Scanning Microscopy

Volume 1 Number 1 Article 34

10-12-1986

Early Scanning Electron Microscopic Studies of Hard Tissue Resorption: Their Relation to Current Concepts Reviewed

A. Boyde University College London

S. J. Jones University College London

Follow this and additional works at: https://digitalcommons.usu.edu/microscopy

Part of the Life Sciences Commons

Recommended Citation Boyde, A. and Jones, S. J. (1986) "Early Scanning Electron Microscopic Studies of Hard Tissue Resorption: Their Relation to Current Concepts Reviewed," Scanning Microscopy: Vol. 1 : No. 1 , Article 34. Available at: https://digitalcommons.usu.edu/microscopy/vol1/iss1/34

This Article is brought to you for free and open access by the Western Dairy Center at DigitalCommons@USU. It has been accepted for inclusion in Scanning Microscopy by an authorized administrator of DigitalCommons@USU. For more information, please contact [email protected]. Scanning Microscopy, Vol. 1, No. 1, 1987 (Pages 369-381) 0981-7035/87$3.oo+.oo Scanning Microscopy International, Chicago (AMF O'Hare), IL 60666 USA

EARLY SCANNINGELECTRON MICROSCOPIC STUDIESOF HARDTISSUE RESORPTION: THEIR RELATIONTO CURRENTCONCEPTS REVIEWED

A Boyde * and SJ Jones

Department of Anatomy and Embryology, University College London, Gower St., London WClE 6BT, England

(Received for publication June 06, 1986, and in revised form October 12, 1986)

Abstract Introduction

This paper highlights some observations This paper was presented as the lead-off made by the authors in SEM studies of hard tissue talk at a meeting (Biomineralization: Edward J. resorption and considers their significance in Reith Memorial Meeting on Scientific Investiga­ relation to current concepts. All mammalian tion of Vertebrate ) held at mineralised tissues may undergo physiological New Orleans, May 5-7, 1986. This meeting was resorption, the resulting surface reflecting the organized with the purpose of reviewing density of mineralisation and the organic matrix some of the major areas in which the SEM has chemistry, organisation and orientation. contributed to understanding of hard tissue Resorption-repair coupling may follow the structure and function in vertebrates. The resorption of any tissue, but SEM studies first present paper sets out to examine one small facet noted this process in the case of the dental of this microscopic-method-well-suited-to-the­ tissues. The difference between fetal and adult problem area, and from one direction only. We boneformation and resorption provided evidence have taken a second look at our own earlier SEM against the concept of osteocytic osteolysis. SEM data to see how it has affected our viewpoints stereophotogrammetric methods for the quantita­ concerning the natural history of resorption, and tion of individual resorption lacunae are now how the SEM study slots in with other recent cell much quicker and have been extended to the study biological studies. of in vitro resorption by mammalian and avian osteocl asts isolated from and seeded into The morphology£!:_ resorption new substrates. Experimental studies using SEM were first The phenomenon of osteocl as tic resorption occurs in all the calcified tissues: not only conducted on the osteotropic hormonal effects on bone, and calcified , but all the dental forming in vivo and extended to the in tissues. The first published SEM study of vitro situation. The effects observed underlined resorption concerned the dentine and enamel of the several actions of PTH on and human deciduous teeth . 3-D images of resor­ indicated their important role in the control of ption surfaces at a range of magnifications bone resorption. Immunological marking techni­ showed that differential rates of removal reflec­ ques monitored by SEM first established that ted local compositional differences in these had no Fe or receptors, although c3 tissues. For example, within the "floors" of other cells in the vicinity did. individual resorption lacunae, peritubular The study of osteoclasts resorbing dentine (PTO) stood proud of the surrounding, substrates other than bone in vitro has increased mineralized- intertubular dentine. In our understanding of the essential components of enamel, prism boundary regions stood above the a resorbable substrate. Experiments growing rest of the enamel structure: at the enamel­ separated bone cells and marrow cells on dentine junction (EDJ), the enamel stood slightly calcified substrates have shown that such cells above dentine (Fig. 1). These studies showed that wi 11 continue to res orb for at least six weeks. the rate of resorption of a hard tissue is influenced by local variations in composition reflecting both the organic matrix chemistry and the density of mineralization (Fig. 2). Referring KEY WORDS:Bone disease, resorption, osteoclasts, to later studies in which it has been surmised osteoblasts, scanning electron microscopy, that collagen breakdown products are chemotactic stereophotogrammetry, enamel, dentine, . for precursors <2> and in which the degradation of the collagen has played a major *Address for correspondence: Alan Boyde role in thought processes <3>, we need to Department of Anatomy and Embryology emphasize that osteoclastic resorption can University College London continue in the absence of a collagenous Gower St., London WClE 6BT, England component to a matrix: there is none of this Phone No: 387 7050 (ext 3315) in enamel or human PTO.

