Osteoclastic Bone Degradation and the Role of Different Cysteine Proteinases and Matrix Metalloproteinases: Differences Between Calvaria and Long Bone

Authors

  • Vincent Everts PhD,

    Corresponding author
    1. Department of Oral Cell Biology, Academic Centre for Dentistry Amsterdam (ACTA), Universiteit van Amsterdam and Vrije Universiteit, Amsterdam, The Netherlands
    • Department of Oral Cell Biology, ACTA, Vrije Universiteit, Van der Boechorststraat 7, 1081 BT Amsterdam, The Netherlands
    Search for more papers by this author
  • Wolf Korper,

    1. Department of Cell Biology and Histology, Academic Medical Centre, University of Amsterdam, Amsterdam, The Netherlands
    Search for more papers by this author
  • Kees A Hoeben,

    1. Department of Cell Biology and Histology, Academic Medical Centre, University of Amsterdam, Amsterdam, The Netherlands
    Search for more papers by this author
  • Ineke DC Jansen,

    1. Department of Oral Cell Biology, Academic Centre for Dentistry Amsterdam (ACTA), Universiteit van Amsterdam and Vrije Universiteit, Amsterdam, The Netherlands
    2. Department of Periodontology, ACTA, Universiteit van Amsterdam and Vrije Universiteit, Amsterdam, The Netherlands
    Search for more papers by this author
  • Dieter Bromme,

    1. Department of Oral Biological and Medical Sciences, Faculty of Dentistry, University of British Columbia, Vancouver, Canada
    Search for more papers by this author
  • Kitty BJM Cleutjens,

    1. Department of Pathology, Cardiovascular Research Institute Maastricht (CARIM), University of Maastricht, Maastricht, The Netherlands
    Search for more papers by this author
  • Sylvia Heeneman,

    1. Department of Pathology, Cardiovascular Research Institute Maastricht (CARIM), University of Maastricht, Maastricht, The Netherlands
    Search for more papers by this author
  • Christoph Peters,

    1. Institut für Molekulare Medizin und Zellforschung, Albert-Ludwigs-Universität, Freiburg, Germany
    Search for more papers by this author
  • Thomas Reinheckel,

    1. Institut für Molekulare Medizin und Zellforschung, Albert-Ludwigs-Universität, Freiburg, Germany
    Search for more papers by this author
  • Paul Saftig,

    1. Biochemisches Institut, University of Kiel, Kiel, Germany
    Search for more papers by this author
  • Wouter Beertsen

    1. Department of Periodontology, ACTA, Universiteit van Amsterdam and Vrije Universiteit, Amsterdam, The Netherlands
    Search for more papers by this author

  • The authors state that they have no conflicts of interest.

Abstract

Osteoclastic bone degradation involves the activity of cathepsin K. We found that in addition to this enzyme other, yet unknown, cysteine proteinases participate in digestion. The results support the notion that osteoclasts from different bone sites use different enzymes to degrade the collagenous bone matrix.

Introduction: The osteoclast resorbs bone by lowering the pH in the resorption lacuna, which is followed by secretion of proteolytic enzymes. One of the enzymes taken to be essential in resorption is the cysteine proteinase, cathepsin K. Some immunolabeling and enzyme inhibitor data, however, suggest that other cysteine proteinases and/or proteolytic enzymes belonging to the group of matrix metalloproteinases (MMPs) may participate in the degradation. In this study, we investigated whether, in addition to cathepsin K, other enzymes participate in osteoclastic bone degradation.

Materials and Methods: In bones obtained from mice deficient for cathepsin K, B, or L or a combination of K and L, the bone-resorbing activity of osteoclasts was analyzed at the electron microscopic level. In addition, bone explants were cultured in the presence of different selective cysteine proteinase inhibitors and an MMP inhibitor, and the effect on resorption was assessed. Because previous studies showed differences in resorption by calvarial osteoclasts compared with those present in long bones, in all experiments, the two types of bone were compared. Finally, bone extracts were analyzed for the level of activity of cysteine proteinases and the effect of inhibitors hereupon.

Results: The analyses of the cathepsin-deficient bone explants showed that, in addition to cathepsin K, calvarial osteoclasts use other cysteine proteinases to degrade bone matrix. It was also shown that, in the absence of cathepsin K, long bone osteoclasts use MMPs for resorption. Cathepsin L proved to be involved in the MMP-mediated resorption of bone by calvarial osteoclasts; in the absence of this cathepsin, calvarial osteoclasts do not use MMPs for resorption. Selective inhibitors of cathepsin K and other cysteine proteinases showed a stronger effect on calvarial resorption than on long bone resorption.

