Proc. Natl. Acad. Sci. USAVol. 83, pp. 4104-4107, June 1986Physiological Sciences

Osteoclasts can be induced in fish having an acellular bony skeleton(osteoclast progenitors/differentiation/sharks/bass/minerals)

J. GLOWACKI*t, K. A. COX*, J. O'SULLIVANt, D. WILKIEO, AND L. J. DEFTOSt*Department of Surgery, Children's Hospital and Harvard Medical School, Boston, MA 02115; tDepartment of Medicine, San Diego Veteran's AdministrationMedical Center and University of California, La Jolla, CA 92161; and tScripps Institute of Oceanography, La Jolla, CA 92037

Communicated by Elizabeth D. Hay, January 22, 1986

ABSTRACT Kelp bass (Paralabrax clathratus) and leopardsharks (Triakis semifasciata) are characterized by an acellular(anosteocytic) bony skeleton and a focally calcified cartilagi-nous endoskeleton, respectively. These skeletal forms are notconsidered to function as mineral reservoirs. Previous studiesshowed that implanted bone particles are resorbed in rats bylarge multinucleated cells with ultrastructural features (ruffledborders) characteristic of osteoclasts. We tested the ability offish to resorb bone matrix and to adapt to reduced salinityconditions. Bone particles were implanted in sharks and bassmaintained in seawater (34 ppt, 40.5 mg of calcium per dl) orin diluted seawater (26 ppt, 28.5 mg of calcium per dl). Seraand elicited tissues were harvested 4 weeks later. In sharks,bone particles were not resorbed, and multinucleated cells werenot evident under either normal or hyposalinity conditions.Shark sera were isoosmolar with the seawater or dilutedseawater, with serum chemistries of the hyposalinity groupreflecting the 23% reduction in environmental minerals andelectrolytes, compared to sharks in normal seawater. Inmarked contrast, bass adapted to diluted seawater resorbedbone particles and maintained normal serum chemistries.Electron microcopy showed that the bone particles weresurrounded by large, foamy multinucleated cells, many withmembrane specializations typical of osteoclasts from highervertebrates, i.e., extensive clear zones apposed to intact bonematrix and active ruffled borders overlying areas of matrixundergoing dissolution. Although osteoclasts had not beendescribed in these fish, this study shows that bass have stemcells that can be stimulated to differentiate into bone-resorbingosteoclasts.

In most vertebrates the skeleton, in addition to its structuralfunction, serves as a reservoir for dietary minerals, especiallycalcium. To maintain calcium homeostasis, bone cells, underthe influence of bone-active hormones such as parathyroidhormone, can mobilize minerals from the bone (1). In thisway serum calcium levels can be tightly regulated betweenmeals or during mineral stress. The ancestors ofmodem fishplacoderms and ostracoderms are thought to have evolved inbrackish waters (2). According to the fossil record, functionalfeatures, namely cellular lacunae and trabeculae, suggest thatthese ancient fish could use their skeleton as a mineralreservoir. As fish have evolved, the efficiency of calciumtransport by the gills has increased considerably, so that inmodem freshwater fish, the skeleton does not play animportant role in mineral homeostasis. Because marine fishlive in a very calcium-rich environment (40 mg/dl) relative toserum calcium (10-14 mg/dl), it is even less advantageousthat they have mineral stores. This may account for theregression to acellular (anosteocytic) bone in many extantteleosts (3). Although acellular bone contains no osteocytes,its formation depended upon the activity of osteoblasts that

either withdrew from or were fatally buried within mineral-ized matrix (4). There is some dispute whether such acellularbone is "alive." Using freshwater species, Moss (5) showedthat acellular-boned fish, in contrast to cellular-boned fish,cannot repair fractures when environmental or dietary cal-cium is restricted. Fleming (6), however, considered thatmineral could be exchanged from acellular bone, but at a veryslow rate. Multinucleated osteoclasts with characteristicruffled borders have not been found in marine fish, even insites of apparent repair or remodeling (5, 7). Tetrapodevolution to terrestrial life required the development ofmechanisms of mineral homeostasis from skeletal stores,including the acquisition of osteocytes, osteoclasts, and para-thyroid glands (8).The subcutaneous implantation of devitalized, mineral-

