A possible role for polyamines in cartilage in the mechanism of calcification

38

Biochimica et Biophysica Aeta

881 (1986)38-45 Elsevier

BBA 22271

A p o s s i b l e r o l e f o r p o l y a m i n e s in c a r t i l a g e in t h e m e c h a n i s m o f c a l c i f i c a t i o n F r a n c o V i t t u r a.., Coiancarlo L u n a z z i a, L u i g i M o r o a, N i c o l a S t a g n i ~, B e n e d e t t o d e B e r n a r d ~, M i c h ~ l e M o r e t t i b, G i o r g i o S t a n t a b, F r a n c o B a c c i o t t i n i c, G i a n c a r l a O r l a n d i n i c, N i n o R e a l i c a n d A m o s C a s t i c Dipartimento di Biochimica, Biofisica e Chimica delle Macromolecole, h lstituto di Anatomia Patologica, Universitit di Trieste, Trieste, and" lstituto di Chimica Biologica, Universitit di Parma, Parma (Italy)

(Received July 10th, 1985) (Revised manuscriptreceivedOctober 21st, 1985)

Key words: Polyamine;Calcification;Proteoglycansubunit; Spermidine;(Cartilage) The role of polyamines in cartilage is not known: they may be somehow related to the mechanism of calcification. In epiphyseal cartilage from calf scapulas, they are more concentrated in the ossifying area, where calcification takes place, than in the resting region. Spermidine is present in greater amounts than spermine and putrescine. Since ornithine decarboxylase (EC 4.1.1.17) is measurable only in the resting region of the tissue, it is in this area that polyamine biosynthesis occurs, while they accumulate in the ossifying area. Immunohistochemical evidence is obtained that only in the ossifying zone is spermidine extracellular. It is at this level that the matrix is rearranged to become calcified, and proteoglycans are dissociated and partially removed. The effect of polyamines on solutions of proteoglycan subunits has been studied in vitro by following variations of turbidity and viscosity. While in the presence of putrescine the specific viscosity decreases to asymptotic values, in the presence of either 30 mM spermidine or 2.5-10 mM spermine, the decrement is more marked. At the same concentrations, increase of the turbidity of proteoglycan subunit solutions was observed. Only spermidine showed the capacity of displacing proteoglycan subunits from a column of Sepharose 4B-type II collagen: at 15 mM concentration, about 90% of proteoglycans were removed from the column. Alkaline phosphatase activity, which plays an important role in calcification, is enhanced by spermidine and spermine. These results obtained in vitro support the hypothesis that polyamines may be related to calcification of preosseous cartilage.

Introduction

Cartilage calcification may be considered a two-step phenomenon which initiates with the formation of early crystallites in matrix vesicles and concludes with the mineralisation of the whole extravesicular territory. An important role in this process is ascribed to tffe enzyme of matrix vesicles, i.e., alkaline phosphatase (EC 3.1.3.1). The mecha* To whomcorrespondenceand requests for reprints should be addressed.

nism of action of the enzyme is not yet clear, although it has been recently demonstrated that alkaline phosphatase is a Ca 2+-binding protein [1] and also that, in vitro, it binds to type I! collagen in competition with proteoglycans [2]. Many experimental data [3-5] suggest that mineralisation is accompanied by dissociation and partial removal of proteoglycans: such an event would therefore expose collagen to alkaline phosphatase, whereby the enzyme, diffused both from cells and matrix vesicles, would act as a promoter of calcium a n d / o r phosphate fixation.

0304-4165/86/$03.50 © 1986 ElsevierScience Publishers B.V. (BiomedicalDivision)

39 Since dissociation of proteoglycans is a process strictly controlled, compounds produced by cells and able to modulate the arrangement of these macromolecules should be identified. Among the various possible compounds, we have focused our attention on polyamines. Greenwald et al. [6] have demonstrated that interactions between collagen and proteoglycans are electrostatic in nature: the possibility therefore exists that polyamines might interfere with such a bonding. Putrescine, spermidine and spermine have been found in many living tissues, including cartilage [7]. Their formation, catalysed by ornithine decarboxylase (EC 4.1.1.17), has been shown by Rath and Reddi [8] during the induction of cartilage transformation in bone. Takigawa et al. [9] have shown that parathyroid hormone, which stimulates the synthesis of glycosaminoglycans, induces ornithine decarboxylase activity and increases polyamine levels in differentiated rabbit costal chondrocytes in culture. The role of these compounds in preosseous cartilage is however unknown. The aim of this work was to assess whether proteoglycan association is influenced by the presence of polyamines and whether these compounds interfere in vitro with the interaction between proteoglycans and collagen. The amounts of polyamines in the resting and ossifying regions of preosseous cartilage were also determined in order to establish a possible correlation between polyamines and the metabolic events preceding calcification. Since alkaline phosphatase appears to be involved, whatever its mechanism of action, the effect of polyamines on the activity of this enzyme was also studied. Polyamines are generally considered intracellular cations, while collagen and proteoglycans are extracellular components of the matrix; it was therefore necessary to assess also the extracellular localization of spermidine by immunohistochemical methods.

