bone tissue remodeling and development: focus on matrix metalloproteinase functions

Archives of Biochemistry and Biophysics 561 (2014) 74–87

Contents lists available at ScienceDirect

Archives of Biochemistry and Biophysics

journal homepage: www.elsevier .com/ locate /yabbi

Review

Bone tissue remodeling and development: Focus on matrixmetalloproteinase functions

http://dx.doi.org/10.1016/j.abb.2014.07.0340003-9861/� 2014 Elsevier Inc. All rights reserved.

⇑ Corresponding author. Address: Extracellular Matrix Biology and Cell Interaction Group, Department of Anatomy, Institute of Biomedical Sciences – ICB III, UnivSão Paulo, Avenida Professor Lineu Prestes, 2415, CEP: 05508-000 São Paulo, SP, Brazil.

E-mail address: [email protected] (K.B.S. Paiva).1 Abbreviations used: MMPs, matrix metalloproteinases; ECM, extracellular matrix; Ihh, Indian hedgehog; PTH, parathyroid hormone; PTHrP, parathyroid hormon

protein; FGF, fibroblast growth factor; VEGF, vascular endothelial growth factor; CTGF, connective tissue growth factor; BMPs, bone morphogenetic proteins; BMmulticellular units; FGFRs, fibroblast growth factor receptors; YAP, Yes-association protein; PAT, palmitoyl acyltransferase; COMP, cartilage oligomeric matrix proteimesenchymal stem cells; AC, articular cartilage; BC, bone collar; BMC, bone marrow cavity; BS, bone spicules; BV, blood vessel; CB, cortical bone; Ch/C, chonchondrocyte; D, diaphysis; E, epiphyses; EL, epiphyseal line; Eo, endosteum; GP, growth plate; HZ, hypertrophic zone; M, metaphysis; MSC, mesenchymal stemosteoblast; Oc, osteoclast; Oo, osteocyte; PC, proliferative chondrocyte; Pc, perichondrium; PHC, prehypertrophic chondrocyte; Po, periosteum; POC (COJ), primary oscentres (chondro-osseous junction); PZ, proliferative zone; RZ, resting zone; TB, trabecular bone; SOC, secondary ossification centres; THC, terminal hypertrophic cho

Katiucia Batista Silva Paiva a,⇑, José Mauro Granjeiro b,c

a Matrix Biology and Cellular Interaction Group (GBMec), Department of Anatomy, Institute of Biomedical Sciences, University of São Paulo, São Paulo, SP, Brazilb National Institute of Metrology (InMetro), Quality and Technology, Head of Bioengineering Program, Xerem, RJ, Brazilc Head of Cell Therapy Center, Unit of Clinical Research, Fluminense Federal University, Niterói, RJ, Brazil

a r t i c l e i n f o

Article history:Received 29 April 2014and in revised form 17 July 2014Available online 23 August 2014

Keywords:Bone developmentBone remodelingExtracellular MatrixMatrix metalloproteinasesTissue inhibitor of matrixmetalloproteinasesRECK, Bone Bioengineering

a b s t r a c t

Bone-forming cells originate from distinct embryological layers, mesoderm (axial and appendicularbones) and ectoderm (precursor of neural crest cells, which mainly form facial bones). These cells willdevelop bones by two principal mechanisms: intramembranous and endochondral ossification. In bothcases, condensation of multipotent mesenchymal cells occurs, at the site of the future bone, which differ-entiate into bone and cartilage-forming cells. During long bone development, an initial cartilaginous tem-plate is formed and replaced by bone in a coordinated and refined program involving chondrocyteproliferation and maturation, vascular invasion, recruitment of adult stem cells and intense remodelingof cartilage and bone matrix. Matrix metalloproteinases (MMPs) are the most important enzymes forcleaving structural components of the extracellular matrix (ECM), as well as other non-ECM moleculesin the ECM space, pericellular perimeter and intracellularly. Thus, the bioactive molecules generatedact on several biological events, such as development, tissue remodeling and homeostasis. Since the dis-covery of collagenase in bone cells, more than half of the MMP members have been detected in bone tis-sues under both physiological and pathological conditions. Pivotal functions of MMPs duringdevelopment and bone regeneration have been revealed by knockout mouse models, such as chondrocyteproliferation and differentiation, osteoclast recruitment and function, bone modeling, coupling of boneresorption and formation (bone remodeling), osteoblast recruitment and survival, angiogenesis, osteo-cyte viability and function (biomechanical properties); as such alterations in MMP function may alterbone quality. In this review, we look at the principal properties of MMPs and their inhibitors (TIMPsand RECK), provide an up-date on their known functions in bone development and remodeling and dis-cuss their potential application to Bone Bioengineering.

� 2014 Elsevier Inc. All rights reserved.

Introduction

The extracellular matrix (ECM)1 is constituted of structural mol-ecules (proteins, proteoglycans, polysaccharides) and enzymessecreted by cells that form a tissue-specific tridimensional macro-molecule network, creating the cellular microenvironment or niches[1,2]. Remodeling of the extracellular and pericellular environment

by proteinases is finely regulated and profoundly affects cellularbehaviors that are essential in many biological processes, such asmaintenance of stem cell properties (stemness) and stem cell fate,embryonic development, morphogenesis, cell migration, cell differ-entiation, apoptosis and tissue remodeling [3]. When the regulationof ECM remodeling is lost, the integrity of tissues is compromised,making the microenvironment propitious for the initiation of

ersity of

e–relatedUs, basicn; MSCs,droblast/cell; Ob,

sificationndrocyte.

http://crossmark.crossref.org/dialog/?doi=10.1016/j.abb.2014.07.034&domain=pdf

http://dx.doi.org/10.1016/j.abb.2014.07.034

mailto:[email protected]

http://dx.doi.org/10.1016/j.abb.2014.07.034

http://www.sciencedirect.com/science/journal/00039861

http://www.elsevier.com/locate/yabbi

K.B.S. Paiva, J.M. Granjeiro / Archives of Biochemistry and Biophysics 561 (2014) 74–87 75

pathological processes, such as connective tissue diseases, cancerand metastasis (tumor microenvironment) [4,5].

Matrix metalloproteinases (MMPs) are the major enzyme familyresponsible for ECM remodeling and, collectively, they are able tocleave all ECM components by sharing common substrates. Tradi-tionally, the biological role of MMPs has been associated only withECM degradation and turnover. Recent studies in MMP gene-knock-out animal models, degradomic and proteomic approaches havebeen important for changing the dogma concerning MMP function,showing overlapping functions (redundant and compensatorymechanisms), doubtful roles (protective or destructive), unexpectedsubstrates and tissue-specific expression, depending on the biolog-ical process involved (proliferation, migration); thus, these enzymesregulate key cell behavior and signaling pathways [6–9]. Thus, theknown specific proteolytic spectrum targets for these MMPs havebeen enlarged to include membrane surface proteins, non-ECMrelated pericellular molecules, other proteinases, intracellular sub-strates, proteinase inhibitors, chemotactic molecules, latent growthfactors, protein-binding growth factors, membrane receptors, cell–cell and cell-ECM adhesion molecules. Accordingly, these enzymesare important for releasing molecules embedded within the ECMor membrane bound and to generate active biomolecules asproducts of ECM component cleavage (matrikins), which are crucialfor physiological and pathological events, as extensively reviewed inthe last decades [10–13].

Bone is a highly dynamic tissue and an important site of contin-uous tissue remodeling during development, homeostasis and tis-sue remodeling/repair. Embryologically, bone develops in twodistinct manners: by intramembranous ossification (direct differ-entiation of mesenchymal cells into osteoblasts) or endochondralossification (mesenchymal stem cells differentiated into chondro-cytes that form a cartilaginous matrix template and are progres-sively replaced by bone matrix). Adult bone is continuallyremodeled throughout life by a physiological process, called boneremodeling or bone metabolism. Bone remodeling occurs via twocoupled mechanisms named bone resorption and bone formation,coordinated by osteoclasts and osteoblasts, respectively. This pro-cess is also initiated after bone lesion or injury, leading to bonerepair or bone regeneration by bone healing (tissue regenerationthat recapitulates bone development without forming fibrousscars). Any imbalance between bone resorption and formationmay lead to metabolic bone diseases, such as osteoporosis. Boneregeneration or bone repair depends on coordinated processesincluding ECM remodeling to form new tissue and to reestablishbone function. In all cases, correct ECM remodeling is requiredand MMPs appear to be essential.

Since the discovery of collagenolytic activity in rat bone cells[14], more than half of the members of the MMP family have beenreported as active during osteogenesis and chondrogenesis in nor-mal bone development in several mammalian species. These MMPsalso contribute to skeletal pathologies involving cartilage and bonedegradation, such as osteoporosis, rheumatoid arthritis, osteoar-thritis, and other diseases [15–20]. Despite the expression of sev-eral MMPs in bone and cartilage cells during normal bonedevelopment, knockout mice models and human genetic diseaseshave been revealed the importance of the MMPs -2, -9, -13, -14,and -16 for skeletal development. It is now also recognized thatthe role of MMPs should be considered in a broader context thanjust for bone and cartilage matrix solubilization and should includeother processes, such as chondrocyte proliferation and differentia-tion, osteoclast recruitment and function, bone modeling, couplingof bone resorption and formation (bone remodeling), osteoblastrecruitment and survival, angiogenesis, and osteocyte viabilityand function (biomechanical properties).

In this review, we will look at the overall properties of MMPsand their inhibitors (TIMPs and RECK), provide an up-date on their

functions in bone development and remodeling and discuss theirpotential applications in Bone Bioengineering.

Overview of the general properties of matrixmetalloproteinases (MMPs)

The first MMP (collagenase) was identified in tadpole tails duringmetamorphosis in 1962 by Gross and Lapière [21]. Most MMPs havebeen studied in vertebrates (25 members), but are also found inlower animals and plants. In humans, the MMPs comprise 24 genes,encoding 23 proteins, as one MMP (MMP-23) is coded by two iden-tical genes at chromosome 1 (MMP-23A and MMP-23B). Mamma-lian MMPs are classified according to; (I) their localization, soluble(secreted into ECM) or insoluble (anchored on cell membrane); (II)their similarities in tridimensional structure and substrate affinity,being usually divided into six subgroups: collagenases (MMPs -1,-8, and -13), gelatinases (MMPs -2 and -9), stromelysins (MMPs -3,-10 and -11), matrilysins (MMPs -7 and -26), membrane-typemetalloproteinase (MMPs -14, -15, -16, -17, and -24 or MMP-MT1,MT2, MT3, MT4, MT5 and MT6, respectively), and others (MMPs -12, -18, -19, -20, -21, -22, -23, -27, and -28); and (III) numericallylisted according to chronological discovery. MMPs -4, -5 and -6 aremissing in the list since they were shown to be identical to othermembers of the family [22] (Fig. 1).

The regulation of MMPs occurs at multiple levels, according tothe cell type involved, in a temporal and spatial manner and quan-tities, by intra and extracellular mechanisms. Inductive or suppres-sive signaling from the ECM (cytokines, growth factors, EMMPRIN,signals from integrins, ECM proteins, cellular stress, morphologicalchanges, etc) and intracellular signal transduction induce the acti-vation or repression of the MMP genes. In the nucleus, these genesmay be transcriptionally controlled by genetic alterations (poly-morphisms or mutations, particularly in promoter regions of theMMP genes) and by epigenetic control (DNA methylation statusand remodeling of chromatin by histone acetylation) as well aspost-transcriptionally through mRNA processing. In the cytoplasm,MMPs may be regulated post-transcriptionally by mRNA stability(microRNAs action and degradation pathway), intracellular activa-tion of furin susceptible MMPs, insertion of prosthetic groups (Nand O-glycosylation and GPI-anchor) or specific domains in thepro-MMP structure, and, finally, by inducible and constitutivepro-enzyme secretion into the ECM. Certain MMPs may be storedin the cytoplasm within granules in specific cell types prior tostimuli, such as inflammatory stimuli, and then secreted. MMPson the cellular membrane may be regulated by their localizationon specialized membrane microdomains (lipid rafts or caveolae),by endocytosis/recycling (clathrin or caveolin-dependent) andintracellular degradation. In the ECM, the MMPs may be controlledby proteolytic processing and inactivation, proteolytic activation ofpro-MMPs, binding of pro and active forms to inhibitors, and inter-action with specific ECM components, leading to specific localiza-tion (pericellular perimeter or distant from cell secretion pointwithin ECM), and allosteric control [23–33].