369 A Boyde and SJ Jones

370 HARDTISSUE RESORPTION figure l. Resorption of human deciduous dentine maturation of the osteoid and its thickness may (top) and enamel prior to shedding a tooth. also be important. Fieldwidth = 296 ,um. The early studies also showed that most figure 2. Rat compact bone o_steocyt~ lacunae human adult bone surfaces entered a "resting" exposed by osteoclastic resorption showing reten­ condition, in which the collagen fibre bundles of tion of separate perilacunar matrix phase, common the matrix were fully mineralized <6, 7,10,12>. in this species. Fieldwidth = 55 ).Jm. Thus intact fibres and bundles could be demon­ figure 3. Human tooth root near apex showing strated in anorganic specimens (Fig. 5). This, patches of resorption and areas_of repair and the fact that much endosteal (Fig. 6) and emanating in prior resorbed areas. Fieldw1dth = trabecular (Fig. 7) resorption occurs in patches 1122 ,um. within fields of well mineralized collagen, figure 4. Horse molar enamel resorbed by osteo­ showed that osteocl asts do not have to contend clasts prior to cementum deposition: another with a barrier of a thin osteoid layer in order example of resorption formation coupling. to recognize real bone <13>. Close examination Fieldwidth = 172 ,um. (Stereo-pair). of our early SfM figures of anorganic prepara­ Figure 5. The lining surface of a Haversian canal tions <6> shows that some endosteal surfaces in adult rat mandible, made anorganic with 1,2 showed nearly, but not quite complete mineraliza­ ethane diamine, showing complete mineralisation tion: some resorption patches can be found ending of matrix surface fibres in this resting surface. within such areas. However, as just pointed out, Fieldwidth = 7. 3 ,A.Jm. some end in osteoid proper. It is obvious that one of the great advan­ Studies of human and other mammalian tages of the SfM is the apparent 3-D nature of cementum resorption <4> showed that some resorp­ the 2-D projection of a surface, making it abun­ tion of the roots of human permanent teeth is a dantly clear as to the morphological complexity perfectly normal occurrence. Indeed, the growth of a resorbing surface. However, a 2-D projection of a second-phase of cementum - which may or may can deceive, and we have always felt the need not contain included cells (cementocytes), but for a real 3-D image in our work. Stereopair which otherwise usually includes a high propor­ images not only convey a vivid impression of tion of extrinsic (Sharpey) fibre bundles - was relationships, but can be formally analysed to shown to occur in relation to small patches of reconstruct profile sections and contour resorption in the apical portion of the root maps <14-16>. We have continued to develop (Fig. 3). This was the first demonstration of stereophotogrammetric methods for the purpose of resorption-formation coupling in an extra­ providing a quantitative description of bone skeletal site. This now popular concept was first resorption <17-19>. proposed for bone on the basis of human metabolic data <5>. (Another dental example of resorption­ formation coupling will be considered later - see Fig. 4 ). Horses have high crowned teeth which erupt The first SEM studies of adult and fetal bone and function before a root is formed. The mecha­ resorption <6, 7> again showed how the nature of nism of tooth attachment is that the crown SEM observations could bring to our attention enamel is covered with cementum, which serves to some important facts. It was found that attach the fibres of the periodontal ligament. resorption lacunae in bone were much srnal ler than This route is adopted by many mammals, but it is in deciduous tooth root removal . It was noted usual for cement to attach to a naturally rough, that osteoclastic resorption left pitted surface of enamel. In horses, however, the lacunar walls intact. This observation indicated completed enamel surface, after maturation­ that survived until release by mineralization, is superficially resorbed by resorption: they did not themselves remove either osteoclasts (Fig. 4). The reduced, post­ mineral or matrix as was hypothesised in the then maturation ameloblasts have somehow to move or be popular concept of osteocytic osteolysis moved out of the way for this purpose. This <9,10>. Another early SfM finding was that resorption is repaired immediately by the resorbed areas not uncommonly finished within formation of cement. This is another example of fields with mineralizing fronts indicating that resorption-formation coupling. However, this is resorption had occurred into bone covered with most unusual in cement coupling to enamel <20>. osteoid <6,7,10>: the latter point was, and is sti 11, at odds with current notions of bone Hormone experiments on bone pathologists who felt, and feel, that osteoid does not resorb. The most recent studies have Experiments using the SEM to observe the confirmed the initial findings - osteoid can be effects of the administration of parathyroid resorbed byosteoclasts <11>. The most conservative hormone (as parathyroid extract, PTf) and explanation of this which we can offer today is calcitonin (CT) to rats <21> showed both the to hypothesize that osteoclasts are able to power of the SEM as an observational tool, and resorb osteoid, but do not prefer to do so. Thus the rapidity with which dramatic changes can be active activated osteoclasts continue resorbing induced in bone matrix and mineral morphology. a surf~ce, perhaps until they reach an ost~oid Resorption induced by parathyroid hormone (PTH) covered patch; when the "stimulation" to continue was observed to begin, at flat bone surfaces, resorption is reduced - but only after some close to the places where these were penetrated osteoid has been tested or tasted. The degree of by blood vessel canals (Fig. 8). A great deal of