Experiment Materials. Putrescine, spermidine, spermine and thyroglobulin (type I) were purchased from Sigma, MO, U.S.A. Coupling reagent water-soluble carbodiimide (EDAC) was from Bio-Rad; [14C]spermidine. 3 HC1 was from Amersham.

Cartilage. Calf scapulas, excised from the animals at the slaughterhouse, immediately after their death, were transferred in ice to the laboratory. Two zones were carefully selected from the cartilaginous portion of scapulas, the resting and the ossifying regions, as reported by Vittur et al. [10]. Interaction of polyamines with proteoglycan subunits. This study was carried out by measuring both viscosity and turbidity of proteoglycan subunit solution at various concentrations of polyamines. Viscosity was determined at 25 _+ 0.05°C with an Ubbelholde semimicro dilution viscosimeter (Cannon Instrument Co., State College, PA). For each polyamine concentration, proteoglycan subunit specific viscosity was calculated as follows: ~lsp=(t-to)/to where t o was the flow-time of the solvent (50 mM Tris-HCl/5 mM CaC12, pH 7.4) and t the flow-time of the proteoglycan subunit/polyamine solution through the viscosimeter. The effect of polyamines on the turbidity of proteoglycan subunit solutions was followed spectrophotometrically at 400 nm. Effect of polyamines on the binding of proteoglycans to a column of Sepharose 4B-collagen. Type II collagen and proteoglycan subunits were extracted from resting cartilage as described by Vittur et al. [2]. Collagen prepared by papain digestion of cartilage, according to the procedure of Lee-Own and Anderson [11], was coupled to CNBr-Sepharose 4B (Pharmacia) (1.6 mg of collagen/ml wet gel) according to the procedure described by Bauer et al. [12]. Proteoglycans were obtained by treating cartilage slices with 4 M guanidinium chloride in 50 mM Tris-HCl buffer (pH 7.4) in the presence of the usual mixture of proteinase inhibitors (100 mM e-aminohexanoic acid/5 mM benzamidine chloride/10 mM Na 2EDTA/1 mM sodium iodoacetate/1 mM phenylmethylsulfonyl fluoride). Proteoglycan subunit D 1 fractions were purified by CsCI density gradient centrifugation in dissociative conditions, according to the one-step procedure of Sajdera and Hascall [13]. 10-ml columns containing Sepharose 4B-collagen, equilibrated with 50 mM Tris-HCl buffer (pH 7.4)/5 mM CaCI, were loaded with 0.2 ml of the buffer solution containing 3.75 mg of proteoglycan subunit/ml. Elution was carried out in sequence with (a) 20 ml of buffer, (b) 20 ml of Tris