The balance between MMPs and their inhibitors is required forphysiological ECM remodeling and imbalance in these enzymesleads to pathological states. In the tissues, MMPs are mainlyreversibly inhibited in the ECM by their physiologic tissue inhibi-tors (TIMPs), while the cell surface MMPs are inhibited by the RECKglycoprotein [34] (Fig. 1).

Overview of the principal cellular and molecular aspects ofbone development and remodeling

The mammalian skeleton has three different embryological ori-gins: (I) paraxial mesoderm, which gives rise to the axial skeleton;

Fig. 1. Schematic representation of general properties of MMPs and their inhibitors (TIMPs and RECK). (A) MMP members are classified into five types of structure accordingto their conserved domains (active form): minimal (catalytic domain), basic (catalytic domain + hinge domain + hemopexin domain), additional domain within catalyticdomain (basic + fibronectin type II domain), insertion in cell membrane by C-terminal transmembrane type I (MT-loop/special catalytic domain for only MT-MMPs + hingedomain + hemopexin domain + transmembrane domain + cytoplasmic tail) [164] or anchored on cell membrane by glycosylphosphatidylinositol/GPI anchor (only on externalface) (basic + GPI domain) and insertion in cell membrane by N-terminal transmembrane type II (basic + cysteine array domain + Ig-like domain – replacing the hinge andhemopexin domains) [165]. (B) Secreted/soluble pro-MMPs are generally activated in the ECM space by specific proteolytic cleavage of the zinc-thiol interaction betweencysteine on the pro-domain and Zn+2 on the catalytic domain by serine proteases or active MMPs, denominated as the cysteine-switch mechanism [166,167], except for pro-MMP-2. MMP-2 requires a more sophisticated activation mechanism on the cell surface by tertiary complex formation: pro-MMP-2/TIMP-2/MMP-14 (stepwise activationmechanism) [168]. Initially, a cell surface MMP-14 binds to the TIMP-2 (N-terminal), thus its inhibition, and the free C-terminal domain of TIMP-2 acts as a receptor for thehemopexin domain of pro-MMP-2 (72 kDa). An adjacent second active MMP-14 cleaves and partially activates pro-MMP-2 (MMP-2 intermediary – 68 kDa), a residual portionof the MMP-2 pro-domain is removed by the autocatalytic process or another active MMP-2 molecule to yield a fully active, mature form of MMP-2 (66 kDa) [169]. Virtually,all MT-MMPs may form a ternary complex to activate pro-MMP-2 [170–174], except MMP-17 [175]. Low concentrations of TIMP-2 stimulate the activation of pro-MMP-2,whereas high concentrations of TIMP-2 inhibit activation of pro-MMP-2 [176,177]. MT-MMPs and some secreted MMPs contain a motif between the pro and catalyticdomains, which serves as a target sequence for proprotein convertases or furins present in the trans-Golgi network, and, thus, are activated intracellularly and secreted intoECM (MMPs -11, -21 and MMP-28) [178] or anchored on the cell surface (MT-MMPs and MMP-23) [179]. (C) MMPs are translated as multidomain enzymes: (I) signalsequence, for targeting to secretory vesicles and it is removal at the rugous endoplasmic reticulum and (II) pro-domain, which regulates latency of the proteolytic activity. (D)TIMPs are tissue MMP inhibitors and are secreted into the ECM space (except TIMP-3, which is found at the cell surface associated with membrane proteins or with ECMmolecules) [180] and can inhibit all active MMPs (direct interaction, non-covalent and reversible, between the N-terminal of TIMP and zinc at the catalytic site of the enzyme,resulting in the high-affinity complex at a 1:1 stoichiometric ratio). RECK is a unique cell-associated MMP inhibitor and can specifically inhibit MMP-9 [34], -2 and -14 [127]at various levels, such as at the transcriptional level, inhibition of pro-enzyme secretion into ECM, conversion of the pro to the active form, catalytic activity, sequestration ofMMPs at the cell surface and internalization [34,181–185]. (E) Once activated, both anchored and secreted MMPs can cleave their specific ECM substrates and generatebioactive molecules (matrikins), modifying the microenvironment.

76 K.B.S. Paiva, J.M. Granjeiro / Archives of Biochemistry and Biophysics 561 (2014) 74–87

(II) the lateral plate mesoderm, which originates the appendicularskeleton; and (III) ectodermal neural crest, which forms the facialskeleton. Although bones are formed by osteoblasts and secrete aspecific bone matrix, their different ontogeny may reflect distinctsignaling and functions [35,36]. Osteogenesis occurs by two mainmechanisms: Intramembranous ossification, signifying the directdifferentiation of multipotent mesenchymal cells into osteoblasts(which form bones of the skull, parts of the craniofacial skeleton,part of clavicles, sesamoid bones, and the periosteal bone that

forms around long bones to promote lateral growth) (Fig. 2) andendochondral ossification, signifying the differentiation of multi-potent mesenchymal cells into the chondrocyte and posteriorreplacement by osteoblasts (from which axial and appendicularskeleton are generated) (Fig. 3), which both depend on extensiveECM remodeling.

In both cases, condensation of multipotent mesenchymal cellsoccurs at the site of the future bone piece and cells are committedto differentiate into osteochondral progenitor cells. Molecular

Fig. 2. Schematic representation of intramembranous ossification and bone phenotype developed by MMP knockout mice. (A-B) Migrating neural crest cells (namelyectomesenchymal cells, which exhibit mesenchymal features) undergo mesenchymal condensation at the site of the future bone. (C) Cells differentiate into osteochondralprogenitor cells, which become pre-osteoblasts, and, finally, osteoblasts [36]. (D) Osteoblasts secrete non-mineralized bone matrix (osteoid), mineralization occurs, and thesecells are embedded the matrix; final osteoblast maturation then occurs (osteocyte), leading to the formation of an osteocyte lacunocanalicular network (osteocyte bodieswithin lacunae and dendritic process within canaliculae). This organization allows osteocytes to communicate with each other and osteoblasts/osteoclasts at bone surfaces, inboth the periosteum and endosteum, by gap junction-coupled cell processes within canaliculi. (E–F) Bone phenotype for MMP-2 [57,61], MMP-14 [87–89] and MMP-14/MMP-16 [109] knockout mice, respectively.


differences exist during differentiation that drive cell fate, originat-ing bone-forming cells (osteoblasts) or cartilage-forming cells(chondroblasts) [37]. Moreover, metabolic and morphogenetic dif-ferences are noted between these two modes of osteogenesis [38],as well as different osteoblast fates and functions [39] (Fig. 4).Although endochondral ossification is dependent on the time–space coordinated actions of different cell types and systemic cir-culating factors (hormones), it is the chondrocyte that conductsthe process. The chondrocytes regulate cartilage to bone transi-tions, representing the main mechanism for normal bone longitu-dinal growth. Accumulating evidence indicates that the growthplate chondrocyte orchestrates the invasion by the ossificationfront not only via the preparation of the cartilage matrix, but also

by secreting soluble molecules that regulate the behavior of theinvading cells. In addition, chondrocytes regulate their prolifera-tion, differentiation and death by several signaling pathways,including via the Indian hedgehog (Ihh), parathyroid hormone(PTH), parathyroid hormone–related protein (PTHrP) and theirreceptors fibroblast growth factor (FGF) FGF18 and its receptorFGFR3, vascular endothelial growth factor (VEGF) and connectivetissue growth factor (CTGF), bone morphogenetic proteins (BMPs),and Wnt proteins [40–43].

During ossification and skeletal growth, two distinct mecha-nisms, coordinated by bone cells, maintain the bone tissue: model-ing (uncoupled) and remodeling (coupled). Bone modeling isrequired for generation and conservation of the shape and size of

Fig. 3. Schematic representation of endochondral ossification and bone phenotype developed by MMP knockout mice. (A) Multipotent mesenchymal cells condense at thesite of future bone formation, assuming the format of the bone to be formed. (B) Cells in the core then differentiate into chondrocytes, which express characteristiccartilaginous matrix genes while thin layers of cells at the periphery differentiate into fibroblast-like perichondrial cells. (C) The chondrocytes proliferate rapidly, formingcartilage anlage. The cells at the center of the core stop proliferating and enlarge to become prehypertrophic, further differentiate into mature hypertrophic chondrocytes andmineralize their surrounding matrix, and undergo programmed cell death. At the same time, perichondrial cells adjacent to the mineralized hypertrophic chondrocytes turninto osteoblasts and form the bone collar (precursor of periosteum); this process is denominated perichondral ossification, characterized by the differentiation of theperichondrium into periosteum [186]. (D) The mineralized cartilaginous matrix is invaded by blood vessels (at the last transverse septa – the thin layer of cartilage matrixthat separates the hypertrophic chondrocytes from invading capillaries) [187], chondroclast/osteoclasts (cartilage- and bone-resorbing cells), as well as bone marrow andosteoblast precursors (hematopoietic and mesenchymal stem cells, respectively), which then invade the matrix from the bone collar and proceed to form the primary centerof ossification (ossification front or chondro-osseous junction). (E) The primary center expands towards the ends of the cartilage model (constituting the two epiphyses), thechondroclasts/osteoclasts remove the cartilage matrix and the osteoblasts deposit woven bone on the cartilage remnants at the chondro-osseous junction (where the growthplate forms at the location of deposition of new trabecular or cancellous bone; this area also named the metaphysis) and bone marrow cavity is formed to produce blood cells(where the diaphysis constitutes the site between the two epiphyses) [40,133,188–190]. The growth plate is divided into zones, according to the stage of chondrocytedifferentiation and molecules secreted into the cartilage matrix: (I) resting or germinal, where disperse undifferentiated chondrocyte progenitors are found; (II) proliferating,in which chondrocytes became flattened, highly proliferative and demonstrate secretory activities, and are disposed at longitudinal columns; and (III) hypertrophic, wherechondrocytes undergo final maturation, with increased cell volume, decreasing secretory activity, driving mineralization and programmed cell death. (F) The secondaryossification centers form at each epiphysis via new invasion of blood vessels after birth, decreasing the cartilaginous growth plate between the primary and secondaryossification centers, as well as forming permanent articular cartilages at each end of the bone. The growth plate is responsible for longitudinal growth of long bones(trabecular bone). On the other hand, lateral growth occurs by appositional mineralization of osteoid, secreted by osteoblasts at the periosteum (cortical bone), resemblingintramembranous ossification. (G) Skeletal maturity occurs when the expanding primary center of ossification meets the secondary centers of ossification, thus obliteratingthe growth plate (epiphyseal line) and growth ceases in early adulthood, when all residual cartilage is replaced by bone, except articular cartilage within synovial joints. (H–M) Bone phenotype for MMP-2 [57,61], MMP-9 [71], MMP-13 [79,80], MMP-9/MMP-13 [85], MMP-14 [87–89], MMP-14/MMP-16 knockout mice [109], respectively. AC:articular cartilage, BC: bone collar, BMC: bone marrow cavity, BS: bone spicules, BV: blood vessel, CB: cortical bone, Ch/C: chondroblast/chondrocyte, D: diaphysis, E:epiphysis, EL: epiphyseal line, Eo: endosteum, GP: growth plate, HZ: hypertrophic zone, M: metaphysis, MSC: mesenchymal stem cell, Ob: osteoblast, Oc: osteoclast, Oo:osteocyte, PC: proliferative chondrocyte, Pc: perichondrium, PHC: prehypertrophic chondrocyte, Po: periosteum, POC (COJ): primary ossification centres (chondro-osseousjunction), PZ: proliferative zone, RZ: resting zone, TB: trabecular bone, SOC: secondary ossification centres, THC: terminal hypertrophic chondrocyte.


