Buried alive: How osteoblasts become osteocytesucgatma/Anat3048/PAPERS...Buried Alive: How...

REVIEWS–A PEER REVIEWED FORUM

Buried Alive: How Osteoblasts BecomeOsteocytesTamara A. Franz-Odendaal,1* Brian K. Hall,1 and P. Eckhard Witten1,2*

During osteogenesis, osteoblasts lay down osteoid and transform into osteocytes embedded in mineralizedbone matrix. Despite the fact that osteocytes are the most abundant cellular component of bone, little isknown about the process of osteoblast-to-osteocyte transformation. What is known is that osteoblastsundergo a number of changes during this transformation, yet retain their connections to preosteoblasts andosteocytes. This review explores the osteoblast-to-osteocyte transformation during intramembranousossification from both morphological and molecular perspectives. We investigate how these data supportfive schemes that describe how an osteoblast could become entrapped in the bone matrix (in mammals) andsuggest one of the five scenarios that best fits as a model. Those osteoblasts on the bone surface that aredestined for burial and destined to become osteocytes slow down matrix production compared toneighbouring osteoblasts, which continue to produce bone matrix. That is, cells that continue to producematrix actively bury cells producing less or no new bone matrix (passive burial). We summarize whichmorphological and molecular changes could be used as characters (or markers) to follow the transformationprocess. Developmental Dynamics 235:176–190, 2006. © 2005 Wiley-Liss, Inc.

Key words: osteoblast; osteocyte; osteogenesis; intramembranous ossification; preosteoblast; osteoid-osteocyte;preosteocyte

Received 24 March 2005; Revised 24 August 2005; Accepted 31 August 2005

INTRODUCTIONIt has been known, for almost one anda half centuries, that osteocytes arederived from osteoblasts (Gegen-bauer, 1864). Osteoblasts (bone form-ing cells) are of mesenchymal origin,secrete non-mineralized bone matrix(osteoid), and finally become incorpo-rated as osteocytes in mineralizedbone matrix. Osteocytes are by far themost abundant cellular component ofmammalian bones, making up 95% ofall bone cells (20,000 to 80,000 cellsper mm3 bone tissue) that cover 94%of all bone surfaces (Frost, 1960;Marotti, 1996); there are approxi-

mately ten times more osteocytes thanosteoblasts in an individual bone(Parfitt, 1990). In humans, osteocytescan live long. Frost (1963) estimatesthe average half-life of a human osteo-cyte as 25 years. However, when weconsider an overall bone-remodellingrate of between 4 to 10% per year (Del-ling and Vogel, 1992; Manolagas,2000), the life of many osteocytes maybe shorter (Marotti et al., 1990). Fur-thermore, the lifespan of osteocytesgreatly exceeds that of active osteo-blasts, which is estimated to be onlythree months in human bones (Mano-lagas, 2000) and 10–20 days in mouse

alveolar bone (McCulloch and Heer-sche, 1988). Osteocytes communicatewith one another and with cells at thebone surface via a meshwork of cellprocesses that run through canaliculiin the bone matrix (Palumbo et al.,1990). Thus, bone cells form a func-tional network within which cells atall stages of bone formation frompreosteoblast to mature osteocyte re-main connected.

The literature provides us with anastounding number of terms concern-ing the transition from osteoblast toosteocyte, such as “osteocytes are en-cased in mineralized matrix” (Holtrop,

1Biology Department, Dalhousie University, Halifax, Nova Scotia, Canada2Zoological Institute and Zoological Museum, University of Hamburg, Hamburg, GermanyGrant sponsor: Canadian-German Science and Technology Cooperation; Grant sponsor: Deutsche Forschungsgemeinschaft; Grant sponsor:European COST ACTION B23-Oral Facial Development and Regeneration; Grant sponsor: NSERC.*Correspondence to: Tamara Franz-Odendaal or P. Eckhard Witten, Biology Department, Dalhousie University, Halifax, NovaScotia, B3H 4J1 Canada. E-mail: [email protected] or [email protected]

DOI 10.1002/dvdy.20603Published online 28 October 2005 in Wiley InterScience (www.interscience.wiley.com).

DEVELOPMENTAL DYNAMICS 235:176–190, 2006

© 2005 Wiley-Liss, Inc.

1990); “osteoblasts are buried withinthe bone matrix” (Manolagas, 2000);and “osteoblasts merge progressivelyinto the bone matrix” (Meunier, 1989).Despite the many different phrasesthat exist to describe this process, thefundamental question of how osteo-blasts are buried remains largely un-answered. This could relate to the factthat most of the research on bone cellsto date has been done on osteoblastsand osteoclasts. An electronic data-base (WebOfScience) search using “os-teoblast*,” “osteocyte*,” and “oste-oclast*” as keywords revealed thatduring the last 30 years, less than 5%of publications mention “osteocyte*,”although work on osteocytes is accel-erating.

Twenty-five years ago, Knese (1979)proposed two possible mechanisms ofosteoblast entrapment: (1) self-en-trapment or (2) becoming embeddedby neighbouring cells. We explore thetransformation by evaluating andsynthesizing knowledge regarding themechanism(s) by which osteoblastsbecome embedded in bone matrix. Weexamine different scenarios of how os-teoblasts can be turned into osteocytesand discuss morphological and molec-ular markers that may prove to beimportant when examining the trans-formation process. Finally, we suggesta model for the osteoblast-to-osteocytetransformation in mammalian in-tramembranous bone formation butwe also emphasize that there is notonly one mechanism for transformingosteoblasts into osteocytes since dif-ferent mechanisms exist in differentbones, different types of bone forma-tion, different positions within a bone,and different vertebrate species.

MODES OF OSSIFICATIONHow osteoblasts transform into osteo-cytes is dependent on the mode of os-sification—intramembranous, peri-chondral, endochondral (see Hall andWitten, 2005, for additional modes ofbone formation)—and the type of bonethat is being generated (woven or la-mellar bone). It may also depend onthe location of bone formation, on thespecies, and on the age and/or genderof the individual.

During intramembranous bone for-mation, mesenchymal cells differenti-ate into osteoblasts and bone is

formed without replacing a cartilagi-nous model. Bones that form by thismethod are called membrane bones.Perichondral ossification is the mostcommon mode of bone formation if acartilaginous precursor is present. Peri-chondral ossification usually startswith the transformation of a perichon-drium into a periosteum (Scott-Savageand Hall, 1980).

Although found in most vertebrategroups, endochondral ossification istypically studied in mammalian longbones. Endochondral ossification in-volves a cartilaginous template that isreplaced by or remodelled into bone byprocesses that involve several co-ordi-nated sequential steps that can in-clude calcification of the cartilage ma-trix, hypertrophy of chondrocytesfollowed by apoptosis or transdifferen-tiation of chondrocytes into osteo-blasts (in some long bones), resorptionof calcified cartilage, recruitment ofosteoblasts, and the deposition of wo-ven bone (and later lamellar bone) onthe surface of mineralized cartilageresidues (Roach, 1990, 1992; Thesinghet al., 1991; Gerstenfeld and Shapiro,1996; Buxton et al., 2003; Eames etal., 2003). In advanced bony fish (Ac-anthomorpha with about 16,000 spe-cies), intramembranous, perichondral,and endochondral bone formation re-sult in a bone that contains no osteo-cytes and so is known as acellularbone (Witten and Huysseune, 2005).The bone of less derived bony fish andtetrapods contains osteocytes and so iscellular bone (Witten et al., 2001; Wit-ten and Hall, 2003). Acellular boneformation includes osteoid and bonematrix deposition but not osteoblastentrapment (Ekanayake and Hall,1988).

Bone elements can arise via addi-tional and less commonly knownmodes of ossification. Bone can formby metaplasia of other tissue types,and many tissues have modes of for-mation that result in structures thatare intermediate between dentine andbone or between bone and cartilage(Ørvig, 1951; Huysseune and Verraes,1990; Hall, 2005; Vickaryous and Ol-son, 2005; Witten et al., 2005). Theseintermediate modes of bone tissue for-mation are not discussed here, but seeBeresford (1981, 1993), Hall (1990,2005), Taylor et al. (1994), Witten and

Hall (2002), and Hall and Witten(2005).

OSTEOBLAST FATES ANDFUNCTIONSOsteoblasts are involved in bone ma-trix mineralization (for a review, seeBoskey, 1996). At the end of the bone-forming phase, osteoblasts can haveone of four different fates: (1) becomeembedded in the bone as osteocytes,(2) transform into inactive osteoblastsand become bone-lining cells, (3) un-dergo programmed cell death (apopto-sis), or in some situations (4) transdif-ferentiate into cells that depositchondroid or chondroid bone (Manol-agas, 2000; Noble et al., 1997; Jilka etal., 1998; Li et al., 2004).

The proportion of osteoblasts follow-ing each fate is not the same in allmammals and is not conserved amongall taxa or all types of bone. Parfitt(1990) report that in human cancel-lous bone, 65% of the osteoblasts un-dergo apoptosis and only about 30%transform into osteocytes. Aubin andLiu (1996) give a figure of 10–20% forthe number of osteoblasts transform-ing into osteocytes. Banks (1974) esti-mated that 10% of osteoblasts trans-form to osteocytes in the antlers of thewhite-tailed deer (Odocoileus virgin-ianus). In advanced bony fishes withacellular bone, the number of osteo-blasts that turn into osteocytes is, ofcourse, zero (Ekanayake and Hall,1987; Witten et al., 2004).

The age of the animal may also in-fluence the number of osteoblasts thattransform into osteocytes. In trabecu-lar bone of ageing beagles, the numberof bone-lining cells decreases (Milleret al., 1980) suggesting, either thatfewer osteoprogenitors differentiateinto osteoblasts (i.e., there is a smallerpool of osteoblasts) or that the propor-tion of cells following each develop-mental fate is altered.

Once embedded into the bone ma-trix, the former osteoblasts, now os-teocytes, cease their activity. An im-portant role of osteocytes and theirnetwork of cell processes is to functionas strain and stress sensors, signalsthat are very important for maintain-ing bone structure (see Burger et al.,2003; Knothe Tate et al., 2004, amongothers). Osteocytes communicate withneighbouring osteocytes and with

HOW OSTEOBLASTS BECOME OSTEOCYTES 177

cells on the bone surface via a mesh-work of cell processes, which are lo-cated inside canaliculi within the bonematrix (Palumbo et al., 1990) (Fig. 1).While cross-talk between osteoblasts(bone deposition) and osteoclasts(bone resorption) (Lacey et al., 1998;Yasuda et al., 1998) has been estab-lished, there is also the likelihood ofcross-talk between osteocytes and os-teoclasts (Burger et al., 2003). Thus, afunctional network of bone cells thatextends from preosteoblast to matureosteocyte is important for maintainingthe integrity of bone as a tissue.

