THE ANATOMICAL RECORD 228:383-391 (1990)
Organization and Development of the Mineral Phase During Early Ontogenesis of the Bony Fin Rays of the Trout Oncorhynchus mykiss WILLIAM J. LANDIS AND JACQUELINE GERAUDIE Department of Orthopedic Surgery, Harvard Medical School, Children's Hospital, Boston, Massachusetts (W.J.L.);and Laboratoire d'Anatomie Comparke, Universitt! Paris VII, Paris Cedex 05, France
ABSTRACT Characterization of mineral deposition has been studied by electron optical methods during early ontogenesis of lepidotrichia, the bony fin rays, of the trout Oncorhynchus mykiss (the former Salmo gairdneri). The fin rays consist of a n extracellular granular ground substance containing in part a network of collagen fibrils within the basal lamella of the fin dermoepidermal interface. Growth of individual rays proceeds in a proximodistal direction. The mineral phase appears as electron-dense needle or plate-like particles and is associated with the collagenous matrix. On analysis of progressively maturing tissue, the mineral was characterized as a poorly crystalline hydroxyapatite with Ca/P molar ratios in the range of 1.0-1.4, corresponding to distal and proximal areas, respectively. With selected-area electron diffraction and dark field imaging of lepidotrichia, the mineral particles were found to be about 3-10 nm thick and 12-20 nm in length (along their crystallographic c-axes), possibly aggregated into larger crystals 35-40 nm long observed with bright field microscopy. No definitive relation was found between either the c- or a,b-axes images of the crystals and the periodic structure of collagen, which forms the framework for mineral deposition in this and in other vertebrate calcifying tissues.
The identification and mode of deposition of the mineral phase in the fin rays or lepidotrichia of the growing trout fry are of interest for modeling the pattern of mineralization for various vertebrate bone tissues. It is known from previous work (Geraudie and Landis, 1982) t h a t lepidotrichial ontogenesis occurs by direct mineralization of a collagenous basal lamella of the skin without passage through a cartilaginous precursor. Consequently, this bone is strictly of dermal origin. Another characteristic of this bone is a progressive apical growth during the life span of the fish. Elongation of the bony rays is achieved by the distal addition of new segments, which calcify successively. This unusual continuous and sequential spatial and temporal process of mineralization has been studied to determine the biophysical quality of the initial mineral deposits in the tissue and their maturation along the growing new segments. The present work was accomplished with conventional electron microscopy combined with specific electron optical and analytical techniques after anhydrous preparation of bone samples, a method that prevents mineral loss and translocation (Landis et al., 1977). The results contribute to a more complete description of the interaction between mineral and collagen and form the basis for understanding initial events in this example of vertebrate bone mineralization. A portion of this study has been presented in preliminary form (Geraudie and Landis, 1989). 0 1990 WILEY-LISS, INC.
MATERIALS AND METHODS Materials
The present study was carried out on the rainbow trout Oncorhynchus mykiss (new identification of the former Salmo gairdneri; Kendall, 1988). Larvae were raised from fertilization in running tap water kept at a temperature of 12 2 1°C. Hatching occurs about 1 month after fertilization. Growing fins were dissected from anesthetized fish (1% MS222:ethyl m-aminobenzoate in 100 ml of tap water) 1 month after hatching. Methods
A number of electron optical and analytical techniques were applied to examine the developing fin rays. These included the following. Electron microscopy under conventional aqueous fixation and processing
In order to observe the morphology of the growing distal lepidotrichium segments containing newly deposited mineral particles, isolated pieces of different fins (pectoral, pelvic, caudal) were immersed in 2.5% glutaraldehyde in 0.1M cacodylate buffer, pH 7.2, at Received November 15, 1989; accepted March 22, 1990. Address reprint requests to Dr. William J. Landis, Department of Orthopedic Surgery, Enders Bldg-Room SB 20, 300 Longwood Avenue, Children's Hospital, Boston, MA 02115.
384
W.J. LANDIS AND J. GERAUDIE
4°C for 2 hours. Fully dehydrated specimens were embedded in Epon 812 for longitudinal sectioning. Thick sections (1pm) were stained with toluidine blue in 1% borax. Thin sections (-80 nm) were obtained with a diamond knife and floated on distilled water for subsequent collection on copper grids. No attempt was made to reduce mineral loss in handling these sections, which were contrasted with aqueous uranyl acetate and lead citrate. Sections were observed on a Philips 300 electron microscope operated a t 80 keV. Electron microscopy under anhydrous conditions of tissue preparation
Anhydrous methods were used in parallel with the conventional techniques above. Anhydrous approaches minimize or prevent artifacts in the mineral phase of bone tissues (Landis et al., 1977). Isolated pieces of fins were dissected in cold (4°C)ethylene glycol (100%)and then kept under vacuum in the same solvent for 3 hours at room temperature with constant mild agitation. Ethylene glycol was changed, and the specimens remained under the same conditions for an additional 24 hours, after which they were immersed at 4°C in Cellosolve, replaced twice over the next 24 hour period. Tissues were then kept at room temperature in a mixture of propylene oxide-Epon 812 (1:l)for 1 week and finally embedded in Epon for transverse or longitudinal sectioning of the fins. Once appropriate regions of thick sections were identified for thin sectioning, the block was trimmed to remove excess Epon. Sections (80-100 nm) were cut with a diamond knife and floated onto 100% ethylene glycol. Thin sections were examined either unstained or contrasted with uranyl acetate and lead citrate in 100% ethylene glycol or in distilled water, depending on the observations to be made. Conventional bright field electron microscopy and electron diffraction
Stained and unstained thin sections were observed with a Philips EM 300 operated a t 60 or 80 keV or a JEOL lOOC operated a t 60-100 keV. Selected-area electron diffraction was carried out on the JEOL lOOC at 80 or 100 keV. Evaporated gold, aluminum, and thallous chloride were used as electron diffraction calibration standards (Landis and Glimcher, 1978). Selected-area electron diffraction-dark field imaging
Application of this technique was based on that described by Arsenault (1988) to determine the size, orientation, and distribution of the mineral crystals present in a tissue specimen. Epon-embedded samples were thin-sectioned (80 nm) so that longitudinal profiles of fin rays were obtained. Unstained sections were examined at 80 or 100 keV in a Philips EM 300, and early mineralizing regions of the tissues were studied by selected-area electron diffraction. Diffraction ref lections corresponding to the a- and b-axes together and to the c-axis alone of the apatite crystals comprising the ray mineral were isolated in a small (1-4 pm) objective aperture (Energy Beam Sciences, Agawam, MA) and then imaged. The a,b-axial orientations of the crystals were determined from the (211), (112), (300), (202), and (301) lattice planes of apatite. The c-axial orientation was determined from the (002) lattice
plane. Images of the respective planes were observed in the dark field mode of the microscope and thereby appeared in reverse contrast compared to the images of conventional bright field microscopy. High spatial resolution electron probe x-ray microanalysis
The same sections studied by electron diffractiondark field imaging and conventional bright field microscopy were used for electron probe microanalysis in a JEOL JSM 50A scanning electron microscope modified for high spatial resolution (6-10 nm) (Landis and Glimcher, 1982). Accelerating volta e was 25 keV, with a fixed-beam current of 1 x 10- 15 amp, and x-ray spectra were recorded during 100-600 seconds integrated detecting time periods. Molar ratios of calcium to phosphorus, corrected for background, were found by interpolation from a calibration curve of x-ray intensities generated from a number of calcium phosphate standards analyzed in the same manner as the fin bone tissue (Landis, 1979). RESULTS Morphological Evolution of the Developing Bone