Conclusions: Our findings suggest that (1) cathepsin K–deficient long bone osteoclasts compensate the lack of this enzyme by using MMPs in the resorption of bone matrix; (2) cathepsin L is involved in MMP-mediated resorption by calvarial osteoclasts; (3) in addition to cathepsin K, other, yet unknown, cysteine proteinases are likely to participate in skull bone degradation; and finally, (4) the data provide strong additional support for the existence of functionally different bone-site specific osteoclasts.

INTRODUCTION

The osteoclast, one of the few multinucleated cells in our body, is unique in its ability to resorb mineralized tissues like bone. To accomplish this, the osteoclast adheres to the mineralized surface, secludes it from the extracellular environment, and initiates resorption in the area opposite to the ruffled border. After dissolution of mineral crystallites caused by a lowered pH, proteolytic enzymes are released, and the bulk of the collagenous matrix is digested in the secluded area. Fragments of the matrix are taken up by the cell and further digested intracellularly and/or excreted at the baso-lateral membrane.(1)

Proteolytic enzymes crucial for the digestion of matrix constituents by osteoclasts primarily belong to two families of proteinases: the cysteine proteinases and the matrix metalloproteinases (MMPs). During the past decade, it has convincingly been shown that an essential enzyme in the digestion of bone matrix by osteoclasts is the cysteine proteinase cathepsin K. This conclusion is based on the following sequence of findings. In patients suffering from the rare osteopetrosis-like disease pycnodysostosis, adjacent to actively resorbing osteoclasts, large areas of nondigested bone matrix were present.(2) In an in vitro model system, it was shown that this phenomenon was caused by a decreased activity of cysteine proteinases.(3) Subsequently, it became apparent that osteoclasts express high levels of cathepsin K,(4,5) and genetic analyses of pycnodysostosis patients showed their deficiency for this enzyme.(6,7) Finally, it was shown in cathepsin K knockout mice that osteoclasts were able to dissolve the mineral, but resorption of the collagenous matrix was greatly disturbed.(8,9) In support of the important role of cathepsin K in osteoclast-mediated bone matrix digestion are studies in which (1) the activity of the enzyme was blocked by selective inhibitors(10–13) or (2) the expression of the enzyme was blocked.(14) However, several data suggest that, apart from cathepsin K, other proteolytic enzymes, either belonging or not belonging to the class of cysteine proteinases, are involved in matrix degradation by osteoclasts.(15)

In addition to the involvement of cysteine proteinases, ample data indicate that MMPs may participate in the osteoclast-mediated resorption of bone matrix.(16) By using selective inhibitors of MMPs, it was shown that resorption of matrix by osteoclasts was strongly inhibited.(16–19) Surprisingly, however, such an effect was only seen in calvarial bone explants and not or far less in long bones. Thus, participation of MMPs in bone resorption shows site-specific variations.(20) Osteoclasts of flat bones like skull or scapula used these enzymes for resorption, whereas osteoclasts in long bones did not depend on the activity of MMPs. Cysteine proteinases, however, are used by both subsets of osteoclasts.(21,22) It is of interest to note that skull-derived osteoclasts have lower levels of cathepsin K activity than those of long bones.(22) These data together indicate that osteoclasts at different sites of the skeleton use different enzymes for the resorption of bone matrix. The question arises whether different cysteine proteinases are also used.

By using inhibitors that are considered specific for certain members of the cysteine proteinase family, Hill et al.(23) showed that inhibition of cathepsin B had a strong inhibitory effect on the resorption of bone. An inhibitor thought to be more selective for cathepsin L also blocked resorption. These findings were confirmed in our group(18) and suggest that indeed other cysteine proteinases are involved in osteoclastic bone degradation. Hill et al. suggested that cathepsin B is of importance for intracellular digestion, whereas cathepsin L is assumed to exert its activity in the extracellular compartment, the ruffled border area. Various other data support the view that more than one cysteine proteinase participates in osteoclastic bone degradation.(24–30)

To study whether, next to cathepsin K, other cysteine proteinases are involved in osteoclast-mediated bone resorption and to find out whether MMPs are somehow linked to involvement of cysteine proteinases, this study was undertaken. We approached these questions in three different ways. First, we analyzed bones obtained from mice deficient for the cathepsins K, L, or B and from those deficient for both K and L. Explants of long bones and calvariae were analyzed electron microscopically to assess the effect of enzyme deficiency on osteoclastic bone resorption. In addition, explants of the two types of bone were obtained from cathepsin K– and L–deficient mice and cultured in the presence or absence of inhibitors of cysteine proteinases and MMPs. To assess whether matrix resorption by osteoclasts was affected, the amount of nondigested bone matrix adjacent to the osteoclasts was quantitatively determined using an approach previously described.(18,19) Second, selective low molecular weight inhibitors of cysteine proteinases, in particular those that were shown to inhibit cathepsin K,(31) were tested for their effect on osteoclastic bone degradation. Finally, the activity of cysteine proteinases was analyzed in extracts of both flat and long bone, and the effect of the inhibitors of cysteine proteinases was assessed in these extracts.