Another function of embedded os-teoblasts (osteocytes) within the bonecell network is the ability of some os-teocytes to deposit and resorb bonearound the osteocyte lacuna in whichthey are housed, thus changing theshape of the lacuna. This process,called osteocytic osteolysis, is oftennot regarded as characteristic of hu-man osteocytes, but evidence for itsoccurrence has been observed in manyvertebrate species, such as bats (Dotyand Nunez, 1985; Kwiecinski, 1985;Kwiecinski et al., 1987), hamsters(Steinberg et al., 1981), squirrels

(Haller and Zimny, 1978), rats (Be-langer, 1977a; Tazawa et al., 2004);rabbits (Zhang et al., 2000), snakes(Alcobendas et al., 1991), eels (Lopezet al., 1980), salmon (Hughes et al.,1994a), carp (Witten et al., 2000) andan unidentifiable Cretaceous reptile(Belanger, 1977b). Osteocytic osteoly-sis may be limited to situations suchas lactation, hibernation, or preg-nancy that require increased mobili-zation of minerals from the skeleton(see Haller and Zimny, 1978; Stein-berg et al., 1981; Kwiecinski, 1985;Hall, 2005).

THE CELLS INVOLVEDMany researchers consider the trans-formation process to involve three celltypes: preosteoblasts differentiate intoosteoblasts, which become trapped asosteocytes. All agree that the transfor-mation involves a range of morphologi-cal changes such as decrease in cellbody size, increase in cell processes, andchanges in intracellular organelles(Marotti, 1977; Palumbo, 1986; and seeKnothe Tate et al., 2004 for a recentreview). Palumbo (1986) estimates a to-tal reduction in cell body volume of!70% between the osteoblast and theosteocyte stage. We first describe thesemore familiar cell types and then dis-cuss various intermediate cell typesthat have been proposed.

Three types of osteoblasts are usu-ally identified at the bone surface,based on function and morphology:preosteoblasts, active osteoblasts, andinactive (or resting) osteoblasts (Hol-trop, 1990). Mammalian preosteo-blasts have been described as lesscuboidal in shape than active osteo-blasts. They are located at a distancefrom the bone surface, do not depositbone matrix, and can still divide (Fig1). Preosteoblasts already produce col-lagen type I precursor molecules,

References from Table 1.1. Aubin and Liu (1996); 2. Candeliere et al. (2001); 3. Nakashima and de Crombrugghe (2003); 4. Liu et al. (1997); 5. Roach (1994); 6.Nampei et al. (2004); 7. Wetterwald et al. (1996); 8. Hadjiargyrou et al. (2001); 9. Schulze et al. (1999); 10. Litvin et al. (2004); 11. Foxand Chow (1998); 12. Burger et al. (2003); 13. Kalajzic et al. (2004); 14. Bianco et al. (1993); 15. Chen et al. (1993); 16. Sandberg et al.(1988); 17. Bianco et al. (1990); 18. Horiuchi et al. (1999); 19. Grzesik and Gehron Robey (1994); 20. Heersche et al. (1992); 21. Beckeret al. (1986); 22. Hughes et al. (1994b); 23. Clauss et al. (1993); 24. Middleton et al. (1995); 25. Martineau-Doize et al. (1988); 26. Wanget al. (1995); 27. Dodds et al. (1994); 28. Jamal and Aubin (1996); 29. Machwate et al. (1995); 30. Rouleau et al. (1988); 31. Ishidou etal. (1995); 32. Mark et al. (1988); 33. Turksen and Aubin (1991); 34. Lee et al. (1993); 35. Kobayashi and Kronenberg (2005); 36. Kamiyaet al. (2001); 37. Lazowski et al. (1994); 38. Mizoguchi et al. (1997); 39. Sasano et al. (2000); 40. Vakeva et al. (1990); 41. Romanowskiet al. (1990); 42. Bianco et al. (1988); 43. Ikeda et al. (1992); Hall and Miyake (1995); 45. Inada et al. (1999); 46. Ducy et al. (1997); 47.Ducy et al. (2000); 48. Nah et al. (2000); 49. Wong et al. (1992).

Fig. 1. Diagram showing the transitional cell types between preosteoblasts and osteocytes duringosteoblast transformation and their relationships to one another during the second phase ofintramembranous ossification (i.e., when transformation occurs). The diagram is not to scale. Thepreosteoblast (1) layer consists of proliferating cells. Gap junctions are present between all cells fordirect communication. Enlargement shows gap junction between the cell process of an osteocyteand an embedding osteoblast. Arrow indicates osteoid deposition front; arrowhead indicatesmineralization front. During the transformation process, cellular organelles decrease and the totalcell body volume decreases substantially. 1. preosteoblast, 2. preosteoblastic osteoblast, 3.osteoblast, 4. osteoblastic osteocyte (Type I preosteocyte), 5. osteoid-osteocyte (Type II preos-teocyte), 6. Type III preosteocyte, 7. young osteocyte, 8. old osteocyte. Diagram drawn by TimFedak (www.figs.ca).

178 FRANZ-ODENDAAL ET AL.

TABLE 1. Molecular Markers (mRNA, protein) Involved During Endochondral (EC) andIntramembranous (IM) Ossificationa

aOnly 10–20% of osteoblasts follow this fate. No data from cell lines have been included below. Much of the molecular expressiondata presently available do not include the transitional stages between osteoblast and osteocyte. Some of the expression data fortransitional stages can be inferred from the expression of cell stages before and after the transformation period (gray shaded fields).The expression data of some key regulatory factors that act upstream of these markers are included. hXBP-1, human X box bindingprotein 1; IGF, insulin-like growth factor; EGF, epidermal growth factor; ALP, Alkaline phosphatase; TIMP, tissue inhibitor ofmetalloproteinase; MEPE, matrix extracellular phosphoglycoprotein; Osf2, osteoblast specific factor 2; DMP1, dentin matrix acidicphosphoprotein 1; *Negative in human bone, "" present, " weak expression, "/# variable expression (positive or negativedepending on study), # absent or present at levels below detection


which, after post-translational modifi-cation, assemble into collagen fibrils(Manolagas, 2000). A number of bone-specific markers are reliably ex-pressed by preosteoblasts, namelyosteonectin, alkaline phosphatase,several IGFs, hXBP-1, TIMP, tenascinC, EGF-R, IL-6, PTH/PTHrP receptor,several integrins, and periostin (osf2)(Table 1). Preosteoblasts differentiateinto active bone matrix-secreting os-teoblasts, which are typically cuboidalin shape (in mammals) and ultimatelyresponsible for depositing organicbone matrix.

Osteoblasts have a large eccentricnucleus with one to three nucleoli, andprominent RER (rough endoplasmicreticulum) and Golgi areas. Theyextend cellular protrusions or pseu-dopodia towards the osteoid seam(Palumbo, 1986) (Fig.1), do not divide,and can be distinguished from preos-teoblasts by the upregulation of asuite of bone markers, the main onesbeing bone sialoprotein, osteocalcin,E11, BMP-R1, vitamin D3 receptor,vitronectin, thrombospondin, decorin,several BMPs, and, of course, collagentype I (Table 1).

Many researchers have reportedextensive heterogeneity in osteo-blast gene and protein expressionpatterns (Aubin et al., 1992, 1993;Heersche et al., 1992; Chen et al.,1993; Ikeda et al., 1995; Liu et al.,1997; Candaliere et al., 2001) (Table1). For example, in adult rat bone mar-row stromal cell cultures, adjacent os-teoblasts that appear identical morpho-logically express very different levels ofosteoblast-associated markers such asosteocalcin, osteonectin (SPARC), andgalectin-3 (Malaval et al., 1994). Thisheterogeneity is not regarded as the re-sult of the cell cycle stage and is stillpresent when analyses are restricted topost-proliferative osteoblasts (Liu et al.,1997). A recent study of the osteoblastsof 21-day-old fetal rat calvaria foundthat despite all cells appearing histolog-ically similar, the only markers ex-pressed by all osteoblasts, irrespectiveof their position in the calvaria, werealkaline phosphatase and the pth/pthrpreceptor (Candeliere et al., 2001) (Table1). Both markers are also expressed bypreosteoblasts, although the levels arelower in preosteoblasts than in osteo-blasts (for pth/pthrp receptor), and arenot detectable above background stain-

ing in osteocytes. Other osteoblastmarkers (osteocalcin, msx-2, c-fos, para-thyroid hormone-related protein) aredifferentially expressed at both mRNAand protein levels in subsets of osteo-blasts depending on their location or en-vironment within calvaria (Candeliereet al., 2001). All markers tested in thisstudy are present in post-proliferactivecells so cell cycling cannot account forthe differential expression patterns.Anatomical site, developmental age,species, and mode of ossification can allinfluence the gene expression profile ofosteoblasts (Heersche and Aubin, 1990;Aubin et al., 1993; Gehron Robey et al.,1993; Liu et al., 1997). Candeliere et al.(2001) also found that preosteoblastsand osteocytes differentially express arepertoire of genes.

As bone matrix deposition contin-ues, osteoblasts become embedded inthe cells secretory product, the os-teoid. Cells at this early stage of os-teoblast to osteocyte differentiationhave been called “large osteocytes,”“young osteocytes,” or osteoid-osteo-cytes (Baud, 1968, Semba et al.,1966). These cells are larger thanmature (“older”) osteocytes and havea well-developed Golgi apparatus forcollagen storage. At the transmis-sion electron microscope level, colla-gen fibrils can be seen surroundingthese matrix-producing cells (Han-cox and Boothroyd, 1965).

On mineralization of the osteoid, os-teocyte ultrastructure undergoes fur-ther changes, a reduction in the ERand Golgi apparatus corresponding toa decrease in protein synthesis andsecretion (Dudley and Spiro, 1961).Now, many of the previously ex-pressed bone markers are down regu-lated or switched off in the osteocyte(e.g., osteocalcin, bone sialoprotein,collagen type I, alkaline phosphatase,IGFs, integrins, periostin, and others)(Table 1). Some studies have shownthat depending on the type of boneformed and the activity and size of thecommitted osteoblast, the newly em-bedded osteocyte may be variable insize and shape in comparison witholder, more mature, osteocytes al-ready embedded in bone matrix(Boyde, 1980; Marotti et al., 1990).Furthermore, the shape of embeddedosteocytes may depend on the bonetype. For example, in woven bone,which is laid down rapidly with ran-

domly oriented collagen fibres, the os-teocytes are isodiametric (Currey,2003). In lamellar bone, however,which is laid down more slowly, osteo-cytes are flattened and oblate withtheir short axis parallel to the thick-ness of the lamella (Currey, 2003).Mature osteocytes are finally situatedwithin lacunae in the bone matrix, arestellate-shaped, and have long cellprocesses (Fig.1).

The Transitional Cell StagesSince osteocytes derive from osteo-blasts, transitional cell stages be-tween differentiated osteoblasts andosteocytes should be identifiablebased on both, morphological and mo-lecular characters (Fig. 1, Table 1).