A morphological description of the developing mineral in the fin ray begins with observations from the growing distal extremities progressively toward more proximal regions of the tissue. Calcification, based on the spatial location and disposition of mineral crystal particles of different sizes, proceeds in a proximodistal direction (Geraudie and Landis, 1982). Figures 1-6 depict the onset and sequential deposition of mineral in a single lepidotrichium. The first indication of a demiray is the appearance of an electron-dense ground substance within the collagen network of the basal lamella of the dermoepidermal interface (Figs. 1,2). Within the interface, this granular material does not extend to the basal lamina of the epidermis on one side and is not directly contiguous with the cells of the dermis on the other side. The ground substance does not fully cover the width of the collagen network in even more proximal regions of the tissue. The collagenous network changes only gradually in width, and within it the ground substance follows as a demiray matures. Near the distal regions of the demiray (Figs. 2-4) isolated needle- or plate-like particles of high electron density are found; these are associated, apparently exclusively, with collagen fibrils. Such particles are the first observed mineral crystals in the tissue (Figs. 7-10). Most mineral-associated fibrils lie within the confines of the granular ground substance (Figs. 3-6). Collagen between epidermal and mesenchymal cells is loosely packed, has narrow diameters measuring about 25-30 nm, and is oriented generally along the longitudinal axis of the ray. It does not exhibit the periodic banding pattern typical of the type I collagen of other vertebrate skeletal or dental tissues or comprising the actinotrichia nearby the demirays in extreme distal regions of the fin (Geraudie and Landis, 1982). Actinotrichia and the dermal type I fibrils adjacent to lepidotrichia do not mineralize; these collagen fibrils are of larger diameter than those of lepidotrichia. No evidence of large cellular processes is apparent in the premineralizing and mineralizing ray regions. A few extracellular membrane-bound vesicles are present at
MINERALIZATION OF TROUT LEPIDOTRICHIA
385
Fig. 1. An electron micrograph illustrating a t relatively low magnification a typical distal region of a demiray as selected for the present study from a growing lepidotrichium, fixed in glutaraldehyde and stained with uranyl acetate and lead citrate. The bony fin ray will arise within an extracellular collagenous matrix (C) in which dense patches of a granular ground substance (GS1 appear. The demiray is located between the fin epidermis (E) and mesodermal mesenchyme
(Mt. It thickens proximally and mineralization occurs progressively in a distal-to-proximal (+I direction. A portion of a n actinotrichium (A), a slender, unmineralized rod of elastoidin in the distal margin of the fin (G6raudie and Meunier, 1980),is shown adjacent to two mesenchymal cells. The region here represents the zone of which Figure 2 is a part, at the beginning of the mineralization sequence in a demiray. 22,500 x .
extreme distal regions of lepidotrichia, but they are not apparently associated with mineral deposits. Progressive distal-to-proximal observations (Figs. 3 6) show that the collagen fibrils and ground substance become increasingly obscured by mineral, but it is difficult in these more mature regions to determine if the mineral is associated only with collagen, the granular ground substance, or both extracellular components. The disposition of the mineral seems to be in random small clusters with individual needles or plates apparently oriented in various directions. Although the precise mineral-organic relation is unclear as mineralization proceeds, the overall mass of mineral increases as well as the width of deposition of mineral within the organic regions of the ray, particularly within the dense granular ground substance. The largest dimensions of the mineral fall into a 35-40 nm range. Figures 7-8 present selected-area electron diffraction patterns generated at 100 keV from different sites of mineral deposition along a single lepidotrichium. Newly deposited mineral a t distal portions of rays (Fig.
7) yields no distinct diffraction reflections; that is, the mineral is amorphous to electrons. Proximal to such regions, progressively more mature mineral produces a very weak reflection a t d - 3.4 A and a stronger but broader reflection in the range of d - 2.8 A (Fig. 8); these reflections correspond to the major lines ofa very poorly crystalline hydroxyapatite taken as a reference standard (Landis and Glimcher, 1978). An elemental spectrum characteristic of the dense mineral deposits is shown in Figure 9. This was obtained by electron probe microanalysis over a small region (6-10 nm) of the growing tip of a single fin ray. Only phosphorus and calcium were detected as elements intrinsic to the deposits a t this location and more proximal sites. From an integration of the total number of x-ray counts under the major peaks of calcium and phosphorus, estimated x-ray intensity ratios of CaIP were obtained, converted to CaIP molar ratios (Landis, 19791, and correlated with the progressively developing length of the ray from its distal growing region toward older proximal zones. Figure 10 illus-
386
W.J. LANDIS AND J. GERAUDIE
Figs. 2-4
MINERALIZATION OF TROUT LEPIDOTRICHIA
387
Figs. 2-6. A series of micrographs depicting the onset of mineral deposition in a single lepidotrichium of the trout. The tissue has been fixed in glutaraldehyde and stained with uranyl acetate and lead citrate. The sequence begins (Fig. 2) near the relatively completely unmineralized fin bone apex (corresponding to the more proximal portion of Fig. 1; see also the bony fin ray of Geraudie and Landis, 1982), proceeds (Fig. 3) to additional sites of initial mineral appearance, and continues (Figs. 4-6) proximally toward more mature tissue and increasing mineral mass. The micrographs represent a near montage of the fin bone sectioned longitudinally. There are numbers of observations to be noted over this relatively short distance representing the beginning of mineral formation: 1) The lepidotrichium itself is formed by a meshwork of collagen fibrils (C) at the basal lamella of the dermo-epidermal interface of the fin bud (Fig. 2). The width of the bony ray is the region limited between the epidermal
basal plasmalemma (BP)and the mesenchymal plasmalemma (MP) of the dermis. As the ray develops, this width increases progressively. 2) An electron-dense granular ground substance (GS)forms in the central portion of the collagen network (Fig. 2) and the first observable mineral, highly electron dense by microscopy (arrowheads, Figs. 2-3), is located within the ground material and collagen fibrils. 3 ) As maturation of the ray occurs, the width of the ground substance increases (but does not yet reach the limit in width between epidermal and mesenchymal cell plasmalemma), and the mass of mineral increases. 4) The mineral is disposed randomly, and deposition is apparently restricted to the zone containing the dense ground substance. 5 ) The collagen is thin, with no obvious 64-70 nm periodic banding. LD, Lamina densa of the dermoepidermal interface; E, epidermal cell; M, mesenchymal cell. x 53,000.
trates the change in Ca/P molar ratio over a 100 pm distance in a single ray. The ratios increase from approximately 1.0 distally to about 1.4 measured proximally. The localization of hydroxyapatite within the collagenous matrix of the lepidotrichium is presented in Figures 11-13. In a region containing multiple mineral crystals near the distal third of a ray (Fig. 111, the
c-axial crystallographic planes, imaged in dark field (Fig. 12), are not especially prominent and show little preferential orientation within or associated with collagen. Likewise, the combined a- and b-axial images of the mineral are vague and widely dispersed in the organic matrix of the ray (Fig. 13). The longest dimension of c-axial images is approximately 12-20 nm, that of a,b-axial images about 3-10 nm.
388
W.J. LANDIS AND J. GERAUDIE
ICak.
iI .
e0
16a
240
. 320
'-
r 400
CHRNNEL N O -
488
1
uo
Molar
La)
W P ux)
" C 560
6-10
9
W25keV 1=10-" arrp Spot size=6-lOrm C m t h g thE=100 SeC
10
Fig. 7. Selected-area electron diffraction (100 keV) from distal initial sites of mineral deposition along a lepidotrichium treated by completely anhydrous means without staining yields no distinct diffraction reflections (the mineral is amorphous to electrons). Magnification of transmission image was X 15,000.
was obtained a t 25 keV with a stationary beam spot of current of 1X lo-" amp. Detection time was 200 seconds. The abscissa is calibrated relative to x-ray energy. Calcium and phosphorus are the only two detectable elements intrinsic to the mineral. Copper originates from the grid supporting the tissue sections.