bones to enable normal and consistent growth, as well as adapta-tion to withstand mechanical stress. This event is achieved via theaction of both osteoblasts and osteoclast in different bone sites andat different times, in a non-cooperative manner; i.e., bone deposi-tion does not depend on prior bone resorption (starting at theonset of bone formation and terminating upon the cessation of lon-gitudinal growth in long bones). On the other hand, bone remodel-ing is essential for calcium homeostasis, replacement of old bone

and repair of damaged bone. Remodeling is mediated by osteoclastrecruitment, differentiation and bone matrix resorption, followedby the recruitment and new bone deposition of osteoblasts at spe-cific bone sites (basic multicellular units – BMUs) in a time-quan-tity balanced manner, and finally maturation of osteoblasts,embedded in mineralized ECM, into osteocytes. This imbalanceleads to osteoporosis (bone resorption > bone formation) and oste-opetrosis (bone resorption < bone formation) [44,45]. In both cases,

Fig. 4. Schematic representation of cytodifferentiation and main molecular markers expressed during commitment of multipotent mesenchymal cells into osteoblastic andchondrogenic lineages.


osteocytes participate actively in forming and resorbing bone atthe lacunae level via a process denominated osteocytic osteolysisor osteocytic perilacunar/pericanalicular remodeling [46,47]. Theosteocyte is a mechanosensitive cell and coordinates the biome-chanical regulation of the bone mass and architecture, accordingto the external mechanical load, by transmitting a signal to effectorcells (osteoblasts and osteoclasts) [48,49]. The osteocyte networkformed in bones arises from intramembranous and endochondralossification and displays differences in microarchitecture thatreflect the mechano-transduction of physical loads to the bone[50]; several transduction pathways are involved in mediatingextracellular mechanical stimuli and intracellular signaling [51].

Roles of MMPs and their inhibitors during bone development,modeling and remodeling

Mice models for MMP gene depletion are a crucial tool for iden-tifying MMP function in vivo. To date, knockout mice for 16 differ-ent MMPs, as well as some double depletion mice have beengenerated and evaluated. Most of these mice survive at least afew weeks after birth and demonstrate distinct phenotypes, pre-senting alterations in the normal development of various organs,the development of various types of pathologies (including cancer)and even protection of tissues from damage. Studies on depletedMMP genes in mice indicate roles for MMPs in proteolytic redun-dancy, enzymatic compensation and adaptative development[52–54].

Despite the knowledge of the expression of several MMPs inbone and cartilage cells during normal bone development, knock-out mice models and human genetic diseases have demonstrateda prominent importance of MMPs -2, -9, -13, -14, and -16 in skel-etal development (Figs. 2 and 3), acting on different cells and bio-logical processes, as summarized in Table 1.

MMP-2 contributes to both intramembranous and endochon-dral ossification and bone metabolism. Itoh et al. [55] generated

the first MMP-2 knockout mice, however no overt phenotypewas described, except for a smaller body size at birth (about15%), compared to wild-type mice. In the early 2000s, a geneticmutation of MMP-2 was found in the multicentric osteolyses orvanishing bone syndromes (MOA, Winchester and Torg syn-dromes) [56–60]. These syndromes are a group of hereditary auto-somal dominant and recessive skeletal disorders characterized bymarked and progressive bone loss and joint destruction, resultingin skeletal deformities and significant functional impairment. Withthis new information, the previously generated knockout micewere re-analyzed. A similar skeletal phenotype was observed inMMP-2 null mice to those seen in the described syndromes, includ-ing craniofacial abnormalities such as short and broad snouts,hypertelorism and taller skulls (observed from 10 days after birth,and often persisting throughout life) [57,61] and articular cartilagedestruction (similar to the rheumatoid pannus present in rheuma-toid arthritis) and progressive loss of bone mineral density (from5 weeks of age, similar to osteopenia) [57]. Defective bone remod-eling described in these animals has been suggested to reduce thenumber of osteoblasts and osteoclasts, in vivo [57], as well as todisruption the osteocyte canicular network [61]. Similarly, Col1a1r/r

homozygous mice (collagenase-resistant Col1a1tm1 Jae) [62] presentabnormalities in the craniofacial bones and, importantly, demon-strate an empty osteocyte canalicular network. Double-deficientmice for MMP-2/Col1a1 have a phenotype that demonstrates moresimilarity to syndromes associated with the MMP-2 mutation thanthe single MMP-2 null mice (severely runted and osteopenic bones,resembling osteolysis). Furthermore, MMP-2 is able to cleave colla-gen type I in these animals (MMP-8, MMP-13 and MMP-14 do notshow collagenolytic activity) even if they are resistant to collage-nase degradation [63]. As such, MMP-2 may represent the principalcollagenase in vivo and its biochemical interaction with collagentype I is important during bone formation.

Cells harvested from both bone marrow and calvaria of MMP-2-null mice were cultured in vitro and demonstrated a reducedproliferation ratio, as well as deficient osteoblastic differentiation

Table 1General skeletal phenotypes described for MMP null mice during bone development and remodeling.

MMPknockoutmice

Main features Modulation of other MMPs References

MMP-2 - Craniofacial abnormalities: short and broad snouts, hypertelorism and taller skulls- Articular cartilage destruction (similar to the rheumatoid pannus present in- rheumatoid arthritis)- Progressive loss of bone mineral density (from 5 weeks of age, similar to osteopenia)- Loss of osteocyte canalicular network architecture- Long bones: shorter, reduced number and proliferation ratio of osteoblasts and osteoclasts

(about 50%) on trabecular area, trabecular bone (metaphysis) with fewer osteocytes andlower density, less cortical bone (diaphysis)

- Reduced potential of osteoblasts and osteoclasts for expansion in culture (cells harvestfrom both bone marrow and calvaria)

- Does not affect either MMP-9and MMP-14 expression

[57,61]

MMP-9 - Long bones: shorter, lengthening of hypertrophic zone due to impaired vascular and osteo-clast/chondroclast recruitment, delayed death of terminal hypertrophic chondrocytes, lessprimary ossification and increased chondro-osseous junction at growth plate

- Increased expression of MMP-13by hypertrophic chondrocytes inchondro-osseous junctionat growth plate

[71]

MMP-13 - Long bones: enlargement of hypertrophic zone by accumulation of terminal hypertrophicchondrocytes (normal chondrocyte proliferation and differentiation) and increased trabec-ular bone formation in chondro-osseous junction at growth plate

[79,80]

MMP-9/MMP-13

- Dramatically shortened skeletal elements and exacerbated phenotypes compared to sin-gle-depleted mice

- Long bones: expanded growth plates, a disorganized architecture of the hypertrophic zone,increased number of terminally differentiated hypertrophic cells, delayed vascular recruit-ment, reduction of trabecular bone formation at chondro-osseous junction (in contrast toMMP-13 and similarly to MMP-9 single depletion), delayed bone marrow cavity formation,and delayed cartilage-bone transition at secondary ossification centers (in contrast toMMP-9 and MMP-13 single depletion mice)

[86]

MMP-14 - Most severe skeletal phenotype among MMP knockout mice- Craniofacial abnormalities: short snout, orbital protrusions, hypertelorism and unclosed

cranial sutures- Persistent ‘‘ghost cartilage’’ and fibrosis at the sites of calvarial bones due to incomplete

cartilage remodeling- Osteopenia (excessive osteoclastic resorption close to the bone-periskeletal soft tissue

interfaces)- Arthritis in all joints (ankylosis)- Dwarfism- Generalized fibrosis of soft tissue (inadequate collagen turnover)- Long bones: impaired secondary ossification (epiphysis), disorganized growth plate with a

reduced number of proliferative chondrocytes (normal differentiation and maturation)and fibrosis in the periosteum

- Increased osteocyte mechanosensitivity- Change mesenchymal stem cell fate, commitment with adipogenesis and chondrogenesis

instead osteogenesis

- Deficient activation of pro-MMP-2,but not of pro-MMP-9

- Decreased MMP-16 transcripts

[88–91,104,105]

MMP-14/MMP-2

- Die at birth- Bone phenotype is more similar to that of the MMP-14-null mice than MMP-2-null mice

[93]

MMP-16 - Inhibition of growth associated with decreased viability of mesenchymal cells in skeletaltissues

- Decreased MMP-14 mRNA [110]

MMP-14/MMP-16

- Die at birth- More severe phenotype compared to MMP single depletion- Craniofacial abnormalities: severe dysfunction in palatal shelf formation leading to cleft

palate, thinner cranial vault bones and deficient development of parietal, frontal and nasalbones

- Long bones: shorter, thin cortical bone, altered growth plate due to decreased proliferationof chondrocytes, extended hypertrophic zone and significantly impaired bone formation atprimary ossification centre and abnormal hypertrophic chondrocyte islands at the middleof the bone marrow

[110]


[57]. Furthermore, both human (SaOS2) and mouse (MC3T3) osteo-blast cell lineages lacking MMP-2 expression (knock-down) alsodemonstrated dramatic reductions in cell proliferation rate, butno effect on MMP-9 and MMP-14 gene expression [57]. Morerecently, the same MMP-2(�/�) bone marrow cells were inducedto osteoblast differentiation, in vitro, for 21 days and gene expres-sion levels of molecules related to bone and MMP-2 were evalu-ated (Cbfa1, osteopontin, bone sialoprotein, collagen type 1a,parathyroid hormone receptor – PTHR, MMP- 9, TIMP-2 andLRP5). Only osteopontin and bone sialoprotein were found to beupregulated, where maximum expressions were observed at 7and 14 days post-induction, respectively. Indeed, siRNA-mediated

downregulation of MMP-2 in SaOS2 cells also demonstrated oste-opontin and sialoprotein upregulation [64]. Both osteopontin andbone sialoprotein were reported to upregulate transcription, acti-vation and functional localization of MMP-2 and osteopontin hasbeen identified as a substrate for MMP-2 [65]. Taken together,these authors suggest that the MMP-2-osteopontin-bone sialopro-tein circuit may be an important regulator of bone homeostasis.

Double-knockout adult mice for MMP-2 and uPARAP/Endo180(endocytic collagen receptor of collagen and collagen fragmentsfor degradation in the lysosomes) revealed shorter long bones,reduced bone mineral density and inferior trabecular bone quality,when compared to single depletion mice [66]. Double knockout


mice for MMP-14 and uPARAP/Endo180 have also been generated;however these animals die soon after birth due to severelyimpaired bone development [67]. In wild type animals, the expres-sion of uPARAP/Endo180 is detected in bone-forming cells in thechondro-osseous junction, in resting and proliferative chondro-cytes in the growth plate, in bone-lining cells in the periosteumand endosteum and in osteocytes in the cortical bone, suggestinga function for this enzyme in collagen turnover in both intramem-branous and endochondral ossification. In vitro assays demon-strated that uPARAP/Endo180 has a greater affinity for collagenfragments than the native molecule [68] and that these fragmentscan also be subsequently cleaved by MMP-2 [69]. Furthermore,downregulation of uPARAP/Endo180 by siRNA increases the enzy-matic activity of MMP-2 [70], suggesting a functional overlap andcoordinated and subsequent events for collagen degradation asuPARAP/Endo180 and MMP-2 act by different mechanisms: intra-cellular and pericellular, respectively.

MMP-9-null mice only present altered bone development dur-ing endochondral ossification [71]. Alterations have been observedafter birth in cartilage-bone replacement, where impaired terminalchondrocyte differentiation affects ossification and vascularizationmechanisms, and alters trabecular bone formation. This phenotypeis due to the delayed death of hypertrophic chondrocytes or alteredossification (reduced trabecular bone and osteoblast recruitment atthe chondro-osseous junction) in the primary ossification center,leading to an expanded hypertrophic zone, and reduced vasculari-zation in the growth plates. Moreover, aberrant death of hypertro-phic chondrocytes and surrounding ossification has been observedin the middle of the long bone (about 3 weeks), leading a transitoryosteopetrotic phenotype. This aberrant skeletal development is,however, compensated for 4 weeks postnatally so that the animalsultimately have a normal skeletal phenotype. Further analysis ofMMP-9-null mice indicated a high expression of MMP-13 in thehypertrophic chondrocytes close to the chondro-osseous junctionof the growth plate, suggesting that this metalloproteinase maytry to compensate for the absence of MMP-9 [72].