40 buffer containing different concentrations (0.5-30 mM) of polyamines and finally 20 ml of buffered 1 M NaC1. The amount of proteoglycan subunits eluted from the column was followed by measuring uronic acid according to Bitter and Muir [14]. Effect of polyamines on alkaline phosphatase activity. The enzyme was partially purified from matrix vesicles, prepared from cartilage, according to the technique developed by Ali et al. [15]: matrix vesicles were first washed with 0.15 M KC1 at 4°C for 30 min, spun down in a Spinco SW-50.1 rotor (50000 rpm for 20 min) and then extracted with a deoxycholate/butanol mixture. The aqueous phase of the extract was used as the source of the enzyme [16]. The catalytic activity was measured at pH 7.4 according to Stagni et al. [17] with and without polyamines in the assay. DNA measurement. Resting or ossifying cartilage was homogenized for 3 rain in 30 mM phosphate buffer (pH 7.4)/containing 2 M NaC1, with the Ultra-Turrax homogenizer, at the highest speed, followed by 2-min sonication in ice, at 3 A, using the Branson sonifier, model S-75 equipped with the microtip. DNA content was measured according to the procedure of Labarca and Paigen [18]. Polyamines determination. Polyamines were extracted according to Raina and Cohen [19]. Tissues were homogenized in an Ultra-Turrax homogenizer in 3% ( w / v ) HC104. A known amount of 1,8-diaminooctane • 2 HCI was added to the homogenate as internal standard. Supernatants were alkalinized to pH 10 with 4 M KOH, and KCIO 4 was removed by centrifugation at 10000 × g for 15 min. The polyamines were extracted with butanol; the extracts were evaporated to dryness and then dissolved in 0.1 M HC1. Polyamines in these acid extracts were determined by HPLC of their dansyl derivatives by following the method of Seiler and Kn~dgen [20]. Separation of the DNSpolyamines was carried out in a column of Ultrasphere ODS 5 /.tm (150 x 4.5 mm). The fluorescence intensity was monitored at the emission wavelength of 550 nm with the excitation wavelength of 368 nm. Elution was performed for 15 min by water (A):mdthanol (B) gradient which changed from 40% of solvent A to 100% of solvent B and for another 10 min at the final concentration (flow-rate: 1 ml/min). Measurement of ornithine decarboxylase activity.

Tissues were homogenized in the Ultra Turrax with 8 vol. of 10 mM Tris-HCl buffer (pH 7.1). The homogenate was centrifuged at 20 000 × g for 30 min and the supernatant was used for the enzyme assay and for the protein determination. Ornithine decarboxylase activity was followed by the amount of CO 2 released from DL-[1laC]ornithine, according to J~inne and WilliamsAshman [21]. The incubation mixture comprised, in a final volume of 0.5 ml, 50 mM Tris-HC1 (pH 7.1)/1 mM dithiothreitol/0.2 mM pyridoxal 5'phosphate/32 # M L-ornithine/0.75 /~Ci DL-[I14C]ornithine/about 1.5 mg protein of the highspeed supernatant. The incubation was carried out for 60 min at 37°C. Proteins were determined by the method of Lowry et al. [22]. Immunohistochemical methods. Antigen was prepared by coupling spermidine tb bovine thyroglobulin using carbodiimide reaction according to the method described by Bartos and Bartos [23]. A conjugated product was obtained with a molar ratio hapten to carrier of 180/1. Antiserum against spermidine was raised in four rabbits. Each rabbit was intramuscularly injected with 1 ml of an antigen emulsion (350/~g protein/0,25 ml saline/0.75 ml Freund's complete adjuvant). Booster injections of the same amount of conjugate were administered at biweekly intervals. Only those antisera which gave positive results in the Ouchterlony tests against spermidine/thyroglobulin conjugate at a dilution of 1 : 32 were used. Cartilage blocks were dissected from calf scapula perpendicularly to the calcification front. Blocks were then fixed overnight in 10% formaldehyde, buffered with 0.1 M phosphate (pH 7.4), and, finally paraffin-embedded. 10-/xm thick sections were then cut and mounted on slides for immunochemical studies. The conventional unlabeled antibody method of peroxidase antiperoxidase [24] was used. Primary serum dilution was 1 : 250 and the following controls were employed: primary antibody was omitted; primary antiserum was replaced with a primary antiserum to casein; primary antiserum was absorbed with the spermidine/thyroglobulin conjugate; incubation with antisera was omitted and the slides were incubated with diaminobenzidine solution only.

41

R ~

Content of polyamines in preosseous cartilage Table I summarizes the results of the analysis of polyamine distribution in the resting and ossifying regions of scapula cartilage. The following information is obtained from these measurements: (a) resting cartilage is devoid of putrescine or the amount is below the limit of sensitivity of the method of measurement; (b) ossifying cartilage contains more polyamines than the resting zone both on the basis of tissue weight and DNA content. In the former, the amount of spermidine is about 5-fold higher and that of spermine about double; (c) the spermidine over spermine ratio is 1.70 in the ossifying cartilage and 0.69 in the resting. No significant difference in the content of polyamines was found in the transforming region as compared to the ossifying zone (data not shown). Ornithine decarboxylase activity This enzyme activity was detectable only in the resting cartilage: 20.2 + 5.8 (S.D.) pmol CO2/h per mg protein (n = 6).

