Differential Effects of Parathyroid Hormone Fragments on Collagen Gene Expression in Chondrocytes
Silke Erdmann, Wolfgang Müller, Safarali Bahrami, Silvia I. Vornehm, Hubert Mayer, Peter Bruckner, Klaus von der Mark, and Harald Burkhardt
Department of Internal Medicine III, Institute for Clinical Immunology and Rheumatology, University of Erlangen-Nürnberg,

Institute for Experimental Medicine, University of Erlangen-Nürnberg,
Institute for Physiology, Section Neurophysiology, Charité, Berlin

Institute for Physiological Chemistry and Pathobiochemistry, University of Münster,
Gesellschaft für Biotechnologische Forschung mbH, Braunschweig, Germany

– Overview
– Materials and Methods
– Results
– Discussion
– References

The effect of parathyroid hormone (PTH) in vivo after secretion by the parathyroid gland is mediated by bioactive fragments of the molecule.

To elucidate their possible role in the regulation of cartilage matrix metabolism, the influence of the amino-terminal (NH2-terminal), the central, and the carboxyl-terminal (COOH-terminal) portion of the PTH on collagen gene expression was studied in a serum free cell culture system
of fetal bovine and human chondrocytes. Expression of a1 (I), a1 (II), a1 (III), and a1 (X) mRNA
was investigated by in situ hybridization and quantified by Northern blot analysis. NH2-terminal
and mid-regional fragments containing a core sequence between amino acid residues 28-34 of PTH induced a significant rise in a1 (II) mRNA in proliferating chondrocytes. In addi-tion, the COOHterminal portion (aa 52-84) of the PTH molecule was shown to exert a stimulatory effect on a1 (II) and a1 (X) mRNA expression in chondrocytes from the hypertrophic zone of bovine epiphyseal cartilage.
PTH peptides harboring either the functional domain in the central or COOH-terminal region of PTH can induce cAMP independent Ca2+ signaling in different subsets of chondrocytes as assessed by microfluorometry of Fura-2/AM loaded cells. These results support the hypothesis that different hormonal effects of PTH on cartilage matrix metabolism are exerted by distinct effector domains and depend on the differentiation stage of the target cell.

PARATHYROID hormone plays a predominant role in the regulation of calcium homeostasis by acting mainly on its target tissues in the renal cortex and bone (15, 42).
Soon after secretion the parathyroid hormone (PTH) molecule undergoes rapid proteolysis in the liver resulting in multiple fragments (7). Since most of the calcium regulatory functions could be mapped to the NH2-terminal portion (PTH l-34) of PTH, it was thought that this fragment contains all structural requirements for biological activity of the entire molecule (43, 52).

The other fragments were regarded as inactive metabolites whose functional importance was confined to processing and intracellular transport events during hormone secretion by the cells of the parathyroid gland (PTH 53-84, references 35, 46).
However, there is now increasing evidence for a broader spectrum of target tissues, including cartilage (26, 33, 34), processes (8,13) which are mediated by additional functional domains on the mid-regional (23) and COOHterminal portion (39,40,44) of PTH. For example, PTH (53-
84) increases alkaline phosphatase activity in osteoblastic cell lines (39) and more recent studies showed that PTH (39-84) and PTH (53-84) dose dependently stimulate the differentiation of osteoclast precursors into osteoclast-like cells (25). Moreover, for two domains of PTH their
functional role in the induction of second messenger pathways has been elucidated: the first two NH2-terminal amino acids of PTH are needed for adenylate cyclase Stimulation via the “classical” PTH receptor (24, 43,52), whereas the mid-regional part, aa 28-34, is responsible for
induction of protein kinase C activation in target cells (23).
This domain in the central part of PTH also seems to be critical for the mitogenic effect of the fragments PTH (28- 48) and (1-34) in primary cultures of sternal embryonic chicken chondrocytes (48). A more complex pattern of PTH effects has been demonstrated in a culture System of neonatal murine mandibular condylar explants exposed simultaneously to PTH fragments 1-34, 28-48, and 53-84 (51). Each of these fragments was shown to exert distinct
biological effects (51) on the cartilage morphology, indicating a potential critical role for PTH in normal endochondral ossification.
In contrast to these well characterized PTH effects on cell numbers, cell shape and extracellular matrix morphology, little is known about the ability of the different PTH fragments to induce
quantitative and/or qualitative changes in collagen gene expression by chondrocytes. Therefore, it was the aim of this study to elucidate the regulatory potential of different PTH fragments on collagen metabolism of chondrocytes, which might be a critical underlying mechanism of PTH action on cartilage.