MATERIALS AND METHODS

Materials

The general cysteine proteinase inhibitor E-64 was obtained from Sigma Chemical (St Louis, MO, USA). The more selective cathepsin B inhibitor CA074 was obtained from Calbiochem (San Diego, CA, USA). Medium 199 and FCS were from Gibco (Gibco Laboratory, Grand Island, NY, USA). The cysteine proteinase inhibitors Mu-Leu-hPhe-VS-Ph, Mu-Np-hPhe-VS-Np (both inhibitors were kindly donated by Celera Pharmaceuticals, San Francisco, CA, USA), Z-Gly-Pro-Gln-VS-Ph, and Z-Gly-Pro-Leu-CHO were kindly synthesized by Dr J Powers (Georgiatech, Atlanta, GA, USA) and were described previously.(31,32) The MMP inhibitor CT1166 was a kind gift from Dr AJ Docherty (R&D Celltech, Slough, UK) and has been described previously.(33,34) The cathepsin B substrate Z-Ala-Arg-Arg-4-methoxy-β-naphtylamide, cathepsin L substrate Z-Phe-Arg-methoxy-β-naphtylamide-HCl, and cathepsin K substrate Z-Leu-Arg-4-methoxy-β-naphtylamide were from Bachem AG (Bubendorf, Switzerland).

Mice

The cathepsin K−/− mice have been described by Saftig et al.(8) The cathepsin L−/− and B−/− mice were described by Roth et al.(35) and Halangk et al.,(36) respectively. Mice deficient for both cathepsin K and L were also generated. These mice survived normally but, as shown for the L−/− mice, lost their hair. Control mice were used with the same C57B6×129SV mixed genetic background.

Tissue processing

Mice of different ages (5 days to 7 months) were processed for microscopic analysis. The older mice were perfused with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 6.9). After perfusion, tissue was prepared and fixed overnight in the same fixative. The tissues obtained from these mice were metacarpals, tibias, and calvariae. The bone samples were decalcified in 0.1 M EDTA with 2.5% glutaraldehyde in 0.1 M sodium cacodylate (pH 7.4). Decalcification was performed at 4°C for ∼2 weeks. After decalcification, the bones were washed in cacodylate buffer, and postfixation was carried out for 1 h in 1% OsO4 in cacodylate buffer. The tissue samples were dehydrated in a series of ethanol and subsequently embedded in epoxy resin.

Bone samples of the younger mice (5 days old) were collected, fixed, and processed as indicated below.

Tissue culture

Calvariae (frontal and parietal bones) and metacarpalia from 5-day-old mice were aseptically removed and cultured in 250 μl medium 199, containing 2.5% FCS. The following proteinase inhibitors were added to the culture medium: (1) the broad-spectrum cysteine proteinase inhibitors E-64 (40 μM) or Mu-Leu-hPhe-VS-Ph (0.1 μM); (2) the more selective cathepsin B inhibitor CA074(37); (3) an inhibitor for cathepsins L, S, and B, Mu-Np-hPhe-VS-Np (1 μM(31)); (4) inhibitors more selective for cathepsin K, Z-Gly-Pro-Gln-VS-Ph and Z-Gly-Pro-Leu-CHO (both inhibitors were used in concentrations of 1 and 10 μM); and (5) the MMP inhibitor CT1166 (10 μM(22)). The explants were cultured for up to 24 h and processed for microscopic examination.

Microscopy

Calvarial and long bone explants were fixed in 1% glutaraldehyde and 4% formaldehyde in 0.1 M sodium cacodylate buffer (pH 7.4) and without decalcification were further processed for light and electron microscopic examination as described previously.(3)

Semithin transverse sections (1–2 μm) were cut and stained with methylene blue or with a modified Goldner Trichrome staining method. Ultrathin sections were cut with a diamond knife, contrasted with uranyl and lead, and examined in a Zeiss EM-10 electron microscope.

Morphometric analyses

Noncultured and cultured bone samples were morphometrically analyzed. These analyses included (1) the number of osteoclasts adjacent to the bone surface and (2) the surface area of the resorption pits as described previously.(19) In various studies, we showed that osteoclastic bone resorption is strongly inhibited by selective inhibitors of proteolytic enzymes.(3,18–20,22,38) Under the influence of inhibitors of cysteine proteinases or MMPs, the cells are able to demineralize the bone but fail to digest the created demineralized bone matrix. These demineralized areas were quantitatively assessed as described.(18–20) In short, random ultrathin sections of the midsagittal plane of the long bone explants and cross-sections of skull explants were analyzed. Of each section, all osteoclasts attached to bone were micrographed. In addition, micrographs were obtained from all areas of nondigested demineralized bone matrix. The micrographs were coded and randomized, and the surface areas of demineralized areas were analyzed by using a digitized X-Y tablet. Data were expressed as mean micrometers squared demineralized area per osteoclast ± SD or SE of five to seven explants.