Some authors distinguish interme-diate or transitional stages betweenosteoblasts and osteocytes, namely os-teoid-osteocytes (Palumbo, 1986), os-teocytic osteoblasts (Nijweide et al.,1981), or preosteocytes (Holtrop, 1990)for the bone cell completely surroundedby osteoid. More recently, Nefussi et al.(1991) distinguish between osteoblasticosteocytes and osteoid osteocytes. Theosteoblastic osteocyte is younger andwill become an osteoid osteocyte. Theseauthors also distinguished an interme-diate pool of cells between preosteo-blasts and osteoblasts that they termpreosteoblastic osteoblasts. In sum-mary, in the transition from preosteo-blast to osteocyte, they identify six celltypes (Fig. 1). The preosteoblastic osteo-blast will not necessarily transform intoan embedded osteocyte and merely re-places the “lost” osteoblast from the os-teoblast layer.

Palumbo et al. (1990) distinguishthree cell types from osteoblast to ma-ture osteocyte. The type I preosteocyteis also known as the osteoblastic os-teocyte. Type II preosteocytes are os-teoid-osteocytes, while type III preos-teocytes are cells that are partlysurrounded by mineralized matrix.Despite the fact that many workersrefer to any bone cell surrounded byosteoid or matrix as an osteocyte, it isimportant to remember that the ini-tial enclosure of an osteoblast by os-teoid is not the end of the process ofosteoblast-to-osteocyte transformation.This stage, as all other stages, is partof a continuum of differentiation,which is why different authors can


identify distinct features at eachstage. We summarize the stages frompreosteoblast to osteocyte in Figure 1.

It is not only the presence or ab-sence of bone markers that alters dur-ing the transformation process butalso their levels of expression. Natu-rally, quantitative characters, such asmRNA and protein expression levels,are more difficult to describe thanqualitative characters. A molecularmarker that has been proposed tocharacterize the intermediate celltypes in the osteoblast-to-osteocytetransformation is an epitope thatbinds an antibody called E11 (Table1). Wetterwald et al. (1996) originallyidentified the E11 antibody, whichrecognizes osteoblasts, preosteocytes,and osteocytes. Liu et al. (1997) thenused this marker to distinguish be-tween mature osteoblasts on osteoid(i.e., osteoblastic osteocytes) and thoseembedded in mineralizing osteoid(preosteocytes). The expression pat-tern of this marker was, however, con-tradicted by Schulze et al. (1999).They detected E11 expression in os-teocytes and their processes but not inosteoblasts. More recently, Hadjiargy-rou et al. (2001) confirmed E11 expres-sion in osteoblasts and osteocytes dur-ing fracture healing in rat femora. Thevariable expression pattern of thismarker may be the result of workingwith different populations of osteo-blasts and osteocytes by these re-searchers, as this is the situation formany of the other osteoblast-associ-ated markers (Aubin, 1998).

In addition to the great repetoire ofgenes expressed at each of the abovestages, there are also differences be-tween bones of different ages. Differ-ential expression of non-collagenousmatrix proteins was observed in ratfemora and calvaria by Ikeda et al.(1992). For example, osteocalcin isweakly expressed in newborn rat bone(2 day) compared to older rat bones (8weeks).

In summary, although a number ofmolecular markers for preosteoblasts,osteoblasts, and osteocytes are known,their level of expression is variable atdifferent stages during osteogenesis,and because of this there is great het-erogeneity in the expression profileswith some lineage markers (Aubin,1998) (Table 1). Consequently, theidentification of intermediate cell

types still depends largely on morpho-logical characteristics (Fig. 1, Table1). A combined effort using a number(or all) of these markers and lookingspecifically at the osteoblast-osteocytetransformation could prove to be ex-tremely useful in our understandingof this dynamic process. From ouranalysis, those that may prove to behelpful are alkaline phosphatase, theinsulin-like growth factors and theirreceptors, TIMP, hxBP-1, osf2, nitricoxide, and DMP1 (Table 1). To thisend, Bonewald and colleagues have es-tablished several osteocyte-like mousecell lines from long bones (called MLOcell lines) (Aarden et al., 1996; Caplanet al., 1983; Kato et al., 2001). Theadvantage of cell culture methods isthat one can manipulate cell culturefactors to influence proliferation, dif-ferentiation, matrix synthesis, andfunction (Tenenbaum and Heersche,1982; Aubin et al. 1992; Stein andLian, 1993). One of the difficultieswith culturing bone cells, however, isthat osteocytes do not divide in situand require the presence of their ex-tracellular mineralized matrix tomaintain their differentiated stateand normal function (Kalajzic et al.,2004). In addition, differences in mat-uration of bone cells in vivo comparedto in vitro have been observed (Litvinet al., 2004). For this reason we havenot included expression data from cellculture work in Table 1. The lack ofmolecular data for transitional cellstages is highlighted in Table 1, al-though some data can be inferred fromthe expression pattern before and af-ter the intermediate stages (greyshading in Table 1).

THE TRANSFORMATIONPROCESS

The transformation from osteoblast toosteocyte occurs during several of theossification processes described. Weconcentrate on intramembranous os-sification, which although a continu-ous process, can be subdivided intotwo broad phases: (1) condensation ofcells and initial synthesis of collagenfibrils, and (2) polarized secretion ofbone matrix (Fig. 2).

Initially, a condensation of mesen-chymal cells (committed to an osteo-blast fate) develops, within which os-teoblasts differentiate and begin

depositing collagen type I fibres withsome type II and III. Collagen type IIImay form an initial framework onwhich bone deposition takes place(Carter et al., 1991). During fracturehealing, collagen type III forms a scaf-fold for the migration of osteoprogeni-tors and capillary in-growth, onlylater being replaced by collagen type I(Bierbaum et al., 2003). A view out-lined in most textbooks (e.g., Windle,1976, figs. 6–16), present in the mindsof most researchers, and based onstudies in mammals, is that the osteo-blasts within the condensation depositthis collagen in a polarized manner(i.e., matrix secretion occurs from onecell surface only) (Fig. 2B–D). What isnot clear is whether these polarizedosteoblasts are organized (Romer,1970; Windle, 1976) or not (Bloom andFawcett, 1969, fig. 10.20, also sug-gested by Ferretti et al., 2002) (com-pare Fig. 2B,C). It is also unknownwhether the very first cells to depositcollagen fibrils at the centre of an os-teogenic condensation (before theyline up along the bone surface) arealso polarized; they may secrete fibresfrom all cell surfaces (Fig. 2A). Never-theless, during this (first) phase of in-tramembranous ossification, osteo-blasts appear to undergo self-burial(Parfitt, 1990).

Recently, bone acidic glycopro-tein-75 was found to be expressed dur-ing the very early stages of intramem-branous osteogenesis and may play arole in defining condensed mesen-chyme regions (Gorski et al., 2004),together with tenascin C and otherextracellular matrix molecules (Halland Miyake, 2000).

Once this initial matrix (osteoid) islaid down, osteoblasts line up alongthe edge of the bone spicule to depositmore bone matrix, thus increasing thesize of the bone. Once the osteoid hasreached a particular thickness, addi-tional osteoblasts become trapped.This second phase of intramembra-nous ossification is discussed in theschemes below.

Possible Schemes of theTransformation ProcessWe consider four schemes for how anosteoblast could become trappedwithin bone matrix (Fig. 3A–D). Afifth scheme (Fig. 3E) represents the


likely way in which osteoblasts avoidbecoming trapped in bone matrix inthe osteocyte-deprived (acellular)bone of higher teleosts (Ekanayakeand Hall, 1987, 1988; Witten et al.,2004). These schemes are outlined be-low.

1. Osteoblasts are unpolarized andlay down bone in all directions.That is, the cells become trapped bytheir own secretions (Fig. 3A).

2. Individual osteoblasts are polar-ized (i.e., lay down bone in one di-rection only), but those within thesame generation (layer) are polar-ized differently to those within ad-jacent layers. Consequently, boneis deposited in all directions andosteoblasts become trapped (Fig.3B).

3. Osteoblasts of each generation arepolarized in the same direction: onegeneration “buries” the precedingone in bone matrix (Fig. 3C). Weare unaware of any evidence hav-ing been provided for this mecha-nism.

4. Within one generation, some osteo-blasts slow down their rate of bonedeposition or stop laying down boneso that they become trapped by thesecretion of their neighbouringcells (Fig. 3D). This scheme wasproposed by Palumbo et al. (1990)and Nefussi et al. (1991).

5. Osteoblasts are highly polarizedand function as a unit to lay downbone synchronously. All cells moveaway from the osteogenic front asbone matrix is deposited, ulti-

mately resulting in acellular bone(Fig. 3E) (Ekanayake and Hall,1987, 1988; Witten, 1997).

The Decision to TransformInto an Osteocyte:Morphological PerspectiveCandaliere et al. (2001) raised thepoint that differences in the develop-mental age of osteoblasts may accountfor the expression pattern of gene het-erogeneity between osteoblasts, suchthat, at least in calvaria, osteoblastsexpress different sets of genes as theyprogress along bone surfaces awayfrom the bone suture. This hypothesisimplies that osteoblasts of differentdevelopmental stages are presentalong bone surfaces; we know thatonly some osteoblasts make the tran-sition to osteocytes. The decision tofollow the osteocyte fate could verywell depend on or be controlled by thegene expression profile of the surfaceosteoblast (which is in contact withunderlying osteocytes). The tools totrack individual osteoblasts as theytransform into osteocytes are avail-able but need to be applied to the tran-sitional cell stages as we are only be-ginning to understand osteoblast geneheterogeneity.

It has also been proposed that oste-oclast signals may modulate osteo-blasts given the cross-talk betweenthese two cell types (Candaliere et al.,2001). Thus, a combination of commu-nication between embedded osteo-blasts, osteocytes, and osteoclasts, to-gether with unique gene profiles,could decide the fate of an osteoblast.Throughout the whole process oftransformation, the differentiatingcell remains in contact with cells inboth the osteoblast layer and with em-bedded osteocytes, even as morpholog-ical characters change (Palumbo etal., 1990).

According to Palumbo et al. (1990),the decision to transform into an os-teoblast is as follows. During bone for-mation, processes on the vascular sur-face of the osteocytes continue to growto enable osteocytes to remain in con-tact with the active osteoblast layerand to modulate their activity. Whenthese vascular-facing processes stopgrowing, they produce a signal thatinduces the recruitment of those os-teoblasts with which they are losing

Fig. 2. A diagram of the possible ways in which the first collagen fibres are laid down by cellswithin a condensation during the first phase of intramembranous ossification prior to lining up ona bone surface. Secreted collagen fibrils are shown in grey shading and the arrows indicate thedirection of secretion. A: The cells secrete collagen in all directions and are not polarized. B: Theindividual cells are polarized but secrete collagen in an apparently random fashion. C: Theindividual cells are polarized and secrete collagen in an organized fashion towards the centre of thecondensation. D: The situation is similar to C except that the cells are lined up when they secretecollagen fibres. See text for literature cited to support these possibilities.


contact. The committed osteoblastsare then transformed into osteoblasticosteocytes. The signal to stop growinga vascular process may by issued bythe osteoblasts with which they havecontact or it may be due to the gradualreduction in the vascular supply to theosteocytes as new layers of bone arelaid down on the osteogenic front. Ineither event, the active lifespan of theosteoblast (prior to entrapment) is in-dependent of the amount of bone ma-trix it produced (as shown in themouse periodontium by McCullochand Heersche, 1988).