371 A Boyde and SJ Jones

372 HARDTISSUE RESORPTION

Figure 6. Resorbed Iamellar bone matrix surface, Bone eel ls adult rabbit scapula covering cells removed by trypsin. Fieldwidth = 110 .)Jm. The first SEM observations of bone cells Figure 7. Human femur endosteal trabecular bone were made in a convenient location where all surface showing resorption into mineralised other overlying tissue layers could be easily matrix. Some parts of the matrix were still stripped away with apparently minimal disturbance mineralising as judged by the incomplete to the pavement <23>. Such a site is appearance of the collagen fibres in this the endocranial surface of the skul 1 cap (cal va­ anorganic preparation. Fieldwidth = 250 ).Jm. rium: Fig. 11). In young rats, most of this Figure 8. Endosteal surface of rat bone from surface is formative, and is covered with a tes­ animal given parathyroid extract 3 daily doses of selation of cells <24> which are "cuboidal" and Parathormone which stimulates resorption and in close lateral contact prior to the induction mineralisation of free matrix surfaces. of drying-associated, shrinkage artefacts <25>. Fieldwidth = 176 ,um. The shrinkage artefact problem provoked us Figure 9. Surface of bone matrix from similarly into studying its origins, which we now more treated animal showing hypermineralisation of clearly understand <26, 27>. If we use the resting bone matrix surface. Fieldwidth = 17 ).Jm. advanced drying techniques of freeze drying (FD) Figure 10. Another area from similar specimen or critical point drying (CPD), we are left, at showing cleavage of well mineralised layer of one near-fina 1 stage in either process, with a matrix from less mineralised sub surfaces layer specimen with no free water, but some firmly­ after the specimen has been made anorganic. bound structural water. This stage occurs after Fieldwidth = 78 ).Jm. the sublimation of ice in FD, and after the Figure 11. Tesselation of osteoblasts covering pressure is reduced after the critical point in the endocrania l surface of the rat cal varium. CPD. In both cases, the drying of the dry sample The small gaps between these cells are induced by induces separation of osteoblasts from each CPD shrinkage. Fieldwidth = 176 .AJm. other. The well mineralized bone matrix hardly shrinks at all in comparison <25-27>. evidence has since been collected which supports Measurements of the areas covered by osteo­ the idea that osteoclasts or their precursors are blasts showed that they fel 1 within a closely transported via the blood stream: these studies, defined range - the secretory territory of the including the key experiments of G Gothlin and osteoblast <22,23>. The same type of preparation JLE Ericsson, and SC Marks and DG Walker are was used to provide details of the normal morpho­ referenced and reviewed in ref. (22>. We logical polarisation of osteoblasts <24>. interpret our SEM findings as possibly indicating that osteoclastic resorption stimulated by PTH Hormone experiments on bone cells occurred with the minimum requirement for migration of these cells. After showing that bone cells in this system Secondly, it was found that when compared with would survive well in culture conditions, the control animals, large areas of bone surface were behaviour of these cells in response to physical converted to what we called the "prolonged, and hormonal stimuli was investigated. These resting" condition, identified by the progression studies showed that the effect of the resorption of mineralization beyond the limits of the super­ inducing hormone, PTH, was primarily on the ficial collagen fibrils and bundles of free bone osteoblast <28-31), and not on the osteoclast as surfaces; thus leaving a smooth, less detailed had been believed up to that time. A direct surface in anorganic preparation. In other words, response of osteoblasts to the vitamin D PTH in the whole animal promoted mineralization metabolite l,25(0H) 2D3 was also demonstrated in of the bone matrix, particularly at its surface this way <28,29>. (Fig. 9 ). These studies also showed that osteoblasts In some instances, clear cases were are actively migratory cells, spreading on to identified where this surface mineralization deliberately denuded portions of the bone, and on overlay areas in which the osteoid had not to surrogate substrates, such as pieces of glass mineralized fully. Upon making such specimens or tooth placed in contact with the osteoblast anorganic, patches of the superficial mineralized layer or with <30): this tendency of layer could be seen to break away (Fig. 10). This bone cells to migrate has since been used as the confirmed the impression that the PTH effect basis of an important method for isolating bone and/or combined with the expected hypercalcaemia, cells in culture <32>. could result in a "dumping" of mineral from blood Returning to the hormonal responsiveness of into the nearest available receptive location in the bone cells as monitored by SEM (28-31>, bone, and this happened to be osteoid <21>. osteoblasts showed no response to calcitonin The mineralization of superficial osteoid (CT). The dorsal surface ruffling which is an promoted by PTH in these experiments is, of adaptation shown by osteoblasts in culture was course, abnormal and the wrong way round - unchanged by addition of CT. However, osteoblasts osteoid normally mineralizes at its deep surface. responded dramatically to PTH in several diffe­ However, such a mechanism would provide the means rent ways, which led to the suggestion that these for al lowing osteoclasts to "see" mineral at the are the cells which regulate bone resorption free bone surface if these eel 1 s are dependent PTH induced the formation of gaps in the for their normal function upon the recognition of otherwise nearly continuous osteoblastic pave­ a calcified matrix. ment, which, in the short term, were greater than

373 A Boyde and SJ Jones

374 HARDTISSUE RESORPTION

Figure 12. Osteoblasts on endocranial surface of nodules or calcospherites of the mineralizing rat calvarium cultured with added PHI for 24 h. front exposed to view . Trypsin is known not This induces marked eel l shape changes. to affect collagen under these conditions, so Fieldwidth = 100 ..um. what was the explanation? At the same time, Vaes Figure 13. Rat fetal bone treated with trypsin: <35> reported the presence of latent collagenase Idea was to remove the osteoblasts but trypsin in fetal mouse bones which could be activated by dissolves fetal osteoid. Thus this surface shows various enzymes including trypsin. The idea that mineral clusters at the mineralising front, The trypsin was acting in this way was given further overlying matrix having been removed, possibly by weight in an experiment which showed that the trypsin activation of latent collagenase. destruction of avian fetal bone by trypsin was Fieldwidth = 22 ,um. prevented by the action of the trypsin inhibitor Figure 14. Adult rabbit scapula near to area cysteine <36>. This evidence provided morphologi­ undergoing resorption: Matrix cleaned by cal grounds supporting the view that the inactive trypsinization as in figure 13. Note that pro-collagenase <37,38> might really be incorpo­ collagen fibrils are intact. Fieldwidth = 23 ..um. rated in fetal bone matrix as a kind of built-in, Figure 15. Osteocytes at endocranial surface of self-destruct mechanism. However, the SEM ev id­ rat calvarium released by osteoclastic resorption enc e so far has not shown that a similar migrating from their lacunae after 24 h in vitro. mechanism exists in the mature form of bone. Fieldwidth = 48 ,um. Recent work which purports to show the Figure 16 (inset). Light micrograph of toluidine degradation of lamel lar bone osteoid under the blue stained preparation of osteoclasts on rat action of the osteoblasts has only shown the cal varium. Fieldwidth = 92 )Jm. effects of the cell culture conditions used <39>. Figure 17. Same field as figure 16 after Because fibroblasts are able to participate in preparing by CPD for SEM. Fieldwidth = 92 )Jm. the degradation of the matrix which they have Figure 18. Same field as figs. 16 and 17 after made, we cannot exclude the possibility that removing the cells by treatment with NaOCl. osteoblasts could undertake the same activity and Fieldwidth = 92 ..um. remove osteoid in the preparation of bone matrix for osteoclastic degradation. We do not believe those occurring in controls. Later, however, PTH that, under some non-physiological circumstances, caused an important change in osteoblast morpho­ osteoblasts could not release matrix degradative logy and a reduction in the loss of cells in enzymes. However, evidence for this occurring in culture conditions so that they, as a whole, more normal resorption is lacking. effectively covered the bone surface. Part of the If osteoblasts were to be able to degrade change in cell morphology was for the osteoblasts non-mineralized bone matrix, why should not their to elongate considerably, and to align themselves derivatives, the osteocytes, do likewise? We have in parallel arrays - more parallel, in fact, than mentioned above, and reviewed previously <40> the arrays of patches of similarly elongated control reasons for believing that this does not occur. culture eel ls (Fig. 12). PTH treated eel ls were Culture experiments also showed another also able to migrate with apparently undiminished character of osteocytes. Although encapsulated in vigor. solid, mineralized bone matrix until their rescue Another very interesting finding was that and liberation by osteoclastic resorption, osteo­ many osteoblasts would go into mitosis following cytes retain the ability to migrate (Fig. 15). the removal of PTH containing medium and its Indeed, liberated ex-osteocytes are certainly a replacement with control medium. This mitogenic significant component of the singly nucleated activity of PTH, by leading to an increase in the cell population left on a resorbing surface, and potential number of bone forming ce 11 s, might be they may be the "helper eel ls" <41>. However, part of the explanation of the stimulatory effect although it cannot be excluded that osteocytes of pulsed doses of PTH on bone formation which may produce enzymes which help to degrade the has been established clinically <33>. residual demineralized collagen fringe left on a PTH influences on the cytoskeleton of cells resorbed bone surface, it should be noted that AR of osteoblast-like lines have recently been docu­ Ten Cate (personal communication) has seen no mented in detail by Rodan and co-workers <34>. similar cells in resorbing kitten deciduous den­ tine. At least in the latter case, then, 1vhen no Other than osteoclastic bone destruction new reparative hard tissue formation need take mechanisms? and the fetal ~adult bone problem place, no "helper" cells can be envisaged, and, of course, there were no osteocytes present to be In searching for a method to remove released from the dentine. In this respect, we osteoblasts from bone matrix surfaces, we tried might propose that one role of osteoblastic "col­ to use trypsin at neutral pH, since this enzyme lagenase" may be to remove the residual deminera- is widely used in cell culture technology to 1 i zed matrix fringe after osteoclastic resor­ dissociate eel ls and to dislodge them from ption. It is known that even in adult lamellar artificial substrates (Figs.13 and 14). This bone, the first layer of new bone deposited on an works well with the more mature forms of bone ex-resorbed surface, tends towards the fetal with oriented collagen, leaving the collagen condition <42>. We hypothesize that "fetal" fibres of the matrix nicely exposed to view osteoblasts can liberate a matrix degradative (Fig. 14). However, in the case of fetal or woven enzyme: they may do this to ensure that a "clean bone (Fig. 13), we found that the osteoid itself start" is made by removing old, partly degraded was largely destroyed, leaving the mineralizing matrix, or this may just occur fortuitously.