The amount of internalized bone collagen in osteoclasts was analyzed in metacarpals of cathepsin K−/− and K/L−/− mice. All osteoclasts of bone sections of five mice of each genotype were micrographed in their entirety, and the surface area of these cells was analyzed. Subsequently, the surface area of lysosomal structures containing cross-banded bone collagen fibrils was assessed, and the data were expressed as mean ± SD volume density of internalized collagen per osteoclast.

Cysteine proteinase activity

Long bones (humeri) and calvariae were collected from wildtype and the various cathepsin-deficient mice. The bones were cleaned from adhering periosteal tissue, chopped into small pieces, sonicated (3 × 5 s) on ice, overnight extracted in sodium acetate buffer (0.1 M, pH 5.3; 0.2% Triton X-100) at 4°C, again sonicated, and finally centrifuged at 15,000 rpm (15 minutes). The supernatant was collected and frozen at −20°C. Extracts were subjected to a fluorometric assay to determine cathepsin B, L, and K activity. Ten-microliter aliquots of extracts were incubated in 50 μl 100 mM phosphate buffer (pH 6.0) containing 1 mg/ml of enzyme substrate (Z-Ala-Arg-Arg-4-methoxy-β-naphtylamide for cathepsin B, Z-Phe-Arg-methoxy-β-naphtylamide-HCl for cathepsin L, and Z-Leu-Arg-4-methoxy-β-naphtylamide for cathepsin K), 1.3 mM EDTA, 1 mM dithiothreitol, and 2.67 mM l-cysteine. Incubation was performed at 37°C, and readings were every 10 minutes. Substrate hydrolysis was determined using a multilabel counter (λex = 355 nm; λem = 430 nm).

The extracts were analyzed for their protein content using a BCA-kit, and the analyses of enzyme activity were performed with samples equalized according to their protein level.

Statistical analysis

Data were statistically analyzed using the Kruskal-Wallis nonparametric ANOVA test followed by Tukey-Kramer's multiple comparison tests. Effects were considered statistically significant when p < 0.05 (two-tailed).

RESULTS

Cathepsin-deficient mice

In vivo observations:

In line with data presented recently by others,(39,40) we found that the number of osteoclasts in long bones was higher in cathepsin K−/− mice. Analyses of the calvariae of these mice revealed that such a difference was not apparent in this type of bone. In bones from cathepsin L–deficient mice, the number of osteoclasts was lower than in control bones.(41) This effect was not only seen in long bones but also in the skull. Deficiency of both cathepsin K and L resulted in a number of osteoclasts similar to control bones.

In flat and long bones that lacked cathepsin K (K−/− and K/L−/−), large areas of demineralized nondigested bone matrix were found adjacent to the osteoclasts (Fig. 1A). Such areas were not found in wildtype mice or in mice deficient for only cathepsin B or L. Quantitative analysis showed that, in bones of the cathepsin K–deficient mice, significantly higher levels were present in long bones than in calvariae (Fig. 2).

Figure Figure 1.

(A) Demineralized nondigested bone matrix (DBM) adjacent to an osteoclast (OC) of a metacarpal of a cathepsin K–deficient mouse. Bar: 10 μm. (B) Intracellular accumulation (*) of nondigested cross-banded bone collagen fibrils (arrows) in an osteoclast of a metacarpal from a mouse deficient for both cathepsin K and L. Bar: 0.2 μm.

Figure Figure 2.

Metacarpals and calvariae were collected form 5-day-old cathepsin-deficient mice and immediately processed for microscopic analysis. The amount of demineralized nondigested bone matrix (DBM) adjacent to the osteoclasts was assessed and the data are expressed as mean micrometers squared DBM per osteoclast ± SD of bones obtained from five mice of each phenotype. The amount of DBM present in calvariae was compared with the amount in metacarpals, and this proved to be significantly higher in metacarpals for both cathepsin K−/− and K/L−/− mice (p < 0.02).

Another characteristic phenomenon of osteoclasts lacking the activity of certain cysteine proteinases is the intracellular presence of internalized undigested, yet demineralized, bone collagen fibrils.(2,42) These intracellular fibrils were seen in many osteoclasts of mice lacking cathepsin K.

It was of interest to note that osteoclasts in the double knockout mice (K/L−/−) contained relatively high levels of internalized collagen (Fig. 1B). By comparing the accumulated amount seen in the osteoclasts of these animals with mice deficient for cathepsin K only, the amount of intracellular bone collagen proved to be higher in the cells of the double knockout animals (cathepsin K−/−: 1.3 ± 0.7%; cathepsin K/L−/−: 5.6 ± 4.3%). The data represent the mean ± SD volume density of intracellular collagen per whole osteoclast (n = 5 mice per genotype).