Fig. 8. Selected-area electron diffraction (100 keV) from a more mature (proximal) region of the same fin ray segment of Figure 7. The pattern is different from distal ray portions in the appearance of a weak and narrow reflection at d - 3.4 (inner arrowhead) and a broad reflection at d 2.8 (outer arrowhead). These correspond to the major reflections of a reference standard of a poorly crystalline hydroxyapatite (Landis and Glimcher, 1978). Magnification of transmission image was x 15,000.
Fig. 10.A plot of the change in molar CaiP determined from electron probe microanalysis as a function of length along a typical unstained lepidotrichium treated anhydrously. The ratios were obtained in a longitudinal thin tissue section progressively along the lepidotrichium from the distal initial sites of mineral deposition toward more proximal regions of the fin bud. Molar ratios were found by interpolation from a calibration curve based on electron probe x-ray intensities of known calcium phosphate standards. The relatively lower ratios of CaiP measured distally correspond to the amorphous electron diffraction patterns generated from the same tissue areas, whereas the relatively higher ratios 50-100 pm distant correspond to those patterns of a very poorly crystalline hydroxyapatite.
-
Fig. 9.Energy-dispersive electron probe elemental spectrum generated from a small (6-10 nm) region of dense mineral within the distal portion of an unstained bony ray treated anhydrously. The spectrum
DISCUSSION
The morphogenesis and ultrastructure of the developing pelvic fin dermal skeleton have been described in detail previously by Geraudie and Landis (1982). Results of that investigation have been extended here to include aspects of the characterization of the mineral phase of this tissue. As in other vertebrate mineralizing systems, the inorganic component of the fin bone is a poorly crystalline hydroxyapatite that associates with the constituents of an organic matrix synthesized
and secreted by specialized cells. One of these matrix components is a collagen of narrow diameter (25-30 nm), recently identified as type I (L. Cohen-Solal and J. Gkraudie, unpublished results), although the fibrils themselves do not exhibit the distinct periodicity otherwise characteristic of bone collagen from different species. The fibrils are principally oriented along the longitudinal axis of the ray. At the outset and during further development of calcification, the mineral is found deposited along or within the collagen in a rela-
MINERALIZATION O F TROUT LEPIDOTRICHIA
tively electron-dense central zone between the epidermal basal plasmalemma and mesenchymal plasmalemma. This region also contains a fine granular ground substance whose composition is not precisely known but has been shown (J.Geraudie, unpublished results) to be positive for periodic acid-Schiff (PAS) staining, an indication of the presence of polysaccharides associated with certain proteins, lipids, or phospholipids. Mineralization occurs only in portions of the fin rays containing both collagen and the dense ground substance. The mineral itself consists of fine needles or platelets. The microscopic needle-like appearance of mineral in the fin rays may be misleading, however, as this shape has been shown to occur as a result of observing thin platelets on-edge. Such a conclusion is based on earlier reports with other vertebrates utilizing a goniometer to tilt the sections examined (Landis et al., 1977; Landis and Glimcher, 1982) or, even more recently, a direct three-dimensional viewing technique, topographic imaging (Landis et al., in press). It is probable, but yet to be determined, that the needle-shaped mineral of lepidotrichia is of the same thin plate-like nature suggested by these other studies. The size of the mineral, which is critical to understanding the basis of crystal deposition within collagen, is approximately 3 x 40 nm in its shortest and longest dimension, respectively, as determined by dark field or bright field microscopy. The 3 nm estimate agrees with values derived from other transmission microscope studies (Robinson and Watson, 1952, 1955; Weiner and Price, 1986; Arsenault and Grynpas, 1988) and new reports using small-angle x-ray scattering (Fratzl et al., 1989) or topographic imaging (Landis et al., in press). The distance would correspond to the thickness of an individual platelet. The 40 nm dimension is also consistent with many published results of the crystallographic c-axial length of vertebrate mineral particles. On the other hand, the present analysis by selected area electron diffraction-dark field imaging suggests the c-axial image of the mineral from lepidotrichia is in the range of 12-20 nm, about a third to a half that determined from the microscopic bright field mode. Such values are similar to those of Arsenault and Grynpas (1988), who used the same technique to measure the dimensions of mineral crystals present in calcifying cartilage and bone of rat. A resolution of the discrepancy in the c-axial length of the mineral requires further study. It may simply lie in the different technical approaches used and the extremely sensitive nature of the dark field method: images in this mode of the microscope will be detectable only if a crystal axial plane is within 3" to the normal of the electron beam (W.J. Landis, unpublished results). Thus, if a crystal is not lying flat on its supporting grid for microscopy or is twisted or otherwise deviates beyond the 3" limit, it may appear shortened in planar dimension or may not be observed at all. With regard to the spatial relation between individual crystals and the collagen fibrils with which they associate in the fin rays, both bright field and dark field microscopy indicate that the c-axial mineral planes are oriented in the general direction of the collagen long axis. However, the orientation and distribution of the mineral c- and a,b-axial dark field images bear no consistent and specific relation with either the
389
hole or the overlap zones of the collagen in the fin bones. This feature is quite distinct from that reported in calcifying tendon (Arsenault, 1988; Landis and Arsenault, 1988) in which the dark field mineral crystal axes are characteristically prominent in appearance and are easily recognized as corresponding to hole and overlap zones of the protein. The precise reason for this difference in the images reflecting mineral-collagen interaction is not clear but may likely be connected with the relatively random orientation of the collagen fibrils themselves within the lepidotrichia compared with the highly oriented manner of collagen alignment in the tendon. Again, as discussed above, under the constraints of large angular deviations from the normal to the electron beam as would occur in the ray with mineral in randomly oriented collagen, the crystal axial images may not appear. Their absence, however, would not be indicative of differences in the fundamental association of mineral with hole and overlap zones of collagen as found in the fin rays and in all other normal vertebrate calcifying tissues except enamel. With the progression of mineralization, as noted previously (Geraudie and Landis, 1982), there is an increasing mass of mineral deposited in the lepidotrichium, and the mineral particles increase in number with both increasing distance from the distal extremes of the ray (longitudinally) and increasing distance from the mesenchymal cell membrane up to the basal lamina of the epidermis (transversely). Ultimately, the collagen and particularly the dense ground substance are obscured by mineral along the full extent of the lepidotrichia; but at earlier developmental times, the collagen and dense ground material adjacent to the mesenchymal plasmalemma are not obliterated by inorganic deposits. These features of mineral deposition suggest two different temporal and spatial patterns occurring during mineral phase maturation, one along the length and the other along the width of the ray regardless of the proximodistal location (Geraudie and Landis, 1982). This concept has some support in the current results from electron diffraction and electron probe microanalysis. The electron diffraction pattern changes with length along the ray from a very vague halo at -2.8 b in regions of newly deposited mineral to that having two or three more discrete reflections in the fin where an increasing mineral mass is present. Correspondingly, the sites of recent deposition of mineral in distal fin regions generate low Ca/P ratios, whereas sites in proximal fin bone areas have higher Ca/P ratios. The maximum ratios fall in the range of about 1.4, a value in general close to that obtained by microanalysis of early mineral in other tissues, including embryonic chick bone (Landis and Glimcher, 1978), rat growth plate cartilage (Landis and Glimcher, 1982), and rat incisor enamel (Landis et al., 1988). Indeed, the deposition and maturity of rat enamel assessed by diffraction and microanalysis follow two different directions along the incisors (Landis et al., 19881, as the present data also suggest for lepidotrichia. There appear to be no other structures besides the collagen fibrils and dense ground substance associated with the mineral in the lepidotrichia. Membranebound extracellular vesicles are present at extreme distal portions of the tissue, but they stain poorly and do
390
W.J. LANDIS AND J. GERAUDIE
Figs. 11-13.
