Periosteal BMP2 activity drives bone graft healing

Bone graft incorporation depends on the orchestrated activation of numerous growth factors and cytokines in both the host and the graft. Prominent in this signaling cascade is BMP2. Although BMP2 is dispensable for bone formation, it is required for the initiation of bone repair; thus understanding the cellular mechanisms underlying bone regeneration driven by BMP2 is essential for improving bone graft therapies. In the present study, we assessed the role of Bmp2 in bone graft incorporation using mice in which Bmp2 has been removed from the limb prior to skeletal formation (Bmp2(cKO)). When autograft transplantations were performed in Bmp2cKO mice, callus formation and bone healing were absent. Transplantation of either a vital wild type (WT) bone graft into a Bmp2(cKO) host or a vital Bmp2(cKO) graft into a WT host also resulted in the inhibition of bone graft incorporation. Histological analyses of these transplants show that in the absence of BMP2, periosteal progenitors remain quiescent and healing is not initiated. When we analyzed the expression of Sox9, a marker of chondrogenesis, on the graft surface, we found it significantly reduced when BMP2 was absent in either the graft itself or the host, suggesting that local BMP2 levels drive periosteal cell condensation and subsequent callus cell differentiation. The lack of integrated healing in the absence of BMP2 was not due to the inability of periosteal cells to respond to BMP2. Healing was achieved when grafts were pre-soaked in rhBMP2 protein, indicating that periosteal progenitors remain responsive in the absence of BMP2. In contrast to the requirement for BMP2 in periosteal progenitor activation in vital bone grafts, we found that bone matrix-derived BMP2 does not significantly enhance bone graft incorporation. Taken together, our data show that BMP2 signaling is not essential for the maintenance of periosteal progenitors, but is required for the activation of these progenitors and their subsequent differentiation along the osteo-chondrogenic pathway. These results indicate that BMP2 will be among the signaling molecules whose presence will determine success or failure of new bone graft strategies.

Differential regulation of metzincins in experimental chronic renal allograft rejection: potential markers and novel therapeutic targets

Chronic renal allograft rejection is characterized by alterations in the extracellular matrix compartment and in the proliferation of various cell types. These features are controlled, in part by the metzincin superfamily of metallo-endopeptidases, including matrix metalloproteinases (MMPs), a disintegrin and metalloproteinase (ADAM) and meprin. Therefore, we investigated the regulation of metzincins in the established Fisher to Lewis rat kidney transplant model. Studies were performed using frozen homogenates and paraffin sections of rat kidneys at day 0 (healthy controls) and during periods of chronic rejection at day +60 and day +100 following transplantation. The messenger RNA (mRNA) expression was examined by Affymetrix Rat Expression Array 230A GeneChip and by real-time Taqman polymerase chain reaction analyses. Protein expression was studied by zymography, Western blot analyses, and immunohistology. mRNA levels of MMPs (MMP-2/-11/-12/-14), of their inhibitors (tissue inhibitors of metalloproteinase (TIMP)-1/-2), ADAM-17 and transforming growth factor (TGF)-beta1 significantly increased during chronic renal allograft rejection. MMP-2 activity and immunohistological staining were augmented accordingly. The most important mRNA elevation was observed in the case of MMP-12. As expected, Western blot analyses also demonstrated increased production of MMP-12, MMP-14, and TIMP-2 (in the latter two cases as individual proteins and as complexes). In contrast, mRNA levels of MMP-9/-24 and meprin alpha/beta had decreased. Accordingly, MMP-9 protein levels and meprin alpha/beta synthesis and activity were downregulated significantly. Members of metzincin families (MMP, ADAM, and meprin) and of TIMPs are differentially regulated in chronic renal allograft rejection. Thus, an altered pattern of metzincins may represent novel diagnostic markers and possibly may provide novel targets for future therapeutic interventions.