Several studies (8, 13, 30) have indicated that the response of chondrocytes to PTH depends on the source and developmental stage of the cartilage. In embryonic transient cartilage, such as
epiphyseal cartilage of long bone rudiments, chondrocytes rapidly proliferate and undergo a series of differentiation steps.

These stages of chondrocyte differentiation are aligned sequentially from the epiphyseal surface down to the growth zone in the diaphysis and are characterized by the expression of
different collagen types as specific markers (for review see reference 55). Preembryonic cells in the superficial layer express collagen type I, proliferating chondrocytes in the middle
zone synthesize collagen II, VI, IX, and XI, whereas hypertrophic chondrocytes of the growth plate can be unequivocally identified on the basis of their collagen type X expression (16,
49). The expression of this characteristic collagen in the deep zone of the epiphysis seems to be functionally related to endochondral ossification processes (16, 49) in the matrix preceeding cartilage resorption by osteoclasts and replacement by endochondral bone. Since PTH is known to promote endochondral ossification (50), it was the aim of our study to investigate
modulatory effects of PTH fragments on the expression of collagen II and X mRNA in epiphyseal chondrocytes of different developmental stages.

Chemicals and Supplies

• Abstract
• Results
• Discussion
• References

Bovine (b) and human (h) PTH fragments were obtained from Sigma (St. Louis, MO): h, b(PTH) bPTH (1-34), (Nie 8,18,Tyr 34) bPTH (3-34), hPTH (13-34), hPTH (28-48), hPTH (39-68),
(Tyr52, Asn 76) hPTH (52-84), and hPTH (64-84).

Tissue culture supplies were purchased from Becton Dickinson (NJ), and PCS from PAA-Labor, Forschungs GmbH (Linz, Austria).

Northern Blot
Chondrocytes were lysed in 4 M guanidinium thiocyanate, 25 mM Nacitrate, 0,5% Na-sarcosyl, and 0.7% ß-mercaptoethanol. RNA was extracted using the cesium-chloride density centrifugation method of Chirgwin et al. (9). Total RNA (10 u-g) was subjected to formaldehyde
gel electrophore-sis, blotted onto nylon filters and cross-linked by exposure to UV light for 5 min. For analysis of RNA, cDNA probes were labeled with [32P]dATP or dCTP by random priming and hybridized in 50% formamide, 5x SSC, 5x Dehnhardt’s solution (4), 0.5% SDS, and 100
(ig/ml herring sperm DNA at 42°C for 16 h. After hybridization, filters were washed twice in 2X SSC at room temperature for 5 min, twice in 2x SSC/0.1% SDS each at 50°C for 30 min, and once in 0.1 X SSC/0.1% SDS at room temperature for 2 min.

The washed filters were exposed to Kodak X-OMAT™ X-ray films (Eastman Kodak).
Microfluorometry of Ca2+-Signaling For Ca2+ imaging, chondrocytes were seeded on 35-mm tissue culture dishes and incubated in Harn F-12 for 24 h before loading with Fura-2/AM (5 u,M) for 30 min at 37°C. Ca2+ was imaged with an upright microscope (Zeiss Axioskop FS, Jena, Germany) and a 40X water Immersion objec-tive. A CCD camera System (Photometrics Ltd., Tuscon AZ) (12, 37, 38) was used to acquire digitized Images of Fura-2 fluorescence. Free Ca2+ concentrations were determined from background corrected image pairs at 350 and 380 nm excitation with the ratio method (18).

The responsive-ness of the calcium signaling machinery of the chondrocyte population was controlled by the Ca2+ response to 5 u,l FCS.
Subsequently, the return of intracellular free Ca2+ to stable baseline levels was recorded for at least 10 min before the application of PTH fragments.
Cells were continuously superfused with saline containing NaCl 140 mM, KC15 mM, MgSO4 2 mM, NaH2PO4 l mM, glucose 5.5 mM, Hepes 20 mM, pH 7.4. The PTH pep-tides were applied by microdrop application of concentrated stock solution into the bath to give final concentrations äs
indicated. When different PTH fragments were sequentially tested on the same chondrocyte population, a return to baseline Ca2+ fluorescence and stability for 10 min was imperative before application of a new peptide.