containing bone particles in rats elicits the recruitment anddifferentiation of multinucleated cells that resorb the implant-ed bone particles (9). These cells are evident in lacunae on thesurfaces of the implanted particles, which are too large(75-250 ,um) to be digested by phagocytosis. Many of themultinucleated cells have ultrastructural features that arecharacteristic of in osso osteoclasts, including ruffled bor-ders, clear zones of attachment to the bone substrate, andabundant mitochrondria, cytoplasmic granules, and vacuoles(10). These cells and the resorption of bone particles havebeen shown to be regulated by bone-active hormones anddrugs, including calcitonin (11), heparin (12), and gluco-corticoids (13). Using this technique, we have demonstratedthat marine bass (Paralabrax clathratus), which have anacellular bony skeleton, and cartilaginous leopard sharks(Triakis semifasciata) did not resorb implanted bone parti-cles, when the fish were kept in normal seawater. In dramaticcontrast, bone particles were resorbed by elicited osteoclastsin bass, but not in sharks, after adaptation to hyposalinityconditions. These bass were able to maintain normal serumchemistries in spite of a 23% reduction in environmentalminerals and electrolytes.

MATERIALS AND METHODSLeopard sharks (Triakis semifasciata), 3-4 kg each, and kelpbass (Paralabrax clathratus), 2-3 kg each, were netted innear-shore waters and maintained at the Scripps Aquarium,La Jolla, CA, on a diet of mackerel and squid. Somespecimens were adapted to open tanks containing seawatergradually mixed with fresh water until the salinity wasreduced from 33.7 to 26.0 parts per thousand (ppt). Theanimals were equilibrated to reduced salinity for 1-2 weeksbefore implantation with test materials (bone particles).Bone particles (75-250 ,m) were prepared from rat long

bones and from bass vertebrae as described (9). Cellular andpotentially inflammatory components were removed fromthe devitalized bone particles by extractions in ethanol andanhydrous ether (9). Fish were anesthetized with tricainemethanesulfonate (Crescent Research Chemicals, ParadiseValley, AZ) and bone particles were inserted into multiple

4104

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

July

14,

202

0

Proc. Natl. Acad. Sci. USA 83 (1986) 4105

Table 1. Composition of tank water and sera from T. semifasciata (leopard sharks) and P. clathratus (kelp bass)Salin- Osmo- Chloride, Phos- Alkalineity, larity, Sodium, Potassium, Calcium, mmol/ phorus, BUN, Protein, phosphatase,*ppt mosM mmol/liter mmol/liter mg/dl liter mg/dl mg/dl mg/ml units/liter

ControlSeawater 33.7 1034 460 9.6 40.5 475 0.1Leopard shark

sera (7)t 1008 ± 11 260 ± 17 2.2 ± 0.1 15.0 ± 0.9 252 ± 14 6.0 ± 1.0 1035 ± 21 2.7 ± 0.7 21.0 ± 8.0Bass sera (3) 378 ± 21 188 ± 9 3.0 ± 0.3 13.9 ± 1.7 166 ± 4 13.0 ± 5.0 5.3 ± 0.5 6.8 ± 0.9 217 ± 98

ExperimentalDiluted

seawater 26.0 701 312 6.7 28.5 289 0Leopard shark

sera (4) 766 ± 19t 223 ± 3t 2.8 ± 0t 12.1 ± 0.6t 216 ± 3t 6.0 ± 1.6 718 ± 40t 2.2 ± 0.3 15.8 ± 6.2Bass sera (3) 392 ± 28 181 ± 3 2.9 ± 0.5 14.9 ± 2.0 153 ± 6O 11.5 ± 3.0 5.3 ± 0.5 7.7 ± 1.0 279 ± 36

Values for sera are mean ± SD. BUN, blood urea nitrogen.*Alkaline phosphatase is total serum alkaline phosphatase.tNumbers in parentheses are numbers of fish in each group.tSignificant difference from control, P < 0.01.