Influence of diurnal hyperosmotic loading on the metabolism and matrix gene expression of a whole-organ intervertebral disc model

It is generally agreed that the mechanical environment of intervertebral disc cells plays an important role in maintaining a balanced matrix metabolism. The precise mechanism by which the signals are transduced into the cells is poorly understood. Osmotic changes in the extracellular matrix (ECM) are thought to be involved. Current in-vitro studies on this topic are mostly short-term and show conflicting data on the reaction of disc cells subjected to osmotic changes which is partially due to the heterogenous and often substantially-reduced culture systems. The aim of the study was therefore to investigate the effects of cyclic osmotic loading for 4 weeks on metabolism and matrix gene expression in a full-organ intervertebral disc culture system. Intervertebral disc/endplate units were isolated from New Zealand White Rabbits and cultured either in iso-osmotic media (335 mosmol/kg) or were diurnally exposed for 8 hours to hyper-osmotic conditions (485 mosmol/kg). Cell viability, metabolic activity, matrix composition and matrix gene expression profile (collagen types I/II and aggrecan) were monitored using Live/Dead cell viability assay, tetrazolium reduction test (WST 8), proteoglycan and DNA quantification assays and quantitative PCR. The results show that diurnal osmotic stimulation did not have significant effects on proteoglycan content, cellularity and disc cell viability after 28 days in culture. However, hyperosmolarity caused increased cell death in the early culture phase and counteracted up-regulation of type I collagen gene expression in nucleus and annulus cells. Moreover, the initially decreased cellular dehydrogenase activity recovered with osmotic stimulation after 4 weeks and aggrecan gene down-regulation was delayed, although the latter was not significant according to our statistical criteria. In contrast, collagen type II did not respond to the osmotic changes and was down-regulated in both groups. In conclusion, diurnal hyper-osmotic stimulation of a whole-organ disc/endplate culture partially inhibits a matrix gene expression profile as encountered in degenerative disc disease and counteracts cellular metabolic hypo-activity.

Synergistic inhibition in cell-cell fusion mediated by the matrix and nucleocapsid protein of canine distemper virus

Canine distemper virus (CDV) causes a chronic, demyelinating, progressive or relapsing neurological disease in dogs, because CDV persists in the CNS. Persistence of virulent CDV, such as the A75/17 strain has been reproduced in cell cultures where it is associated with a non-cytolytic infection with very limited cell-cell fusion. This is in sharp contrast to attenuated CDV infection in cell cultures, such as the Onderstepoort (OP) CDV strain, which produces extensive fusion activity and cytolysis. Fusion efficiency may be determined by the structure of the viral fusion protein per se but also by its interaction with other structural proteins of CDV. This was studied by combining genes derived from persistent and non-persistent CDV strains in transient transfection experiments. It was found that fusion efficiency was markedly attenuated by the structure of the fusion protein of the neurovirulent A75/17-CDV. Moreover, we showed that the interaction of the surface glycoproteins with the M protein of the persistent strain greatly influenced fusion activity. Site directed mutagenesis showed that the c-terminus of the M protein is of particular importance in this respect. Interestingly, although the nucleocapsid protein alone did not affect F/H-induced cell-cell fusion, maximal inhibition occurred when the latter was added to combined glycoproteins with matrix protein. Thus, the present study suggests that very limited fusogenicity in virulent CDV infection, which favours persistence by limiting cell destruction involves complex interactions between all viral structural proteins.

Tenascin-W is found in malignant mammary tumors, promotes alpha8 integrin-dependent motility and requires p38MAPK activity for BMP-2 and TNF-alpha induced expression in vitro