Agarose Cell Culture
Juvenile human costal cartilage obtained from funnel chest operations was dissected free of surrounding tissues and cut into 0.5-mm slices. Chondrocytes were released by collagenase digestion and cultured in agarose gels under serum-free conditions as described previously (6, 53). Briefly, matrix-free cells suspended in media containing 0.5% of low melting agarose
were seeded into prewarmed culture dishes coated with 1% high melting agarose gels. The cultures were maintained at 37°C to keep the low melting agarose in the liquid state and, thus, to allow the cells to sediment at the interface of the two agarose layers. Thereafter, the low
melting agarose was allowed to gel by brief exposure of the cultures to 4°C. Cultures were then supplemented with additional medium.

Cells were grown at densities of 2 X 106/ml in DMEM (Grand Island Biologicals Corp., Basel, Switzerland) containing 60 mg/ml of ßaminoproprionitrile fumarate, 50 mg/ml sodium ascorbate, l mM cysteine, l mM pyruvate, 100 U/ml penicillin, and 100 mg/ml streptomycin. Where applicable, PTH fragments were added for 24 h. Subsequently, the media were exchanged and the cultures were maintained for another 48 h in the presence of l mCi [14C]proline (uniformly labeled,285 mCi/mmol, Amersham International) and analyzed for collagen synthesis.

The radiolabeled agarose cultures were homogenized and newly synthesized collagens extracted by digestion with pepsin (6). The total extracted proteins were precipitated with ethanol and analyzed by SDSPAGE (4.5-15% gradient gel) followed by fluorography.

3.1. In costal cartilage, collagen X is expressed much
earlier than alkaline phosphatase

• Abstract
• Materials and Methods
• Discussion
• References

The expression of collagen X in normal human costal cartilage tissue from donors of different ages
was examined by extracting the tissues with neutral buffers containing high concentrations of guanidinium hydrochloride, i.e. under conditions denaturing soluble collagens.
The crude extraction mixtures were subjected to polyacrylamide gel elcctrophoresis in SDS and collagen X was probed by immunoblotting with an antiserum specifically reacting with human collagen X (Kirsch and von der Mark, 1991).

Equal amounts of total protein from each donor were loaded

Fig. 1. Collagen X is a component of costal cartilage matrix already in childhood. Collagens were extracted with 6 M guanidinium hydrochloride from costal cartilage of normal donors of several
ages. Equal amounts of total protein were run on a 4.5-15% polyacrylamide gradient gel. An immunoblot is shown and the donor age is indicated below the lanes.

Fig. 2. Collagen X immunostaining of human rib cartilage from a 3-year-old donor. Costal cartilage and adjacent connective tissue were recovered from a central location remote from the bony parts of both the sternum and ribs, as well as the costal growth plate.
Staining was with a rabbit antiserum to human collagen X, detected by an antibody peroxidase anti-peroxidase kit. P, perichondrium; Sand I, surface- and inner zone, respectively. Note: staining occurred in the extracellular cartilage matrix in the vicinity of all chondrocytes,
but appeared somewhat stronger in locations rich in cells close to the perichondrial boundaries.
Tissues surrounding cartilage were negative (upper left corner).

onto the gels and collagen X was identified as an immunoreactive band with an apparent molecular mass of approximately 60 kDa. Collagen X was clearly detectable in tissues from 3-year-old donors and reached maximal levels in 7-year-old children, i.e. well before puberty (Fig. 1).

Collagen X-producing chondrocytes occur throughout the costal cartilage and deposit the protein predominantly into the extracellular matrix in their vicinity (Fig. 2). Thus, regardless of the donor age, costal cartilage contains chondrocytes having reached the most advanced stages of late differentiation.

Gels examined by staining with coomassie blue displayed strong bands of al(II)-chains
of collagen II, but failed to produce evidence for a2(I)-chains of collagen I (not shown).
Rib cartilage was also extracted with neutral saltbuffer without guanidinium chloride and alkaline
phosphatase activity was determined in the extracts.
Similar levels of enzyme activity were recovered from rib cartilage of subjects older than 17 years, but not pre-pubertal children (Fig. 3). Alkaline phosphatase activity was not demonstrable by enzyme histology in rib cartilage of a 3-year-old child (not shown). Therefore, unlike in all other cartilaginous tissues undergoing late differentiation, a clear separation occurs in
rib cartilage in the expression of the two markers of chondrocyte hypertrophy. The occurrence of alkaline phosphatase activity, but not collagen X, coincides with tissue mineralization and ossification.