The Decision to TransformInto an Osteocyte: MolecularPerspectiveSeveral factors have been reported tomodulate osteoblast function and/orare involved in controlling the deci-sion to transform into an osteocyte.Here we discuss some of the key play-ers.

The transcription factors, Runx2and Osterix, are crucial for osteoblastdifferentiation during both intramem-branous and endochondral ossification(Nakashima and de Crombrugghe,2003; Inohaya and Kudo, 2000; Koba-yashi and Kronenberg, 2005). Runx2/Cbfa1 directly activates a number ofosteoblast/osteocyte markers such astype I collagen, osteopontin, bone sia-loprotein, and osteocalcin (Aubin,1998; Inohaya and Kudo, 2000; forfurther discussions, see Eames et al.,2003). A homozygous mutation of thisgene in mice causes a complete lack ofbone formation with arrested osteo-blast differentiation (Komori andKishimoto, 1998; Otto et al., 1997).This transcription factor is, however,not expressed by preosteoblasts, nor isit expressed in osteo-chondroprogeni-tor cells (Nakashima and de Crom-brugghe, 2003), which are commonprecursor cells that have the potentialto differentiate along either the osteo-blast/osteocyte or chondroblast/chon-drocyte lineages (for more informationon this cell type, see Hall, 2005). Osxis probably downstream of Runx2since no Osx transcripts are detect-able in Runx2 null mice (Nakashimaand de Crombrugghe, 2003). Severalother transcription factors are also in-volved in osteoblast proliferation anddifferentiation—Msx1, Msx2, Dlx5,

Dlx6, Twist, Runx2, Osx, Sp3—buthave no bearing on our discussion onthe transformation from osteoblast toosteocyte.

Some important insights can begained from clinical studies on osteo-porosis and craniosynostoses. Leptinis a gene product synthesized by adi-

pocytes but that may serve an impor-tant signal to modulate osteoblastfunction (Gordeladze et al., 2002) andcould explain how obesity has a pro-tective effect on osteoporosis. Mice ho-mozygous for the obese gene have ab-normal glucose and fat metabolism,reduced stature, and do not produce

Fig. 3. A schematic of the four different schemes (A–D) by which osteoblasts could becomeentrapped in bone matrix as osteocytes and one scheme (E) that leads to bone without osteocytes(acellular bone). The column on the left is the situation before osteoblasts become embedded inosteoid and the column on the right depicts their positions in the osteoid during the next depositionphase. Arrows indicate the direction of matrix deposition by osteoblasts towards the osteogenicfront. Solid line indicates the bone surface. Black shaded cells indicate those osteoblasts that willbecome trapped in the next phase of matrix deposition as osteocytes (grey shading). The cell in D(column 1) stops or slows down matrix secretion and becomes “buried” by its neighbours (passive).The dotted line in E indicates the previous bone surface, which was displaced by the deposition ofacellular bone. In all five schemes osteoblasts line up along the bone surface. A: Osteoblastssecrete matrix in all directions. B: Each osteoblast is polarized in a different direction but stillsecretes matrix in one direction only. C: One generation of osteoblasts buries the next generation.D: The osteoblast to be embedded slows down its matrix production compared to neighbouringosteoblasts. E: Matrix secretion does not embed cells. See text for detailed discussion.


leptin. Recently, it has been shownthat leptin protects cells (including os-teoblasts) from apoptosis and facili-tates the transformation from osteo-blast to preosteocyte. A widely heldview amongst researchers who studyosteoporosis is that a decrease in boneformation is the result of a reductionin the lifespan of the osteoblast popu-lation. This reduced lifespan could bethe result of an early incorporationinto bone matrix (hence increase intransformation rate to becoming os-teocytes [Lips et al., 1978; Darby andMeunier, 1981; Dempster et al.,1983]).

A recent investigation by Borton etal. (2001) suggested that attenuatingTGF$-related signaling mechanismscan increase the propensity of an os-teoblast to mature into an osteocyte,and decrease the duration of its pro-ductive functioning by shortening itslifespan, ultimately resulting in os-teopenia. TGF$ is also produced byosteoblasts and is incorporated intobone matrix (Seyedin et al., 1986;Robey et al., 1987). The model thathas been put forward is that osteo-blast matrix production rate and os-teoblast incorporation into matrix asosteocytes is regulated in part byTGF$-related signaling mechanisms(Jordan et al., 2003).

Researchers studying craniosynos-toses (premature fusion of skull su-tures) have identified genes involvedin blocking osteoblast differentiation,namely FGFs and Wnt genes andSox2 (Masukhani et al., 2005). It isthought that FGFs may act via Wntgenes to inhibit osteoblast differentia-tion and that Sox2 is FGF-inducible.

Genetic models of bone-specificmarkers (such as knockouts, gain- andloss-of-function mutants, transgenicmice, etc.) can make important contri-butions to our understanding of thetransformation process. Recent stud-ies have identified some factors (e.g.,LRP5, sclerostin, Fgfr2c, matrix met-alloproteases, and glucocorticoids)that have an impact on osteoblast ap-optosis and/or activity (Babij et al2003; Winkler et al., 2003; Eswaraku-mar et al., 2004; Karsdal et al., 2004;O’Brien et al., 2004) and thus indi-rectly on the decision process to be-come an osteocyte (or not). Kalajzic etal. (2004) generated a transgenicmouse with a dentin matrix protein 1

cis-regulatory system that drivesgreen fluorescent protein as a markerfor living osteocytes. These studies(and others) provide us with potentialfuture models in which to explore thetransformation process in more detail.However, at present none of thesestudies shed light on the specificmechanism of how the transformationfrom osteoblast to osteocyte takesplace.

How Long Does It Take?The transformation from osteoblast toosteocyte takes about three days inthe femoral metaphysis of 2-week-oldrabbits (Owen, 1963). Young (1962)gives an estimate of 2 days for new-born rat tibiae and ribs, while Kember(1960) estimates 5 days for the sameanimals. In newborn mice, labelledwith tritiated thymidine, labelled os-teocytes are only seen from day 10 inthe periodontal surfaces of the molarroot, but in the periodontal alveolarbone osteoblast-osteocyte tranforma-tion is much longer, about 19 days. Itis clear from this study that the timeto transform is not consistent in allbones/sites of the same animal (Mc-Culloch and Heersche, 1988). At bothsurfaces, bone apposition rates weresimilar. That is, transformation rateis independent of bone appositionrate. Interestingly, Ten Cate andMills (1972) have shown that thesource of the progenitor cells for theperiodontal surface of the alveolarbone and of the osteoblasts lining theendosteal and periosteal surfaces inalveolar processes is not the same.That is, the origin of the osteoblastsmay influence the transformationtime to osteocytes. More recently, dif-ferent rates of bone formation havebeen reported along the human iliumof pre-menopausal women (Parfitt,1990) and during calvarial develop-ment in rats (Candaliere et al., 2001).

It is clear from the above studiesand others that even within one boneelement in a restricted population, dif-ferent rates of bone formation occur. Ithas also been shown that bone depo-sition rates are significantly affectedby environmental/experimental condi-tions, skeletal element, and age (Starkand Chinsamy, 2002). In summary,many recent studies investigate bonedeposition rates but few focus on the

osteoblast-osteocyte transformationother than the early studies men-tioned above, which indicate that thisprocess is site-specific depending onthe origin of the osteoblasts, the bone,and possibly the age and sex of theanimal but independent of the bonedeposition rate.

How Does It Happen?It is thought that osteoblasts controlthe initiation of mineralization of bonematrix by leaving behind matrix vesi-cles in the osteoid. As noted above,E11 is one marker expressed duringthe transition from osteoblast to os-teocyte. Recently it has been sug-gested that E11 is necessary for theformation of fully mineralized vesicleson the developing cellular processes ofosteoid-osteocytes (Barragan et al.,2004). Lengthening and narrowing ofcell processes results in the mineral-ized vesicles becoming associated withcollagen-mediated mineralization.

It has also been suggested that os-teoid-osteocytes may participate inmatrix secretion and mineralizationand may be involved in the orientationof collagen fibres (as are osteoblasts)(Palumbo, 1986). Palumbo (1986) alsosuggested that these functions of theosteoid-osteocyte are performed fromthe mineral-facing side; the vascularside does not have any cell processesat this stage of the osteoblast-to-osteo-cyte transformation (Fig. 1). This ob-servation implies that mineralizationof any osteoid matrix situated be-tween the osteoblast layer and theosteoid-osteocyte is dependent on theactivity of osteoblasts (or osteoblastic-osteocytes) and not on osteoid-osteo-cytes (Palumbo, 1986).

Perhaps one of the most detailedstudies on how osteoblasts become os-teocytes stems from the examinationof intramembranous bone formationin rat calvaria by transmission elec-tron microscopy. Based on their obser-vations, Nefussi et al. (1991) hypothe-sized the following mechanism forosteoblast entrapment: one of thealigned polarized osteoblasts de-creases its activity and loses its align-ment with neighbouring active osteo-blasts (now called osteoblasticosteocytes) (Fig. 1). These cells slowlybecome embedded in the matrixformed by the aligned osteoblasts but


maintain lateral cell contacts withthese cells. Concurrently, the preos-teoblast above the cell that is beingtransformed into an osteocyte (calledthe preosteoblastic osteoblast), movesin to occupy the space in the activeosteoblast layer so that the alignedlayer of active osteoblasts is not dis-rupted (Fig. 1). In other words, accord-ing to Nefussi et al. (1991), embeddingof an osteoblast is a passive process.

The osteoblastic osteocyte continuesits passive embedding procedure sothat by the end of this stage, this cellis completely out of the osteoblastlayer. This cell is now morphologicallydistinct—reduced cell size, decreasedorganelles—and is called an osteoid-osteocyte (Fig. 1). These authors couldnot deduce whether the osteoid be-tween this cell type and the osteoblastlayer was solely synthesized by theactive osteoblast layer or by cells inthe osteoid layer. Palumbo et al.(1990) concluded that the osteoblasticosteocyte (type I preosteocyte) and theosteoid osteocyte (type II preosteo-cyte) are still polarized towards themineral surface and produce matrix.Matrix production by type III preos-teocytes is minimal since mineralizingmatrix vesicles are never found on thevascular surface.

Palumbo et al. (1990), examiningthe parallel-fibred bone of newbornrabbit tibiae, also concluded that en-trapment was a passive process.Cellular changes previously reportedduring osteoblast-to-osteocyte trans-formation, such as decrease in cell sizeand organelle number (Palumbo,1986; Aarden et al., 1994), are indica-tive of a cell whose activity has de-clined and appear to support this hy-pothesis.