375 A Boyde and SJ Jones

376 HARDTISSUE RESORPTION figure 19. Part of the ruff led border zone of an The origin of osteoclasts osteoclast peeled separated from bone. Fieldwidth = 18 )..Im. A few years ago, it was fashionable to figure 20. Surface of slice of sperm whale assume that osteoclasts arose by the fusion of dentine cultured with chick osteoclasts for 19 mononuclear cells of the monocyte-macrophage days, showing, extensive resorption which has series (44>. We used immunological marking techn­ occurred in this period. Note the wide range of iques to examine this hypothesis, using the SEM size of the resorption pits. Some new osteo­ as the means of observing whether rosettes of clasts may have differentiated from marrow cells RBCs formed in the assay system. The results were in this period. Surface decorated with 10 micron negative - they demonstrated that osteoclasts did polystyrene latex spheres. Fieldwidth = 480 ,urn. not bear the Fe or C3 receptors characteristic of figure 21. Lower part of field shows edge of all cells of the macrophage-monocyte lineage resorption bay in human compact bone made by <45,46>, suggesting a separate origin at an chick osteoclast, revealing the alternate earlier stage of stem cell differentiation. This orientation of lamellae. Fieldwidth = 48 )..Im. finding was at first treated with considerable Figure 22. Chick osteoclastic resorption of human scepticism, but has now found acceptance compact bone. This field is selected to show following recent work which has extended and apparent selection of densely mineralised areas confirmed the original observations <47,48>. showing as white in this BSE image. However, careful study shows equal selection for all The isolated osteoclast resorption assay <49>. density phases in bone and unmineralised osteoid. Fieldwidth = 450 ,um. Most previous assays of the bone resorption Figure 23. Internal surface of mollusc shell process in vitro have been based on the measure­ resorbed by chick osteoc 1 asts. ment of biochemical parameters based upon all Fieldwidth = 104 ,um. events affecting all parts of all the matrix and Figure 24. Sperm whale dentine resorbed by osteo­ all the cells in a limited culture system clasts originating from chick marrow in 19 day i .g. 50 53>. The measurement of the release of culture. Sample broken and tilted to show 5Ca or 3H praline from prelabelled bones is a profile of fracture line through resorption sensitive approach: results are usually expressed lacuna. Fieldwidth = 90 )..Im. as the ratio of experimental to control values <50-52>. The release of calcium can also be The osteocytic osteolysis concept measured by simple chemical, colorimetric proce­ dures <53>. The problems with all these methods From our own viewpoint, osteocytic are that they do not establish that the release osteolysis became an unsupportable concept when of matrix or mineral components is the result we were unable to discover degradative changes in of osteoclastic activity, or to what extent the matrix surrounding osteocyte lacunae <6>. osteoclastic activity is control led by osteo­ Other evidence discrediting this hypothesis was blasts or furthered by other cell types, for considered by AM Parfitt in his influential example, macrophages. Further, these assays review <43>. Every case of "osteocytic osteoly­ always use fetal bone, and fetal bone matrix is sis" so far described in the literature can be degraded by non-specific proteases which may traced back to the failure to recognize the spec­ activate la tent coll agenase <10, 36>. Thus, as trum of bone and bone cell types from woven discussed earlier, osteoblasts may themselves be (fetal) to adult (lamellar). SEM studies clearly important in the degradative process in cultured emphasize the points of difference in these fetal bone. Analysis of sections of bones tissue types: large eel ls incorporated in large prepared for light microscopy has also been used numbers, often close to each other, with the in resorption assays, but this poses a sampling mineralization of juxta-lacunar collagen often and measurement problem in a rapidly changing incomplete in the back wall of the forming osteo­ tissue. We therefore wished to establish a system cyte lacuna in the mineralizing front region, are in which osteoclasts could be tested in isolation the hallmarks of the rapidly forming, fetal bone and on a known starting platform (Figs.20 - 24). type. Such characters are found in various bone The system has been successful and has been diseases, particularly those in which bone adopted by other groups <49, 54-61>. turnover rate is increased. Thus irregularly The measurement of the volume of bone or mineralized lacunar walls can be found in all dentine removed per resorptive episode in vivo forms of rickets and hyperparathyroidism <40>, might in itself be a useful parameter in charact­ but they are all formed as such in the first erizing this process under different normal and instance. There is no established evidence for pathological circumstances, and this is where we osteocytic involvement in bone matrix destruction began using SEM stereophotogrammetry <8, 14-19>. or demineralization. The converse is not true To provide the instrumental back-up for an assay however. of bone resorption based on the activity of Rat osteocytes may make perilacunar matrix individual cells, we developed computer aided (Fig. 2) which mineralizes to a high degree. If systems for the measurement of the volume of they do not encourage the mineralization of old resorption pits <8, 18, 19>. XYZdata acquisition bone matrix, they certainly permit it to occur - from the RS3 instrument built for us by Ross as is demonstrated by the maturation (i.e. Instruments <18> was controlled by a software increase of mineralization) phenomenon found in package (OC3D) for a Sharp MZ80K microcomputer. all lamellar bone. This instrument has now been superceded by an