Only in costal cartilage from post-pubertal donors, are matrix vesicles containing apatite mineral (Fig. 4) observed by electron microscopy on unstained sections with zero-loss filtering.

The mineral of the matrix vesicles was further investigated by electron diffraction. Bragg
reflections or reflection rings were recorded (Fig. 4, inset) and corresponded to a lattice spacing of 0.344 and 0.272 nm, respectively, which are characteristic for biological apatite consisting of a mixture of hydroxyl apatite and dahlite.

3.2. Proliferation and maturation of costal chondrocytes in culture
The capacity of costal chondrocytes to advance to late stages of differentiation was further investigated in long-term suspension culture in agarose gels. Under these conditions, chondrocytes from the cranial portion of chick embryo sterna proliferated and became
overtly hypertrophic within 1-2 weeks if the media contained 100 ng / ml of IGF-1 or insulin, or 50ng / ml of thyroxine or FBS (Böhme et al., 1995).
Costal chondrocytes from a 7-year-old proband were exposed to media containing FBS, or other factors without serum, and proliferation was assessed by direct counting of the cells in several representative microscopic fields revisited throughout the cultures. 710 a1 (X)

The cells only slightly proliferated (approx. 1.1-fold after 3 weeks) after stimulation by 10% FBS. Cell division was apparent in a few, but not all cells (arrows in Fig. 5, panels Q T). Lesser amounts of serum, as well as insulin, IGF-1, thyroxine, or PTH at maximal concentrations compatible with cell survival did not stimulate cell division (Fig. 5, panels E-P, and results not shown). These results indicated that most costal chondrocytes do not divide frequently during childhood.
However, cells cultured with 10% FBS expressed hypertrophy markers (Figs. 6 and 7) and increased their size (arrowheads in Fig. 5, panels Q-T). Exposure to the carboxyterminal fragment (residues 53-84) and, to a lesser extent, the aminoterminal portion (residues 1-34)
of parathyroid hormone, also raised synthesis of collagen X well above the detection limit (Fig. 6), but the cells did not increase their size (Fig. 5, panels I-P).

By contrast, collagen X was not synthesized in the presence of other signals, including insulin, IGF-1, and thyroid hormone (not shown), whereas the cell size was increased by treatment with IGF-1 (Fig. 5, panels M-P) or insulin (not shown). Therefore, collagen X-synthesis was not uniformly associated with a large cellular volume as seen in cartilage undergoing late
differentiation during endochondral ossification, such as in growth plates. Alkaline phosphatase activity was generated by the cells only under the influence of the carboxyterminal fragment of PTH for 48 h or 10% FBS.

All other factors in the defined culture medium were ineffective (Fig. 7). Thus, similarly to authentic tissue, the two hypertrophy markers were not simultaneously expressed by cultured costal chondrocytes under the direction of the aminoterminal domain of PTH (residues
1-34). This difference was not apparent in cells under the control of 10% FBS or, to a lesser extent, the carboxyterminal domain of PTH (residues 53-84). Analogously to native tissues, however, mineralized matrix vesicles similar to those shown in Fig. 4 only appeared in cultures expressing both hypertrophy markers (not shown).

Fig. 3. Alkaline phosphatase activity occurs only in post-pubertal costal cartilage. Proteins were extracted under non-denaturing conditions from costal cartilage fragments from donors of different ages. Alkaline phosphatase activity was determined by hydrolysis of pnitrophenyl-

phosphate in aliquots of the extracts containing equal amounts of total protein and values are given in arbitrary relative

Fig. 4. Matrix vesicles containing biological apatite mineral occur in post-pubertal costal cartilage. Electron micrograph of an unstained section of costal cartilage-matrix from a 17-year-old male donor. The tissue fragment was derived from a central location remote from
overtly osseous tissues. Inset: diffraction pattern of the mineral within the matrix vesicle shown in main panel. Several crystalline reflections (arrowhead) and a diffraction ring (arrow) are visible, corresponding to lattice spacings of 0.344 and 0.272 nm, respectively. Both patterns are characteristic for biological apatite mineral. Bar, 0.1 mm.