This MMP-9 phenotype may be caused by two different mech-anisms: first, an inefficient cartilage matrix proteolysis due to theabsence of MMP-9, resulting in a limited bioavailability of VEGFfrom ECM and consequences on osteoclast and endothelial cellrecruitment into the mineralized cartilage [73]; secondly, thegrowth plates of MMP-9-null mice are enriched by galectin-3 (lec-tin with anti-apoptotic activity that is cleaved by MMP-9 tobecome inactivated in the ECM), while knockout mice for galec-tin-3 present the opposite phenotype to MMP-9-deficient mice(premature death of hypertrophic chondrocytes), suggesting thatcleavage of galectin-3 by MMP-9 may control terminal differentia-tion of hypertropic chondrocytes [74].

MMP-13 has been found expressed exclusively in bone forma-tion in vivo [75–78] and mice depleted for MMP-13 demonstratealterations only in endochondral ossification, bone modeling andremodeling [79,80]. MMP-13(�/�) mice show progressive abnormalskeletal development, in relation to growth plate expansion (butnormal proliferation and death ratios in chondrocytes) during thefirst 4-months of life, marked VEGF, collagen type X and osteopon-tin expression by hypertrophic chondrocytes, increased trabecularbone mass but irregular bone spicules (affecting both bone model-ing and remodeling). Subsequently, the hypertrophic zone of thegrowth plate returns to a normal size as well as bone mass. No dif-ferences have been observed in the cortical bone of long bones orangiogenesis [79,80]. MMP-13 is a downstream target of parathy-roid hormone (PTH)-related protein (PTHrP) [81] and the transcrip-tion factor Cbfa1/Runx2 in hypertrophic chondrocytes [82,83], andthis is confirmed by the report that Cbfa1(�/�) fails to express MMP-13 during fetal development [79]. The absence of another tran-scriptional factor, c-maf, resulted in the prolongation of the

chondrocyte hypertrophic state and marked suppression ofMMP-13 activity, suggesting that c-maf was, in part, responsiblefor initiating terminal differentiation [84].

Double-knockout mice for MMP-9(�/�)/MMP-13(�/�) presentdramatically shortened skeletal elements and exacerbated pheno-types compared to single-depleted mice, as well as expandedgrowth plates, a disorganized architecture of the hypertrophicchondrocyte zone, increased number of terminally-differentiatedhypertrophic cells, reduction of trabecular bone formation at chon-dro-osseous junction at the growth plate (in contrast to MMP-13and similarly to MMP-9 single depletion), delayed bone marrowcavity formation, and delayed cartilage-bone transition at second-ary ossification centers (in contrast to MMP-9 and MMP-13 singledepletion mice). Recent studies have reported roles for both MMP-9 and -13 during chondrocyte canal formation for guiding invadingblood vessels in secondary ossification centers [85]. Takentogether, these data indicate a strong synergy between MMP-13and MMP-9 in last transverse septa degradation and both primaryand secondary ossification centers during endochondral ossifica-tion by cleavage of aggrecan and by initial cleavage of both colla-gens type I and II by MMP-13 followed by the clearance of thesefragments by MMP-9, as well as a cooperation in trabecular boneformation. Furthermore, these results suggest that MMP-13 is themajor collagenase in collagen type II proteolysis and that thisenzyme is necessary to allow normal blood vessel invasion at thetransverse septa. As such, specific cleavage of ECM substrates byMMP-9 and MMP-13 may be crucial for adequate bone formationdue to endochondral ossification [80].

An autosomal dominant disorder, spondyloepimetaphyseal dys-plasia-Missouri type, is caused by a missense mutation in theMMP-13 gene [86]. As such, the observation that MMP-13-nullmice present impaired endochondral ossification was expected[79,80], in contrast to observations in MMP-9-deficient mice. Inter-estingly, these abnormalities, in both affected patients and knock-out mice, resolve themselves by adolescence or by 12 weeks of age,respectively. They appear to be caused by the late exit of chondro-cytes from the growth plate, despite normal differentiation.

MMP-14-null mice present both intramembranous and endo-chondral ossification, representing the most severe skeletal defor-mities among MMP-deficient animals, and play a pivotal role in cellfate and remodeling of the connective tissue associated with theskeleton [87–89]. Newborn mutant mice are viable and indistin-guishable from wild-type mice at birth; however, disturbances ingrowth are seen within days (smaller body size and weight andbone abnormalities) and significant mortality before weaning(about 33%). Surviving knockout mice develop progressive cranio-facial dysmorphism (short snout, orbital protrusions, hypertelor-ism, and unclosed cranial sutures), osteopenia, arthritis in alljoints (resulting in ankylosis), dwarfism and generalized fibrosisof soft tissue, due inadequate collagen type I and II turnover; mostof these animals die at around 50–90-days of age. During endo-chondral ossification, secondary ossification in epiphyses wasimpaired, while growth plates became thick and disorganized witha reduced number of proliferative chondrocytes, without affectingchondrocyte differentiation and maturation. Osteopenia was char-acterized by excessive osteoclastic resorption close to the bone-periskeletal soft tissue interfaces and fibrosis in the periosteumof long bones [87,89]. During cranial development, the transienthyaline cartilage (denominated calvarial cartilages) does notundergo classical endochondral ossification nor require osteoclastactivity for cartilage resorption, rather they are remodeled and dis-appear before the onset of intramembranous ossification. In theseanimals, cranial abnormalities are attributed to incomplete carti-lage remodeling and persistent ‘‘ghost cartilage’’ at the sites of cal-varial bones [87]. Therefore, MMP-14 may operate differentlyduring the remodeling of cartilage of long bones and cranial vault.


Additionally, the ossification of Meckelss cartilage also providesfurther evidence for such a mechanism. This cartilage forms themandible by both intramembranous and endochondral ossifica-tion. The rostral portion undergoes a classical endochondral pro-gram and the posterior portion undergo two different fates:forms fibrous tissue, called sphenomandibular ligament, and formsbone (mandible and malleus), due to a remodeling mechanism thathas not yet been defined [90]. In MMP-14-null mice, the rostralportion of Meckelss cartilage is replaced by bone via endochondralossification, but with persistent ‘‘ghost cartilage’’ on the posteriorportion. Accordingly, apoptosis of these nonhypertrophic chondro-cytes has been associated with the participation of MMP-14 duringcartilage matrix degradation in wild-type mice [88]. Explants fromMMP-14-null mice also present a deficient activation of pro-MMP-2, but not of pro-MMP-9, suggesting that MMP-14 is essential forits activation in vivo [89]; however altered MMP-2 activity mayalso be due to other MT-MMP activators [89,91]. Double-knockoutmice for MMP-2 and MMP-14 are not viable and die at birth [92].These animals develop a bone phenotype that is more similar tothat of MMP-14-null mice than MMP-2-null mice indicating func-tional overlap.

Recently, calvarial development was further evaluated in MMP-14-knockout mice and a novel pathway involving MMP-14/ADAM-9/FGFR2 was proposed. Fibroblast growth factor receptors (FGFRs)are pivotal for normal intramembranous ossification in cranialbones and FGFR2 gene depletion results in several craniofacialabnormalities [93]. FGFR2 can be shed by metalloproteinases, suchas ADAM-9 [94]. ADAM-9 is upregulated in MMP-14(�/�), leading tothe accumulation of the cleaved form of FGFR2. These authors sug-gested that impaired FGF signaling in MMP-14-null mice may alsocontribute to the severe bone phenotype developed by MMP-14-knockout mice.

Both MMP-14 and MMP-13 have been detected in pericellularosteocyte processes and are co-localized with typical fragmentsproduced by the cleavage of collagen type I. In MMP-14 adult nullmice (30–40-days old), osteocytes are not able to cleave collagentype I, contributing a loss in osteocyte processes; however collag-enolytic activity has been detected on the periosteal surfaces ofcortical bone, and is independent of MMP-14 As such, both colla-genases (MMP-14 and MMP-13) present an overlap in collagentype I turnover and other MMPs may be also candidates for carry-ing out collagenolytic activity [95]. In this context, MMP-14 func-tions as a sheddase for releasing CD44 from the cell membrane[96]. CD44-knockout mice display short long bones, thin corticalbone, reduced bone marrow cavity area and decreased boneresorption by osteoclasts (mild osteopetrotic phenotype). In thesecells, there was an increased secretion of MMP-14 and reducedactivation of pro-MMP-9 [97]. In addition, CD44 appears to act asa ‘‘receptor’’ for MMP-14 and may activate pro-MMP-9 at the cellmembrane [98]. CD44 is also known to be important for adhesion,invasion and cell–cell signaling in osteocytes [99]. Focal adhesionshave been considered to be candidate structures involved in signaltransduction by converting extracellular mechanical signals intobiological responses in osteocytes [100,101]. Considering the pecu-liar structure of MMP-14 among MMPs (it has an extracellulardomain and an intracellular domain), this enzyme is a candidatefor transducing signals from ECM into the cell and it is also knownthat the cytoplasmic domain interacts with molecules of the trans-duction cascade [102]. This hypothesis was evaluated in osteocytesfrom MMP-14-knockout mice and osteocyte cell lineages inresponse to mechanical loading. The absence of MMP-14 increasedosteocyte mechanosensitivity after mechanical loading due toincreased nitric oxide production, upregulation of c-jun and c-fosgene expression and larger focal adhesions, suggesting a novelfunction for MMP-14 in the cellular mechanosensing in osteocytes[103].

Szabova et al. [104] showed that re-expression of MMP-14, inskeletal cells that were depleted for MMP-14 and express collagentype II, is able to ameliorate bone phenotype by the partial rescueof chondrocyte proliferation and increases in both intramembra-nous and endochondral ossification. Furthermore, collagen type IIwas detected in bone and cartilage-forming cells and MMP-14was expressed in these cells and also in bone marrow stromal pre-cursor cells (skeletal stem cells or mesenchymal stem cells –MSCs). The authors suggested that precursor cells may definewhich molecule and its correspondent proteolytic enzyme issecreted into the ECM before cell commitment. As such, a recentstudy has demonstrated that depletion of MMP-14 in MSCschanges the cell fate, where instead of undergoing osteogenesis,these cells are committed to adipogenesis and chondrogenesis.Furthermore, MMP-14 regulates cell morphology by the b1-inte-grin/RhoGTPase pathway and affects the transcriptional regulator(Yes-association protein – YAP) and transcriptional coactivatorwith PDZ-binding motif (TAZ), leading to the control of MSC line-age commitment. The authors suggest that the MMP-14/integrin/YAP/TAZ axis is crucial for defining MSC commitment duringdevelopment [105].

Recently, a study showed the importance of the correct localiza-tion and attachment of MMP-14 at the cell membrane by palmitoy-lation (the only reversible post-translational lipid modificationinvolved in protein traffic, protein stabilization, protein–proteininteractions and signal transduction control). A family of proteinswith palmitoyl acyltransferase (PAT) activity is responsible foradding 16-carbon palmitate to proteins. ZDHHC13-knockout mice,for example, develop osteoporosis (reduced bone mass) due todelayed secondary ossification centers and a disorganized growthplate after birth [106]. In addition, MMP-14 was identified as adirect substrate of ZDHHC13 in these animals and reductions inMMP-14 palmitoylation lead to its alternative subcellular distribu-tion, as well as decreased expressions of VEGF and osteocalcin inboth chondrocytes and osteoblasts. This is the first evidence thatpost-translational control of MMP-14 may affect bone metabolism[107]. Moreover, an MMP-14 mutation also been found in a multi-centric osteolyses or vanishing bone syndromes (MOA, Winchesterand Torg syndromes) and arthritis, previously described for MMP-2mutations, and impairments of pro-MMP-2 activation [108],strengthening an important physiological MMP-2/MMP-14 axisin vivo.