More recently, osteoblast entrap-ment was discussed by Ferretti et al.(2002) in the context of a TEM studyof osteoblasts in the calvariae andlong bones of newborn rabbits andembryonic chickens. They considerintramembranous bone formation asstatic osteogenesis because it is per-formed by stationary osteoblaststhat transform into osteocytes at thesite where they differentiated. Theydescribe cells irregularly arrangedin cords of two to three cell layers atthe site of new bone trabeculae for-mation. Each of these cells is polar-ized in a different direction with re-

spect to adjacent ones (as revealedby the position of cell organelles withrespect to the nucleus and the pres-ence of newly secreted type 1 colla-gen fibrils). Movable osteoblasts, po-larized in the same direction, thendifferentiate along the trabeculaepreviously laid down by the “station-ary” osteoblasts. So, according tothis study, intramembranous ossifi-cation is biphasic, consisting of a sta-tionary and a dynamic phase. Sur-prisingly, they do not describe acondensation stage prior to bone for-mation so it is difficult to interprettheir descriptions. Ferretti et al.(2002) found no significant struc-tural or ultrastructural differencesin stationary compared with dy-namic osteoblasts. In addition theyconclude that two different mecha-nisms of osteoblast-to-osteocytetransformation takes place: “station-ary” osteoblasts transform into os-teocytes by self-burial, whereas “dy-namic” osteoblasts are selected totransform into osteocytes by the se-cretory activity of neighbouring os-teoblasts.

CHOOSING A MODELDuring the first phase of intramem-branous bone formation (when preos-teoblasts are within condensations)(Dunlop and Hall, 1995; Hall, 2005),osteoblasts may deposit collagen in alldirections (Fig. 2A) (Ham and Cor-mack, 1979), may be polarized but ori-entated differently and thus depositcollagen in various directions (Fig. 2B)(Bloom and Fawcett, 1969, also sug-gested by Ferretti et al., 2002), may bepolarized and directed towards thecentre of the condensation (Fig. 2C),or may be polarized and aligned, thusdepositing collagen in one directiononly (Fig. 2D) (Romer, 1970; Windle,1976). The remaining question thatwe cannot as yet answer is whethercells committed to become osteoblastsmust also become polarized andaligned. To answer this question,more detailed histological studies onearly osteogenic condensations are re-quired. Molecular markers for earlycondensations such as Osterix, os-teopontin, osteonectin, and alkalinephosphatase (Liu et al., 1997; Aubin,1998), although also expressed duringlater stages, can certainly help to

identify cells within a condensationthat develop into osteoblasts. Dunlopand Hall (1995) showed that alkalinephosphatase is even activated beforethe condensation of osteoblast precur-sors during osteogenesis in the firstmandibular arch in chicks. At present,alkaline phosphatase remains the ear-liest marker for osteogenic condensa-tions and demonstration of the en-zyme accompanied by structural andultrastructural studies should revealif osteoblast precursors and early os-teoblasts are polarized and into whichdirection the first matrix is secreted(Fig. 2).

The second phase of intramembra-nous ossification is better understoodand appears to us far more coordi-nated in time and space. Osteoblastsare definitely polarized at this stage;the polarized activity of osteoblastsundergoing transformation into osteo-cytes has been shown based on theposition of three main cell organelles:nucleus, endoplasmic reticulum, andGolgi apparatus (Dudley and Spiro,1961; Pritchard, 1972; Palumbo,1986). But the question remains, areosteoblasts simply buried by the nextgeneration of osteoblasts (Fig. 3C) ordo they contribute to their own burialby stopping and/or slowing down ma-trix deposition before becoming em-bedded in the bone matrix (Fig. 3D).Nefussi et al. (1991) suggest that, in-deed, a combination of both processesmay occur.

Once a certain amount of osteoidhas been deposited, and we do notknow whether the amount is critical,but a minimum of osteoid is perhapsrequired to ensure the spacing of os-teocytes (osteocyte generations) in thebone matrix, osteoblasts becometrapped, probably by the mechanismfirst proposed by Nefussi et al. (1991)and outlined in scheme D (Fig. 3D).After all, it is evident that osteoblastscan slow down their matrix produc-tion, for example, when they trans-form into bone-lining cells or beforeundergoing apoptosis. Thus, it ap-pears reasonable to assume an osteo-blast destined to become an osteocyteslows down matrix production in com-parison to neighbouring osteoblastsand in this way would become pas-sively entrapped in the bone matrix.Even if a polarized osteoblast in theprocess of becoming trapped does not


stop matrix production completely butcontinues to secrete matrix at a slowerrate, the entrapment would still be apassive process, in the sense thatneighbouring cells are responsible forthe entrapment. As we know from ma-trix production by odontoblasts andosteoblasts, which form dentine andacellular bone, respectively, in ad-vanced bony fish (Acanthomorpha), apolarized and synchronized produc-tion of matrix by aligned cells does notlead to the embedding of cells into thematrix. If we accept scheme four (Fig.3D), then the developmental heteroge-neity of osteoblasts provides the em-bedding of some osteoblasts into thebone matrix.

PERSPECTIVEFuture research that focuses on themechanisms that underlie osteoblastheterogeneity might not only providea better understanding of how osteo-blasts become embedded in bone ma-trix but also contribute to our under-standing of synchronized processes,such as dentine and acellular bone for-mation, when cells do not become em-bedded in the matrix. There is so farno evidence to suggest that some os-teoblasts are predetermined to be-come embedded (and hence slow downmatrix production) within the newlysecreted osteoid, although Lanyon(1993) suggested that the osteoblaststhat secrete less matrix and, ulti-mately, become osteocytes may bethose with prior connections to under-lying osteocytes. The key to under-standing the transformation of osteo-blasts to osteocytes and how the latterare buried alive may reside in viewingall bone cells (preosteoblasts, osteo-blasts, bone-lining cells, osteocytes) asa continuum, as a tissue, with cells indifferent developmental stages. Tocharacterize these different develop-mental stages and to understand theirfunction in time and space, it will cer-tainly be helpful to close the gaps, re-garding the expression of molecularmarkers during the transformationprocess.

Currently, bone is often still viewedas a hard substance (although we callit hard tissue), with the focus on howthe mineralized material is depositedby osteoblasts. However, bone is a dy-namic tissue composed of living cells,

95% of which are osteocytes. In thisreview, we raised the question “howthe tissue called bone develops” in or-der to increase attention to osteocytes.

ACKNOWLEDGMENTSWe thank two anonymous reviewersfor critical comments on a previousversion of this manuscript. P.E.W.and B.K.H. acknowledge funding fromthe Canadian-German Science andTechnology Cooperation. P.E.W. ac-knowledges funding from the Deut-sche Forschungsgemeinschaft andfrom the European COST ACTIONB23-Oral facial development and re-generation. B.K.H. and T.F.O. ac-knowledge funding from NSERC.

REFERENCES

Aarden EM, Burger EH, Nijweide PJ.1994. Function of osteocytes in bone.J Cell Biochem 55:287–299.

Aarden EM, Wassenaar AM, Alblas MJ,Nijweide PJ. 1996. Immunocytochemicaldemonstration of extracellular matrixproteins in isolated osteocytes. Histo-chem Cell Biol 106:5:495–501.

Alcobendas M, Baud CA, Castanet J. 1991.Structural changes of the preosteocyticarea in Vipera aspis (L.) (Ophidia, Viper-idae) bone tissue in various physiologicalconditions. Calc Tissue Int 49:1:53–57.

Aubin JE. 1998. Bone stem cells. J CellBiochem Suppl 30/31:73–82.

Aubin JE, Liu F. 1996. The osteoblast lin-eage. In: Bilezikian JP, Raisz, LG, RodanGA, editors. The principles of bone biol-ogy. Toronto: Academic Press. p 51–67.

Aubin JE, Bellows CG, Turksen K, Liu F,Heersche JNM. 1992. Analysis of the os-teoblast lineage and regulation of differ-entiation. In: Slavkin H, Price P, editors.Chemistry and biology of mineralizedtissues. Amsterdam: Elsevier Science. p267–276.

Aubin JE, Turksen K, Heersche JNM 1993.Osteoblastic cell lineage. In: Noda M, ed-itor. Cellular and molecular biology ofbone. San Diego: Academic Press. p1–45.

Babij P, Zhao W, Small C, Kharode Y, Ya-worsky PJ, Bouxsein ML, Reddy PS,Bodine PVN, Robinson JA, Bhat B, Mar-zolf J, Moran RA, Bex F. 2003. High bonemass in mice expressing a mutant LRP5gene. J Bone Miner Res 18:960–974.

Banks WJ. 1974. The ossification process ofthe developing antler in the white-taileddeer (Odocoileus virginianus). Calc Tis-sue Res 14:257–274.

Barragan C, Nicolella D, Dusevich V, Dal-las M, Erick JD, Bonewald LF. 2004. Os-eoid-osteocyte MLO-A5 cells producemineralized spherical structures that as-sociate with, expand and engulf collagenfibrils: A proposed model for the miner-alisation of bone. Eighth International

Conference on the Chemistry and Biol-ogy of Mineralized Tissues. Alberta,Canada. October 17–24, 2004.

Baud CA. 1968. Submicroscopic structureand functional aspects of the osteocyte.Clin Orthop Rel Res 56:227–236.

Becker J, Schuppan D, Benzian H, Bals T,Hahn EG, Cantaluppi C, Reichart P.1986. Immunohistochemical distributionof collagens types IV, V, and VI and ofprocollagens Types I and III in humanalveolar bone and dentine. J HistochemCytochem 34:1417–1429.

Belanger LF. 1977a. The skeletal tissues.In: Weiss L, Greep RO, editors. Histol-ogy, 4th ed. New York: McGraw-HillBook Company.

Belanger LF. 1977b. Osteocytic osteolysisin a Cretaceous reptile. Rev Can Biol36:1:71–73.

Beresford WA. 1981. Chondroid bone, sec-ondary cartilage and metaplasia. Balti-more: Urban and Schwarzenberg.

Beresford WA. 1993. Cranial skeletal tis-sues: diversity and evolutionary trends.In: Hanken J, Hall BK, editors. Theskull, Vol. 2: Patterns of structural andsystematic diversity. Chicago: Univer-sity of Chicago Press. p 9–130.

Bianco P, Silvestrini G, Termine JD, Bo-nucci E. 1988. Immunohistochemical lo-calization of osteonectin in developinghuman and calf bone using monoclonalantibodies. Calcif Tissue Int 43:3:155–161.

Bianco P, Fisher LW, Young MF, Ter-mine JD, Gehron Robey P. 1990. Ex-pression and localization of the twosmall proteoglycans biglycan anddecorin in developing human skeletaland non-skeletal tissues. J HistochemCytochem 38:11:1549 –1563.

