo logic basis of bone histomorphometry; theory of quantitative microscopy and reconstruction of the third dimension. In: Bone Histomorphometry: Techniques and Interpretation, ed. by Recker, R. R., CRC Press, Boca Raton, FL, 1983, pp. 53-88. Poole, A. R.: Enzymatic degradation: cartilage destruction. In: Cartilage Changes in Osteoarthritis, ed. by Brandt, K. D., Indiana School of Medicine, Indianapolis, IN, 1990, pp.63-72. Poole, A.R., Pidoux, 1., Reiner, A. and Rosenberg, L.: An immunoelectron microscope study of the organization of proteoglycan monomer, link protein and collagen in the matrix of articular cartilage.]. Cell Bioi. 93: 921-937, 1982. Roth, V. and Mow, V. e.: The intrinsic tensile behaviour of bovine articular cartilage and its variation with age. ]. Bone Jt. Surg. 62: 1102-1117,1980. Shimizu, S., Shiozawa, S., Shiozawa, K., Imura, S. and Fujita, T.: Quantitative histologic studies on the pathogenesis of periarticular osteoporosis in rheumatoid arthritis. Arthr. Rheum. 28: 25-31,1985. Thomson, B.M., Saklatvala, J. and Chambers, T.].: Osteoblasts mediate interleukin-1 stimulation of bone resorption by rat osteoclasts.]. Exp. Med. 164: 104-112, 1986. van der Rest, M. and Mayne, R.: Type IX collagen proteoglycan from cartilage is covalently cross-linked to type II collagen. ]. Bioi. Chem. 263: 1615 -1618,1988. Woolley, D. E.: Mammalian collagenases. In: Extracellular Matrix Biochemistry, ed. by Piez, K.A. and Reddi, A.H., Elsevier, New York, 1984, pp. 119-157. WU, J.-J., Lark, M. W., Chun, L. E. and Eyre, D. R.: Sites of stromylesin cleavage in collagen types II, IX, X, and XI of cartilage. J. Bioi. Chem. 266: 5625 -5628,1991. Dr. A. R. Poole, Joint Disease Laboratory, Shriners Hospital, 1529 Cedar Avenue, Montreal, Quebec H3G 1A6, Canada.