Spm r[sp

Spd 0.1

Put

0

I 200

lOO

POLYAMINE

[mMJ

Fig. 1. Effect of polyamines on the specific viscosity of proteoglycan subunit solutions. The specific viscosity of proteogly'can subunit (fraction DI) solutions (1 mg.ml - I in 50 mM TrisHCI/5 mM CaCl 2 (pH 7.4)) was measured in the presence of 0.5-175 mM polyamines at 25 ±0.05°C. Spm, spermine; Spd, spermidine; Put, putrescine.

Interaction of polyamines with proteoglycan subunits Interaction of polyamines with proteoglycans has been studied by following the variations in viscosity and turbidity of the proteoglycan subunit solutions. Figs. 1 and 2 illustrate the results of these experiments. As shown by Fig. 1, putrescine, spermine and spermidine interact with proteoglycan subunits, but according to three different patterns. While in the presence of putrescine, the specific viscosity of proteoglycan subunits decreases to asymptotic values, in the presence of

either spermidine or spermine, the decrement is more pronounced, at the critical concentrations of 30 and 2.5-10 mM, respectively. At higher concentrations, the specific viscosity of proteoglycan subunits almost reaches the initial values. The data obtained by turbidity measurements are illustrated by Fig. 2. Different solutions of proteoglycan subunits (0.5-5.0 mg/ml) are not precipitated by putrescine, at least within the tested concentrations (maximal 150 mM). On the contrary, an increase of turbidity occurs in. the presence of the

TABLE 1 POLYAMINES IN RESTING AND OSSIFYING ZONES OF PREOSSEOUS CARTILAGE Results are means ± S.D. of the number of experiments reported in parentheses; n.d., not detectable. Results, reported as n m o i / m g DNA, were calculated after determination of DNA content of resting cartilage (879 ± 88 (4) # g / g wet tissue) and ossifying cartilage (1 164± 182 (4) # g / g wet tissue). Putrescine

Resting Ossifying

Spermidine

Spermine

nmol/g wet tissue

nmol/mg DNA

nmol/g wet tissue

nmol/mg DNA

nmoi/g wet tissue

nmol/mg DNA

n.d. 33.8 + 21.9 (7)

n.d. 30.4

27.7 :i: 5.7 (10) 168.9 :t: 21.9 (7)

31.5 145.2

39.9 + 11.4 (9) 100.0 + 10.4 (6)

45.4 85.9

42

2

100

a~ E 50 m

<

0 -

04

20

10 SPERMIDINE

gO POLYAMINE

100 tmM I

Fig, 2. Effect of polyamines on the turbidity of proteoglycan subunit solutions. The turbidity of proteoglycan subunit (fraction D1) solutions (5 mg-m1-1 in 50 mM Tris-HCl/5 mM CaCI 2 (pH 7.4)) in the presence of various polyamines (2.5-110 mM) was measured at 400 nm. Spm, spermine; Spd, spermidine; Put, putrescine.

other two bases, the effect being more marked at 30 mM spermidine and at 2.5-10.0 mM spermine. Polyamine-proteoglycans interaction is obviously influenced by the ionic strength of the reaction mixture. The experiments described have been carried out at low ionic strength; however, the addition of 100 mM NaCI does not prevent the interaction. The only difference is a modest shift of the concentration for the maximal activity of polyamines (spermidine from 30 to 40 mM; spermine from 10 to 15 mM). On the contrary, the increment of proteoglycan concentrations from 0.5 to 10 mg/ml did not affect the results obtained upon addition of polyamine (data not shown).

Effect of polyammes on the interaction of proteoglycan subunits with collagen This effect was studied by following the elution of proteoglycans from a column of Sepharose 4Bcollagen loaded with proteoglycan subunits. While

30

( mM)

Fig. 3. Effect of spermidine on the binding of proteoglycans to a column of Sepharose 4B-collagen, 10-ml columns of Sepharose 4B-collagen, equilibrated with 50 mM Tris-HC1/5 mM CaC12 (pH 7.4) were loaded with 0.2 ml of proteoglycan subunit (PGS) solutions (3.75 mg-ml 1). Columns were then eluted with (a) 20 ml of buffer, (b) 20 ml of buffer containing spermidine at various concentrations, (c) 20 ml of buffer containing 1 M NaC1. Results are presented as percent of total proteoglycan subunits eluted in the solvent (b). Each point represents the mean of three experiments.

putrescine and spermine were without effect, spermidine showed a strong capacity in displacing proteoglycan subunits, as shown in Fig. 3. At 15 mM concentration, about 90% of the proteoglycan subunits were removed from the column.