MMP-16-knockout mice display altered bone development dueto a decreased viability of mesenchymal cells in bone tissues [109].In long bones, MMP-16 was co-localized with MMP-14 in restingand proliferative chondrocytes, in the chondro-osseous junctionof the growth plate and on the surface of the periosteum. InMMP-14-knockout mice, the mRNA level of MMP-16 in long boneswas lower than that of wild type mice (about 20%), but the distri-bution of MMP-16 was not altered. On the other hand, MMP-16knockout mice are shorter than wild type mice and the mRNA levelof MMP-14 in long bones was also lower than that of wild typemice (about 37%); however, the distribution of MMP-14 was notaltered. Interestingly, a single allelic depletion of the MMP-16 geneis sufficient to develop the same phenotype as that of homozygousdepletion. These observations indicate that depletion of MMP-14does not affect local expression of MMP-16 and vice versa, suggest-ing an overlap in function.

Double-knockout mice for MMP-14 and MMP-16 have alsobeen generated and develop a more severe bone phenotype thana single depletion for each MMP, affecting both intramembranousand endochondral ossification. These mice demonstrate craniofa-cial abnormalities, severe dysfunction in palatal shelf formationleading to cleft palate, thinner cranial vault bones and deficientdevelopment of parietal, frontal and nasal bones. These animalsalso displayed shorter long bones than their littermates, thin


cortical bone, altered growth plate due to the decreased prolifera-tion of chondrocytes, an extended hypertrophic zone and signifi-cant impairment of bone formation in the primary ossificationcenter and abnormal hypertrophic chondrocyte islands in the mid-dle of the bone marrow. Taken together, MMP-16 depletionreduces mesenchymal cell involvement in skeletal formation, andthis may be the most important collagenase during bone develop-ment, where it seems to work in cooperation with MMP-14 [109].

In the same context of MMPs, bone cells also secrete inhibitors ofMMPs [110], posteriorly, identified as TIMPs. The most expressedTIMP during bone development is TIMP-2 [111,112]. However,other TIMP members are expressed in bone elements during mouseand human skeletal development [113,112,114–121], mainly bychondrocytes and bone-lining cells [117,118,122,123]. In cartilage,their high concentration may contribute to anti-angiogenic proper-ties, however TIMP-1, TIMP-2 and TIMP-4 null mice did not developabnormal bone phenotypes [124]. TIMP-3-knockout mice developcartilage degradation similar to that observed in age-associatedosteoarthritis due to increased degradation of collagen type II andaggrecan [125]. Furthermore, double-knockout mice for COMP (car-tilage oligomeric matrix protein) and matrilin-3, proteins abundantin the cartilage matrix, develop an abnormal bone quality [126].These animals are normal at birth but show reduced body size,short long bones, and an increase in trabecular bone density atmetaphysis, delayed degradation of aggrecan in the chondro-osse-ous junction at the growth plate, and increased deposition of TIMP-3. Authors suggested that excess TIMP-3 may inhibit MMPs in thecartilage, leading to the reduced cleavage of aggrecan.

It has been demonstrated that mice lacking RECK(�/�) die in ute-ro (E10.5) before onset of bone development. These animals reveala severe disarray of mesenchymal tissues, deprivation of fibrillarcollagen type I, and abnormal organogenesis, suggesting essentialroles for RECK in the remodeling of the ECM in mammalian devel-opment [127]. Kondo et al. [128] localized RECK expression duringdevelopment in chondrocytes in the Meckels cartilage, vertebrae,ribs, and chondrocyte zones in the growth plate in long bone, aswell as high expression in chondrocytic lineage cells and co-local-ization with collagen type II. Moreover, RECK upregulation wascorrelated with MMPs -2, -9 and -14 downregulation during laterchondrocyte differentiation. As such, RECK may be important forendochondral ossification. Our research group has identified RECKexpression during normal mouse development in the alveolar bonesurrounding tooth germ, connective tissue surrounding Meckelscartilage, in hypertrophic chondrocytes and osteoblasts at thechondro-osseous junction in the growth plate, and at the perichon-drial surface. Furthermore, we have also observed an inverseexpression of RECK and MMPs, similarly to a previous report fromKondo (KBS Paiva – unpublished data). In human osteoarthriticcartilage, high expression of RECK has been detected comparedto normal cartilage and RECK silencing in chondrocytes lead to adecrease in proliferation and an increase in migration [129]. Takentogether, RECK may constitute a novel player in the complex mech-anism of craniofacial formation, endochondral ossification and inthe control of joint pathologies.

Roles of MMPs and their inhibitors for bone regeneration

Bone regeneration is required for the restoration of bone tissuelost by trauma, factures or surgical removal due to locally invasivepathologies. Bone repair is a mechanism that resembles bonedevelopment (intramembranous or endochondral ossification) forgenerating new bone at the site of injury. It is characterized by fourdistinct and consecutive phases: (I) inflammation, where a hema-toma is immediately formed after fracture and subsequent induc-tion of immune system, recruitment of MSCs and angiogenesis

(3 days postfracture) and granulation tissue formation, whereTGF-b1 release from the hematoma controls cell recruitment (mac-rophages and bone-forming cells) at the fracture site, with mito-genic effect of MSCs, osteoprogenitor cells, osteoblasts, andchondrocytes, stimulating ECM production and inhibition of prote-olytic enzymes (3–6 days postfracture); (II–III) callus formation(soft callus: cartilage and hard callus: intramembranous bone),where cartilage is remodeled to form woven bone by endochondralossification and direct bone formation by intramembranous ossifi-cation (4–14 days postfracture); finally, (IV) trabecular boneformed by these two basic mechanisms is remodeled into second-ary bone from 14-days postfracture for a few months [130–133].

The evaluations of MMPs expression and MMP-knockout micemodels have also demonstrated that these enzymes are involvedin different stages during the healing of bone fractures and criti-cal-size bone defects and are essential for the success of bonerepair (Fig. 5). Among them, MMPs -2, -7, -9, -8, -12, -13, and -14, which coordinate the remodeling mechanism of target ECMproteins due to the their expression at certain time and in distinctsubsets (MMP-2/MMP-14 and MMP-9/MMP-13) [134–145]. In awell-established rat model for evaluating bone regeneration, alve-olar bone regeneration after tooth extraction was found to be char-acterized by three distinct steps (exudative, proliferative andreparative); we observed that gelatinases and RECK are expressedat all stages, mainly in the osteoblasts facing new bone, connectivetissue and endothelial cells. This may be important for the replace-ment of the blood clot by connective tissue, and in the formation,maturation and remodeling of new bone [146]. In this context,MMPs are suggested to be biomarkers of the progression of frac-ture healing (measurement of MMPs -9 and -13 in urine by ELISA)[147] and the status of bone remodeling after arthroscopic shoul-der acromioplasty surgery (measurement of MMPs -2 and -9 insubacromial space and peripherical blood by ELISA) [148].

The mechanism of bone healing can fail due to extensive boneloss. As an alternative, scaffolds or biomaterials have beenemployed to help this process in vivo, act by stimulating MSC dif-ferentiation (osteoinduction), migration of resident cells and guid-ance of bone growth (osteoconduction) and formation of new boneby cells at the injury site (osteogenic) [149]. Implantation of bio-materials desired for bone regeneration also requires a coordinatedand controlled tissue and matrix remodeling. Implants of deminer-alized (mainly collagen type I) or inorganic bone (hydroxyapatite)require MMPs for resorption of biomaterial or remodeling sur-rounding tissue, where the MMPs are secreted by macrophagesin contact with the graft [150,151]. A recent study reported thatMMP-2, -9 and VEGF expression were induced by sintered anor-ganic bone during the healing of critical-size defects in rats[152]. Furthermore, in the same study, in the experimental groupthat employed the use of biomaterial within the critical defect, cal-varial remodeling was faster than in the control group. The authorssuggested that stimulated expression of gelatinases and VEGFleads to bone formation/remodeling with successful healing. In thiscontext, the description of MMPs involved in biological responses,triggered by biomaterials, is relevant to understand their stabilityin the host.

Currently, bone quality is desired during new bone formationand during regeneration and the quality of the bone formed isrelated to the prediction of fracture risk. This term encompassesall features besides bone mass, in association with the toleratedbone load (biomechanical properties). As such, few studies haveassociated bone quality with MMP expression. The architectural,composition and biomechanical properties of bone have been com-pared in MMP-2(�/�) and MMP-9(�/�) mice [153]. At 16 weeks ofage, these animals presented altered microarchitectural parame-ters (which characterize the morphology of trabecular bone) andmechanical properties. MMP-2-null mice present few trabeculae

Fig. 5. Schematic representation of main steps of bone healing and bone phenotype developed by MMP knockout mice. (A–C) Bone phenotype for MMP-2 [140], MMP-9[134], and MMP-13 [138] knockout mice, respectively.


(Tb.N), low trabecular connectivity density (Conn.D), but normaltrabecular bone volume fraction (BV/TV) and mineral density ofthe trabeculae (Tb.TMD), thin and porous diaphyseal cortex (espe-cially near the endosteal surface), decreased mineral-collagenratio, less nanoindentation modulus and hardness cortex, leadingto weakened bone. This effect was associated with a decrease inmineralization density of the tissue and an increase in porosity,

Fig. 6. Schematic representation of potential areas of the study of MMPs/inhibi

suggesting that MMP-2 appears important to bone strength. Incontrast, MMP-9-null mice presented more trabeculae (Tb.N) andthin trabeculae (Tb.Th), high trabecular connectivity density(Conn.D), lower mineral density of the trabeculae (Tb.TMD), nor-mal diaphyseal cortex, trabecular bone volume fraction (BV/TV),mineral-collagen ratio and nanoindentation modulus and hardnesscortex, which did not affect the bending strength of the whole

tors in Mesenchymal Stem Cell Biology and Bone Bioengineering concepts.


bone, despite causing a decrease in structure, suggesting thatMMP-9 appears important to bone toughness. The exact causefor this bone brittleness is presently unknown, but the state ofbone collagen type I is known to affect brittleness [154] andMMP-9 may directly influence the organization of proteins foundin the bone extracellular matrix. Although gelatinases share similarsubstrates they are expressed by different bone cells: MMP-2 ispreferentially found in osteoblasts and MMP-9 in osteoclasts. Thus,these MMPs may be associated with bone formation and boneresorption, respectively, possibly explaining the differences foundbetween MMP-2 and MMP-9-knockout mice. Furthermore, arecent study showed that MMP-13 plays an important role in themaintenance of cortical bone quality [155]. MMP-13-knockoutmice have a reduced resistance to fracture in their long bonesdue to the modulation of the osteocyte lacunar–canalicular net-work remodeling in the cortical bone, suggesting that MMP-13 isnecessary for the normal distribution of mineral density in corticalbone. Since both MMP-2 and MMP-9 are proteolytically activatedby MMP-13 [156] and depletion of both gelatinases has been asso-ciated with decreased bone quality [153], the bone phenotypedeveloped by MMP-13-null mice may reflect in inefficient bonematrix remodeling due to the loss of MMP-2 and MMP-9 activity,particularly MMP-2.

When currently available approaches for bone regeneration arenot sufficient, bone bioengineering, employing concepts that com-bine autologous cells (currently, the MSCs are the most investi-gated cells), biomaterials and signaling molecules to create atridimensional bone graft to deliver into the fractured bone site,may open new possibilities to design modern protocols for boneregeneration.