Bianco P, Riminucci M, Bonucci E, Ter-mine JD, Gehron Robey P. 1993. Bonesialoprotein (BSP) secretion and osteo-blast differentiation: relationship to bro-modeoxyuridine incorporation, alkalinephosphatase, and matrix deposition.J Histochem Cytochem 41:2:183–191.

Bierbaum S, Hempel U, Geißler U, HankeT, Scharnweber D, Wenzel K, Worch H.2003. Modification of Ti6AL4V surfacesusing collagen I, III, and fibronectin. II.Influence on osteoblast responses.J Biomed Mat Res 67A:2:431–438.

Bloom W, Fawcett D.W. 1969. A textbookof histology. Toronto: WB SaundersCompany. p 858.

Borton A, Frederick J, Datto M, Wang X-F,Weinstein R. 2001. The loss of SMAD3results in lower rate of bone formationand osteopenia through dysregulation ofosteoblast differentiation and apoptosis.J Bone Miner Res 16:1754–1764.

Boskey AL. 1996. Matrix proteins and min-eralisation: an overview. Connect TissueRes 35:357–363.

Boyde A. 1980. Evidence against “osteo-cytic osteolysis.” Metab Bone Dis Rel Res25:239–255.

Burger EH, Klein-Nulend J, Smit TH.2003. Strain-derived canalicular fluidflow regulates osteoclast activity in a re-


modelling osteon: a proposal. J Biomech36:1453–1459.

Buxton PG, Hall BK, Archer CW, Francis-West P. 2003. Secondary chondrocyte-de-rived Ihh stimulates proliferation ofperiosteal cells during chick develop-ment. Development 130:4729–4739.

Candaliere GA, Liu F, Aubin JE. 2001. In-dividual osteoblasts in the developingcalvaria express different gene reper-toires. Bone 28:351–361.

Caplan AI, Syftestad G, Osdoby P. 1983.The development of embryonic bone andcartilage in tissue culture. Clin OrthopRel Res 174:243–263.

Carter DH, Sloan P, Aaron JE. 1991. Im-munolocalization of collagen types I andIII, tenascin and fibronectin in in-tramembranous bone. J Histochem Cyto-chem 39:599–606.

Chen J, Singh K, Mukherjee BB, Sodek J.1993. Developmental expression of os-teopontin (OPN) mRNA in rat tissues:Evidence for a role for osteopontin inbone formation and resorption. Matrix13:113–123.

Clauss IM, Gravallese EM, Darling JM,Shapiro F, Glimcher MJ, Glimcher LH.1993. In situ hybridization studies sug-gest a role for the basic region-leucinezipper protein hxBP-1 in exocrine glandand skeletal development during mouseembryogenesis. Dev Dyn 197:146–156.

Currey JD. 2003. The many adaptations ofbone. J Biomech 36:1487–1495.

Darby AJ, Meunier PJ. 1981. Mean wallthickness and formation periods of tra-becular bone packets in idiopathic osteo-porosis. Calcif Tissue Int 33:199–204.

Delling G, Vogel M. 1992. Patomorpholo-gie der Osteoporose. In: Schild HH,Heller M, editors. Osteoporose. Stutt-gart: Georg Thieme Verlag.

Dempster DW, Arlot MA, Meunier PJ.1983. Mean wall thickness and forma-tion periods of trabecular bone packetsin corticosteroid-induced osteoporosis.Calcif Tissue Int 35:410–417.

Dodds RA, Merry K, Littlewood A, GowenM. 1994. Expression of mRNA for IL1$,IL6 and TGFß1 in developing humanbone and cartilage. J Histochem Cyto-chem 42:6:733–744.

Doty SB, Nunez EA. 1985. Activation ofosteoclasts and the repopulation of bonesurfaces following hibernation in the bat,Myotis lucifugus. Anat Rec 213:481–495.

Ducy P, Zhang R, Geoffrey V, Ridall AL,Karsenty G. 1997. Osf2/Cbfa1: s tran-scriptional activator of osteoblast differ-entiation. Cell 89:747–754.

Ducy P, Schinke T, Karsenty G. 2000. Theosteoblast: a sophisticated fibroblast un-der central surveillance. Science 289:1501–1504.

Dudley HR, Spiro D. 1961. The fine struc-ture of bone cells. J Biophys BiochemCytol 11:627–649.

Dunlop LT, Hall BK. 1995. Relationshipsbetween cellular condensation, preosteo-blast formation and epithelial-mesen-chymal interactions in initiation of osteo-genesis. Int J Dev Bio 39:357–371.

Eames BF, de la Fuente L, Helms JA.2003. Molecular ontogeny of the skele-ton. Birth Defects Res Part C 69:93–101.

Ekanayake S, Hall BK. 1987. The develop-ment of acellularity of the vertebral boneof the Japanese Medaka, Oryzias latipes(Teleostei,Cyprinidontidae).JMorph193:253–261.

Ekanayake S, Hall BK. 1988. Ultrastruc-ture of the osteogenesis of acellular ver-tebral bone in the Japanese Medaka,Oryzias latipes (Teleostei, Cyprinidonti-dae). Am J Anat 182:241–249.

Eswarakumar VP, Horowitz MC, LocklinR, Morris-Kay GM, Lonai P. 2004. Again-of-function mutation of Fgfr2c dem-onstrates the roles of this receptor vari-ant in osteogenesis. PNAS 101:12555–12560.

Ferretti M, Palumbo C, Contri M, MarottiG. 2002. Static and dynamic osteogene-sis: two different types of bone forma-tion. Anat Embryol 206:21–29.

Fox SW, Chow JWM. 1998. Nitric oxidesynthase expression in bone cells. Bone23:1–6.

Frost HM. 1960. In vivo osteocyte death.J Bone Joint Surg 42A:138–143.

Frost HM. 1963. Bone remodelling dynam-ics. Springfield, IL: Thomas Publishers.

Gegenbauer C. 1864. Untersuchungen zurvergleichenden Anatomie der Wirbelt-iere. Heft 1. Carpus und Tarsus. Leipzig:Schultergtel.

Gehron Robey P, Fedarko NS, HefferanTE, Bianco P, Vetter UK, Grzesik W,Friedenstein A, van der Plum G, MintzKP, Young MF, Kerr JM, Ibaraki K, Hee-gaard A. 1993. Structural and molecularregulation of bone matrix proteins.J Bone Miner Res 8:S2:S483–S487.

Gerstenfeld LC, Shapiro FD. 1996. Expres-sion of bone-specific genes by hypertro-phic chondrocytes: implication of thecomplex functions of the hypertrophicchondrocyte during endochondral bonedevelopment. J Cell Biochem 62:1–9.

Gordeladze JO, Drevon C, Syversen U, Re-seland JE. 2002. Leptin stimulates hu-man osteoblastic cell proliferation, denovo collagen synthesis, and mineralisa-tion: impact on differentiation markers,apoptosis, and osteoclastic signaling.J Cell. Biochem 85:825–836.

Gorski J, Wang A, Lovitch D, Law D, Pow-ell K, Midura RJ. 2004. Extracellularbone acidic glycoprotein-75 defines con-densed mesenchyme regions to be min-eralized and localizes with bone sialopro-tein during intramembranous boneformation. J Biol Chem 279:25455–25463.

Grzesik WJ, Gehron Robey P. 1994. Bonematrix RGD glycoproteins: Immunolo-calization and interaction with humanprimary osteoblastic bone cells in vitro.J Bone Min Res 9:487–496.

Hadjiargyrou M, Rightmire EP, Ando T,Lombardo FT. 2001. The E11 osteoblastlineage marker is differentially ex-pressed during fracture healing. Bone 29:49–154.

Hall BK, editor. 1990. Bone, Volume 1: theosteoblast and osteocyte. Boca Raton,FL: Telford Press and CRC Press. p 494.

Hall BK. 2005. Bones and cartilage. Devel-opmental and evolutionary skeletal biol-ogy. London: Elsevier/Academic Press.

Hall BK, Miyake T. 1995. Divide, accumu-late, differentiate: cell condensation inskeletal development revisited. Int J DevBiol 39:881–893.

Hall BK, Miyake T. 2000. All for one andone for all: condensations and the initia-tion of skeletal development. BioEssays22:138–147.

Hall BK, Witten PE. 2005. The origin andplasticity of skeletal tissues in verte-brate evolution and development. In:Anderson JS, Sues H-D, editors. Majortransitions in vertebrate evolution.Festschrift for Dr. Robert L. Carroll.Bloomington, IN: Indiana UniversityPress (in press).

Haller AC, Zimny ML. 1978. Effects of hi-bernation on interradicular alveolarbone. J Dent Res 56:1552–1557.

Ham AW, Cormack DH. (eds.). 1979. His-tophysiology of cartilage, bone andjoints. Toronto: JB Lippincott Company.p 485.

Hancox NM, Boothroyd B. 1965. Electronmicroscopy of the early stages of osteo-genesis. Clin Orthopaed Relat Res 40:153–161.

Heersche JNM, Aubin JE. 1990. Regula-tion of cellular activity of bone-formingcells. In: Hall BK, editor. Bone, Vol. 1:the osteoblast and osteocyte. Caldwell,NJ: Telford Press. p 327–350.

Heersche JNM, Reimers SM, Wrana JL,Waye MM, Gupta AK. 1992. Changes inexpression of alpha 1 type 1 collagen andosteocalcin mRNA in osteoblasts andodontoblasts at different stages of matu-rity as shown by in situ hybridization.Proc Finn Dent Soc 88(Suppl.1):173–182.

Holtrop ME. 1990. Light and Electron mi-croscopical structure of bone formingcells. In: Hall BK, editor. Bone, Vol. 1:the osteoblast and osteocyte. Caldwell,NJ: The Telford Press. p 1–40.

Horiuchi K, Amizuka N, Takeshita S,Takamatsu H, Katsuura M, Ozawa, H,Toyama Y, Bonewald LF, Kudo A. 1999.Identification and characterization of anovel protein, periostin, with restrictedexpression to periosteum and periodon-tal ligament and increased expression bytransforming growth factor $. Bone MinRes 14:1239–1249.

Hughes DR, Bassett JR, Moffat LA. 1994a.Histological identification of osteocytesin the allegedly acellular bone of the seabreams Acanthopagrus australis, Pagrusauratus and Rhabdosargus sarba (Spari-dae, Perciformes, Teleostei). Anat Em-bryol Berl 190:163–179.

Hughes DE, Salter DM, Simpson R. 1994b.CD44 expression in human bone: a novelmarker of osteocytic differentiation.J Bone Min Res 9:39–44.

Huysseune A, Verraes W. 1990. Carbohy-drate histochemistry of mature chon-droid bone in Astatotilapia elegans. (Te-leostei: Cichlidae) with a comparison to


acellular bone and cartilage. Ann Sci Na-tur Zool Paris 13:29–43.

Ikeda T, Nomura S, Yamaguchi A, Suda T,Yoshiki S. 1992. In situ hybridisation ofbone matrix proteins in undecalcifiedadult rat bone sections. J Histochem Cy-tochem 40:1079–1088.