Tenascins represent a family of extracellular matrix glycoproteins with distinctive expression patterns. Here we have analyzed the most recently described member, tenascin-W, in breast cancer. Mammary tumors isolated from transgenic mice expressing hormone-induced oncogenes reveal tenascin-W in the stroma around lesions with a high likelihood of metastasis. The presence of tenascin-W was correlated with the expression of its putative receptor, alpha8 integrin. HC11 cells derived from normal mammary epithelium do not express alpha8 integrin and fail to cross tenascin-W-coated filters. However, 4T1 mammary carcinoma cells do express alpha8 integrin and their migration is stimulated by tenascin-W. The expression of tenascin-W is induced by BMP-2 but not by TGF-beta1, though the latter is a potent inducer of tenascin-C. The expression of tenascin-W is dependent on p38MAPK and JNK signaling pathways. Since preinflammatory cytokines also act through p38MAPK and JNK signaling pathways, the possible role of TNF-alpha in tenascin-W expression was also examined. TNF-alpha induced the expression of both tenascin-W and tenascin-C, and this induction was p38MAPK- and cyclooxygenase-dependent. Our results show that tenascin-W may be a useful diagnostic marker for breast malignancies, and that the induction of tenascin-W in the tumor stroma may contribute to the invasive behavior of tumor cells.

Experimental heart transplantation: effect of cyclosporine on expression and activity of metzincins

Metzincins, such as matrix metalloproteases (MMP), and extracellular matrix (ECM) proteins are differentially regulated in inflammation. We hypothesised that metzincins are also dysregulated in experimental acute cardiac allograft rejection. We investigated the Dark Agouti-to-Lewis (DA-to-Lew) rat model of acute cardiac allograft rejection. Cyclosporine (CsA) (7.5 mg/kg/d) was given from transplantation to sacrifice (day +5). At that time, mRNA levels were analysed by Affymetrix genechip and quantitative reverse transcription polymerase chain reaction (qRTPCR). MMP protein and activities were analysed by immunohistology, fluorometry, zymography and Western blots. In untreated rejected DA allografts, mRNA levels of MMP-2/-7/-9/-/12-/14, a disintegrin and metalloprotease (ADAM)-17, tissue inhibitor of metalloprotease (TIMP)-1/-3 were increased, whereas MMP-11/-16/-24 and TIMP-2/-4 were lowered compared to native DA hearts. With respect to these untreated allografts, CsA lowered mRNA levels of MMP-7, TIMP-1/-3 (TIMP-2/-4 remained relatively low) and ADAM17, but augmented mRNA levels of MMP-11/-16/-23 and of many ECM genes. Immunohistology showed increased staining of MMP-2 in acute rejection (AR). Overall MMP activity was augmented in both transplanted groups, but CsA reduced MMP-9 activity and MMP-14 production. Taken together, MMP and TIMP were upregulated during acute AR. CsA ameliorated histology of rejection but showed potential pro-fibrotic effects. Thus, MMP and TIMP may play a role in acute cardiac allograft rejection, and beneficial modification of the MMP-ECM balance requires interventions beyond CsA.

Cleavage of extracellular matrix in periodontitis: gingipains differentially affect cell adhesion activities of fibronectin and tenascin-C

Gingipains are cysteine proteases that represent major virulence factors of the periodontopathogenic bacterium Porphyromonas gingivalis. Gingipains are reported to degrade extracellular matrix (ECM) of periodontal tissues, leading to tissue destruction and apoptosis. The exact mechanism is not known, however. Fibronectin and tenascin-C are pericellular ECM glycoproteins present in periodontal tissues. Whereas fibronectin mediates fibroblast adhesion, tenascin-C binds to fibronectin and inhibits its cell-spreading activity. Using purified proteins in vitro, we asked whether fibronectin and tenascin-C are cleaved by gingipains at clinically relevant concentrations, and how fragmentation by the bacterial proteases affects their biological activity in cell adhesion. Fibronectin was cleaved into distinct fragments by all three gingipains; however, only arginine-specific HRgpA and RgpB but not lysine-specific Kgp destroyed its cell-spreading activity. This result was confirmed with recombinant cell-binding domain of fibronectin. Of the two major tenascin-C splice variants, the large but not the small was a substrate for gingipains, indicating that cleavage occurred primarily in the alternatively spliced domain. Surprisingly, cleavage of large tenascin-C variant by all three gingipains generated fragments with increased anti-adhesive activity towards intact fibronectin. Fibronectin and tenascin-C fragments were detected in gingival crevicular fluid of a subset of periodontitis patients. We conclude that cleavage by gingipains directly affects the biological activity of both fibronectin and tenascin-C in a manner that might lead to increased cell detachment and loss during periodontal disease.