MINERALIZATION OF TROUT LEPIDOTRICHIA
391
not contain mineral particles (Geraudie and Landis, and the Bristol Myers/Zimmer Corporation. The au1982). As such, they differ from the vesicles sequester- thors are grateful for the technical assistance of Karen ing mineral as reported in other tissues, including cal- J. Hodgens (Boston) and Francoise Allizard (Paris). cifying cartilage (Anderson, 1969; Bonucci, 1971) or Electron microscopy was performed in part a t the Centendon (Landis, 1986). In the developing bony fin rays tre d'accueil de Microscopie Electronique, Universite of the trout, it is interesting that the diameter of the Pierre et Marie Curie-CNRS, 105, Boulevard Raspail, type I collagen fibrils where mineralization occurs is 75005 Paris. smaller than that of the adjacent dermal type I fibrils, LITERATURE CITED which remain unmineralized in the regions between lepidotrichia. In addition, the small fibrils exhibit a Anderson, H.C. 1969 Vesicles associated with calcification in the matrix of epiphyseal cartilage. J . Cell Biol., 41:59-72. weak periodicity on staining compared with the larger fibrils. The reasons for mineral deposition apparently Arsenault, A.L. 1988 Crystal-collagen relationships in calcified turkey leg tendons visualized by selected area dark field electron associated with a specific type I fibril and for differmicroscopy. Calcif. Tissue Int., 43:202-212. ences in sizes and banding of collagen, as well as the Arsenault, A.L., and M.D. Grynpas 1988 Crystals in calcified epiphyseal cartilage and cortical bone of the rat. Calcif. Tissue Int., source of these fibrils, are not known. It may be that 43t219-225. the fin bud exemplifies a structure in which mineral- Bonucci, E. 1971 The locus of initial calcification in cartilage and ization is critically influenced by the yet unidentified bone. Clin. Orthop., 78t108-139. ground substance, for instance, by collagen cross- Fratzl, P., N. Fratzl-Zelman, K. Klaushofer, 0. Hoffman, G. Vogl, and K. Koller 1989 Age related changes of crystal size and orientation linking and/or the origin of the fibrils themselves. With in bone tissue: A small-angle x-ray scattering study. Calcif. Tisregard to the latter, the fibrils may derive from two sue Int., 44:S91. distinct mesenchymal-like cells in the fin or from a Geraudie, J., and W. J. Landis 1989 Mineral deposition in the develsingle cell whose type I procollagen may be subject to oping pelvic fin bud of the trout, Salmo guL'rdneri. Connect. Tissue Res., 22t224. post-translational changes to produce fibrils of differJ., and W.J. Landis 1982 The fine structure of the developent diameters and potential to mineralize. These vari- Geraudie, ing pelvic fin skeleton in the trout, Salmo guirdneri. Am. J. ous possibilities of cellular or matrix control on calciAnat., 163t141-157. fication are under investigation. Geraudie, J., and F.J. Meunier 1980 Elastoidin actinotrichia in coelacanth fins: A comparative study with teleosts. Tissue Cell, 12: Electron probe microanalysis demonstrates only cal637-645. cium and phosphorus as the elements intrinsic to the Kendall, R.L. 1988 Taxonomic changes in North American trout fin bones. Because the ground substance of the bones names. Trans. Am. Fish. SOC.,117:321. appears to contain polysaccharides, sulfur could also be Landis, W.J. 1979 Application of electron probe x-ray microanalysis to calcification studies of bone and cartilage. In: Scanning Electron present. Its detection, however, as well as that of other Microscopy. Vol. 11.0. Johari, ed. SEM, Inc., AMF OHare, IL, pp. elements such as magnesium, silicon, chlorine, and so555-570. dium, possibly associated with mineralization of this Landis, W.J. 1986 A study of calcification in the leg tendons from the tissue, is likely to be below instrumentation limits. The domestic turkey. J. Ultrastruct. Mol. Struct. Res., 94r217-238. application of ion microscopy (Landis et al., 1986) or Landis, W.J., and A.L. Arsenault 1989 Vesicle- and collagen-mediated calcification in the turkey leg tendon. Connect. Tissue Res., 22; alternative techniques more sensitive to these minor 35-42. constituents putatively within lepidotrichia is being Landis, W.J., G.Y. Burke, J.R. Neuringer, M.C. Paine, A. Nanci, P. pursued in conjunction with additional approaches to Bai, and H. Warshawsky 1988 Earliest enamel deposits of the rat incisor examined by electron microscopy, electron diffraction, and define more fully the organic matrix-mineral interacelectron probe microanalysis. Anat. Rec., 220:233-238. tions in this vertebrate bone model. Landis, W.J., and M.J. Glimcher 1978 Electron diffraction and elecACKNOWLEDGMENTS
This work was supported by grants from the National Institutes of Health (AR 34078 and AR 34081) and the CNRS (UA 041137), the Peabody Foundation, Figs. 11-1 3. A series of micrographs illustrating, respectively, the bright field image, dark field image of mineral c-axes, and dark field image of mineral u,b-axes taken from the same extracellular region of a newly developing lepidotrichium treated by anhydrous means without staining. The mineral phase is associated with a collagenous matrix. The mineral particles, appearing in bright field (Fig. 11)as fine, electron-dense needle- and plate-like material about 35-40 nm in its longest dimension, are found in dark field (Fig. 12) to be composed of smaller crystals 12-20 nm along their c-axis lengths (in dark field, the crystals are the bright images set against the dense background). Both the c-axis and +axes images of the mineral are localized along collagen in a seemingly random manner. Unlike their distribution in a regular fashion within the hole and overlap zones of collagen fibrils in the turkey tendon (Arsenault, 1988; Landis and Arsenault, 1989), the crystal axes dark field images do not appear repetitively along the same specific regions of the collagen fibrils in lepidotrichia. x 32,000.
tron probe microanalysis of the mineral phase of bone tissue prepared by anhydrous techniques. J . Ultrastruct. Res., 63r188-223. Landis, W.J., and M.J. Glimcher 1982 Electron optical and analytical observations of rat growth plate cartilage prepared by ultracryomicrotomy. J . Ultrastruct. Res., 78r227-268. Landis, W.J., D.D. Lee, J.T. Brenna, S. Chandra, and G.H. Morrison 1986 Detection and localization of silicon and associated elements in vertebrate bone tissue by imaging ion microscopy. Calcif. Tissue Int., 38:52-59. Landis, W.J., J. Moradian-Oldak, and S. Weiner 1990 Topographic imaging of mineral and collagen in the calcifying turkey tendon. Connect. Tissue Res. (in press). Landis, W.J., M.C. Paine, and M.J. Glimcher 1977 Electron microscopic observations of bone tissue prepared anhydrously in organic solvents. J. Ultrastruct. Res., 59:l-30. Robinson, R.A., and M.L. Watson 1952 Collagen-crystal relationships in bone as seen in the electron microscope. Anat. Rec., 114:383409. Robinson, R.A., and M.L. Watson 1955 Crystal-collagen relationships in bone as seen in the electron microscope. 111. Crystal and collagen morphology as a function of age. Ann. N.Y. Acad. Sci., 60: 596-628. Weiner, S., and P.A. Price 1986 Disaggregation of bone into crystals. Calcif. Tissue Int., 39:365-3 75.