Ikeda T, Nagai Y, Yamaguchi A, Yokose S,Yoshiki S. 1995. Age-related reduction inbone matrix protein mRNA expression inrat bone tissues: application of histomor-phometry to in situ hybridization. Bone16:17–23.

Inada M, Yasui T, Nomura S, Miyake S,Deguchi K, Himeno M, Sato M, Yam-agiwa H, Kimura T, Yasui N, Ochi T,Endo N, Kitamura Y, Kishimoto T,Komori T. 1999. Maturational distur-bance of chondrocytes in Cbfa1-deficientmice. Dev Dyn 214: 279–290.

Inohaya K, Kudo A. 2000. Temporal andspatial expression of cbfa1 expressionduring embryonic development in the te-leost, Oryzias latipes. Dev Genes Evol210:570–574.

Ishidou Y, Kitajima I, Obama H, Ma-ruyama I, Murata F, Imamura T,Yamada N, Ten Dijke P, Miyazono K,Sakou T. 1995. Enhanced expression oftype I receptors for bone morphogeneticproteins during bone formation. J BoneMiner Res 10:1651–1659.

Jamal HH, Aubin JE. 1996. CD44 expres-sion in fetal rat bone: in vivo and in vitroanalysis. Exp Cell Res 223:467–477.

Jilka RL, Weinstein RS, Bellido T, ParfittAM, Manolagas SC. 1998. Osteoblastprogrammed cell death (apoptosis): mod-ulation by growth factors and cytokines.JBMR 13:5:793–802

Jordan GR, Loveridge N, Power J, ClarkeMT, Parker M, Reeve J. 2003. The ratioof osteocytic incorporation to bone ma-trix formation in femoral neck cancellousbone: an enhanced osteoblast work ratein the vicinity of hip osteoarthritis. Cal-cif Tissue Int 72:190–196.

Kalajzic I, Braut A, Guo D, Jiang X, Kro-nenberg MS, Mina M, Harris MA, HarrisSE, Rowe DW. 2004. Dentin matrix pro-tein 1 expression during osteoblastic dif-ferentiation, generation of an osteocyteGFP-transgene. Bone 35:74–82.

Kamiya N, Shigemasa K, Takagi M. 2001.Gene expression and immunohistochem-ical localization of decorin and biglycanin association with early bone formationin the developing mandible. J Oral Sci43:179–188.

Karsdal MA, Andersen TA, Bonewald L,Christiansen C. 2004. Matrix metallo-proteinases (MMPs) safeguard osteo-blasts from apoptosis during transdiffer-entiation into osteocytes: MT1-MMPmaintains osteocyte viability. DNA CellBiol 23:155–165.

Kato Y, Boskey A, Spevak L, Dallas M,Hori M, Bonewald LF. 2001. Establish-ment of an osteoid preosteocyte-like cellMLO-A5 that spontaneously mineralisesin culture. J Bone Miner Res 16:1622–1635.

Kember NF. 1960. Cell division in endo-chondral ossification: a study of cell pro-

liferation in rat bones by the method oftritiated thymidine autoradiography.J Bone Joint Surg 42B:824–39.

Knese KH. 1979. Stuzgewebe und Skelett-system. Handbuch der mikroskopischenAnatomie des Menschen. Bd. 2. DieGewebe, Teil 5. Berlin: Springer Verlag.p 938.

Knothe Tate ML, Adamson JR, Tami AE,Bauer TW. 2004. The osteocyte. IntJ Biochem Cell Biol 36:1–8.

Kobayashi T, Kronenberg H. 2005. Minire-view: transcriptional regulation in devel-opment of bone. Endocrinology 146:1012–1017.

Komori T, Kishimoto T. 1998. Cbfa1 inbone development. Curr Opin Genet Dev8:494–499.

Kwiecinski GG. 1985. Bone remodellingand its regulation in Myotis lucifugus.Prog Zool 30:45–47.

Kwiecinski GG, Krook L, Wimsatt WA.1987. Annual skeletal changes in the lit-tle brown bat, Myotis lucifungus lucifun-gus, with particular references to preg-nancy and lactation. Am J Anat 178:410–420.

Lacey DL, Timms E, Tan HL, Kelley MJ,Dunstan CR, Burgess T, Elliott R, Co-lombero A, Elliott G, Scully S, Hsu H,Sullivan J, Hawkins N, Davy E, Cappar-elli C, Eli A, Qian YX, Kaufman S, SarosiI, Shalhoub V, Senaldi G, Guo J, DelaneyJ, Boyle WJ. 1998. Osteoprotegerin li-gand is a cytokine that regulates oste-oclast differentiation and activation. Cell93:165–176.

Lanyon LE. 1993. Osteocytes, strain detec-tion, bone modelling and remodelling.Calc Tiss Int 53(Suppl 1):s102–s107.

Lazowski D-A, Fraher LJ, Hodsman A,Steer B, Modrowski D, Han VKM. 1994.Regional variation of insulin-like growthfactor-1 gene expression in mature ratbone and cartilage. Bone 15:563–576.

Lee K, Deeds, JD, Bond, AT, Juppner H,Abou-Samra A-B, Segre GV. 1993. Insitu localisation of PTH/PTHrP receptormRNA in the bone of fetal and youngrats. Bone 14:341–345.

Li M, Amizuka N, Oda K, Tokunaga K, ItoT, Takeuchi K, Takagi R, Maeda T. 2004.Histochemical evidence of the initialchondrogenesis and osteogenesis in theperiosteum of a rib fractured model: im-plications of osteocyte involvement inperiosteal chondrogenesis. Microsc ResTech 64:330–342.

Lips P, Courpron P, Meunier PJ. 1978.Mean wall thickness and formation peri-ods of trabecular bone packets in the hu-man iliac crest: changes with age. CalcifTissue Res 26:13–17.

Litvin J, Selim A-H, Montgomery MO, Le-hmann K, Rico MC, Devlin H, Bednarik,DP, Safadi FF. 2004. Expression andfunction of periostin-isoforms in bone.J Cell Biochem 92:1044–1061.

Liu F, Malaval L, Aubin JE. 1997. Themature osteoblast phenotype is charac-terized by extensive plasticity. Exp CellRes 232:97–105.

Lopez E, Macintyre I, Martelly E, Lallier F,Vidal B. 1980. Paradoxical effect of 1,25

dihydroxycholecalciferol on osteoblasticand osteoclastic activity in the skeletonof the eel Anguilla anguilla L. Calc Tis-sue Int 32:83–87.

Machwate M, Jullienne A, Moukhtar M,Marie PJ. 1995. Temporal variation ofc-fos proto-oncogene expression duringosteoblast differentiation and osteogene-sis in developing rat bone. J Cell Biol57:62–70.

Malaval L, Modrowski D, Gupta AK, AubinJE. 1994. Cellular expression of bone-related proteins during in vitro osteogen-esis in rat bone marrow stromal cell cul-tures. J Cell Physiol 158:555–572.

Manolagas SC. 2000. Birth and death ofbone cells: basic regulatory mechanismsand implications for pathogenesis andtreatment of osteoporosis. EndocrinolRev 21:115–137.

Mark MP, Butler WT, Prince CW, Finkel-man RD, Ruch JV. 1988. Developmentalexpression of 44-kDa bone phosphopro-tein (osteopontin) and bone: carboxyglu-tamic acid (Gla)-containing protein (os-teocalcin) in calcifying tissues of rat.Differentiation. 37:123–136.

Marotti G. 1977. Decrement in volume ofosteoblasts during osteon formation andits effect on the size of the correspondingosteocytes. In: Meunier PJ, editor. Bonehistomorphometry. Paris: Armour Mon-tagu. p 385–397

Marotti G. 1996. The structure of bone tis-sues and the cellular control of their dep-osition. Italian J Anat Embryol 101:25–79.

Marotti G, Cane V, Palazzini S, PalumboC. 1990. Structure-function relation-ships of the osteocyte. Italian J MinerElectrolyte Metab 4:93–106.

Masukhani A, Ambrosetti D, Holmes G,Cornivelli L, Basilico C. 2005. Sox2 in-duction by FGF and FGFR2 activitingmutations inhibit Wnt signaling and os-teoblast differentiation. J Cell Biol 168:1065–1076.

Martineau-Doize B, Lai WH, WarshawskyH, Bergeron JJM. 1988. In vivo demon-stration of cell types in bone that harborepidermal growth factor response. Endo-crinology 123:841–858.

McCulloch CAG, Heersche JNM. 1988.Lifetime of the osteoblast in mouse peri-odontium. Anat Rec 222:128–135.

Meunier FJ. 1989. The acellularisationprocess in Osteichthyan bone. FortschrZool 35:443–446.

Middleton J, Arnott N, Walsh S, BeresfordJ. 1995. Osteoblasts and osteoclasts inadult human osteophyte tissue expressmRNAs for insulin-like growth factors Iand II and the type 1 IGF receptor. Bone16:287–293.

Miller SC, Bowman BM, Smith JM, JeeWSS. 1980. Characterisation of en-dosteal bone-lining cells from fatty mar-row bone sites in adult beagles. Anat Rec198:163–173.

Mizoguchi I, Takahashi I, Sasano Y, Ka-gayama M, Kuboki Y, Mitani H. 1997.Localization of types I, II and X collagenand osteocalcin in intramembranous, en-


dochondral and chondroid bone of rats.Anat Embryol 196:291–297.

Nah HD, Pacifici M, Gerstenfeld LC, Ad-ams SL, Kirsch T. 2000. Transient chon-drogenic phase in the intramembranouspathway during normal skeletal develop-ment. J Bone Miner Res 15:522–533.

Nakashima K, de Crombrugghe B. 2003.Transcriptional mechanisms in osteo-blast differentiation and bone formation.Trends Genet 19:458–466.

Nampei A, Hashimoto J, Hayashida K,Tsuboi H, Shi K, Tsuji I, Miyashita H,Yamada T, Matsukawa N, MatsumotoM, Morimoto S, Ogihara T, Ochi T, Yo-shikawa H. 2004. Matrix extracellularphosphoglycoprotein (MEPE) is highlyexpressed in osteocytes in human bone.J Bone Miner Metab 22:176–184.

Nefussi JR, Sautier JM, Nicolas V, ForestN. 1991. How osteoblasts become osteo-cytes: a decreasing matrix forming pro-cess. J Biol Buccale 19:75–82.

Nijweide PJ, Van der Plas A, Scherft JP.1981. Biochemical and histological stud-ies on various bone cell preparations.Calc Tiss Int 33:529–540.

Noble BS, Stevens H, Loveridge N, ReeveJ. 1997. Identification of apoptoticchanges in osteocytes in normal andpathological human bone. Bone 20:73–282.

O’Brien CA, Jia D, Plotkin Li, Bellido T,Powers CC, Stewart SA, Manolagas SC,Weinstein RS. 2004. Glucocorticoids actdirectly on osteoblasts and osteocytes toinduce their apoptosis and reduce boneformation and strength. Endocrinology154:1835–1841.