377 A Boyde and SJ Jones improved one called SFS3 and a new software composition of the substrate (Fig. 21) by deter­ Jackage (STEREOSHIPS)written for a Sirius micro­ mining this afterwards, primarily by the use of ~omputer <19, 49). BSE A+ B imaging of samples which were well In our earlier SEM studies of in vivo polished before initial use (Fig. 22). Reid <11> ~esorption we could assign accurate values to the has, in this way, been able to show that avian volume, areas, radius of curvature, and depth of osteoclasts have no predisposition to res rb bone individual "scoops" , but these were minimal of different density due to different degrees of ,alues because such resorption almost invariably maturation-mineralization. Nevertheless, the 3tarted in areas which had had curved surfaces. local depth of resorption is affecte by the fhere was no reference plane against which to degree of mineralization; just like the situation neasure, and the edge of the resorption pits were in tooth resorption ( & Fig. 1). Thus within "requently affected by later resorption in an any one resorption 1 acuna, the depth of destruc­ 3djacent "episode" (Figs. 16-19). These problems tion is less in the more mineralized bone phase Jid not prevent us from finding large differences <62>. This observation has validated our earlier Jetween the volumes of resorption pits in normal decision that a more uniform substrate would be a Jone and normal deciduous dentine. wiser choice. Cachalot dentine. In developing our in Other substrates. Our in vitro studies also vitro, single ce 11, resorption assay system, we showed that other mineral and other organic ~hose to use dentine as the surrogate-for-bone matrix substances could be tackled by isolated substrate. Further, we chose to use male sperm osteoclasts; and further, that mineral-alone and ~hale (in old English and modern French •~acha• matrix-alone substrates are resorbed <58>. Consi­ lot") dentine. The reasons for these choices were dering the various materials we have examined: that dentine is a more homogeneous material than dental enamel, avian egg shells and molluscan Jone, avoiding the pitfalls of variations in shells (Fig. 23) contain no collagen, but have ~onposition from one micro-area in the substrate other specific and protein-carbohydrate to another. There are no holes in a cut piece of complexes. Avian egg shell and mollusc shells jentine through which our osteoclasts can fall contain no phosphate: these are calcium carbonate (Fig. 20). There are plenty - the original Haver­ (calcite or aragonite) containing materials. As sian canals (Figs.21 and 22) - in any piece of would be expected from the higher acid solubility Jone. Male sperm whale mandibular teeth are so of carbonate compared to calcium phosphates (inc­ large that we can tailor large numbers of large luding ), these shells are more flat pieces from one tooth. We have many cells on prone to osteoclastic demineralization and can be ::ine substrate, and the substrates have the same used as sensitive markers to show where an osteo­ ::irigin in many different experiments <19, 49, 54 clast has migrated during an experimental period. - 58>. Pure calcite and pure apatite synthetic sub­ strates were also attacked, showing that no matrix need be present. Further, we have shown that demineralized dentine and bone and as yet Solo. The in vitro system shows that iso­ unmineralized predentine and osteoid are also lated osteoclasts, which can be observed to be resorbed - though with an evident reduction in Nell out of contact with other types of cells incidence compared with the mineralized natrices. Nhich settle on to S~ID, bone or other surrogate Hormones. Our in vitro studies ha1 e shown ubstrata, are perfectly capable of resorbing all that calcitonin (CT), the hormone which the phases present in these materials (54-63). suppresses bone resorption in vivo and osteo­ This type of experiment showed, therefore, that clastic motility in vitro, does not stop activated osteoclasts can dissolve all phases of resorption by osteoclasts in culture at the exact oone (see also <3>). The absolute requirement for doses which eliminate peripheral cytcplasmic the involvement of osteoblasts, macrophages or movement on plastic <55,56>. This result shows other eel l types can be excluded. These experi­ that demonstration of cessation of movement is nents also showed that osteoclasts in vitro may not necessarily a demonstration of cessation of nave away from a resorbed area of bone or den­ resorption. Levels of resorption are, however, tine, leaving a "demineralized collagen fringe" measurably reduced and at higher concentrations in each lacuna. The depth of these fringes is may be suppressed entirely; there is an obvious apparently more extensive than we have encoun­ correlation between movement, the cytoskeleton, tered in studies of in vivo bone resorption using and exocytosis. Resorption also occurs in serum­ the SEM. Something in the resorptive process is free medium, without any added hormones. then different in the in vitro system. Amongst Longevity. More recent experiments have various possibilities <63> are that the enzymes shown that osteoclasts originating both from the necessary to complete the degradation of bone minced - chopped bone fraction and the marrow­ matrix may disperse too easily in vitro, or be only fraction of neonatal long bones of chicks eficient in quantity relative to the amount of are able to survive and to continue to be able to acid produced in vitro. They may also be produced resorb for periods of at least 6 weeks <63>. This by other eel l types, such as osteobl asts, which longevity in culture provides an estimate which are to lay down the reparative bone, or they may is larger than most previous deductions <64-66>. be of fibroblastic, macrophagic or other origin. However, we consider it likely that precursor Bone. We have also used human bone samples cells differentiated to become osteoclasts within in in vitro experiments, and compensated for our this experimental period <63). A proportion of lack of knowledge about the variations in the the resorption lacunae found during only 3 week