• Abstract
• Materials and Methods
• Results
• References

The results presented in this study show that different fragments of the PTH molecule stimulate collagen type II and X gene expression in chondrocytes under serum-free conditions. In initial experiments using postnatal human costal chondrocytes, PTH (1-34) and PTH (52-84) stimulated synthesis of collagen type II and X in a serum-free agarose culture System. Due to the restricted availability of appropriate amounts of human chondrocytes and the disadvantages of the agarose culture System for studies of gene expression at the mRNA level, further experimentation was performed using a serum-free culture System of bovine growth plate chondrocytes, separated into cells from the resting and proliferating zone, and in cells from the hypertrophic zone.

In agreement with the Stimulation of collagen type II and X at the protein level, a rise in a1 (II) and a1 (X) mRNA level was detected in response to different PTH petides. The results clearly show that the stimulatory PTH effect on collagen mRNA levels is dependent on the differentiation stage of the cells and induced by at least two different functional domains of PTH.

The first domain is located in the central part of the PTH molecule between aa 28-34 and is capable of stimulating a1 (II) expression in resting and proliferating fetal chondrocytes.
A second domain is located in the COOH-terminal part of PTH between aa 52-84.

This domain is recognized only by cells which are differentiating towards the hypertrophic stage; it is not active on proliferating chondrocytes.
Fragments lacking the genuine NH2-terminus (aa residues 1-3) of the hormone, which is indispensible for activation of the PTH receptor associated adenylate cyclase (15, 43, 52) also stimulate type II collagen expression.

This indicates that cAMP does not play a critical role in the signaling pathway of PTH-mediated upregulation of type II collagen gene expression.
All functionally active fragments are capable of inducing a rise in calcium concentration in the chondrocytes as shown in Fig. 9. Similarly, a cAMP independent Ca2+- signaling, involving protein kinase-C activation (57) has been demonstrated in other experimental Systems of PTH action (48). The activation of this signal transfer cascade by PTH is dependent on a region (aa 28-34) in the central part of PTH (23). We mapped the functional domain for Stimulation of collagen type II gene expression in proliferating chondrocytes to the same region. It is, therefore, very likely that a Ca2+ signal induced by the protein kinase-C domain of PTH (23) is also involved in the upregulation of a1 (II) expression in proliferating chondrocytes.

Figure 9. Ratio imaging of intracellular free calcium in bovine fetal chondrocytes as determined by Fura-2 fluorescence. Monolayer cultures of chondrocytes from the hypertrophic zone were stimutated by a bolus application (final concentration: 10-8 M) of PTH (52-84)

(A2 PTH5 and B6 PTH5), PTH (28-48) (A4 PTH2 and B2 PTH2), or PTH (1-34) (B4 PTH1).
A shows representative ratio Images obtained by a Stimulation experiment with a sequential application of PTH peptides 52-84 and 28-48. A1 is the control image before Stimulation with PTH (52-84). A3 shows the return to a stable baseline level and represents the control image preceding Stimulation with PTH (28-48). B shows the images derived from a sequential Stimulation experiment with three different PTH fragments: PTH (28-48) (B2 PTH2), PTH (1-34) (B4 PTH1), and PTH (52-84) (B6 PTH5). B1 Cont is the control image before Stimulation; B3 Cont and B5 Cont demonstrate the return to stable baseline after the preceeding stimulations with the respective PTH peptides. Bar, 20 mm.

In chick chondrocytes, it has been shown (48) that this central PTH domain mediates an EGTA-sensitive mitogenic effect on the cells. However, under the experimental conditions of this study which are high plating density of the cells (1.3-1.91 05/cm2), a 24 h period of hormone treatment, and strict serum-free conditions, no mitogenic effect was detectable for any PTH
fragment.

These culture conditions account for the absence of any proliferative response of the chondrocytes to PTH, which is in accordance with the observations made by Schlüter et al. (48) that high cell densities impede the mitogenic effect of PTH. Therefore, the same functional domain of PTH (amino acid residue 28-34) can exert quite different effects, either a mitogenic response or an increase in differentiated function (a1 (II) expression), depending on the duration of the hormonal Stimulus and the cell density.
Table I. Ca2+ Response Frequency in Fetal Bovine Chondrocyte from the Hypertrophic Zone of the Epiphysis: Ratio Image Analysis PTH (1-34) PTH (28-48) PTH (52-84) FCS Responses (n) 19 24 35 127 Total cells (n) 50 85 119 150 Response 26.0 28.2 29.4 84.6

Fura-2 loaded chondrocytes from the hypertrophic zone of cartilage were examined by microfluorometry. Changes of intracellular free Ca2+ were determined in individual cells by ratio image analysis.