Ørvig T. 1951. Histological studies of pla-coderms and fossil elasmobranches: I.The endoskeleton with remarks on thehard tissues of lower vertebrates in gen-eral. Arkin Zool 2:321–456.

Otto F, Thornell AP, Crompton T, DenzelA, Gilmour KC, Rosewell IR, StampGWH, Beddington RSP, Mundlos S, Ol-sen BR. 1997. Cbfa1, a candidate genefor cleidocranial dysplasia syndrome, isessential for osteoblast differentiationand bone development. Cell 89:765–771.

Owen M. 1963. Cell Population kinetics ofan osteogenic tissue II. J Cell Biol 19:19–32.

Palumbo C. 1986. A three-dimensional ul-trastructural study of osteoid-osteocytesin the tibia of chick embryos. Cell TissueRes 246:125–131.

Palumbo C, Palazzini S, Zappe D, MarottiG. 1990. Osteocyte differentiation in thetibia of newborn rabbits: an ultrastruc-tural study of the formation of cytoplas-mic processes. Acta Anat 137:350–358.

Parfitt AM. 1990. Bone-forming cells inclinical conditions. In: Hall BK, editor.Bone, Vol. 1: the osteoblast and osteo-cyte. Boca Raton, FL: Telford Press andCRC Press. p 351–429.

Pritchard JJ. 1972. General histology ofboneIn, In: Bourne GH, editor. The bio-chemistry and physiology of bone. NewYork: Academic Press. p 1–20.

Roach HI. 1990. Long-term organ cultureof embryonic chick femora: a system for

investigating bone and cartilage forma-tion at an intermediate level of organiza-tion. J Bone Miner Res 5:85–100.

Roach HI. 1992. Trans-differentiation ofhypertrophic chondrocytes into cells ca-pable of producing a mineralized bonematrix. Bone Miner 19:1–20.

Roach HI. 1994. Why does bone matrix con-tain non-collagenous proteins? The pos-sible roles of osteocalcin, osteonectin, os-teopontin, and bone sialoprotein in bonemineralization and resorption. Cell BiolInt 18:617–628.

Robey PG, Young MF, Flanders KC, RocheNS, Kondaiah P, Reddi AH, Termine JD,Sporn MB, Roberts AB. 1987. Osteo-blasts synthesize and respond to trans-forming growth factor-type beta (TGF$)in vitro. J Cell Biol 105:457–463.

Romanowski R, Jundt G, Termine JD, vonder Mark K, Schulz A. 1990. Immuno-electron microscopy of osteonectin andtype I collagen in osteoblasts and bonematrix. Calcif Tissue Int 46:353–360.

Romer AS. 1970. The vertebrate body. To-ronto: WB Saunders Company. p 601.

Rouleau MF, Mitchell J, Goltzman D.1988. In vivo distribution of parathyroidhormone receptors in bone: evidence thata predominant osseous target cell is notthematureosteoblast.Endocrinology123:87–191.

Sandberg M, Autio-Harmainen H, VuorioE. 1988. Localization of the expression oftypes I, III, and IV collagen, TGF-$1 andc-fos genes in developing human calvar-ial bones. Dev Biol 130:324–334.

Sasano Y, Li H-C, Zhu J-X, Imanaka-Yo-shida K, Mizoguchi I, Kagayama M.2000. Immunohistochemical localizationof type I collagen, fibronectin, and tena-scin during embryonic osteogenesis inthe dentary of mandibles and tibias ofrats. Histochem J 32:591–598.

Schulze E, Witt M, Kasper M, Lowik CW,Funk RH. 1999. Immunohistochemicalinvestigations on the differentiationmarker protein E11 in rat calvaria, cal-varia cell culture and the osteoblasticcell line ROS 17/2.8. Histochem Cell Biol111:61–69.

Scott-Savage P, Hall BK. 1980. Differentia-tive ability of the tibial periosteum fromthe embryonic chick. Acta Anat 106:129–140.

Semba T, Tolnal S, Belanger LF. 1966. Ob-servations on the fine structure of chickembryo osteocytes: effects of parathyroidextract. In: Sixth International Confer-ence for electron microscopy, Kyoto. To-kyo: Maruzen Co, Ltd. p 569

Seyedin SM, Thompson AY, Bentz H,Rosen DM, McPherson JM, Conti A, Sie-gel NR, Galluppi GR, Piez KA. 1986.Cartilage-inducing factor-A. Apparentidentity to transforming growth factor-beta. J Biol Chem 261:5693–5695.

Stark JM, Chinsamy A. 2002. Bone micro-structure and developmental plasticityin birds and other dinosaurs. J Morphol254:232–246.

Stein GS, Lian JB. 1993. Molecular mech-anisms mediating proliferation/differen-tiation interrelationships during pro-

gressive development of the osteoblastphenotype. Endocr Rev 14:424–442.

Steinberg B, Singh IJ, Mitchell OG. 1981.The effects of cold-stress, hibernationand prolonged inactivity on bone dynam-ics in the Golden Hamster, Mesodricetusauratus. J Morph 167:43–51.

Taylor LH, Hall BK, Miyake T, Cone DK.1994. Ectopic ossicles associated withmetacercariae of Apophallus brevis(Trematoda) in yellow perch, Perca fla-vescens (Teleostei): development andidentification of bone and chondroidbone. Anat Embryol 190:29–46.

Tazawa T, Hoshi K, Kawamoto S, TanakaM, Ejiri S, Ozawa H. 2004. Osteocyticosteolysis observed in rats to which para-thyroid hormone was continuously ad-ministered. J Bone Miner Metab 22:6:524–529.

Ten Cate AR, Mills C. 1972. The develop-ment of the periodontium: the origin ofalveolar bone. Anat Rec 173:69–78.

Tenenbaum HC, Heersche JNM. 1982. Dif-ferentiation of osteoblasts and formationof mineralized bone in vitro. Calcif Tis-sue Int 34:76–79.

Thesingh CW, Groot CG, Wassenaar AM.1991. Transdifferentiation of hypertro-phic chondrocytes into osteoblasts in mu-rine fetal metatarsal bones, induced byco-cultured cerebrum. Bone Miner Res12:25–40.

Turksen K, Aubin JE. 1991. Positive andnegative immunoselection for enrich-ment of two classes of osteoprogenitorcells. J Cell Biol 114:373–384.

Vakeva L, Mackie E, Kantomaa T, ThesleffI. 1990. Comparison of the distributionpatterns of tenascin and alkaline phos-phatase in developing teeth, cartilage,and bone of rats and mice. Anat Rec 228:69–76.

Vickaryous MK, Olson WM. 2005. Over-looked skeletal elements of the limb skel-eton; sesamoids, lunulae and other ap-pendicular ossicles. In: Hall BK, editor.Fins and limbs. Chicago: Chicago Uni-versity Press (in press).

Wang E, Wang J, Chin E, Zhou J, BondyCA. 1995. Cellular patterns of insulin-like growth factor system gene expres-sion in murine chondrogenesis and os-teogenesis. Endocrinology 136:2741–2751.

Wetterwald A, Hoffstetter W, CecchiniMG, Lanske B, Wagner C, Fleisch H,Atkinson M. 1996. Characterization andcloning of the E11 antigen, a marker ex-pressed by rat osteoblasts and osteo-cytes. Bone 18:125–132.

Windle WF. 1976.Textbook of histology,5th ed. Toronto: McGraw-Hill Book Com-pany. P 561.

Winkler DG, Sutherland MK, GeogheganJC, YU CP, Hayes T, Skonier JE, Shpe-ktor D, Jonas M, Kovacevich BR, Staeh-ling-Hampton K, Appleby M, BrunkowME, Latham JA. 2003. Osteocyte controlof bone formation via sclerostin, a novelBMP antagonist. EMBO J 22:6267–6276.

Witten PE. 1997. Enzyme histochemicalcharacteristics of osteoblasts and mono-


nucleated osteocytes in a teleost fishwith acellular bone (Oreochromis niloti-cus, Cichlidae). Cell Tissue Res 287:591–599.

Witten PE, Hall BK. 2002. Differentiationand growth of kype skeletal tissues inanadromous male Atlantic salmon (Salmosalar). Int J Dev Biol 46:719–730.

Witten PE, Hall BK. 2003. Seasonalchanges in the lower jaw skeleton inmale Atlantic salmon (Salmo salar L.):remodelling and regression of the kypeafter spawning. J Anat 203:435–450.

Witten PE, Huysseune A. 2005. Mecha-nisms of chondrogenesis and osteogene-sis in fins. In: Hall BK (ed.) Fins andLimbs. Chicago Univ. Press, Chicago. Inpress.

Witten PE, Villwock W, Peters N, Hall BK.2000. Bone resorption and bone remod-eling in juvenile carp (Cyprinus carpio).J Appl Ichthy 16:254–261

Witten PE, Hansen A, Hall BK. 2001. Fea-tures of mono and multinucleated boneresorbing cells of the zebrafish Danio re-rio and their contribution to skeletal de-velopment, remodeling and growth. JMorph 250:197–207.

Witten PE, Huysseune A, Franz-OdendaalTA, Fedak T, Vickaryous M, Cole A, HallBK. 2004. Acellular teleost bone: primi-tive or derived, dead or alive? PalaeontolNewslett 55:37-41.

Witten PE, Gil-Martens L, Hall BK,Huysseune A, Obach A. 2005. Com-pressed vertebrae in Atlantic salmon(Salmo salar): evidence for metaplasticchondrogenesis as a skeletogenic re-sponse late in ontogeny. Dis Aquat Or-gan 64:237–246.

Wong M, Lawton T, Goetinck PF, Kuhn JL,Goldstein SA, Bonadio J. 1992. Aggrecancore protein is expressed in membranousbone of the chick embryo. Molecular and

biochemical studies of normal andnanomelic embryos. J Biol Chem 267:5592–5598.

Yasuda H, Shima N, Nakagawa N, Mo-chizuki SI, Yano K, Fujise N, Sato Y,Goto M, Yamaguchi K, KuriyamaM, Kanno T, Murakami A, Tsuda E,Morinaga T, Higashio K. 1998. Iden-tity of osteoclastogenesis inhibitoryfactor (OCIF) and osteoprotegerin(OPG): a mechanism by which OPG/OCIF inhibits osteoclastogenesis invitro. Endocrinology 139:1329 –1337.

Young RW. 1962. Cell proliferation andspecialisation during endochondral os-teogenesis in young rats. J Cell Biol 14:357–370.

Zhang X, Bao K, Dai K. 2000. Stress-relaxation plates and the remodeling ofcallus and cortex under the plate inrabbits. Chin Med J (Engl) 113:9:805–809.


Buried alive: How osteoblasts become osteocytesucgatma/Anat3048/PAPERS...Buried Alive: How...

Documents

Transcript of Buried alive: How osteoblasts become osteocytesucgatma/Anat3048/PAPERS...Buried Alive: How...