Z o6O > o 40 < 0~

/

0

Spd

llo

I

POLYAMINE

310

I

510

ImM ]

Fig. 4. Effect of polyamines on alkaline phosphatase activity. Spm, spermine; Spd, spermidine; Put, putrescine.

43 munoprecipitates are generally localized along the longitudinal septa of the columnar cell zone, but they are also present in transversal septa. Discussion

Fig. 5. Localizationof spermidine in epiphyseal cartilage with peroxidase-antiperoxidasecomplex method. (1) In the resting zone, an intense staining for spermidine is evident only in the cells. (2) A diffuse staining (black dots) of the intercellular matrix is present only at the limit of the zone of columnar cells where hypertrophyof chondrocytesinitiates. (Original magnification: x 100).

Effect of polyarnines on alkaline phosphatase activity As shown in Fig. 4, spermine and spermidine increase the activity of the alkaline phosphatase. Putrescine appears to be less effective.

Irnmunohistochemical analysis Immunohistochemical localization of spermidine is presented in Fig. 5. Positive reaction for spermidine is evident: (a) in the cells of the resting zone of preosseous cartilage. Only occasionally cell clusters are present in resting cartilage with positive staining for spermidine in the pericellular matrix; (b) cell staining disappears, approaching the zone of proliferating and columnar cells; (c) staining for spermidine is markedly evident in the matrix only at the limit of columnar cells when hypertrophy of chondrocytes initiates. Im-

A growing amount of data demonstrate that polyamines are involved in the control of many metabolic reactions (for a review on the biological effects of polyamines see, for example, Ref. 25). They are considered growth factors and stabilizers of whole cells and membranes; they are associated with the structure and metabolism of nucleic acids and are considered also modulators of protein synthesis. Numerous are also the enzymatic reactions where polyamines appear to exert their control. From the data presented in this paper, a new role of polyamine is suggested, at least in vitro, that of reacting with proteoglycans (Figs. 1 and 2). That this effect of polyamines depends upon the base employed suggests that the size and not only the net charge of the molecules are important in the mechanism leading to the decrement of viscosity and precipitation of proteoglycan subunits. Spermine appears to be the most active compound. The fact that beyond the critical concentrations of the two polyamines, the original viscosity and turbidity values of proteoglycan subunit solutions are restored, can be explained by a more favorable polyamine to p'roteoglycan subunit ratio in stabilizing the colloidal aggregates formed by the two polyionic substances. Interestingly, spermidine is the only compound which shows the ability to displace proteoglycan subunits from collagen, as illustrated in Fig. 3. The other two polyamines are inert in this regard, and also lysozyme is inefficient, in spite of its positive net charge (data not shown). On the basis of these observations, one may conclude that the displacement of proteoglycan subunits from collagen by spermidine is not only to be ascribed to an interference of the ionic bonds between proteoglycan subunit and collagen, but also to a specific action of the structure and conformation of spermidine on the interaction between the core protein of proteoglycan subunit and collagen [6]. An observation of interest is the activation of

44 alkaline phosphatase by polyamines. Also this effect is more pronounced in the presence of spermine and spermidine (Fig. 4). On the basis of the present data, it is difficult to explain the phenomenon: it is perhaps useful to remember that the enzyme is a glycoprotein [26] and that it is activated by Ca 2÷ and Mg 2÷. Polyamines may mimic this effect by a direct action on the conformation of the enzyme. Analyses of the amount of polyamines in the different regions of cartilage do not allow discrimination between intra- and extracellular distribution of the compounds. However, immunohistochemical data show that only at the level of ossifying cartilage, polyamines are extracellular. This area is the zone of maximal polyamine concentration, as revealed by the chemical analysis, and is also the zone where mineralization takes place. The extracellular matrix in this zone represents about 10% of total cartilage volume: since a polyelectrolyte (proteoglycan subunit) can concentrate in its vicinity an electrolyte of opposite charge (polyamine), it is evident that the polyamine concentrations we have employed are of the same order of magnitude as those existing in vivo. Consequently, the effects we describe in vitro may be of interest also for the in vivo situation. The amounts of polyamines we have found in cartilage are of the same order of magnitude as those reported by Conroy et al. [7]. Table I shows that the highest concentration of 15olyamines is to be found at the level of the ossifying region; at the resting zone, putrescine is not even detectable. This is not surprising, since in the ossifying area many cellular compounds are intensely synthesized and accumulate [10,27]. Among the three polyamines, spermidine is the most abundant: a high molar ratio spermidine/spermine has been taken as an index of rapid growth [25]. The highest amount of spermidine in the ossifying region is maintained by putrescine formed in the resting region and accumulated in the ossifying area. The fact that among the polyamines, spermidine is the only compound which exerts in vitro a strong influence on the interaction of proteoglycan subunits with collagen and is also the predominant polyamine in the ossifying region of cartilage suggests that the compound is involved in preparing the matrix for calcification.