Organization and Development of the Mineral Phase During Early Ontogenesis of the Bony Fin Rays of the Trout Oncorhynchus mykiss WILLIAM J. LANDIS AND JACQUELINE GERAUDIE Department of Orthopedic Surgery, Harvard Medical School, Children's Hospital, Boston, Massachusetts (W.J.L.);and Laboratoire d'Anatomie Comparke, Universitt! Paris VII, Paris Cedex 05, France
ABSTRACT Characterization of mineral deposition has been studied by electron optical methods during early ontogenesis of lepidotrichia, the bony fin rays, of the trout Oncorhynchus mykiss (the former Salmo gairdneri). The fin rays consist of a n extracellular granular ground substance containing in part a network of collagen fibrils within the basal lamella of the fin dermoepidermal interface. Growth of individual rays proceeds in a proximodistal direction. The mineral phase appears as electron-dense needle or plate-like particles and is associated with the collagenous matrix. On analysis of progressively maturing tissue, the mineral was characterized as a poorly crystalline hydroxyapatite with Ca/P molar ratios in the range of 1.0-1.4, corresponding to distal and proximal areas, respectively. With selected-area electron diffraction and dark field imaging of lepidotrichia, the mineral particles were found to be about 3-10 nm thick and 12-20 nm in length (along their crystallographic c-axes), possibly aggregated into larger crystals 35-40 nm long observed with bright field microscopy. No definitive relation was found between either the c- or a,b-axes images of the crystals and the periodic structure of collagen, which forms the framework for mineral deposition in this and in other vertebrate calcifying tissues.
The identification and mode of deposition of the mineral phase in the fin rays or lepidotrichia of the growing trout fry are of interest for modeling the pattern of mineralization for various vertebrate bone tissues. It is known from previous work (Geraudie and Landis, 1982) t h a t lepidotrichial ontogenesis occurs by direct mineralization of a collagenous basal lamella of the skin without passage through a cartilaginous precursor. Consequently, this bone is strictly of dermal origin. Another characteristic of this bone is a progressive apical growth during the life span of the fish. Elongation of the bony rays is achieved by the distal addition of new segments, which calcify successively. This unusual continuous and sequential spatial and temporal process of mineralization has been studied to determine the biophysical quality of the initial mineral deposits in the tissue and their maturation along the growing new segments. The present work was accomplished with conventional electron microscopy combined with specific electron optical and analytical techniques after anhydrous preparation of bone samples, a method that prevents mineral loss and translocation (Landis et al., 1977). The results contribute to a more complete description of the interaction between mineral and collagen and form the basis for understanding initial events in this example of vertebrate bone mineralization. A portion of this study has been presented in preliminary form (Geraudie and Landis, 1989). 0 1990 WILEY-LISS, INC.
MATERIALS AND METHODS Materials
The present study was carried out on the rainbow trout Oncorhynchus mykiss (new identification of the former Salmo gairdneri; Kendall, 1988). Larvae were raised from fertilization in running tap water kept at a temperature of 12 2 1°C. Hatching occurs about 1 month after fertilization. Growing fins were dissected from anesthetized fish (1% MS222:ethyl m-aminobenzoate in 100 ml of tap water) 1 month after hatching. Methods
A number of electron optical and analytical techniques were applied to examine the developing fin rays. These included the following. Electron microscopy under conventional aqueous fixation and processing
In order to observe the morphology of the growing distal lepidotrichium segments containing newly deposited mineral particles, isolated pieces of different fins (pectoral, pelvic, caudal) were immersed in 2.5% glutaraldehyde in 0.1M cacodylate buffer, pH 7.2, at Received November 15, 1989; accepted March 22, 1990. Address reprint requests to Dr. William J. Landis, Department of Orthopedic Surgery, Enders Bldg-Room SB 20, 300 Longwood Avenue, Children's Hospital, Boston, MA 02115.
384
W.J. LANDIS AND J. GERAUDIE
4°C for 2 hours. Fully dehydrated specimens were embedded in Epon 812 for longitudinal sectioning. Thick sections (1pm) were stained with toluidine blue in 1% borax. Thin sections (-80 nm) were obtained with a diamond knife and floated on distilled water for subsequent collection on copper grids. No attempt was made to reduce mineral loss in handling these sections, which were contrasted with aqueous uranyl acetate and lead citrate. Sections were observed on a Philips 300 electron microscope operated a t 80 keV. Electron microscopy under anhydrous conditions of tissue preparation
Anhydrous methods were used in parallel with the conventional techniques above. Anhydrous approaches minimize or prevent artifacts in the mineral phase of bone tissues (Landis et al., 1977). Isolated pieces of fins were dissected in cold (4°C)ethylene glycol (100%)and then kept under vacuum in the same solvent for 3 hours at room temperature with constant mild agitation. Ethylene glycol was changed, and the specimens remained under the same conditions for an additional 24 hours, after which they were immersed at 4°C in Cellosolve, replaced twice over the next 24 hour period. Tissues were then kept at room temperature in a mixture of propylene oxide-Epon 812 (1:l)for 1 week and finally embedded in Epon for transverse or longitudinal sectioning of the fins. Once appropriate regions of thick sections were identified for thin sectioning, the block was trimmed to remove excess Epon. Sections (80-100 nm) were cut with a diamond knife and floated onto 100% ethylene glycol. Thin sections were examined either unstained or contrasted with uranyl acetate and lead citrate in 100% ethylene glycol or in distilled water, depending on the observations to be made. Conventional bright field electron microscopy and electron diffraction
Stained and unstained thin sections were observed with a Philips EM 300 operated a t 60 or 80 keV or a JEOL lOOC operated a t 60-100 keV. Selected-area electron diffraction was carried out on the JEOL lOOC at 80 or 100 keV. Evaporated gold, aluminum, and thallous chloride were used as electron diffraction calibration standards (Landis and Glimcher, 1978). Selected-area electron diffraction-dark field imaging
Application of this technique was based on that described by Arsenault (1988) to determine the size, orientation, and distribution of the mineral crystals present in a tissue specimen. Epon-embedded samples were thin-sectioned (80 nm) so that longitudinal profiles of fin rays were obtained. Unstained sections were examined at 80 or 100 keV in a Philips EM 300, and early mineralizing regions of the tissues were studied by selected-area electron diffraction. Diffraction ref lections corresponding to the a- and b-axes together and to the c-axis alone of the apatite crystals comprising the ray mineral were isolated in a small (1-4 pm) objective aperture (Energy Beam Sciences, Agawam, MA) and then imaged. The a,b-axial orientations of the crystals were determined from the (211), (112), (300), (202), and (301) lattice planes of apatite. The c-axial orientation was determined from the (002) lattice
plane. Images of the respective planes were observed in the dark field mode of the microscope and thereby appeared in reverse contrast compared to the images of conventional bright field microscopy. High spatial resolution electron probe x-ray microanalysis
The same sections studied by electron diffractiondark field imaging and conventional bright field microscopy were used for electron probe microanalysis in a JEOL JSM 50A scanning electron microscope modified for high spatial resolution (6-10 nm) (Landis and Glimcher, 1982). Accelerating volta e was 25 keV, with a fixed-beam current of 1 x 10- 15 amp, and x-ray spectra were recorded during 100-600 seconds integrated detecting time periods. Molar ratios of calcium to phosphorus, corrected for background, were found by interpolation from a calibration curve of x-ray intensities generated from a number of calcium phosphate standards analyzed in the same manner as the fin bone tissue (Landis, 1979). RESULTS Morphological Evolution of the Developing Bone