Cultured explants:

To study whether, in addition to cathepsin K, other proteolytic enzymes are involved in the degradation of bone matrix by osteoclasts, we studied the effects of inhibitors of different classes of enzymes. Bone explants obtained from 5-day-old mice deficient for cathepsin K, L, or both K and L were cultured in the presence of selective general inhibitors for cysteine proteinases (E-64) or MMPs (CT1166) and, in most culture experiments, with a combination of both inhibitors.

Cathepsin K−/− and K/L−/−:

In line with data previously presented,(18,22) morphometric analyses of the bone explants of control (wildtype) mice cultured with the cysteine proteinase inhibitor E-64 revealed for both types of bone a significant increase in the amount of nondigested bone matrix adjacent to the osteoclasts (see WT of Figs. 3A and 3B). In the presence of the MMP inhibitor, such an effect was seen in calvarial (WT in Fig. 3B) but not in long bone explants (WT in Fig. 3A). Also, the latter finding is in line with previous observations.(22) A 24-h incubation of bone samples from cathepsin K– or K/L–deficient mice with E-64 showed that the inhibitor had no additional effect on the amount of nondigested demineralized bone matrix adjacent to osteoclasts (Fig. 3).

Figure Figure 3.

Effect of a general MMP and cysteine proteinase inhibitor (10 μM CT1166 and 40 μM E-64, respectively) on the occurrence of nondigested bone matrix adjacent to the ruffled border of osteoclasts of bone explants obtained from wildtype, cathepsin K−/− and K/L−/− mice. The explants were cultured for 24 h (A and B) with or without the inhibitors and subsequently further processed for microscopic analysis. (C and D) Data are presented of short-term (5 h) cultures. Bone explants were obtained from cathepsin K−/− mice and cultured with or without a cysteine proteinase and/or MMP-inhibitor. The data are expressed as mean micrometer squared DBM per osteoclast ± SE of (A and B) six or (C and D) seven explants per incubation. The data obtained from inhibitor cultured explants were compared with those cultured without inhibitor and statistically analyzed. WT, wildtype; co, control, only DMSO [dissolvent for the MMP-inhibitor] was added; CPi, cysteine proteinase inhibitor (40 μM E-64); MMPi, MMP-inhibitor (10 μM CT1166). (A and C) Metacarpal bone explants. (B and D) Calvarial explants.

The MMP inhibitor, however, seemed to have an effect. In both metacarpal and calvarial bone explants of the cathepsin K−/− mice, a higher amount of demineralized bone matrix was found (Figs. 3A and 3B). Such an effect was absent in bones obtained from the K/L-deficient mice. Although the amount of nondigested bone matrix was higher in the K−/− explants, the effect was not significant. We therefore decided to culture explants of the K−/− mice for a shorter period, 5 h, and found that inhibition of the MMP activity of long bone explants resulted in a significantly increased amount of nondigested bone matrix (Fig. 3C). In addition to the amount of nondigested matrix already present because of the absence of cathepsin K activity, inhibition of MMPs proved to result in a three times larger surface area of nondigested bone matrix. In this type of bone (metacarpal), the cysteine proteinase inhibitor E-64 had no additional effect. The combined presence of both types of inhibitors resulted in an effect similar to that found with the MMP inhibitor alone. In contrast to the effects on long bones, in calvarial explants of K−/− mice, a significant effect was seen with each of the two inhibitors (Fig. 3D).

Finally, it was noted that the difference in the level of demineralized bone matrix as seen in vivo (Fig. 2) was no longer apparent after culturing.

Cathepsin L−/−:

The effect of E-64 on cathepsin L−/− bones (calvariae and long bones) was comparable with that seen in bones obtained from wildtype mice. The MMP inhibitor had no effect on the long bone cultures (data not shown), which is in line with previous findings.(22) In calvarial explants of control mice, the latter inhibitor resulted in an increased amount of nondigested matrix, again similar to what we found previously. However, in calvarial explants obtained from cathepsin L–deficient mice, the MMP inhibitor had no effect on the osteoclast-mediated resorption of bone (Fig. 4).

Figure Figure 4.

Effect of a general cysteine proteinase and MMP inhibitor (40 μM E-64 and 10 μM CT1166, respectively) on the occurrence of nondigested bone matrix adjacent to the ruffled border of osteoclasts of calvarial bone explants obtained from wildtype (WT) or cathepsin L−/− mice. The explants were cultured for 24 h with or without the inhibitors and subsequently further processed for microscopic analysis. Data are expressed as mean micrometer squared DBM per osteoclast ± SD of six explants per incubation. For each culture condition, the values of the cathepsin L−/− were compared with those of the wildtype (WT) explants and statistically analyzed. NS, not significant; CPi, cysteine proteinase inhibitor (40 μM E-64); MMPi, MMP inhibitor (10 μM CT1166).