Perspectives of MMPs function for Bone Bioengineering

With regard to Bone Bioengineering concepts, knowledge of therepertoire of molecules secreted into the ECM and attached on thecell surface, constituting the cell Secretome, is important to designan appropriate protocol to induce osteogenesis. Among the stemcells, mesenchymal stem cells (MSCs) have been widely studiedfor Tissue Bioengineering approaches, since they are able to differ-entiate into multi-lineages and exert trophic effects due to thesecretion of bioactive molecules. As such, the proteolytic enzymearsenal involved in ECM remodeling secreted from employed cells(mainly MSCs) is critical for controlling the degradation ratio ofbioabsorbable biomaterial, and this is essential to ensure that theprocess occurs successfully in certain cases. Furthermore, growthfactors embedded in the host ECM or conjugated to biomaterialscan be activated or released by MMPs and available for tissuerepair or remodeling.

Considering the pivotal functions of MMPs during normal bonedevelopment, MMPs appear to be potential targets in the cell Sec-retome to improve Bone and Cartilage Bioengineering, represent-ing a new approach to study these enzymes. However, little isknown about the MMPs and their inhibitors, with regard to severalpoints in Mesenchymal Stem Cell Biology and Tissue Engineering(Fig. 6):

1) Secretion and function of MMPs/inhibitors by MSCs is notwell known. It is also not known whether these enzymes/proteins can modulate principal stem cell properties (self-renewal and differentiation potential in vitro) [157,158]. Fur-thermore, it is not known whether MMPs/inhibitors areexpressed during MSC commitment to the osteoblastic andchondrocytic lineages and their role during differentiationsteps;

2) It is not known whether MMPs/inhibitors secreted by MSCsmay control ECM remodeling for maintaining self-renewaland differentiation within the tissue niche, for recruitmentor differentiation of the MSCs population from the niche tothe tissue injury site, and for future tissue repair/remodeling;

3) Which bioactive molecules or biomaterials may induce MMPexpression in MSCs and which intracellular pathways areactivated [159].

With regard to biomaterial composition, the mechanism andratio of degradation of polymeric biodegradable biomaterials bythe proteolytic action of MMPs and whether bioactive moleculesmay be released from this biomaterial are also important for thecoordination of biomaterial cleavage and new tissue formation atthe same time [160,161]. On the other hand, it is possible to deliverMMPs to the MSCs for bone regeneration by two distinct gene ther-apy approaches (already used for other gene delivery techniques):first, using vectors containing MMPs (encoding protein fragment orinhibitory small RNA) entrapped in biomaterials, which are thenreleased gradually [162,163]; secondly, MSCs can be geneticallymodified to demonstrate MMP overexpression or protein down-regulation, mediated by short or small RNA.

All of these questions remain open and their investigation willbe important to improve our knowledge regarding the control ofECM remodeling for bone and cartilage regeneration.

Acknowledgments

Our research group is supported by São Paulo Research Founda-tion (FAPESP) Grants no. 2010/08918-9 and K.B.S.P. Young ScientistFellowship no. 2011/00204-0.

References

[1] R. Noguera, O.A. Nieto, I. Tadeo, F. Farinas, T. Alvaro, Histol. Histopathol. 27(2012) 693–705.

[2] T. Rozario, D.W. DeSimone, Dev. Biol. 341 (2010) 126–140.[3] M.F. Brizzi, G. Tarone, P. Defilippi, Curr. Opin. Cell Biol. 24 (2012) 645–651.[4] V. Catalano, A. Turdo, F.S. Di, F. Dieli, M. Todaro, G. Stassi, Semin. Cancer Biol.

23 (2013) 522–532.[5] P. Lu, V.M. Weaver, Z. Werb, J. Cell Biol. 196 (2012) 395–406.[6] B. Bauvois, Biochim. Biophys. Acta 1825 (2012) 29–36.[7] G.S. Butler, C.M. Overall, Biochemistry 48 (2009) 10830–10845.[8] E.I. Deryugina, J.P. Quigley, Biochim. Biophys. Acta 1803 (2010) 103–120.[9] F. Mannello, V. Medda, Prog. Histochem. Cytochem. 47 (2012) 27–58.

[10] F.X. Maquart, G. Bellon, S. Pasco, J.C. Monboisse, Biochimie 87 (2005) 353–360.

[11] G. Murphy, H. Nagase, FEBS J. 278 (2011) 2–15.[12] D. Rodriguez, C.J. Morrison, C.M. Overall, Biochim. Biophys. Acta 1803 (2010)

39–54.[13] E.M. Tam, C.J. Morrison, Y.I. Wu, M.S. Stack, C.M. Overall, Proc. Natl. Acad. Sci.

U.S.A. 101 (2004) 6917–6922.[14] J.F. Woods, G. Nichols, J. Cell Biol. 26 (1965) 747–757.[15] A. Aiken, R. Khokha, Biochim. Biophys. Acta 1803 (2010) 121–132.[16] T.E. Cawston, A.J. Wilson, Best Pract. Res. Clin. Rheumatol. 20 (2006) 983–

1002.[17] S.M. Krane, M. Inada, Bone 43 (2008) 7–18.[18] C.J. Malemud, Front. Biosci. 11 (2006) 1702–1715.[19] N. Ortega, D. Behonick, D. Stickens, Z. Werb, Ann. N. Y. Acad. Sci. 995 (2003)

109–116.[20] N. Ortega, D.J. Behonick, Z. Werb, Trends Cell Biol. 14 (2004) 86–93.[21] J. GROSS, (1962).[22] H. Nagase, R. Visse, G. Murphy, Cardiovasc. Res. 69 (2006) 562–573.[23] I.M. Clark, T.E. Swingler, C.L. Sampieri, D.R. Edwards, Int. J. Biochem. Cell Biol.

40 (2008) 1362–1378.[24] E.E. Gabison, E. Huet, C. Baudouin, S. Menashi, Prog. Retin. Eye Res. 28 (2009)

19–33.[25] E. Hadler-Olsen, B. Fadnes, I. Sylte, L. Uhlin-Hansen, J.O. Winberg, FEBS J. 278

(2011) 28–45.[26] S. Hernandez-Barrantes, M. Bernardo, M. Toth, R. Fridman, Semin. Cancer Biol.

12 (2002) 131–138.[27] L. Li, H. Li, Cancer Biol. Ther. 14 (9) (2013) 796–805.

http://refhub.elsevier.com/S0003-9861(14)00285-9/h0005






































[28] C. Piperi, A.G. Papavassiliou, Curr. Top. Med. Chem. 12 (2012) 1095–1112.[29] H.J. Ra, W.C. Parks, Matrix Biol. 26 (2007) 587–596.[30] A. Rietz, J. Spiers, Br. J. Pharmacol. 166 (2012) 1225–1243.[31] P.E. Van den Steen, G. Opdenakker, M.R. Wormald, R.A. Dwek, P.M. Rudd,

Biochim. Biophys. Acta. 1528 (2001) 61–73.[32] M.P. Vincenti, Methods Mol. Biol. 151 (2001) 121–148.[33] C. Yan, D.D. Boyd, J. Cell Physiol. 211 (2007) 19–26.[34] C. Takahashi, Z. Sheng, T.P. Horan, H. Kitayama, M. Maki, K. Hitomi, Y. Kitaura,

S. Takai, R.M. Sasahara, A. Horimoto, Y. Ikawa, B.J. Ratzkin, T. Arakawa, M.Noda, Proc. Natl. Acad. Sci. U.S.A. 95 (1998) 13221–13226.

[35] U.I. Chung, H. Kawaguchi, T. Takato, K. Nakamura, J. Orthop. Sci. 9 (2004)410–414.

[36] T.A. Franz-Odendaal, Front. Biosci. 16 (2011) 2734–2746.[37] B.F. Eames, J.A. Helms, Dev. Dyn. 231 (2004) 4–13.[38] B.K. Hall, J. Exp. Zool. B Mol. Dev. Evol. 304 (2005) 548–557.[39] T.A. Franz-Odendaal, B.K. Hall, P.E. Witten, Dev. Dyn. 235 (2006) 176–190.[40] M.M. Cohen Jr., Am. J. Med. Genet. A 140 (2006) 2646–2706.[41] H. Hojo, S. Ohba, F. Yano, U.I. Chung, J. Bone Miner. Metab. 28 (2010) 489–

502.[42] E.J. Mackie, L. Tatarczuch, M. Mirams, J. Endocrinol. 211 (2011) 109–121.[43] K.A. Staines, A.S. Pollard, I.M. McGonnell, C. Farquharson, A.A. Pitsillides, J.

Endocrinol. 219 (2013) R1–R12.[44] N.A. Sims, T.J. Martin, Bonekey Rep. 3 (2014) 481.[45] A. Teti, Curr. Osteoporos. Rep. 9 (2011) 264–273.[46] H. Qing, L.F. Bonewald, Int. J. Oral Sci. 1 (2009) 59–65.[47] J.J. Wysolmerski, Bonekey Rep. 1 (2012) 229.[48] J. Klein-Nulend, A.D. Bakker, R.G. Bacabac, A. Vatsa, S. Weinbaum, Bone 54

(2013) 182–190.[49] I. Villemure, I.A. Stokes, J. Biomech. 42 (2009) 1793–1803.[50] A. Himeno-Ando, Y. Izumi, A. Yamaguchi, T. Iimura, Biochem. Biophys. Res.

Commun. 417 (2012) 765–770.[51] D.J. Papachristou, K.K. Papachroni, E.K. Basdra, A.G. Papavassiliou, Bioessays

31 (2009) 794–804.[52] S. Loffek, O. Schilling, C.W. Franzke, Eur. Respir. J. 38 (2011) 191–208.[53] A. Page-McCaw, A.J. Ewald, Z. Werb, Nat. Rev. Mol. Cell Biol. 8 (2007) 221–

233.[54] L. Szabova, M.Y. Son, J. Shi, M. Sramko, S.S. Yamada, W.D. Swaim, P. Zerfas, S.

Kahan, K. Holmbeck, Blood 116 (2010) 5752–5761.[55] T. Itoh, T. Ikeda, H. Gomi, S. Nakao, T. Suzuki, S. Itohara, J. Biol. Chem. 272

(1997) 22389–22392.[56] J.A. Martignetti, A.A. Aqeel, W.A. Sewairi, C.E. Boumah, M. Kambouris, S.A.

Mayouf, K.V. Sheth, W.A. Eid, O. Dowling, J. Harris, M.J. Glucksman, S. Bahabri,B.F. Meyer, R.J. Desnick, Nat. Genet. 28 (2001) 261–265.

[57] R.A. Mosig, O. Dowling, A. DiFeo, M.C. Ramirez, I.C. Parker, E. Abe, J. Diouri,A.A. Aqeel, J.D. Wylie, S.A. Oblander, J. Madri, P. Bianco, S.S. Apte, M. Zaidi, S.B.Doty, R.J. Majeska, M.B. Schaffler, J.A. Martignetti, Hum. Mol. Genet. 16 (2007)1113–1123.

[58] C. Rouzier, R. Vanatka, S. Bannwarth, N. Philip, A. Coussement, V. Paquis-Flucklinger, J.C. Lambert, Clin. Genet. 69 (2006) 271–276.

[59] B. Tuysuz, R. Mosig, G. Altun, S. Sancak, M.J. Glucksman, J.A. Martignetti, Eur. J.Hum. Genet. 17 (2009) 565–572.

[60] A. Zankl, L. Bonafe, V. Calcaterra, R.M. Di, A. Superti-Furga, Clin. Genet. 67(2005) 261–266.

[61] K. Inoue, Y. Mikuni-Takagaki, K. Oikawa, T. Itoh, M. Inada, T. Noguchi, J.S. Park,T. Onodera, S.M. Krane, M. Noda, S. Itohara, J. Biol. Chem. 281 (2006) 33814–33824.

[62] W. Zhao, M.H. Byrne, Y. Wang, S.M. Krane, J. Clin. Invest. 106 (2000) 941–949.[63] M. Egeblad, H.C. Shen, D.J. Behonick, L. Wilmes, A. Eichten, L.V. Korets, F.