lateral intramuscular pockets that were closed with interrupt-ed polypropylene suture. After 4 weeks in the various tankconditions, the animals were anesthetized, blood was col-lected from the caudal hemal canal, and the remaining boneparticles and surrounding tissues were harvested and pre-pared for light (12) and electron microscopy (10). Specimenswere fixed in 2% (wt/vol) paraformaldehyde in 0.1 Mcacodylate, pH 7.5, embedded in glycol methacrylate, sec-tioned without decalcification, and stained with 0.5%toluidine blue. Selected segments (1 mm3) were fixed in 2.5%(vol/vol) glutaraldehyde, 4.0% (wt/vol) paraformaldehyde in0.1 M cacodylate, pH 7.4, decalcified in 7.5% (wt/vol)EDTA, 2.5% (vol/vol) glutaraldehyde, postfixed in 1%osmium tetroxide, and embedded in Spurr's resin. Ultrathinsections (0.05 um) were viewed on a Phillips transmissionelectron microscope.

RESULTSTable 1 presents the composition of water in the two tanksand of sera from sharks and bass at the end of the experiment.Control bass sera had lower values than sharks in osmolarity,blood urea nitrogen (BUN), sodium, calcium, and chlorideand had higher values in phosphorus, protein, and alkalinephosphatase. Serum osmolarity, BUN, most serum electro-lytes, and minerals were significantly reduced in the sharksmaintained in the diluted seawater tank for a total of 6 weeks.These reductions were proportional to the 23% reduction insalinity of the water. In contrast, sera from bass that wereadapted to the experimental (diluted seawater) tank for 6weeks were indistinguishable from controls (seawater) withthe exception of chloride, which showed a small (7%)reduction. Thus the shark sera reflected the environmentalconditions, but the bass homeostatic mechanisms maintainedmost serum components constant.

All incisions healed uneventfully. Light microscopic eval-uation of bone particle specimens from the sharks revealedthe inertness of the bone particles. There was no evidence ofinflammation or foreign body reaction. Bone particles weresurrounded by loose connective tissue and their profilesshowed no evidence of surface resorption. Sections fromcontrol and experimental sharks were identical in thesefeatures. In marked contrast, specimens from bass in theexperimental, hyposalinity environment showed an abun-dance of large, foamy multinucleated cells, evidence ofextensive resorption of the bone particles, and mild fibrosis(Fig. 1 A and B). Electron microscopic analysis of randomlyselected portions of specimens from hyposalinity bass re-vealed very large, multinucleated cells with abundant

mitochondria, granular endoplasmic reticulum, cytoplasmicvacuoles and granules, and microvilli on the free cell surface.In many sections, clear zones of attachment to underlyingbone particle substrate and ruffled borders opposing areas ofthe bone particle substrate that had the appearance of frayedcollagen fibrils were seen (Fig. 2). Tissue reaction was thesame for implanted rat or bass bone particles. Theseultrastructural features are characteristic of in osso osteo-

FIG. 1. Micrographs of tissue response to bone particles 4 weeksafter implantation in kelp bass in seawater (A) and in diluted seawater(B). Arrows indicate multinucleated cells in surface lacunae. Noticethe osteocytic lacunae in bone particles prepared from rat bone in Acompared to the anosteocytic bass bone particles in B. (Toluidineblue; x64.)

Physiological Sciences: Glowacki et al.

Dow

nloa

ded

by g

uest

on

July

14,

202

0

4106 Physiological Sciences: Glowacki et al.

.{3-4 m*

1

X .:#

.Yi -- .ii

V,

.;4

-5'IN.f

W-Al

6A; h

a=,WA XX

J

4

a6

FIG. 2. Electron micrograph of multinucleated cell elicited in kelp bass adapted to diluted seawater. The bass bone particle (BP) showsevidence of dissolution of collagen fibrils beneath the ruffled border (R). Two clear zones (C) of attachment, extensive rough endoplasmicreticulum, and cytoplasmic vacuoles and granules are evident. The linear densities reflect artefactual folds in the sections ofbone particle. Noticethe absence of osteocytic lacunae and canalicular spaces in the bass bone particles. (Bar = 5 um.)