Cathepsin activity in bone extracts:

Analyses of bone extracts for the activity of cathepsins by using different substrates revealed for all phenotypes (wildtype as well as cathepsin-deficient mice) that cysteine proteinase activity was much higher in long bones than in calvarial bone samples (Fig. 5A). It should be noted that using the cathepsin B (Fig. 5A) or the alleged selective cathepsin K substrate(43) extracts obtained from cathepsin K−/− and K/L−/− bones always showed a relatively high cysteine proteinase activity. For both substrates, similar values were noted (compare Figs. 5A–5C).

Figure Figure 5.

(A) Long bone and calvarial extracts obtained from mice deficient for different cathepsins (cathepsin K, L, B, and K/L) were analyzed for the activity of cathepsins. The figure represents the activity values obtained by using a substrate for cathepsin B. Comparable data were obtained with a substrate for cathepsin K (see B and C) and L. A complete inhibition was found in the presence of E-64, indicating the activity was cysteine proteinase mediated. Note the very low level of activity in bones obtained from cathepsin B−/− mice and the difference in activity between extracts of humeri and calvariae. Data are expressed as mean activity in arbitrary units per milligram protein ± SD of bones obtained form at least four mice per phenotype. (B and C) Extracts of humeri and calvariae of wildtype and cathepsin K−/− mice were analyzed for the activity of cysteine proteinases by using a cathepsin K substrate. Note the higher activity in humerus extracts of the K−/− mice compared with the control values of the wildtype mouse. (Note the different scales of the y-axes of B and C.) The cathepsin B inhibitor Ca-074 was tested for its inhibitory activity and proved to block enzyme activity in the calvarial extracts almost completely (85%, C), whereas inhibition of the long bone extracts was only 40% (B). The level of activity of the K−/− bone extracts was compared with those of the wildtype bones and statistically analyzed: *p < 0.05; **p < 0.001; NS, not significant.

Extracts of wildtype and cathepsin K−/− bones were analyzed for enzyme activity using the cathepsin K substrate and the effect of CA-074 hereupon. The inhibitor almost completely blocked the activity in extracts of humerus (Fig. 5B) and calvaria (Fig. 5C) obtained from wildtype mice. A comparable inhibition was seen in extracts of cathepsin K−/− calvariae (Fig. 5C), but in the humerus extracts of this mice inhibition was only 40% (Fig. 5B).

Effect of selective inhibitors on resorption and enzyme activity

Effect on osteoclastic bone matrix degradation:

Calvarial and metacarpal bone samples of control mice were incubated for 24 h in the presence or absence of a series of selective cathepsin inhibitors (Fig. 6). Analysis of the amount of nondigested bone matrix adjacent to the ruffled border of osteoclasts revealed that the broad spectrum cysteine proteinase inhibitor Mu-Leu-hPhe-VS-Ph had an effect on both types of bone comparable with that found with E-64 (Fig. 6, bar 7). Mu-Leu-hPhe-VS-Ph proved to be a very effective inhibitor of the digestion of bone matrix. In the presence of 0.1 μM of this inhibitor, the amount of nondigested bone matrix was comparable to that with 40 μM E-64.

Figure Figure 6.

Effect of different cysteine proteinase inhibitors on the occurrence of demineralized nondigested bone matrix adjacent to osteoclasts of metacarpals or calvariae of normal mice. The following inhibitors in the respective concentrations were used (see Materials and Methods section for the specificity of the compounds): 1. Control, without inhibitor but with DMSO used to dissolve the inhibitors; 2. Mu-Np-hPhe-VS-Np, 1 μM; 3. Z-Gly-Pro-Gln-VS-Ph, 1 μM; 4. Z-Gly-Pro-Gln-VS-Ph, 10 μM; 5. Z-Gly-Pro-Leu-CHO, 1 μM; 6. Z-Gly-Pro-Leu-CHO, 10 μM; 7. Mu-Leu-hPhe-VS-Ph, 0.1 μM. The explants were cultured for 24 h in the absence or presence of the inhibitors and then processed for electron microscopic analysis. Data are presented as mean micrometers squared DBM per osteoclast ± SE of six explants per incubation. Amounts of DBM of the various incubations were compared with the respective controls (1 of metacarpus or calvaria) and statistically analyzed. NS, not significant.

In metacarpal bones, none of the other inhibitors showed a distinct effect on digestion of bone matrix by osteoclasts, whereas in calvarial explants, some inhibitors had an effect. Higher amounts of nondigested bone matrix were found with Z-Gly-Pro-Gln-VS-Ph (10 μM; Fig. 6, bar 4) and Z-Gly-Pro-Leu-CHO (10 μM; Fig. 6, bar 6).