Assessment of the Matrix Degenerative Effects of MMP-3, ADAMTS-4 and HTRA1 injected into a bovine Intervertebral Disc Organ Culture Model

Study Design. In vitro study to develop an intervertebral disc degeneration (IDD) organ culture model, using coccygeal bovine intervertebral discs (IVDs) and injection of proteolytic enzymes MMP-3, ADAMTS-4 and HTRA1.Objective. This study aimed to develop an in-vitro model of enzyme-mediated IDD to mimic the clinical outcome in humans for investigation of therapeutic treatment options.Summary of Background Data. Bovine IVDs are comparable to human IVDs in terms of cell composition and biomechanical behavior. Researchers injected papain and trypsin into them to create an IDD model with a degenerated nucleus pulposus (NP) area. They achieved macroscopic cavities as well as a loss of glycosaminoglycans (GAGs). However, none of these enzymes are clinically relevant.Methods. Bovine IVDs were harvested maintaining the endplates. Active forms of MMP-3, ADAMTS-4 and HTRA1 were injected at a dose of 10μg/ml each. Phosphate buffered saline (PBS) was injected as a control. Discs were cultured for 8 days and loaded diurnally (day 1 to day 4 with 0.4 MPa for 16 h) and left under free swelling condition from day 4 to day 8 to avoid expected artifacts due to dehydration of the NP. Outcome parameters included disc height, metabolic cell activity, DNA content, glycosaminoglycan (GAG) content, total collagen content, relative gene expression and histological investigation.Results. The mean metabolic cell activity was significantly lower in the NP area of discs injected with ADAMTS-4 compared to the day 0 control discs. Disc height was decreased following injection with HTRA1, and was significantly correlated with changes in GAG/DNA of the NP tissue. Total collagen content tended to be lower in groups injected with ADAMTS4 and MMP-3.Conclusion. MMP-3, ADAMTS-4 and HTRA1 neither provoked visible matrix degradation nor major shifts in gene expression. However, cell activity was significantly reduced and HTRA1 induced loss of disc height which positively correlated with changes in GAG/DNA content. The use of higher doses of these enzymes or a combination thereof may therefore be necessary to induce disc degeneration

RNAi reduces expression and intracellular retention of mutant cartilage oligomeric matrix protein.

Mutations in cartilage oligomeric matrix protein (COMP), a large extracellular glycoprotein expressed in musculoskeletal tissues, cause two skeletal dysplasias, pseudoachondroplasia and multiple epiphyseal dysplasia. These mutations lead to massive intracellular retention of COMP, chondrocyte death and loss of growth plate chondrocytes that are necessary for linear growth. In contrast, COMP null mice have only minor growth plate abnormalities, normal growth and longevity. This suggests that reducing mutant and wild-type COMP expression in chondrocytes may prevent the toxic cellular phenotype causing the skeletal dysplasias. We tested this hypothesis using RNA interference to reduce steady state levels of COMP mRNA. A panel of shRNAs directed against COMP was tested. One shRNA (3B) reduced endogenous and recombinant COMP mRNA dramatically, regardless of expression levels. The activity of the shRNA against COMP mRNA was maintained for up to 10 weeks. We also demonstrate that this treatment reduced ER stress. Moreover, we show that reducing steady state levels of COMP mRNA alleviates intracellular retention of other extracellular matrix proteins associated with the pseudoachondroplasia cellular pathology. These findings are a proof of principle and the foundation for the development of a therapeutic intervention based on reduction of COMP expression.