A response was considered to have occured when at least an increase of 0.5 mM from the
baseline Ca2+ was recorded. The number (n) of responsive cells upon Stimulation by different PTH peptides and with FCS is shown.

The response frequency was calculated from the ratio of responsive cells to total cells measured.

Figure 9. Ratio imaging of intracellular free calcium in bovine fetal chondrocytes as determined by Fura-2 fluorescence. Monolayer cultures of chondrocytes from the hypertrophic zone were stimutated by a bolus application (final concentration: 10-8 M) of PTH (52-84)
(A2 PTH5 and B6 PTH5), PTH (28-48) (A4 PTH2 and B2 PTH2), or PTH (1-34) (B4 PTH1). A shows representative ratio Images obtained by a Stimulation experiment with a sequential application of PTH peptides 52-84 and 28-48. A1 is the control image before Stimulation with PTH (52-84). A3 shows the return to a stable baseline level and represents the control image preceding Stimulation with PTH (28-48). B shows the images derived from a sequential Stimulation experiment with three different PTH fragments: PTH (28-48) (B2 PTH2), PTH (1-34) (B4 PTH1), and PTH (52-84) (B6 PTH5). B1 Cont is the control image before Stimulation; B3 Cont and B5 Cont demonstrate the return to stable baseline after the preceeding stimulations with the respective PTH peptides. Bar, 20 mm.

These environmental influences are critical for PTHmediated effects, hence differences in culture conditions may also account for some apparently controversial results found in the literature on the hormone action on Collagen synthesis by chondrocytes (13, 21, 36, 41). In the study of
Crabb et al. (13) articular chick chondrocytes remained totally unresponsive to PTH, whereas a mitogenic effect and Inhibition of collagen synthesis was reported in growth plate chondrocytes. Similar inhibitory effects of PTH (1- 34) and PTH (54-84) on collagen X synthesis by
hypertrophic rabbit chondrocytes were recently reported by Iwamoto et al. (21).

In a detailed analysis of PTH effects on long term cultures of maturing chick sternal and tibial growth plate chondrocytes, Iwamoto et al. (21) demonstrated that PTH has an inhibitory effect on the emergence of collagen X expressing cells and that this inhibitory effect persisted for the whole maturation pathway of the chondrocytes, which is in contrast to the more stage-specific effects of FGF-2 (22).
However, these experiments were performed in long term cultures and in the presence of 5-10% FCS in the culture medium, while all stimulatory effects of the PTH fragments on collagen gene expression reported here depend on strict serum-free conditions. We have shown that PCS dramatically enhances collagen gene expression within 24 h.

As uncontrolled effects by endogenous growth factors in PCS cannot be excluded (10), serum was omitted from all stages of the experiments reported here. In agreement with Iwamoto et al. (21, 22) we found in our culture System that the stimulatory PTH (1-34) effect on collagen gene
expression was not only abolished, but also reverted, when the chondrocytes were exposed to serum during collagenase digestion before the PTH Stimulus (Vornehm, S., manuscript in preparation). Reduced viability of the chondrocytes prepared and cultured under serum-free conditions was excluded by the fact that freshly prepared cells showed strong signals for a1 (II) mRNA and divided normally.
Since prolonged enzymatic digestion of cartilage in the absence of serum might cause cell damage, care was taken to reduce the time of enzyme treatment to a minimum.

Properly treated chondrocytes remained viable and retained their FCS-responsiveness, as well as responsive-ness to PTH. This PTH response of freshly isolated cells remained stable for a culture period of at least 3d under strict serumfree conditions, however, preference was given in this study to an immediate Stimulation of the chondrocytes already 6 h after Isolation in order to exclude any uncontrolled influence of in vitro (de)differentiation.

The stimulatory effect of PTH peptides on collagen types II and X expression is not restricted to mRNA levels and monolayer conditions; identical results were obtained at protein level with human costal chondrocytes cultured in agarose Suspension. Thus, it is likely that the apparent conflict between our data and those published by Iwamoto et al. (21, 22) result from different chondrocyte culture Systems and reflect the biologically relevant sensitivity of PTH effects to modulation by growth factors present in serum.
Analysis of serum factors that modify the PTH effect would help in the understanding of the complex regulation of collagen metabolism during endochondral bone formation.
In this study a new effector domain for chondrocytes was localized in the COOH-terminal region of PTH (amino acid residues 52-84), which exerts a selective effect on collagen type II and X expression in growth plate chondrocytes from the hypertrophic zone. Proliferating chondrocytes did not respond to PTH peptides derived from the COOH terminus.
It has been shown by Murray et al. (40) that cells differentiated towards the osteoblastic lineage (human osteosarcoma SaOS-2 cells) increase type I collagen mRNA levels in response to PTH (1-34), but not to PTH (53-84), although the COOH-terminal fragment stimulated expression of mRNA for osteocalcin, the vitamin Dreceptor and alkaline phosphatase in the same cells.