Cathepsin activity in bone extracts and the effect of inhibitors:

The inhibitors used in the in vitro bone resorption assay were also tested for their effects on cysteine proteinase activity in bone extracts of control mice. The only notable difference in inhibition profile we found was with the general inhibitor Mu-Np-hPhe-VS-Np. A relatively high residual activity with either enzyme substrate was apparent in calvarial extracts with 1 μM of inhibitor (21% versus 3% in long bones; p < 0.01; Fig. 7); a higher concentration resulted for both types of bone extracts in an almost complete inhibition (Fig. 7).

Figure Figure 7.

Effect of the cysteine proteinase inhibitor Mu-Np-hPhe-VS-Np on cysteine proteinase activity (cathepsin B substrate) in extracts of calvariae and humeri of normal mice. Extracts of bones of four mice were analyzed for the activity of cathepsin B, and different concentrations of the inhibitor were present during the incubation. In extracts of long bone, a complete inhibition is apparent with 1 μM of inhibitor, whereas extracts of calvarial bone needed a higher concentration of inhibitor to reach such a level of inhibition. Data are expressed as mean percent inhibition ± SD of extracts of four bones. *p < 0.01 compared with calvaria at 1000 nM.

DISCUSSION

The data presented in this study reveal a series of novel findings regarding the participation of proteolytic enzymes in the degradation of bone matrix by osteoclasts. By making use of bone explants obtained from mice deficient for one or two cysteine proteinases and by using different low molecular weight proteinase inhibitors, we were able to show substantial differences among different sites in the skeleton as to the participation of various proteinases in osteoclast-mediated bone degradation.

The use of low molecular weight cysteine proteinase inhibitors showed that resorption of the different types of bone analyzed—long bone versus skull bone—was differently affected. Two of the five inhibitors tested in the culture experiments proved to (partially) inhibit resorption of skull bone but had no effect on long bone resorption (Fig. 6). The data indicate differences in the cysteine proteinases used at the two skeletal sites. In line with this finding are the observed differences in the activity of these enzymes between skull and long bones. Taken together, our findings indicate that osteoclastic resorption of skull bone depends on cysteine proteinase activity that differs from the activity needed by long bone osteoclasts. The exact nature of these enzymes has yet to be elucidated.

Although in both flat and long bones, cathepsin K is expressed by osteoclasts, the cathepsin K–deficient mice showed an osteopetrotic effect in long bones, whereas the calvariae appear normal.(9) This lack of effect may be explained by the level of this enzyme in the two types of bone. Skull osteoclasts were shown to express much lower levels of cathepsin K than long bone cells,(22) thus suggesting that cathepsin K is more important for resorption of long bones than for resorption of skull bone.(22) However, because both types of osteoclasts seem similarly effective in resorption,(22) the findings also suggest that skull osteoclasts use, in addition to cathepsin K, not only MMPs,(18–20,22) but also other cysteine proteinases.

In the absence of cathepsin K, osteoclasts were strongly hampered in their capacity to digest bone matrix. This was not only apparent because of the extracellular accumulation of nondigested demineralized matrix, but also because of the huge amounts of bone collagen fibrils in lysosomal structures of the osteoclasts. The presence of intracellular collagen fibrils indicates uptake of this collagen by the cells. This is probably caused by an insufficient extracellular digestion because of the absence of cathepsin K. We assume that the uptake by the osteoclast is an attempt of the cell to digest the collagen intracellular in the lysosomal apparatus. Whether the collagen accumulates in the vacuoles or is finally transcytosed to the basolateral membrane is not known. The findings are in line with previous studies in mice(8,9) and humans(2) and underscore the essential role played by cathepsin K in osteoclast-mediated degradation of bone. However, our observations indicate that cathepsin K is perhaps less important in skull bone resorption than in resorption of long bones. Because blockage of the activity of all cysteine proteinases resulted in a significantly higher level of undigested bone matrix in calvarial bone (Fig. 3D), we conclude that, besides cathepsin K, other, yet unknown, cysteine proteinases are involved in the digestion of bone matrix by calvarial osteoclasts.

Quite unexpected was the effect of the MMP inhibitor on resorption by osteoclasts in long bones. Whereas previous studies seem to indicate that these osteoclasts resorb bone without a clear participation of MMPs,(20–22) our data do suggest the involvement of members of this class of proteinases under conditions in which cathepsin K expression is suppressed (Fig. 3C). In this respect, our observations are in line with those of Kiviranta et al.,(39) who showed an increased expression of certain MMPs (MMP-9, −13, and −14) in bones of cathepsin K–deficient mice. The latter authors proposed that this would support the suggestion that bone-lining cells participate in the digestion of demineralized bone matrix,(38) because it was shown that this type of cell digests nondigested bone matrix left by the osteoclast. This activity depends on MMPs.(38) The increased expression of MMPs may relate also to the activity of the osteoclasts; an option supported by recent data presented by Henriksen et al.(44) These authors showed that, because of blockage of the activity of cysteine proteinases, MMPs became involved in osteoclast-mediated matrix resorption.