Ribozyme-mediated reduction of wild-type and mutant cartilage oligomeric matrix protein (COMP) mRNA and protein.

Dominant-negative mutations in the homopentameric extracellular matrix glycoprotein cartilage oligomeric matrix protein (COMP) result in inappropriate intracellular retention of misfolded COMP in the rough endoplasmic reticulum of chondrocytes, causing chondrocyte cell death, which leads to two skeletal dysplasias: pseudoachondroplasia (PSACH) and multiple epiphyseal dysplasia (EDM1). COMP null mice show no adverse effects on normal bone development and growth, suggesting a possible therapy involving removal of COMP mRNA. The goal of this study was to assess the ability of a hammerhead ribozyme (Ribo56, designed against the D469del mutation) to reduce COMP mRNA expression. In COS7 cells transfected with plasmids that overexpress wild-type or mutant COMP mRNA and Ribo56, the ribozyme reduced overexpressed normal COMP mRNA by 46% and mutant COMP mRNA by 56% in a dose-dependent manner. Surprisingly, the use of recombinant adenoviruses to deliver wild-type or mutant COMP mRNA and Ribo56 simultaneously into COS7 cells proved problematic for the activity of the ribozyme to reduce COMP expression. However, in normal human costochondral cells (hCCCs) infected only with adenoviruses expressing Ribo56, expression of endogenous wild-type COMP mRNA was reduced in a dose-dependent manner by 50%. In chondrocytes that contain heterozygous COMP mutations (D469del, G427E and D511Y) that cause PSACH, Ribo56 was more effective at reducing COMP mRNA (up to 70%). These results indicate that Ribo56 is effective at reducing mutant and wild-type COMP levels in cells and suggests a possible mode of therapy to reduce the mutant protein load.

ROLE OF PROSTAGLANDIN E2 IN THE REGULATION OF PANCREATIC STELLATE CELLS HYPER ACTIVITY ASSOCIATED WITH PANCREATIC CANCER

Pancreatic cancer is one of the most lethal type of cancer due to its high metastasis rate and resistance to chemotherapy. Pancreatic fibrosis is a constant pathological feature of chronic pancreatitis and the hyperactive stroma associated with pancreatic cancer. Strong evidence supports an important role of cyclooxygenase-2 (COX-2) and COX-2 generated prostaglandin E2 (PGE2) during pancreatic fibrosis. Pancreatic stellate cells (PSC) are the predominant source of extracellular matrix production (ECM), thus being the key players in both diseases. Given this background, the primary objective is to delineate the role of PGE2 on human pancreatic stellate cells (PSC) hyper activation associated with pancreatic cancer. This study showed that human PSC cells express COX-2 and synthesize high levels of PGE2. PGE2 stimulated PSC migration and invasion; expression of extra cellular matrix (ECM) genes and tissue degrading matrix metallo proteinases (MMP) genes. I further identified the PGE2 EP receptor responsible for mediating these effects on PSC. Using genetic and pharmacological approaches I identified the receptor required for PGE2 mediates PSC hyper activation. Treating PSC with Specific antagonists against EP1, EP2 and EP4, demonstrated that blocking EP4 receptor only, resulted in a complete reduction of PGE2 mediated PSC activation. Furthermore, siRNA mediated silencing of EP4, but not other EP receptors, blocked the effects of PGE2 on PSC fibrogenic activity. Further examination of the downstream pathway modulators revealed that PGE2 stimulation of PSC involved CREB and not AKT pathway. The regulation of PSC by PGE2 was further investigated at the molecular level, with a focus on COL1A1. Collagen I deposition by PSC is one of the most important events in pancreatic cancer. I found that PGE2 regulates PSC through activation of COL1A1 expression and transcriptional activity. Downstream of PGE2, silencing of EP4 receptor caused a complete reduction of COL1A1 expression and activity supporting the role of EP4 mediated stimulation of PSC. Taken together, this data indicate that PGE2 regulates PSC via EP4 and suggest that EP4 can be a better therapeutic target for pancreatic cancer to reduce the extensive stromal reaction, possibly in combination with chemotherapeutic drugs can further kill pancreatic cancer cells.