A morphological description of the developing mineral in the fin ray begins with observations from the growing distal extremities progressively toward more proximal regions of the tissue. Calcification, based on the spatial location and disposition of mineral crystal particles of different sizes, proceeds in a proximodistal direction (Geraudie and Landis, 1982). Figures 1-6 depict the onset and sequential deposition of mineral in a single lepidotrichium. The first indication of a demiray is the appearance of an electron-dense ground substance within the collagen network of the basal lamella of the dermoepidermal interface (Figs. 1,2). Within the interface, this granular material does not extend to the basal lamina of the epidermis on one side and is not directly contiguous with the cells of the dermis on the other side. The ground substance does not fully cover the width of the collagen network in even more proximal regions of the tissue. The collagenous network changes only gradually in width, and within it the ground substance follows as a demiray matures. Near the distal regions of the demiray (Figs. 2-4) isolated needle- or plate-like particles of high electron density are found; these are associated, apparently exclusively, with collagen fibrils. Such particles are the first observed mineral crystals in the tissue (Figs. 7-10). Most mineral-associated fibrils lie within the confines of the granular ground substance (Figs. 3-6). Collagen between epidermal and mesenchymal cells is loosely packed, has narrow diameters measuring about 25-30 nm, and is oriented generally along the longitudinal axis of the ray. It does not exhibit the periodic banding pattern typical of the type I collagen of other vertebrate skeletal or dental tissues or comprising the actinotrichia nearby the demirays in extreme distal regions of the fin (Geraudie and Landis, 1982). Actinotrichia and the dermal type I fibrils adjacent to lepidotrichia do not mineralize; these collagen fibrils are of larger diameter than those of lepidotrichia. No evidence of large cellular processes is apparent in the premineralizing and mineralizing ray regions. A few extracellular membrane-bound vesicles are present at
MINERALIZATION OF TROUT LEPIDOTRICHIA
385
Fig. 1. An electron micrograph illustrating a t relatively low magnification a typical distal region of a demiray as selected for the present study from a growing lepidotrichium, fixed in glutaraldehyde and stained with uranyl acetate and lead citrate. The bony fin ray will arise within an extracellular collagenous matrix (C) in which dense patches of a granular ground substance (GS1 appear. The demiray is located between the fin epidermis (E) and mesodermal mesenchyme
(Mt. It thickens proximally and mineralization occurs progressively in a distal-to-proximal (+I direction. A portion of a n actinotrichium (A), a slender, unmineralized rod of elastoidin in the distal margin of the fin (G6raudie and Meunier, 1980),is shown adjacent to two mesenchymal cells. The region here represents the zone of which Figure 2 is a part, at the beginning of the mineralization sequence in a demiray. 22,500 x .
extreme distal regions of lepidotrichia, but they are not apparently associated with mineral deposits. Progressive distal-to-proximal observations (Figs. 3 6) show that the collagen fibrils and ground substance become increasingly obscured by mineral, but it is difficult in these more mature regions to determine if the mineral is associated only with collagen, the granular ground substance, or both extracellular components. The disposition of the mineral seems to be in random small clusters with individual needles or plates apparently oriented in various directions. Although the precise mineral-organic relation is unclear as mineralization proceeds, the overall mass of mineral increases as well as the width of deposition of mineral within the organic regions of the ray, particularly within the dense granular ground substance. The largest dimensions of the mineral fall into a 35-40 nm range. Figures 7-8 present selected-area electron diffraction patterns generated at 100 keV from different sites of mineral deposition along a single lepidotrichium. Newly deposited mineral a t distal portions of rays (Fig.
7) yields no distinct diffraction reflections; that is, the mineral is amorphous to electrons. Proximal to such regions, progressively more mature mineral produces a very weak reflection a t d - 3.4 A and a stronger but broader reflection in the range of d - 2.8 A (Fig. 8); these reflections correspond to the major lines ofa very poorly crystalline hydroxyapatite taken as a reference standard (Landis and Glimcher, 1978). An elemental spectrum characteristic of the dense mineral deposits is shown in Figure 9. This was obtained by electron probe microanalysis over a small region (6-10 nm) of the growing tip of a single fin ray. Only phosphorus and calcium were detected as elements intrinsic to the deposits a t this location and more proximal sites. From an integration of the total number of x-ray counts under the major peaks of calcium and phosphorus, estimated x-ray intensity ratios of CaIP were obtained, converted to CaIP molar ratios (Landis, 19791, and correlated with the progressively developing length of the ray from its distal growing region toward older proximal zones. Figure 10 illus-
386
W.J. LANDIS AND J. GERAUDIE
Figs. 2-4
MINERALIZATION OF TROUT LEPIDOTRICHIA
387
Figs. 2-6. A series of micrographs depicting the onset of mineral deposition in a single lepidotrichium of the trout. The tissue has been fixed in glutaraldehyde and stained with uranyl acetate and lead citrate. The sequence begins (Fig. 2) near the relatively completely unmineralized fin bone apex (corresponding to the more proximal portion of Fig. 1; see also the bony fin ray of Geraudie and Landis, 1982), proceeds (Fig. 3) to additional sites of initial mineral appearance, and continues (Figs. 4-6) proximally toward more mature tissue and increasing mineral mass. The micrographs represent a near montage of the fin bone sectioned longitudinally. There are numbers of observations to be noted over this relatively short distance representing the beginning of mineral formation: 1) The lepidotrichium itself is formed by a meshwork of collagen fibrils (C) at the basal lamella of the dermo-epidermal interface of the fin bud (Fig. 2). The width of the bony ray is the region limited between the epidermal
basal plasmalemma (BP)and the mesenchymal plasmalemma (MP) of the dermis. As the ray develops, this width increases progressively. 2) An electron-dense granular ground substance (GS)forms in the central portion of the collagen network (Fig. 2) and the first observable mineral, highly electron dense by microscopy (arrowheads, Figs. 2-3), is located within the ground material and collagen fibrils. 3 ) As maturation of the ray occurs, the width of the ground substance increases (but does not yet reach the limit in width between epidermal and mesenchymal cell plasmalemma), and the mass of mineral increases. 4) The mineral is disposed randomly, and deposition is apparently restricted to the zone containing the dense ground substance. 5 ) The collagen is thin, with no obvious 64-70 nm periodic banding. LD, Lamina densa of the dermoepidermal interface; E, epidermal cell; M, mesenchymal cell. x 53,000.
trates the change in Ca/P molar ratio over a 100 pm distance in a single ray. The ratios increase from approximately 1.0 distally to about 1.4 measured proximally. The localization of hydroxyapatite within the collagenous matrix of the lepidotrichium is presented in Figures 11-13. In a region containing multiple mineral crystals near the distal third of a ray (Fig. 111, the
c-axial crystallographic planes, imaged in dark field (Fig. 12), are not especially prominent and show little preferential orientation within or associated with collagen. Likewise, the combined a- and b-axial images of the mineral are vague and widely dispersed in the organic matrix of the ray (Fig. 13). The longest dimension of c-axial images is approximately 12-20 nm, that of a,b-axial images about 3-10 nm.
388
W.J. LANDIS AND J. GERAUDIE
ICak.
iI .
e0
16a
240
. 320
'-
r 400
CHRNNEL N O -
488
1
uo
Molar
La)
W P ux)
" C 560
6-10
9
W25keV 1=10-" arrp Spot size=6-lOrm C m t h g thE=100 SeC
10
Fig. 7. Selected-area electron diffraction (100 keV) from distal initial sites of mineral deposition along a lepidotrichium treated by completely anhydrous means without staining yields no distinct diffraction reflections (the mineral is amorphous to electrons). Magnification of transmission image was X 15,000.
was obtained a t 25 keV with a stationary beam spot of current of 1X lo-" amp. Detection time was 200 seconds. The abscissa is calibrated relative to x-ray energy. Calcium and phosphorus are the only two detectable elements intrinsic to the mineral. Copper originates from the grid supporting the tissue sections.