Kheradmand, Z. Werb, L.M. Coussens, Dev. Dyn. 236 (2007) 1683–1693.[64] R.A. Mosig, J.A. Martignetti, Dis. Model. Mech. 6 (2013) 397–403.[65] C.M. Overall, R.A. Dean, Cancer Metastasis Rev. 25 (2006) 69–75.[66] D.H. Madsen, H.J. Jurgensen, S. Ingvarsen, M.C. Melander, R. Albrechtsen, A.

Hald, K. Holmbeck, T.H. Bugge, N. Behrendt, L.H. Engelholm, PLoS ONE 8(2013) e71261.

[67] R.A. Wagenaar-Miller, L.H. Engelholm, J. Gavard, S.S. Yamada, J.S. Gutkind, N.Behrendt, T.H. Bugge, K. Holmbeck, Mol. Cell. Biol. 27 (2007) 6309–6322.

[68] D.H. Madsen, L.H. Engelholm, S. Ingvarsen, T. Hillig, R.A. Wagenaar-Miller, L.Kjoller, H. Gardsvoll, G. Hoyer-Hansen, K. Holmbeck, T.H. Bugge, N. Behrendt,J. Biol. Chem. 282 (2007) 27037–27045.

[69] D.S. Hipps, R.M. Hembry, A.J. Docherty, J.J. Reynolds, G. Murphy, Biol. Chem.Hoppe Seyler 372 (1991) 287–296.

[70] G. Messaritou, L. East, C. Roghi, C.M. Isacke, H. Yarwood, J. Cell Sci. 122 (2009)4042–4048.

[71] T.H. Vu, J.M. Shipley, G. Bergers, J.E. Berger, J.A. Helms, D. Hanahan, S.D.Shapiro, R.M. Senior, Z. Werb, Cell 93 (1998) 411–422.

[72] T. Kojima, T. Hasegawa, P.H. de Freitas, T. Yamamoto, M. Sasaki, K. Horiuchi,H. Hongo, T. Yamada, N. Sakagami, N. Saito, M. Yoshizawa, T. Kobayashi, T.Maeda, C. Saito, N. Amizuka, Biomed. Res. 34 (2013) 119–128.

[73] N. Ortega, K. Wang, N. Ferrara, Z. Werb, T.H. Vu, Dis. Model. Mech. 3 (2010)224–235.

[74] N. Ortega, D.J. Behonick, C. Colnot, D.N. Cooper, Z. Werb, Mol. Biol. Cell 16(2005) 3028–3039.

[75] K. Fuller, T.J. Chambers, J. Cell Sci. 108 (Pt 6) (1995) 2221–2230.[76] S. Gack, R. Vallon, J. Schmidt, A. Grigoriadis, J. Tuckermann, J. Schenkel, H.

Weiher, E.F. Wagner, P. Angel, Cell Growth Differ. 6 (1995) 759–767.

[77] V. Mattot, M.B. Raes, P. Henriet, Y. Eeckhout, D. Stehelin, B. Vandenbunder, X.Desbiens, J. Cell Sci. 108 (Pt 2) (1995) 529–535.

[78] M. Stahle-Backdahl, B. Sandstedt, K. Bruce, A. Lindahl, M.G. Jimenez, J.A. Vega,C. Lopez-Otin, Lab. Invest. 76 (1997) 717–728.

[79] M. Inada, Y. Wang, M.H. Byrne, M.U. Rahman, C. Miyaura, C. Lopez-Otin, S.M.Krane, Proc. Natl. Acad. Sci. U.S.A. 101 (2004) 17192–17197.

[80] D. Stickens, D.J. Behonick, N. Ortega, B. Heyer, B. Hartenstein, Y. Yu, A.J.Fosang, M. Schorpp-Kistner, P. Angel, Z. Werb, Development 131 (2004)5883–5895.

[81] N. Selvamurugan, M.R. Pulumati, D.R. Tyson, N.C. Partridge, J. Biol. Chem. 275(2000) 5037–5042.

[82] M.J. Jimenez, M. Balbin, J.M. Lopez, J. Alvarez, T. Komori, C. Lopez-Otin, Mol.Cell. Biol. 19 (1999) 4431–4442.

[83] D. Porte, J. Tuckermann, M. Becker, B. Baumann, S. Teurich, T. Higgins, M.J.Owen, M. Schorpp-Kistner, P. Angel, Oncogene 18 (1999) 667–678.

[84] H.E. MacLean, J.I. Kim, M.J. Glimcher, J. Wang, H.M. Kronenberg, L.H.Glimcher, Dev. Biol. 262 (2003) 51–63.

[85] M.J. Blumer, S. Longato, H. Fritsch, Ann. Anat. 190 (2008) 305–315.[86] A.M. Kennedy, M. Inada, S.M. Krane, P.T. Christie, B. Harding, C. Lopez-Otin,

L.M. Sanchez, A.A. Pannett, A. Dearlove, C. Hartley, M.H. Byrne, A.A. Reed, M.A.Nesbit, M.P. Whyte, R.V. Thakker, J. Clin. Invest. 115 (2005) 2832–2842.

[87] K. Holmbeck, P. Bianco, J. Caterina, S. Yamada, M. Kromer, S.A. Kuznetsov, M.Mankani, P.G. Robey, A.R. Poole, I. Pidoux, J.M. Ward, H. Birkedal-Hansen, Cell99 (1999) 81–92.

[88] K. Holmbeck, P. Bianco, K. Chrysovergis, S. Yamada, H. Birkedal-Hansen, J. CellBiol. 163 (2003) 661–671.

[89] Z. Zhou, S.S. Apte, R. Soininen, R. Cao, G.Y. Baaklini, R.W. Rauser, J. Wang, Y.Cao, K. Tryggvason, Proc. Natl. Acad. Sci. U.S.A. 97 (2000) 4052–4057.

[90] J. Muhlhauser, J. Submicrosc. Cytol. 18 (1986) 717–724.[91] S.A. Oblander, Z. Zhou, B.G. Galvez, B. Starcher, J.M. Shannon, M. Durbeej, A.G.

Arroyo, K. Tryggvason, S.S. Apte, Dev. Biol. 277 (2005) 255–269.[92] J. Oh, R. Takahashi, E. Adachi, S. Kondo, S. Kuratomi, A. Noma, D.B. Alexander,

H. Motoda, A. Okada, M. Seiki, T. Itoh, S. Itohara, C. Takahashi, M. Noda,Oncogene 23 (2004) 5041–5048.

[93] N.E. Hatch, Gene Expr. 20 (2010) 295–311.[94] L. Peduto, V.E. Reuter, D.R. Shaffer, H.I. Scher, C.P. Blobel, Cancer Res. 65

(2005) 9312–9319.[95] K. Holmbeck, P. Bianco, I. Pidoux, S. Inoue, R.C. Billinghurst, W. Wu, K.

Chrysovergis, S. Yamada, H. Birkedal-Hansen, A.R. Poole, J. Cell Sci. 118 (2005)147–156.

[96] M. Kajita, Y. Itoh, T. Chiba, H. Mori, A. Okada, H. Kinoh, M. Seiki, J. Cell Biol.153 (2001) 893–904.

[97] V. Samanna, T. Ma, T.W. Mak, M. Rogers, M.A. Chellaiah, J. Cell Physiol. 213(2007) 710–720.

[98] M.A. Chellaiah, T. Ma, Biomed. Res. Int. 2013 (2013) 302392.[99] D.E. Hughes, D.M. Salter, R. Simpson, J. Bone Miner. Res. 9 (1994) 39–44.

[100] A. Santos, A.D. Bakker, B. Zandieh-Doulabi, J.M. de Blieck-Hogervorst, J. Klein-Nulend, Biochem. Biophys. Res. Commun. 391 (2010) 364–369.

[101] S.R. Young, R. Gerard-O’Riley, J.B. Kim, F.M. Pavalko, J. Bone Miner. Res. 24(2009) 411–424.

[102] P.A. Eisenach, C. Roghi, M. Fogarasi, G. Murphy, W.R. English, J. Cell Sci. 123(2010) 4182–4193.

[103] R.N. Kulkarni, A.D. Bakker, E.V. Gruber, T.D. Chae, J.B. Veldkamp, J. Klein-Nulend, V. Everts, Biochem. Biophys. Res. Commun. 417 (2012) 824–829.

[104] L. Szabova, S.S. Yamada, H. Wimer, K. Chrysovergis, S. Ingvarsen, N. Behrendt,L.H. Engelholm, K. Holmbeck, J. Bone Miner. Res. 24 (2009) 1905–1916.

[105] Y. Tang, R.G. Rowe, E.L. Botvinick, A. Kurup, A.J. Putnam, M. Seiki, V.M.Weaver, E.T. Keller, S. Goldstein, J. Dai, D. Begun, T. Saunders, S.J. Weiss, Dev.Cell. 25 (2013) 402–416.

[106] A.N. Saleem, Y.H. Chen, H.J. Baek, Y.W. Hsiao, H.W. Huang, H.J. Kao, K.M. Liu,L.F. Shen, I.W. Song, C.P. Tu, J.Y. Wu, T. Kikuchi, M.J. Justice, J.J. Yen, Y.T. Chen,PLoS Genet. 6 (2010) e1000985.

[107] I.W. Song, W.R. Li, L.Y. Chen, L.F. Shen, K.M. Liu, J.J. Yen, Y.J. Chen, Y.J. Chen,V.B. Kraus, J.Y. Wu, M.T. Lee, Y.T. Chen, PLoS ONE 9 (2014) e92194.

[108] B.R. Evans, R.A. Mosig, M. Lobl, C.R. Martignetti, C. Camacho, V. Grum-Tokars,M.J. Glucksman, J.A. Martignetti, Am. J. Hum. Genet. 91 (2012) 572–576.

[109] J. Shi, M.Y. Son, S. Yamada, L. Szabova, S. Kahan, K. Chrysovergis, L. Wolf, A.Surmak, K. Holmbeck, Dev. Biol. 313 (2008) 196–209.

[110] T.E. Cawston, W.A. Galloway, E. Mercer, G. Murphy, J.J. Reynolds, Biochem. J.195 (1981) 159–165.

[111] S.S. Apte, N. Fukai, D.R. Beier, B.R. Olsen, J. Biol. Chem. 272 (1997) 25511–25517.

[112] L. Blavier, Y.A. DeClerck, Mol. Biol. Cell 8 (1997) 1513–1527.[113] S.S. Apte, K. Hayashi, M.F. Seldin, M.G. Mattei, M. Hayashi, B.R. Olsen, Dev.

Dyn. 200 (1994) 177–197.[114] S. Bord, A. Horner, C.A. Beeton, R.M. Hembry, J.E. Compston, Bone 24 (1999)

229–235.[115] A.M. Flenniken, B.R. Williams, Genes Dev. 4 (1990) 1094–1106.[116] K. Hammani, A. Blakis, D. Morsette, A.M. Bowcock, C. Schmutte, P. Henriet,

Y.A. DeClerck, J. Biol. Chem. 271 (1996) 25498–25505.[117] W. Huang, W.Q. Li, F. Dehnade, M. Zafarullah, J. Cell. Biochem. 85 (2002) 295–

303.[118] M.C. Meikle, A.M. McGarrity, B.M. Thomson, J.J. Reynolds, Bone Miner. 12

(1991) 41–55.[119] L. Rifas, A. Fausto, M.J. Scott, L.V. Avioli, H.G. Welgus, Endocrinology 134

(1994) 213–221.













































































































































































[120] S. Su, J. Grover, P.J. Roughley, J.A. DiBattista, J. Martel-Pelletier, J.P. Pelletier,M. Zafarullah, Rheumatol. Int. 18 (1999) 183–191.

[121] Y. Zeng, R.C. Rosborough, Y. Li, A.R. Gupta, J. Bennett, Dev. Dyn. 211 (1998)228–237.

[122] K. Joronen, H. Salminen, V. Glumoff, M. Savontaus, E. Vuorio, Histochem. CellBiol. 114 (2000) 157–165.

[123] Y. Shen, I.G. Winkler, V. Barbier, N.A. Sims, J. Hendy, J.P. Levesque, PLoS ONE5 (9) (2010). pii: e13086.