378 HARD TISSUE RESORPTION culture periods (Figs. 20 & 24) are exceptionally <9> Belanger LF (1969) Osteocytic osteolysis. large, single and smooth, and therefore presumab­ Calcif.Tiss.Res. 4 1-12. ly single-cell origin scoops formed by the cont­ Boyde A (1972) Scanning electron inued activity of large osteoclasts. Some have microscopic studies of bone. In: The biochemistry been shown to have volumes over 300 pl, hundreds and physiology of bone (ed) Bourne GH, 2nd of times greater than those found within 24 and edn., Vol.! New York:Academic Press, 259-310. 48 h periods <49, 63). In parallel observations <11> Reid S (1986) Effect of mineral content of made by LM, we have shown that there is a small human bone on in vitro resorption. Anat & population of osteoclasts with exceptionally high Embryol. 174: 225-234. numbers of nuclei in the starting population. <12> Boyde A, Jones SJ (1972) Scanning electron These may presumably be added to by the fusion of microscopic studies of the formation of minera­ others during the experimental period. lised tissues. In:Developmental aspects of oral The actual numbers. Measurements of volumes biology (eds.) Slavkin HC, Ba vet ta LA , New York: of SWD resorbed by, for example, chick osteo­ Academic Press, 243-274. clasts indicate mean rates per cell with an <13> Chambers TJ, Fuller K (1986) Osteoblasts average of 8 nuclei of 2200 fl/day over shorter initiate bone resorption by exposing bone mineral time periods <49>. There are good correlations to osteocl as tic contact. Bone 7: 145-146. (r. greater than 0.96) between volume and <14> Boyde A (1966) Photogrammetric studies of area of resorption pits. However, the slope of electron micrographs of replicas of the surface the regression line for area/volume is different of developing dental enamel. Advances in Fluorine at different time periods and for osteoclasts Res. & Dental Caries Prevention. 4: 137-145. originating from different species. It is <15> Boyde A (1968) Height measurements from therefore unsafe to assume that volumetric (3-D) stereopair scanning electron micrographs. Beitr. measurement can be dispensed with in studies of Elektronennikroskop. Direktabb. Oberfl. this type, or that rigorous methods can be short (Munster). 1: 97-105. cut <67>. <16) Howell PGT, Boyde A (1984) Three dimensional analysis of surfaces. In: The analysis of organic Acknowledgements and biological surfaces (ed) Echlin P. John Wiley and Sons Inc.: New York, 325-349. Our unit has been supported by the Medical (17) Boyde A, Ross HF (1975) Photogrammetry and Research Council, the Science and Engineering the SEM. Photogram. Record 8 408-457 Research Council and the Foundation for Age <18> Howell PGT, Boyde A, Ross HF (1986) A three Research. Gilbert Vaes provided the idea of axis stereocomparator for SEM photogrammetry - reviewing our SEM studies in the context of RS3. Scanning 8: 182-186. resorption. We are grateful for the technical <19> Boyde, A., Howell PGT, Franc F (1986) assistance of Elaine Maconnachie, Roy Radcliffe Simple SEM stereophotogrammetric method for three and Carl Prescott, and the secretarial help of dimensional evaluation of features on flat Ann Scott, Sue Kiely and Julie Smith. We thank substrates. J. Microscopy, 143: 257-264. Stephen Reid for Figs. 21 and 22. <20) Jones SJ, Boyde A (1974) Coronal cementa­ genesis in the horse. Arch.oral Bio 1.19: 605-614. References <21> Marks SC, Walker, DG (1976) Mammalian osteopetrosis - a model for studying cellular and Boyde A, Lester KS (1967) The structure humoral factors in bone resorption. In: GH and development of marsupial enamel tubules. z. Bourne,(ed.), The Biochemistry and Physiology of Zel 1 Forsch. 82: 558-576. Bone, Vol. IV, 2nd ed., New York: Academic Press, <2> Mundy GR, Varani J, Orr W, Gondek MD, Ward 227-301. PA (1978) Resorbing bone is chemotactic for mono­ <22) Jones SJ, Boyde A (1970) Experimental cytes. Nature 275: 132-135. studies of the interpretation of bone surfaces <3> Blair HC, Kahn AJ, Crouch EC, Jeffrey JJ, studied with the Scanning Electron Microscope. Teitelbaum SL (1985) Isolated osteoclasts resorb Scanning Electron Microscopy 1970: 193-200. the organic and inorganic components of bone. (23) Jones SJ (1974) Secretory territories and Cale. Tiss Int. 385: 510 (Abst. 38). rate of matrix production of osteoblasts. Cale. <4> Jones SJ, Boyde A (1972) A study of human Tiss. Res. 14: 309-315. root cementum surfaces as prepared for and <24> Jones SJ, Ness AR (1977). A study of the examined in the SEM. Z. Zell Forsch 130 318-337. arrangement of osteoblasts of rat calvarium cul­ <5> Harris WH, Heaney RP (1969) Skeletal renewal tured in medium with, or without, added parathy­ and metabolic bone disease. N. Engl. J. Med. 280: roid extract. J. Cell Sci. 25: 247-263. 193-202. <25> Boyde A, Bailey E, Jones SJ, Tamarin A <6> Boyde A, Hobdell MH (1969) Scanning elec­ (1977) Dimensional changes during specimen prepa­ tron microscopy of lamellar bone z. Zellforsch ration for Scanning Electron Microscopy 1977; 93: 213-231. I:507-518. <7> Boyde A, Hobdel l MH (1969) Scanning elec­ <26> Boyde A, Maconnachie E (1979) Volume tron microscopy of primary membrane bone. z. changes during preparation of mouse embryonic Zell Forsch 99: 98-108. tissue for scanning electron microscopy. Scanning <8> Boyde A, Jones SJ (1979) Estimation of the 2:149-163. size of resorption lacunae in mammalian calcified <27> Boyde A, Maconnachie E (1984) Not quite tissues using SEM stereophotogrammetry. Scanning critical point drying. In: Science of Biological Electron Microscopy 1979;11: 393-402. Specimen Preparation, J-P Revel, T Barnard & GH