Fig. 8. Selected-area electron diffraction (100 keV) from a more mature (proximal) region of the same fin ray segment of Figure 7. The pattern is different from distal ray portions in the appearance of a weak and narrow reflection at d - 3.4 (inner arrowhead) and a broad reflection at d 2.8 (outer arrowhead). These correspond to the major reflections of a reference standard of a poorly crystalline hydroxyapatite (Landis and Glimcher, 1978). Magnification of transmission image was x 15,000.
Fig. 10.A plot of the change in molar CaiP determined from electron probe microanalysis as a function of length along a typical unstained lepidotrichium treated anhydrously. The ratios were obtained in a longitudinal thin tissue section progressively along the lepidotrichium from the distal initial sites of mineral deposition toward more proximal regions of the fin bud. Molar ratios were found by interpolation from a calibration curve based on electron probe x-ray intensities of known calcium phosphate standards. The relatively lower ratios of CaiP measured distally correspond to the amorphous electron diffraction patterns generated from the same tissue areas, whereas the relatively higher ratios 50-100 pm distant correspond to those patterns of a very poorly crystalline hydroxyapatite.
-
Fig. 9.Energy-dispersive electron probe elemental spectrum generated from a small (6-10 nm) region of dense mineral within the distal portion of an unstained bony ray treated anhydrously. The spectrum
DISCUSSION
The morphogenesis and ultrastructure of the developing pelvic fin dermal skeleton have been described in detail previously by Geraudie and Landis (1982). Results of that investigation have been extended here to include aspects of the characterization of the mineral phase of this tissue. As in other vertebrate mineralizing systems, the inorganic component of the fin bone is a poorly crystalline hydroxyapatite that associates with the constituents of an organic matrix synthesized
and secreted by specialized cells. One of these matrix components is a collagen of narrow diameter (25-30 nm), recently identified as type I (L. Cohen-Solal and J. Gkraudie, unpublished results), although the fibrils themselves do not exhibit the distinct periodicity otherwise characteristic of bone collagen from different species. The fibrils are principally oriented along the longitudinal axis of the ray. At the outset and during further development of calcification, the mineral is found deposited along or within the collagen in a rela-
MINERALIZATION O F TROUT LEPIDOTRICHIA
tively electron-dense central zone between the epidermal basal plasmalemma and mesenchymal plasmalemma. This region also contains a fine granular ground substance whose composition is not precisely known but has been shown (J.Geraudie, unpublished results) to be positive for periodic acid-Schiff (PAS) staining, an indication of the presence of polysaccharides associated with certain proteins, lipids, or phospholipids. Mineralization occurs only in portions of the fin rays containing both collagen and the dense ground substance. The mineral itself consists of fine needles or platelets. The microscopic needle-like appearance of mineral in the fin rays may be misleading, however, as this shape has been shown to occur as a result of observing thin platelets on-edge. Such a conclusion is based on earlier reports with other vertebrates utilizing a goniometer to tilt the sections examined (Landis et al., 1977; Landis and Glimcher, 1982) or, even more recently, a direct three-dimensional viewing technique, topographic imaging (Landis et al., in press). It is probable, but yet to be determined, that the needle-shaped mineral of lepidotrichia is of the same thin plate-like nature suggested by these other studies. The size of the mineral, which is critical to understanding the basis of crystal deposition within collagen, is approximately 3 x 40 nm in its shortest and longest dimension, respectively, as determined by dark field or bright field microscopy. The 3 nm estimate agrees with values derived from other transmission microscope studies (Robinson and Watson, 1952, 1955; Weiner and Price, 1986; Arsenault and Grynpas, 1988) and new reports using small-angle x-ray scattering (Fratzl et al., 1989) or topographic imaging (Landis et al., in press). The distance would correspond to the thickness of an individual platelet. The 40 nm dimension is also consistent with many published results of the crystallographic c-axial length of vertebrate mineral particles. On the other hand, the present analysis by selected area electron diffraction-dark field imaging suggests the c-axial image of the mineral from lepidotrichia is in the range of 12-20 nm, about a third to a half that determined from the microscopic bright field mode. Such values are similar to those of Arsenault and Grynpas (1988), who used the same technique to measure the dimensions of mineral crystals present in calcifying cartilage and bone of rat. A resolution of the discrepancy in the c-axial length of the mineral requires further study. It may simply lie in the different technical approaches used and the extremely sensitive nature of the dark field method: images in this mode of the microscope will be detectable only if a crystal axial plane is within 3" to the normal of the electron beam (W.J. Landis, unpublished results). Thus, if a crystal is not lying flat on its supporting grid for microscopy or is twisted or otherwise deviates beyond the 3" limit, it may appear shortened in planar dimension or may not be observed at all. With regard to the spatial relation between individual crystals and the collagen fibrils with which they associate in the fin rays, both bright field and dark field microscopy indicate that the c-axial mineral planes are oriented in the general direction of the collagen long axis. However, the orientation and distribution of the mineral c- and a,b-axial dark field images bear no consistent and specific relation with either the
389
hole or the overlap zones of the collagen in the fin bones. This feature is quite distinct from that reported in calcifying tendon (Arsenault, 1988; Landis and Arsenault, 1988) in which the dark field mineral crystal axes are characteristically prominent in appearance and are easily recognized as corresponding to hole and overlap zones of the protein. The precise reason for this difference in the images reflecting mineral-collagen interaction is not clear but may likely be connected with the relatively random orientation of the collagen fibrils themselves within the lepidotrichia compared with the highly oriented manner of collagen alignment in the tendon. Again, as discussed above, under the constraints of large angular deviations from the normal to the electron beam as would occur in the ray with mineral in randomly oriented collagen, the crystal axial images may not appear. Their absence, however, would not be indicative of differences in the fundamental association of mineral with hole and overlap zones of collagen as found in the fin rays and in all other normal vertebrate calcifying tissues except enamel. With the progression of mineralization, as noted previously (Geraudie and Landis, 1982), there is an increasing mass of mineral deposited in the lepidotrichium, and the mineral particles increase in number with both increasing distance from the distal extremes of the ray (longitudinally) and increasing distance from the mesenchymal cell membrane up to the basal lamina of the epidermis (transversely). Ultimately, the collagen and particularly the dense ground substance are obscured by mineral along the full extent of the lepidotrichia; but at earlier developmental times, the collagen and dense ground material adjacent to the mesenchymal plasmalemma are not obliterated by inorganic deposits. These features of mineral deposition suggest two different temporal and spatial patterns occurring during mineral phase maturation, one along the length and the other along the width of the ray regardless of the proximodistal location (Geraudie and Landis, 1982). This concept has some support in the current results from electron diffraction and electron probe microanalysis. The electron diffraction pattern changes with length along the ray from a very vague halo at -2.8 b in regions of newly deposited mineral to that having two or three more discrete reflections in the fin where an increasing mineral mass is present. Correspondingly, the sites of recent deposition of mineral in distal fin regions generate low Ca/P ratios, whereas sites in proximal fin bone areas have higher Ca/P ratios. The maximum ratios fall in the range of about 1.4, a value in general close to that obtained by microanalysis of early mineral in other tissues, including embryonic chick bone (Landis and Glimcher, 1978), rat growth plate cartilage (Landis and Glimcher, 1982), and rat incisor enamel (Landis et al., 1988). Indeed, the deposition and maturity of rat enamel assessed by diffraction and microanalysis follow two different directions along the incisors (Landis et al., 19881, as the present data also suggest for lepidotrichia. There appear to be no other structures besides the collagen fibrils and dense ground substance associated with the mineral in the lepidotrichia. Membranebound extracellular vesicles are present at extreme distal portions of the tissue, but they stain poorly and do
390
W.J. LANDIS AND J. GERAUDIE
Figs. 11-13.