Proc. Natl. Acad. Sci. USA 83 (1986)

'U ;-, ,- :: ..1, I

.1I

A

"Aimu

i.

lill".;

.Z;6.#fj

1:Y.

a

Z..iMg. _

.i: A;'t

`.

-I

if

.. i.

.:

z:.. f",

'T'.1%,-j:1'. i. f.

*Cl

Dow

nloa

ded

by g

uest

on

July

14,

202

0

Proc. Natl. Acad. Sci. USA 83 (1986) 4107

clasts in other vertebrates but, to our knowledge, have notbeen described in bass bones.

DISCUSSIONThe technique of subcutaneous implantation of bone particlepermits the evaluation of (i) the recruitment of osteoclastprogenitors and their differentiation into bone-resorbingosteoclasts and (ii) the mechanisms and regulation of theirresorption of bone substrate. The recruitment of cells issynchronized and concentrated because of the large surfacearea for cell-substrate interaction. Implants avoid the sam-pling and technical limitations inherent in attempts to analyzethe intact skeleton and reflect, in a more concentrated way,the state of resorption (11-13). In this way, we have shownthe development of bone-resorbing osteoclasts in bony kelpbass, but not in cartilaginous sharks, after the fish had beenacclimated to conditions of reduced salinity. These resultssuggest that osteoclast-mediated bone resorption may occurin marine species if subjected to mineral perturbations.Lopez et al. (14) have presented light micrographic evidencefor modulation of osteoclasts in the cellular bone of silver eels(Anguilla anguilla L.).Known, or as yet unidentified, calcitropic hormones may

regulate osteoclast differentiation and regulation in the bassunder these circumstances. The complex endocrine controlof mineral metabolism in birds and animals with cellular boneis mediated by three major hormones, parathyroid hormone,vitamin D metabolites, and calcitonin. Fish do not haveparathyroid glands, although they may have other sources ofparathyroid hormone-related substances (15). In addition,the role of the hypophysis in calcium regulation in freshwaterfish has been recognized (16). Evidence has been presentedthat prolactin (17) or parathyroid hormone-related hypercal-cin (18) may be the hypercalcemic factors of the hypophysis.Information on the contribution of 1,25-dihydroxyvitamin Din fish is limited, but some studies suggest that the metabolitestimulates intestinal uptake of calcium (19, 20). Our obser-vations show osteoclast-mediated bone resorption in fish andsuggest the existence ofhormones or factors that mediate thisphenomenon. The detection of high levels of circulatingcalcitonin in sharks (21) and fish (22) and the discovery of theunexpected hypercalcemic effect of synthetic calcitonin insharks, but not in bony fish (23), have renewed interest in thetarget cells of mineral-regulating hormones in marine fish.Bone cells responsible for calcium flux out of bone innonaquatic vertebrates are not generally considered essentialfor mineral homeostasis in marine fish because of the abun-dant supply of calcium in seawater. In fish, calcium balancestudies have shown regulated, bidirectional fluxes of calciumover epithelial organs-skin (24), gills (25), the gastro-intestinal tract (26), and kidneys (27). A more extremeexample of the lack of a mineral reserve in the skeleton isexemplified by the Chondrichthyes (sharks and rays) thathave an almost completely cartilaginous skeleton. The car-tilage is only minimally calcified in areas where rigidity isneeded.

Osteoclasts have been recognized as the cellular agents ofbone resorption since 1873 (28). On the basis of labelingexperiments in newts, Fischman and Hay (29) showed in 1962that multinucleated osteoclasts in regenerating forelimbsarise by the fusion of mononuclear leukocytes. With electronmicroscopic radioautographic techniques in parabiotic rats,

Gothlin and Ericsson (30) showed that circulating monocytescan function as preosteoclasts. The precise identity of theosteoclast progenitor, its relationship to the progenitor offoreign body cells, and factors involved in osteoclast recruit-ment and differentiation remain unknown for all species. Thepresent study shows that bass have stem cells that can bestimulated to differentiate to osteoclasts. These progenitorsmay arise from the spleen or liver. Cartilaginous sharks,however, appear to lack the cellular or humoral factorsnecessary for this recruitment. Further studies in thesespecies may provide information about the mechanisms andregulation of osteoclast differentiation.