Acknowledgements Research supported by the Italian National Research Council and by grants from the Italian Ministry of Public Education.

References 1 De Bernard, B., Lunazzi, G.C., Modricky, C., Moro, L., Panfili, E., PolleseUo, P., Stagni, N. and Vittur, F. (1985) in The Chemistry and Biology of Mineralized Tissues (Butler, W.T., ed.), pp. 142-145, EBSCO Media, Birmingham, AL 2 Vittur, F., Stagni, N., Moro, L. and De Bernard, B. (1984) Experientia 40, 836-837 3 De Bernard, B., Stagni, N., Colautti, I., Vittur, F. and Bonucci, E. (1977) Clin. Orthop. Rel. Res. 126, 285-291 4 Larsson, S.E., Ray, R.D. and Kuettner, K.E. (1973) Calcif. Tissue Res. 13, 271-285 5 Blumenthal, N.C., Posner, A.S., Silverman, L.D. and Rosenberg, L.C. (1979) Calcif. Tissue Int. 27, 75-82 6 Greenwald, R.A., Schwartz, C.E. and Cantor, J.O. (1975) Biochem. J. 145, 601-605 7 Conroy, P.D., Simms, D.M. and Pointon, J.J. (1977) Biochem. J. 62, 347-350 8 Rath, N.C. and Reddi, A.H. (1978) Biochem. Biophys. Res. Commun. 81,106-113 9 Takigawa, M., Ishida, H., Takano, T. and Suzuki, F. (1983) Proc. Natl. Acad. Sci. USA 77, 1481-1485 10 Vittur, F., Pugliarello, M.C. and De Bernard, B. (1971) Experientia 27, 126-127 11 Lee-Own, V. and Anderson, J.C. (1976) Biochem. J. 156, 259-264 12 Bauer, E.A., Joffrey, J.J. and Eisen, A.Z. (1971) Biochem. Biophys. Res. Commun. 44, 813-818 13 Sajdera, S.W. and Hascall, V. (1969) J. Biol. Chem. 244, 77-87 14 Bitter, T. and Muir, H.M. (1962) Anal. Biochem. 4, 330-334 15 Ali, S.Y., Sajdera, S.W. and Anderson, H.C. (1970) Proc. Natl. Acad. Sci. USA 67, 1513-1520 16 Stagni, N., Vittur, F. and De Bernard. B. (1983) Biochim. Biophys. Acta 761,246-251 17 Stagni, N., Furlan, G., Vittur, F., Zanetti, M. and De Bernard, B. (1979) Calcif. Tissue Int. 29, 27-32 18 Labarca, C. and Paigen, K. (1980) Anal. Biochem. 102, 344-352 19 Raina, A. and Cohen, S.S. (1966) Proc. Natl. Acad. Sci. USA 55, 1587-1593 20 Seiler, N. and Kn6dgen, B. (1978) J. Chromatogr. 145, 29-39 21 Janne, J. and Williams-Ashman, H.G. (1971) J. Biol. Chem. 246, 1725-1732 22 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275 23 Bartos, F. and Bartos, D. (1978) in Advances in Polyamine Research (Campbell, R.A., Morris, D.R., Bartos, D., Daves, G.D. and Bartos, F., eds.), pp. 65-70, Raven Press, New York

45 24 Mukay, K. and Rosai, J. (1980) in Progress in Surgical Pathology (Fenoglio, C.M. and Wolff, M., eds.), pp. 15-49, Masson, New York 25 J~inne, J., P6s6, H. and Raina, A. (1977) Biochim. Biophys. Acta 473, 241-293

26 De Bernard, B. (1982) Clin. Orthop. Rel. Res. 162, 233-244 27 Lindenbaum, A. and Kuettner, K.E. (1967) Calcif. Tissue Res. 1, 153-165