MINERALIZATION OF TROUT LEPIDOTRICHIA
391
not contain mineral particles (Geraudie and Landis, and the Bristol Myers/Zimmer Corporation. The au1982). As such, they differ from the vesicles sequester- thors are grateful for the technical assistance of Karen ing mineral as reported in other tissues, including cal- J. Hodgens (Boston) and Francoise Allizard (Paris). cifying cartilage (Anderson, 1969; Bonucci, 1971) or Electron microscopy was performed in part a t the Centendon (Landis, 1986). In the developing bony fin rays tre d'accueil de Microscopie Electronique, Universite of the trout, it is interesting that the diameter of the Pierre et Marie Curie-CNRS, 105, Boulevard Raspail, type I collagen fibrils where mineralization occurs is 75005 Paris. smaller than that of the adjacent dermal type I fibrils, LITERATURE CITED which remain unmineralized in the regions between lepidotrichia. In addition, the small fibrils exhibit a Anderson, H.C. 1969 Vesicles associated with calcification in the matrix of epiphyseal cartilage. J . Cell Biol., 41:59-72. weak periodicity on staining compared with the larger fibrils. The reasons for mineral deposition apparently Arsenault, A.L. 1988 Crystal-collagen relationships in calcified turkey leg tendons visualized by selected area dark field electron associated with a specific type I fibril and for differmicroscopy. Calcif. Tissue Int., 43:202-212. ences in sizes and banding of collagen, as well as the Arsenault, A.L., and M.D. Grynpas 1988 Crystals in calcified epiphyseal cartilage and cortical bone of the rat. Calcif. Tissue Int., source of these fibrils, are not known. It may be that 43t219-225. the fin bud exemplifies a structure in which mineral- Bonucci, E. 1971 The locus of initial calcification in cartilage and ization is critically influenced by the yet unidentified bone. Clin. Orthop., 78t108-139. ground substance, for instance, by collagen cross- Fratzl, P., N. Fratzl-Zelman, K. Klaushofer, 0. Hoffman, G. Vogl, and K. Koller 1989 Age related changes of crystal size and orientation linking and/or the origin of the fibrils themselves. With in bone tissue: A small-angle x-ray scattering study. Calcif. Tisregard to the latter, the fibrils may derive from two sue Int., 44:S91. distinct mesenchymal-like cells in the fin or from a Geraudie, J., and W. J. Landis 1989 Mineral deposition in the develsingle cell whose type I procollagen may be subject to oping pelvic fin bud of the trout, Salmo guL'rdneri. Connect. Tissue Res., 22t224. post-translational changes to produce fibrils of differJ., and W.J. Landis 1982 The fine structure of the developent diameters and potential to mineralize. These vari- Geraudie, ing pelvic fin skeleton in the trout, Salmo guirdneri. Am. J. ous possibilities of cellular or matrix control on calciAnat., 163t141-157. fication are under investigation. Geraudie, J., and F.J. Meunier 1980 Elastoidin actinotrichia in coelacanth fins: A comparative study with teleosts. Tissue Cell, 12: Electron probe microanalysis demonstrates only cal637-645. cium and phosphorus as the elements intrinsic to the Kendall, R.L. 1988 Taxonomic changes in North American trout fin bones. Because the ground substance of the bones names. Trans. Am. Fish. SOC.,117:321. appears to contain polysaccharides, sulfur could also be Landis, W.J. 1979 Application of electron probe x-ray microanalysis to calcification studies of bone and cartilage. In: Scanning Electron present. Its detection, however, as well as that of other Microscopy. Vol. 11.0. Johari, ed. SEM, Inc., AMF OHare, IL, pp. elements such as magnesium, silicon, chlorine, and so555-570. dium, possibly associated with mineralization of this Landis, W.J. 1986 A study of calcification in the leg tendons from the tissue, is likely to be below instrumentation limits. The domestic turkey. J. Ultrastruct. Mol. Struct. Res., 94r217-238. application of ion microscopy (Landis et al., 1986) or Landis, W.J., and A.L. Arsenault 1989 Vesicle- and collagen-mediated calcification in the turkey leg tendon. Connect. Tissue Res., 22; alternative techniques more sensitive to these minor 35-42. constituents putatively within lepidotrichia is being Landis, W.J., G.Y. Burke, J.R. Neuringer, M.C. Paine, A. Nanci, P. pursued in conjunction with additional approaches to Bai, and H. Warshawsky 1988 Earliest enamel deposits of the rat incisor examined by electron microscopy, electron diffraction, and define more fully the organic matrix-mineral interacelectron probe microanalysis. Anat. Rec., 220:233-238. tions in this vertebrate bone model. Landis, W.J., and M.J. Glimcher 1978 Electron diffraction and elecACKNOWLEDGMENTS
This work was supported by grants from the National Institutes of Health (AR 34078 and AR 34081) and the CNRS (UA 041137), the Peabody Foundation, Figs. 11-1 3. A series of micrographs illustrating, respectively, the bright field image, dark field image of mineral c-axes, and dark field image of mineral u,b-axes taken from the same extracellular region of a newly developing lepidotrichium treated by anhydrous means without staining. The mineral phase is associated with a collagenous matrix. The mineral particles, appearing in bright field (Fig. 11)as fine, electron-dense needle- and plate-like material about 35-40 nm in its longest dimension, are found in dark field (Fig. 12) to be composed of smaller crystals 12-20 nm along their c-axis lengths (in dark field, the crystals are the bright images set against the dense background). Both the c-axis and +axes images of the mineral are localized along collagen in a seemingly random manner. Unlike their distribution in a regular fashion within the hole and overlap zones of collagen fibrils in the turkey tendon (Arsenault, 1988; Landis and Arsenault, 1989), the crystal axes dark field images do not appear repetitively along the same specific regions of the collagen fibrils in lepidotrichia. x 32,000.
tron probe microanalysis of the mineral phase of bone tissue prepared by anhydrous techniques. J . Ultrastruct. Res., 63r188-223. Landis, W.J., and M.J. Glimcher 1982 Electron optical and analytical observations of rat growth plate cartilage prepared by ultracryomicrotomy. J . Ultrastruct. Res., 78r227-268. Landis, W.J., D.D. Lee, J.T. Brenna, S. Chandra, and G.H. Morrison 1986 Detection and localization of silicon and associated elements in vertebrate bone tissue by imaging ion microscopy. Calcif. Tissue Int., 38:52-59. Landis, W.J., J. Moradian-Oldak, and S. Weiner 1990 Topographic imaging of mineral and collagen in the calcifying turkey tendon. Connect. Tissue Res. (in press). Landis, W.J., M.C. Paine, and M.J. Glimcher 1977 Electron microscopic observations of bone tissue prepared anhydrously in organic solvents. J. Ultrastruct. Res., 59:l-30. Robinson, R.A., and M.L. Watson 1952 Collagen-crystal relationships in bone as seen in the electron microscope. Anat. Rec., 114:383409. Robinson, R.A., and M.L. Watson 1955 Crystal-collagen relationships in bone as seen in the electron microscope. 111. Crystal and collagen morphology as a function of age. Ann. N.Y. Acad. Sci., 60: 596-628. Weiner, S., and P.A. Price 1986 Disaggregation of bone into crystals. Calcif. Tissue Int., 39:365-3 75.