379 A Boyde and SJ Jones

Haggis (ed.), SEM Inc: AMF O'Hare IL, 71-75. <46> Jones SJ, Hogg NM, Shapiro IM, Slusarenko <28> Jones SJ, Boyde A (1977) SEM of bone cells M, Boyde A (1981) Cells with Fe receptors in the in culture. Excerpt a Medi ca Int. Congr.Ser.421: cell layer next to osteoblasts and osteoclasts on 97-104 • bone. Met ab. Bone Dis. Rel. Res. 2: 357-362. <29> Jones SJ, Boyde A, Ness AR (1976) SEM (47> Horton MA, Rimmer EF, Chambers TJ (1986) studies of steeoblasts in organ culture. In: Bone Giant cell formation in rabbit long-term bone Histomorphometry. Meunier PJ, (ed). Paris: marrow cultures: Immunological and functional Armour-Montagu, 275-290. studies. J. Bone Min. Res. 1: 5-14. (30) Jones SJ, Boyde A (1977) The migration of <48> Horton MA, Lewis D, McNaul ty K, Pringle osteoblasts. Cell Tiss. Res.184: 179-185. JAS, Chambers TJ (1986) Characterisation of mono­ <31> Jones SJ, Boyde A (1976) Experimental study clonal antibodies specific for human osteoclasts. of changes in osteoblastic shape induced by J. Bone Min. Res. l: 94 (Abst. 140). ca lei tonin and parathyroid extract in an organ <49> Jones SJ, Boyde A, Ali NN, Maconnachie E culture system. Cell Tiss.Res. 169: 449-465. (1986) Variation in the sizes of resorption <32> Beresford JN, Gallagher JA, Poser JW, lacunae made in vitro. Scanning Electron Russell RGG (1984) Production of osteocalcin by Microscopy 1986; IV: 1571-1580. human bone cells in vitro. Effects of <50> Reynolds JJ, Dingle TJ 1970) A sensitive in l,25(0H) 2 D3 , 24,25(0H) 2D3 , parathyroid hormone vitro method for studying the induction and inhi­ and glucocorticoids. Metab. Bone Dis. Rel.Res. bition of bone resorption. Calcif. Tiss. Res. 4: 5: 229-234. 339-349. (33) Reeve J, Hesp R, Williams D, Hulme P, <51> Raisz LG, Simmons HA, Sandberg AL, Canalis Zanel l i JM, Darby AJ, Tregear GW, Parsons J.A. E (1979) Direct stimulation of bone resorption by (1976) Anabolic effect of low doses of a fragment epidermal growth factor. Endocrinology, of human parathyroid hormone on the in 107:270-272. postmenopausal osteoporosis. The Lancet, 15 May, <52> Mundy GR, Eilon G (1981) Effects of 1976, pp. 1035-1038. inhibition of microtubule assembly on bone <34> Egan JJ, Rodan GA (1986) Cell density mineral release and enzyme release by human dependent disassembly of osteoblast cytoskeleton breast cancer cells. Lab. & Clin. Invest. 67: correlates with increased cAMP levels. J. Bone 69-76. Min.Res. l: 68 (Abst. 39). <53> Ali NN, Auger DW, Bennett A, Edwards DA, <35> Vaes G (1972) The release of collagenase as Harris M. (1979) Effect of Prostacylin and its an inactive proenzyme by bone explants in cul­ breakdown product 6-oxo-PGF 1 on bone resorption tu re. Bi ochem J. 126: 275-289. in vitro. In Prostacylin, (eds). JR Vane, S <36) Boyde A, Banes AJ, Dillaman RM, Mechanic GL Bergstrom Raven Press: New York, 179-185. (1978) A morphological study of an avian bone <54> Boyde A, Ali NN, Jones SJ (1984) disorder caused by myeloblastosis-associated Resorption of dentine by isolated osteoclasts in virus. Met ab.Bone.Dis.Rel.Res. l: 235-242. vitro. Brit. Dent. J. 156: 216-220. <37> Eeckhout Y, Vaes G (1977) Further studies <55> Ali NN, Boyde A, Jones SJ (1984) Motility on the activation of procollagenase, the latent and resorption: Osteoclastic activity in vitro. precursor of bone coll agenase. Bio chem. J. Anat. & Embryo!. 170: 51-56. 166: 21-31. <56) Jones SJ, Boyde A, Ali NN, Maconnachie E <38) Heath JK, Atkinson SJ, Meikle MC, (1985) A review of bone eel l and substratum Reynolds JJ (1984) Mouse osteoblasts synthesize interactions. An illustration of the role of collagenase in response to bone resorbing agents. scanning electron microscopy. Scanning 7: 5-24. Biochim. Biophys. Acta. 802: 151-154. <57> Boyde A, Ali NN, Jones SJ (1985) Optical (39) Chambers TJ, Fuller K (1985) Bone cells and scanning electron microscopy in the single predispose bone surfaces to resorption by osteoclast resorption assay. Scanning Electron exposure of mineral to osteoclastic contact. J. Microscopy 1985; Ill: 1259-1271. Ce! l Sci. 76: 155-165. <58> Jones SJ, Boyde A, Ali NN (1984) The <40) Boyde A (1980) Evidence against osteocytic resorption of biological and non biological osteolysis. Metab. Bone Dis. Rel.Res. 2 Suppl: substrates by cultured avian and mammalian 239-255. osteoclasts. Anat. & Embryo!. 170: 247-256. <41> Heersche JNM (1978) Mechanism of <59> Chambers TJ, Revell PA, Fuller K, Athanasou osteoclastic bone resorption: A new hypothesis. NA (1984) Resorption of bone by isolated rabbit Cale. Tiss.Res. 26: 81-84. osteoclasts. J. Cell Sci. 66: 383-399. <42> Marotti G (1980) Three dimensional study <60> Chambers TJ, Fuller K, McSheehy PMJ, of osteocyte lacunae. Metab. Bone Dis. Rel.Res. Pringle JAS (1985) The effect of calcium 2Suppl: 223-229. regulating hormones on bone resorption by <43> Parfitt AM (1977) The cellular basis of isolated human osteoclastoma eel ls. J.Pathol. bone turnover and bone loss. Clin. Orthop. Rel. 145: 297-305. Res. 127: 236-247. <61> Chambers TJ, McSheehy PMJ, Thomson BM, <44> Chambers TJ (1980) The cellular basis of Fuller K (l9B5) The effect of calcium regulating bone resorption. Clin. Orthop. Rel.Res. 151: 283- hormones and prostaglandins on bone resorption by 293. osteoclasts disaggregated from neonatal rabbit <45> Hogg N, Shapiro IM, Jones SJ, Slusarenko M, bones. Endocrinol. 116: 234-239. Boyde A (1980) The lack of Fe receptors on <62> Reid SA (1986) A study of lamellar organisa­ osteoc lasts. Ce 11. Tiss. Res. 212: 509-516. tion in juvenile and adult human bone. Anat. & Embryo!., 174: 329-338.