[124] K. Brew, H. Nagase, Biochim. Biophys. Acta 1803 (2010) 55–71.[125] S. Sahebjam, R. Khokha, J.S. Mort, Arthritis Rheum. 56 (2007) 905–909.[126] G. Groma, W. Xin, I. Grskovic, A. Niehoff, B. Brachvogel, M. Paulsson, F.

Zaucke, Arthritis Rheum. 64 (2012) 2644–2654.[127] J. Oh, R. Takahashi, S. Kondo, A. Mizoguchi, E. Adachi, R.M. Sasahara, S.

Nishimura, Y. Imamura, H. Kitayama, D.B. Alexander, C. Ide, T.P. Horan, T.Arakawa, H. Yoshida, S. Nishikawa, Y. Itoh, M. Seiki, S. Itohara, C. Takahashi,M. Noda, Cell 107 (2001) 789–800.

[128] S. Kondo, C. Shukunami, Y. Morioka, N. Matsumoto, R. Takahashi, J. Oh, T.Atsumi, A. Umezawa, A. Kudo, H. Kitayama, Y. Hiraki, M. Noda, J. Cell Sci. 120(2007) 849–857.

[129] T. Kimura, A. Okada, T. Yatabe, M. Okubo, Y. Toyama, M. Noda, Y. Okada, Am. J.Pathol. 176 (2010) 2858–2867.

[130] L.C. Gerstenfeld, D.M. Cullinane, G.L. Barnes, D.T. Graves, T.A. Einhorn, J. CellBiochem. 88 (2003) 873–884.

[131] B. Mavcic, V. Antolic, Int. Orthop. 36 (2012) 689–695.[132] N. Rozen, D. Lewinson, T. Bick, S. Meretyk, M. Soudry, Gene Expr. 17 (2007)

197–213.[133] F. Shapiro, Eur. Cell Mater. 15 (2008) 53–76.[134] C. Colnot, Z. Thompson, T. Miclau, Z. Werb, J.A. Helms, Development 130

(2003) 4123–4133.[135] M. Fajardo, C.J. Liu, K. Ilalov, K.A. Egol, J. Orthop. Trauma 24 (2010) 557–563.[136] P. Henle, G. Zimmermann, S. Weiss, Bone 37 (2005) 791–798.[137] T. Itagaki, T. Honma, I. Takahashi, S. Echigo, Y. Sasano, Anat. Rec. 291 (2008)

1038–1046.[138] N. Kosaki, H. Takaishi, S. Kamekura, T. Kimura, Y. Okada, L. Minqi, N. Amizuka,

U.I. Chung, K. Nakamura, H. Kawaguchi, Y. Toyama, J. D’Armiento, Biochem.Biophys. Res. Commun. 354 (2007) 846–851.

[139] W. Lehmann, C.M. Edgar, K. Wang, T.J. Cho, G.L. Barnes, S. Kakar, D.T. Graves,J.M. Rueger, L.C. Gerstenfeld, T.A. Einhorn, Bone 36 (2005) 300–310.

[140] S. Lieu, E. Hansen, R. Dedini, D. Behonick, Z. Werb, T. Miclau, R. Marcucio, C.Colnot, Dis. Model. Mech. 4 (2011) 203–211.

[141] M.M. McDonald, A. Morse, K. Mikulec, L. Peacock, P.A. Baldock, P.J. Kostenuik,D.G. Little, J. Bone Miner. Res. 28 (2013) 1550–1560.

[142] H. Uusitalo, A. Hiltunen, M. Soderstrom, H.T. Aro, E. Vuorio, Calcif. Tissue Int.67 (2000) 382–390.

[143] K. Wang, P. Vishwanath, G.S. Eichler, M.O. Al-Sebaei, C.M. Edgar, T.A. Einhorn,T.F. Smith, L.C. Gerstenfeld, Matrix Biol. 25 (2006) 271–281.

[144] X. Wang, Y.Y. Yu, S. Lieu, F. Yang, J. Lang, C. Lu, Z. Werb, D. Hu, T. Miclau, R.Marcucio, C. Colnot, Bone 52 (2013) 111–119.

[145] H. Yamagiwa, K. Tokunaga, T. Hayami, H. Hatano, M. Uchida, N. Endo, H.E.Takahashi, Bone 25 (1999) 197–203.

[146] T. Accorsi-Mendonca, K.B. Paiva, W.F. Zambuzzi, T.M. Cestari, V.S. Lara, M.C.Sogayar, R. Taga, J.M. Granjeiro, J. Mol. Histol. 39 (2008) 201–208.

[147] N.A. Wigner, N. Kulkarni, M. Yakavonis, M. Young, B. Tinsley, B. Meeks, T.A.Einhorn, L.C. Gerstenfeld, Injury 43 (2012) 274–278.

[148] E. Galliera, P. Randelli, G. Dogliotti, E. Dozio, A. Colombini, G. Lombardi, P.Cabitza, M.M. Corsi, Injury 41 (2010) 1204–1207.

[149] D.G. Boeckel, R.S. Shinkai, M.L. Grossi, E.R. Teixeira, J. Oral Implantol. 38(2012) 538–545.

[150] T. Accorsi-Mendonca, W.F. Zambuzzi, K.B. da Silva, J. Mol. Histol. 36 (2005)311–316.

[151] W.F. Zambuzzi, K.B. Paiva, R. Menezes, R.C. Oliveira, R. Taga, J.M. Granjeiro, J.Mol. Histol. 40 (2009) 301–309.

[152] C.A. Rocha, T.M. Cestari, H.A. Vidotti, G.F. de Assis, G.P. Garlet, R. Taga, J. Mol.Histol. (2014).

[153] J.S. Nyman, C.C. Lynch, D.S. Perrien, S. Thiolloy, E.C. O’Quinn, C.A. Patil, X. Bi,G.M. Pharr, A. Mahadevan-Jansen, G.R. Mundy, J. Bone Miner. Res. 26 (2011)1252–1260.

[154] J.S. Nyman, M. Reyes, X. Wang, Micron 36 (7–8) (2005) 566–582.[155] S.Y. Tang, R.P. Herber, S.P. Ho, T. Alliston, J. Bone Miner. Res. 27 (2012) 1936–

1950.[156] V. Knauper, H. Will, C. Lopez-Otin, B. Smith, S.J. Atkinson, H. Stanton, R.M.

Hembry, G. Murphy, J. Biol. Chem. 271 (1996) 17124–17131.[157] T.P. Lozito, R.S. Tuan, J. Cell Physiol. 226 (2011) 385–396.[158] T.P. Lozito, W.M. Jackson, L.J. Nesti, R.S. Tuan, Matrix Biol. 34 (2014) 132–143,

http://dx.doi.org/10.1016/j.matbio.2013.10.003.[159] C. Laflamme, M. Rouabhia, Biomed. Mater. 3 (2008) 015008.[160] J. Kim, I.S. Kim, T.H. Cho, H.C. Kim, S.J. Yoon, J. Choi, Y. Park, K. Sun, S.J. Hwang,

J. Biomed. Mater. Res. A 95 (2010) 673–681.[161] M.P. Lutolf, J.L. Lauer-Fields, H.G. Schmoekel, A.T. Metters, F.E. Weber, G.B.

Fields, J.A. Hubbell, Proc. Natl. Acad. Sci. U.S.A. 100 (2003) 5413–5418.[162] E.J. Kim, H.J. Cho, D. Park, J.Y. Kim, Y.B. Kim, T.G. Park, C.K. Shim, Y.K. Oh, Mol.

Ther. 19 (2011) 355–361.[163] Y. Lei, Q.K. Ng, T. Segura, Microsc. Res. Tech. 73 (2010) 910–917.[164] W.R. English, B. Holtz, G. Vogt, V. Knauper, G. Murphy, J. Biol. Chem. 276

(2001) 42018–42026.[165] D. Pei, FEBS Lett. 457 (1999) 262–270.[166] E.B. Springman, E.L. Angleton, H. Birkedal-Hansen, H.E. Van Wart, Proc. Natl.

Acad. Sci. U.S.A. 87 (1990) 364–368.[167] H.E. Van Wart, H. Birkedal-Hansen, Proc. Natl. Acad. Sci. U.S.A. 87 (1990)

5578–5582.[168] A.Y. Strongin, I. Collier, G. Bannikov, B.L. Marmer, G.A. Grant, G.I. Goldberg, J.

Biol. Chem. 270 (1995) 5331–5338.[169] E.I. Deryugina, B. Ratnikov, E. Monosov, T.I. Postnova, R. DiScipio, J.W. Smith,

A.Y. Strongin, Exp. Cell Res. 263 (2001) 209–223.[170] G.S. Butler, H. Will, S.J. Atkinson, G. Murphy, Eur. J. Biochem. 244 (1997) 653–

657.[171] E. Llano, A.M. Pendas, J.P. Freije, A. Nakano, V. Knauper, G. Murphy, C. Lopez-

Otin, Cancer Res. 59 (1999) 2570–2576.[172] D. Pei, J. Biol. Chem. 274 (1999) 8925–8932.[173] T. Takino, H. Sato, A. Shinagawa, M. Seiki, J. Biol. Chem. 270 (1995) 23013–

23020.[174] G. Velasco, S. Cal, A. Merlos-Suarez, A.A. Ferrando, S. Alvarez, A. Nakano, J.

Arribas, C. Lopez-Otin, Cancer Res. 60 (2000) 877–882.[175] W.R. English, X.S. Puente, J.M. Freije, V. Knauper, A. Amour, A. Merryweather,

C. Lopez-Otin, G. Murphy, J. Biol. Chem. 275 (2000) 14046–14055.[176] C.M. Overall, E. Tam, G.A. McQuibban, C. Morrison, U.M. Wallon, H.F. Bigg, A.E.

King, C.R. Roberts, J. Biol. Chem. 275 (2000) 39497–39506.[177] A.Y. Strongin, B.L. Marmer, G.A. Grant, G.I. Goldberg, J. Biol. Chem. 268 (1993)

14033–14039.[178] D. Pei, S.J. Weiss, Nature 375 (1995) 244–247.[179] H. Sato, T. Kinoshita, T. Takino, K. Nakayama, M. Seiki, FEBS Lett. 393 (1996)

101–104.[180] K.J. Leco, R. Khokha, N. Pavloff, S.P. Hawkes, D.R. Edwards, J. Biol. Chem. 269

(1994) 9352–9360.[181] T. Miki, Y. Takegami, K. Okawa, T. Muraguchi, M. Noda, C. Takahashi, J. Biol.

Chem. 282 (2007) 12341–12352.[182] A. Omura, T. Matsuzaki, K. Mio, T. Ogura, M. Yamamoto, A. Fujita, K. Okawa,

H. Kitayama, C. Takahashi, C. Sato, M. Noda, J. Biol. Chem. 284 (2009) 3461–3469.

[183] S. Simizu, S. Takagi, Y. Tamura, H. Osada, Cancer Res. 65 (2005) 7455–7461.[184] S. Takagi, S. Simizu, H. Osada, Cancer Res. 69 (2009) 1502–1508.[185] B. Welm, J. Mott, Z. Werb, Curr. Biol. 12 (2002) R209–R211.[186] P. Scott-Savage, B.K. Hall, Acta Anat. 106 (1980) 129–140.[187] E.R. Lee, G. Murphy, M. El-Alfy, M.A. Davoli, L. Lamplugh, A.J. Docherty, C.P.

Leblond, Dev. Dyn. 215 (1999) 190–205.[188] D.J. Behonick, Z. Werb, Mech. Dev. 120 (2003) 1327–1336.[189] N. Dirckx, H.M. Van, C. Maes, Birth Defects Res. C Embryo. Today 99 (2013)

170–191.[190] C. Maes, Calcif. Tissue Int. 92 (2013) 307–323.










































































http://dx.doi.org/10.1016/j.matbio.2013.10.003























































bone tissue remodeling and development: focus on matrix metalloproteinase functions

Documents