The data obtained from the bones of the cathepsin-deficient mice provide new insight in the participation of various cathepsins in osteoclastic bone degradation. In previous studies in which selective inhibitors of various cysteine proteinases were used,(18,23) involvement of both cathepsin B and L was suggested. The localization and activity of these enzymes in and around the osteoclast(24,29) suggested their possible involvement in bone digestion. However, the data presented in this study do not support this. We have not been able to find osteoclast-related effects in bones of mice lacking cathepsin B or L. Moreover, no signs of an osteopetrotic phenotype were noted. In contrast, there was a decrease in trabecular bone in cathepsin L–deficient mice.(41) These observations support the findings of James et al.,(45) who proposed that cathepsin L is not involved in osteoclastic bone degradation. This was only apparent, however, for the long bone system. In calvariae of cathepsin L–deficient mice, resorption proved to differ from bone of wildtype animals. Our data show that inhibition of MMPs had no effect on osteoclastic resorption of cathepsin L–deficient skull bone (Fig. 4), whereas in control bones, an effect was found in line with previous studies.(18,19,22) Therefore, it seems that, in the absence of cathepsin L activity, MMPs are not used by the osteoclast for the resorption of calvarial bone matrix. This could suggest that cathepsin L plays a role in osteoclast-mediated bone matrix resorption by activating MMPs of the osteoclasts. It is well known that certain cysteine proteinases have the capacity to activate MMPs.(46) In line with this view is the finding of the increased amount of internalized collagen by cells lacking both cathepsin K and L. MMPs normally active and involved in matrix degradation(44) are not activated in the absence of cathepsin L, thus leading to an accumulation of nondigested bone collagen. Alternatively, such an accumulation can be explained also by assuming a role played by cathepsin L in the lysosomal apparatus, either intracellularly or extracellularly in the ruffled border area.

Our findings with the low molecular weight cysteine proteinase inhibitors Mu-Leu-hPe-VS-Ph and Mu-Np-hPe-VS-Np are in line with data presented by Xia et al.(31) These authors showed that the more general inhibitor (Mu-Leu-hPe-VS-Ph) strongly inhibited the formation of resorption pits by isolated osteoclasts, whereas the inhibitor for cathepsins S, L, and B (Mu-Np-hPe-VS-Np) had no effect. In our study, however, effects with more selective inhibitors for cathepsin K were only apparent for the skull system and absent in the long bones, suggesting again that the osteoclasts of the different bone types make use of different cysteine proteinases. The latter assumption is in line with the activity data obtained from the bone extracts. Next to a higher level of cysteine proteinase activity in extracts of long bones, we also found different inhibition profiles by making use of the various inhibitors.

The findings presented in this study support the view that bone site–specific osteoclasts may exist.(21,22) It is still unclear, however, whether the differences in activity of the osteoclasts present at these different bone sites are caused by differences in osteoclasts, differences in substrate, and/or differences in osteoblasts that modulate the activity of the osteoclasts. In our initial study on the heterogeneity of osteoclasts, we seeded native calvarial osteoclasts on slices of long bone and vice versa.(22) Under these conditions, the osteoclasts kept their bone site–specific phenotype, strongly suggesting that the osteoclasts are indeed different. However, a more in-depth study is needed to elucidate this interesting topic in more detail.

In conclusion, the data presented in this study indicate that, in addition to cathepsin K, other proteolytic enzymes participate in the osteoclast-mediated digestion of bone. Our data provide additional support for the view that bone site–specific osteoclasts exist and that these osteoclasts make use of a different enzyme repertoire to resorb the bone matrix. Under normal conditions, it seems that cathepsin K is the most essential cysteine proteinase in degrading the matrix of long bones. Calvarial osteoclasts, however, make use of one or more other members of this group of enzymes. Further studies are needed to elucidate the nature of these enzymes. In addition to cathepsins, the calvarial osteoclasts use MMPs. This seems to occur both under normal conditions and in the absence of cathepsin K. Surprisingly, the lack of cathepsin K results in the use of MMPs by long bone osteoclasts, enzymes that are normally not involved in matrix degradation by these osteoclasts. Finally, cathepsin L seems to play a hitherto unrecognized role in the digestion of bone matrix by calvarial osteoclasts. The enzyme may be involved in modulating MMP activity in this type of osteoclast.

Acknowledgements

This study was supported by Grant NIH AR46182 (DB).

Ancillary