Insulin-associated changes in carnitine palmitoyltransferase activity are mediated through the insulin-like growth factor I pathway in the cultured neonatal rat cardiac myocyte

The mitochondrial carnitine palmitoyltransferase (CPT) system is composed of two proteins, CPT-I and CPT-II, involved in the transport of fatty acids into the mitochondrial matrix to undergo $\beta$-oxidation. CPT-I is located outside the inner membrane and CPT-II is located on the inner aspect of the inner membrane. The CPT proteins are distinct with different molecular weights and activities. The malonyl-CoA sensitivity of CPT-I has been proposed as a regulatory step in $\beta$-oxidation. Using the neonatal rat cardiac myocyte, assays were designed to discriminate between these activities in situ using digitonin and Triton X-100. With this methodology, we are able to determine the involvement of the IGF-I pathway in the insulin-mediated increase in CPT activities. Concentrations of digitonin up to 25 $\mu$M fail to release citrate synthase from the mitochondrial matrix or alter the malonyl-CoA sensitivity of CPT-I. If the mitochondrial matrix was exposed, malonyl-CoA insensitive CPT-II would reduce malonyl-CoA sensitivity. In contrast to digitonin, Triton X-100 (0.15%) releases citrate synthase from the matrix and exposes CPT-II. CPT-II activity is confirmed by the absence of malonyl-CoA sensitivity. To examine the effects of various agents on the expression and/or activity of CPT, it is necessary to use serum-free medium to eliminate mitogenic effects of serum proteins. Comparison of different media to optimize CPT activity and cell viability resulted in the decision to use Dulbecco's Modified Eagle medium supplemented with transferrin. In three established models of cardiac hypertrophy using the neonatal rat cardiac myocyte there is a significant increase in CPT-I and CPT-II activity in the treated cells. Analogous to the situation seen in the hypertrophy model, insulin also significantly increases the activity of the mitochondrial proteins CPT-I, CPT-II and cytochrome oxidase with a coinciding increase the expression of CPT-II and cytochrome oxidase mRNA. The removal of serum increases the I$\sb{50}$ (concentration of inhibitor that halves enzyme activity) of CPT-I for malonyl-CoA by four-fold. Incubation with insulin returns I$\sb{50}$ values to serum levels. Incubation with insulin significantly increases malonyl-CoA and ATP levels in the cells with a resulting reduction in palmitate oxidation. Once malonyl-CoA inhibition of CPT-I is removed by permeabilizing the cells, insulin significantly increases the oxidation of palmitoyl-CoA in a manner which parallels the increase in CPT-I activity. Interestingly, CPT-II activity increases significantly only at the tissue culture concentration (1.7 $\mu$M) of insulin suggesting that the IGF-I pathway may be involved. Supporting a role for the IGF-I pathway in the insulin-induced increase in CPT activity is the significant increase in the synthesis of both cellular and mitochondrial proteins as well as increased synthesis of CPT-II. Consistent with an IGF-mediated pathway for the effect of insulin, IGF-I (10 ng/ml) significantly increases the activities of both CPT-I and -II. An IGF-I analogue which inhibits the autophosphorylation of the IGF-I receptor blunts the insulin-mediated increase in CPT-I and -II activity by greater than 70% and virtually eliminates the IGF-I response by greater than 90%. This is the first study to demonstrate the involvement of the IGF-I pathway in the regulation of mitochondrial protein expression, e.g. CPT. ^

TGF-βRI kinase activity mediates Emdogain-stimulated in vitro osteoclastogenesis.