The authors gratefully acknowledge the encouragement of Drs. J.Folkman and M. Glimcher, the technical assistance of MichaelMiller, and secretarial support of Devorah Rosen. This work is theresult of research sponsored in part by National Oceanic andAtmospheric Administration, National Sea Grant College Program,Department of Commerce, Grant NA800AA-D-00120, Project R-MP-28, through the California Sea Grant College Program.

1. Parfitt, A. M. (1976) Metabolism 25, 809-844.2. Romer, A. S. (1946) Q. Rev. Biol. 21, 33-69.3. Moss, M. L. (1961) Acta Anat. 46, 343-362.4. Simmons, D. J. (1971) Clin. Orthop. Relat. Res. 76, 244-280.5. Moss, M. L. (1962) Acta Anat. 48, 46-54.6. Fleming, W. R. (1967) Am. Zool. 7, 835-842.7. Weiss, R. E. & Watabe, N. (1979) Calcif. Tissue Int. 28,

43-56.8. Moss, M. L. (1965) Acta Anat. 60, 262-372.9. Glowacki, J., Altobelli, D. & Mulliken, J. B. (1981) Calcif.

Tissue Int. 33, 71-76.10. Glowacki, J., Jasty, M. & Goldring, S. (1986) J. Bone Miner.

Res. 1, in press.11. Glowacki, J. (1984) in Osteoporosis: Social and Clinical As-

pects, eds. Gennari, C. & Segre, G. (Excerpta Medica,Amsterdam), pp. 103-108.

12. Glowacki, J. (1983) Life Sci. 33, 1019-1024.13. Glowacki, J. (1983) Endocrinology 112, 255 (abstr.).14. Lopez, E., Peignoux-Deville, J., Lallier, F., Martelly, E. &

Milet, C. (1976) Calcif. Tissue Res. 20, 173-186.15. Lopez, E. (1984) Gen. Comp. Endocrinol. 53, 28-36.16. Pang, P. T. K., Griffith, R. W. & Schreibman, M. P. (1973)

Gen. Comp. Encrinol. 20, 358-361.17. Pang, P. T. K. (1981) Comp. Endocrinol. 43, 252-255.18. Bern, H. A. (1975) Am. Zool. 15, 937-948.19. Cartier, M. M., Milet, C., Martelly, E., Lopez, E. & Warrot,

S. (1979) J. Physiol. (Paris) 75, 275-282.20. Fenwick, J. C., Smith, K., Smith, J. & Flik, G. (1984) Gen.

Comp. Endocrinol. 55, 398-404.21. Glowacki, J. & Deftos, L. J. (1984) in Endocrine Control of

Bone and Calcium Metabolism, eds. Cohn, D. V., Potts, J. T.& Fujita, T. (Elsevier, Amsterdam), Vol. 8B, pp. 194-196.

22. Watts, E. G., Copp, D. H. & Deftos, L. J. (1975) Endocrinol-ogy 96, 214-218.

23. Glowacki, J., O'Sullivan, J., Miller, M., Wilkie, D. W. &Deftos, L. J. (1985) Endocrinology 116, 827-829.

24. Mashiko, K. & Jozuka, K. (1964) Annot. Zool. Jpn. 37, 41-50.25. Fenwick, J. C. & So, Y. P. (1974) Exp. Zool. 188, 125-131.26. Odie, M., Utida, S. (1968) Mar. Biol. 1, 172-177.27. Nishimura, H. & Imai, M. (1982) Fed. Proc. Fed. Am. Soc.

Exp. Biol. 41, 2355-2360.28. Kolliker, A. (1873) Die normale Resorption des

Knochengewebes und ihre Bedeutung fur die Enstehung dertypischen Knochenformem (Vogel, Leipzig, G.D.R.).

29. Fischman, D. A. & Hay, E. D. (1962) Anat. Rec. 143,329-337.

30. Gothlin, G. & Ericsson, J. L. E. (1973) Virchows Arch. B 12,318-329.

Physiological Sciences: Glowacki et al.

Dow

nloa

ded

by g

uest

on

July

14,

202

0