380 HARDTISSUE RESORPTION

(63> Jones SJ, Ali NN, Boyde A (1986) Sur vi val Authors: Other cells are able and rescrptive activity of chick osteoclasts in to dismantle the matrix which they have made. culture. Anat. & Embryo1. 174: 265-275. This is amply illustrated in many studies on the <64> Jaworski ZFG, Duck B, Sekaly G (1981) activities of fibroblasts (as an EM based Kinetics of osteoclasts and their nuclei in example, see AR Ten Cate (1972) Morphological evolving secondary Haversian systems. J. Anat. studies of fibrocytes in connective tissues 133: 397-405. undergoing rapid remodelling. J. Anat. 112: 410- <65> Loutit JF, Townsend KMS (1982) Longevity 414). Our argument is that evidence for of osteoclasts in radiation chimaeras of beige osteocytes or osteoblasts performing this and osteopetrotic microphthalmic mice. Brit. J. function during normal resorption in vivo is Exp. Pathol. 63: 214-220. lacking. (66> Marks SC (1984) Congenital osteopetrotic Lacunar margin modification by osteocytes mutations as probes of the origin, structure, and does occur, but this has only been demonstrated function of osteoclasts. Clin. Orthop. Rel. Res. in the sense of the accretion of new matrix and 189: 239-263. mineral in the walls of osteocyte lacunae. <67> Chambers TJ, Thomson BM, Fuller K (1984) Perilacunar bone was not discovered by SEM - Effect of substrate composition on bone that can be attributed to the work of IA Mjor resorption by rabbit osteoclasts. J. Cell Sci.7[J: (Bone matrix adjacent to lacunae and canaliculi. 61-71. Anat Rec. 144, 327-329, 1962) using the TEM. However, it became very much clearer from the Discussion with Reviewers earliest SEM studies that perilacunar bone was an interesting material which can be deposited by Reviewer I: What contribution could the SEMmake osteocytes in some locations and/or in some towards resolving some of the outstanding issues, species. such as resorption-formation coupling and the It remains to be discovered what is the origin of the osteoclasts? nature of the antigen discovered by Nijweide's Authors: The SEM is just one tool which can be antibody to something at the cell surface of used to focus on small facets of the bone chick osteocytes. It is as likely that this resorption problem. As regards resorption represents new synthetic activity (something to formation coupling, we have indicated that it do with peri 1 acunar bone?), as that it may enabled us to see that this was happening in the represent matrix degradative activity. extra-skeletal context of cellular cementum The examples we have chosen here from the formation on the roots of human permanent teeth. study of hard tissue resorption, we state again, It is improbable that the necessary global view are essentially chosen to illustrate the merits would have been obtained by reconstructions from of the SEM in this sphere of investigation. Of serial sections from any other microscopic course, it has also been of great utility in method. The survey potential of the SEM has not studying the formation and the structure of the yet been exploited in examining skeletal hard tissues, including many aspects of resorption-formation coupling at the local scale. mechanical and clinical interference with these As regards the origin of osteoclasts, the tissues in orthopaedics and dentistry. SEM provided the means to observe that the surface antigens of the cells of the monocyte 0. Johari: In stereo-pair (Fig. 4), how can macrophage series were not borne by osteoclasts. stereo-effect be observed? At that time it was widely assumed that the cell Authors: A few persons can detect the stereo types were at least very closely related. So, effect without the aid of any stereo-viewing at least, even if the SEM cannot solve problems, device. For others, we suggest that the page be it can raise them. rotated 90° and viewed under a stereo-viewer. However, we believe that the SEM has a role to play in investigating the function of osteoclasts. An example of this is its sensitivity to the detection of changes in the surface of a substrate using BSE imaging. This may be very important when testing the resorptive ability of putative osteoclast precursor cells, since it is a characteristic ability of an osteoclast to demineralize the tissue it contacts.

Reviewer II: It is not justified to extrapolate from the standpoint that lacunar margin modifica­ tion by osteoclasts does not exist, to conclude that their precursors, the osteoblasts, cannot remove osteoid. The authors have overlooked recent evidence for an osteocyte-speci fie eel l surface antigen (PJ Nijweide & RJP Mulder, A monoclonal antibody specifically directed against osteocytes. Cale. Tiss. Int. 38 suppl, Abst. 28, 1985).

381