OBJECTIVES Emdogain, containing an extract of fetal porcine enamel matrix proteins, is a potent stimulator of in vitro osteoclastogenesis. The underlying molecular mechanisms are, however, unclear. MATERIAL AND METHODS Here, we have addressed the role of transforming growth factor-beta receptor type 1 (TGF-βRI) kinase activity on osteoclastogenesis in murine bone marrow cultures. RESULTS Inhibition of TGF-βRI kinase activity with SB431542 abolished the effect of Emdogain on osteoclastogenesis induced by receptor activator of nuclear factor kappa-B ligand or tumor necrosis factor-alpha. SB431542 also suppressed the Emdogain-mediated increase of OSCAR, a co-stimulatory protein, and dendritic cell-specific transmembrane protein and Atp6v0d2, the latter two being involved in cell fusion. Similar to transforming growth factor-beta1 (TGF-β), Emdogain could not compensate for the inhibition of IL-4 and IFNγ on osteoclast formation. When using the murine macrophage cell line RAW246.7, SB431542 and the smad-3 inhibitor SIS3 blocked Emdogain-stimulated expression of the transcription factor NFATc1. CONCLUSIONS Taken together, the data suggest that TGF-βRI kinase activity is necessary to mediate in vitro effects of Emdogain on osteoclastogenesis. CLINICAL RELEVANCE Based on these in vitro data, we can speculate that at least part of the clinical effects of Emdogain on osteoclastogenesis is mediated via TGF-β signaling.

Endogenous and synthetic MMP inhibitors in CNS physiopathology.

Matrix metalloproteinases (MMPs, including the membrane-type MMPs (MT-MMPs)), a disintegrin and metalloproteinase (ADAM), and ADAM with thrombospondin motifs belong to the metzincins, a subclass of metalloproteinases that contain a Met residue and a Zn(2+) ion at the catalytic site necessary for enzymatic reaction. MMP proteolytic activity is mainly controlled by their natural tissue inhibitors of metalloproteinase (TIMP). A number of synthetic inhibitors have been developed to control deleterious MMP activity. The roles of MMPs and some of their ECM substrates in CNS physiology and pathology are covered by other chapters of the present volume and will thus not be addressed in depth. This chapter will focus (i) on the endogenous MMP inhibitors in the CNS, (ii) on MMP and TIMP regulations in three large classes of neuropathologic processes (inflammatory, neurodegenerative, and infectious), and (iii) on synthetic inhibitors of MMPs and the perspective of their use in different brain diseases.

A Metalloproteinase Mirolysin of Tannerella forsythia Inhibits All Pathways of the Complement System

Recent reports focusing on virulence factors of periodontal pathogens implicated proteinases as major determinants of remarkable pathogenicity of these species, with special emphasis on their capacity to modulate complement activity. In particular, bacteria-mediated cleavage of C5 and subsequent release of C5a seems to be an important phenomenon in the manipulation of the local inflammatory response in periodontitis. In this study, we present mirolysin, a novel metalloproteinase secreted by Tannerella forsythia, a well-recognized pathogen strongly associated with periodontitis. Mirolysin exhibited a strong effect on all complement pathways. It inhibited the classical and lectin complement pathways due to efficient degradation of mannose-binding lectin, ficolin-2, ficolin-3, and C4, whereas inhibition of the alternative pathway was caused by degradation of C5. This specificity toward complement largely resembled the activity of a previously characterized metalloproteinase of T. forsythia, karilysin. Interestingly, mirolysin released the biologically active C5a peptide in human plasma and induced migration of neutrophils. Importantly, we demonstrated that combination of mirolysin with karilysin, as well as a cysteine proteinase of another periodontal pathogen, Prevotella intermedia, resulted in a strong synergistic effect on complement. Furthermore, mutant strains of T. forsythia, devoid of either mirolysin or karilysin, showed diminished survival in human serum, providing further evidence for the synergistic inactivation of complement by these metalloproteinases. Taken together, our findings on interactions of mirolysin with complement significantly add to the understanding of immune evasion strategies of T. forsythia and expand the knowledge on molecular mechanisms driving pathogenic events in the infected periodontium.