This underlines the domain specificity and differentiation stagedependency of the PTH action and supports the concept of a physiological role for PTH metabolites in the hormonal control of matrix metabolism in the growth plate. In accordance with this hypothesis are results from in situ
hybridization studies on fetal rat cartilage (32), showing strong PTH receptor gene expression in a distinct zone of maturing chondrocytes immediately above the layers of hypertrophic cartilage. Moreover, by light microscope autoradiography, Barling and Bibby (5) demonstrated
[3H]PTH binding to hypertrophic chondrocytes; a histological study from 1943 revealed hypertrophy, calcification, and premature closure of the growth plate induced by intraperitoneal administration of PTH to growing mice (50). In an organ culture system of mandibular explants, the COOH-terminal fragment PTH 53-84 exerted a profound change of morphology in the zone of hypertrophic cartilage (51).
In this respect our results suggest that one important facet of PTH action is the Stimulation of collagen type X gene expression in the hypertrophic zone of the epiphysis.
Moreover, modulation of collagen metabolism could be a critical event in calcification of growth plate cartilage since recent data (27, 28, 29) indicate that the interaction of collagen type II and X with the matrix vesicles in the growth plate activate Ca2+ loading of these extracellular microstructures, which are considered the Initiation sites of mineral deposition in cartilage.
Another aspect of the role of PTH in cartilage mineralization is closely related to our finding that PTH metabolites are capable of inducing a rise in intracellular free Ca2+ in chondrocytes from the hypertrophic zone.
Matrix vesicles are formed in chondrocytes by budding from the cytoplasmatic membrane (3, 19) leaving the possibility open that they retain the chondrocytic PTH receptors in their cell membrane. For mineralization of cartilage, it remains to be elucidated whether the matrix
vesicles still respond to PTH fragments after deposition in the extracellular matrix by increasing the intravesicular Ca2+ concentration.
It is not yet clear how the COOH-terminal part of PTH is recognized by the hypertrophic chondrocytes, and why the proliferating chondrocytes remain unresponsive to COOHterminal PTH fragments. PTH and PTH-related peptide (PTHrP) bind to a common heptahelical G-protein coupled receptor molecule (24, 47). Since this classical PTH receptor has a widespread tissue distribution, receptor heterogeneity as a consequence of alternative splicing of the intron-rich PTH receptor gene (23, 31) is an attractive hypothesis for explanation of the observed heterogeneous
PTH responses in the chondrocyte. However, direct proof for the existence of such receptor isoforms in cartilage is lacking. A more ligand selective isoform of the PTH/PTHrP receptor has been identified and characterized for its unresponsiveness to PTH-related peptide (PTHrP), but this PTH 2 receptor seems to be particularly abundant in pancreas and brain and also recognizes the aminoterminal fragment of PTH (54). However, more recently a novel PTH receptor with specificity for the carboxylterminal region of PTH has been characterized in rat osteosarcoma and parathyroid cell lines (20). Moreover, in osteosarcoma cell lines (ROS 17/2.8), this COOH-terminal receptor seemed to be upregulated in response to PTH
Stimuli (PTH 1-34) that are mediated via the common PTH/PTHrP receptor, implying its role in C-receptor expression (20).
Thus, the possibility remains that one of the above mentioned receptors may be involved in the stimulatory effect of the COOH-terminal part of PTH. Receptor isoforms could also explain the differential effects of central vs COOH-terminal PTH peptides on the induction of Ca2+ signaling in distinct cell subsets. Alternatively, Civitelli et al. (11) suggested a nonuniform distribution of functional receptors over the cell surface to explain similar heterogeneous calcium responses to PTH in the osteogenic sarcoma cell line UMR 106. A conformational change in a common receptor molecule could also account for the selective action of the different functional domains of PTH on distinct subpopulation of chondrocytes. Irrespective ofthe receptor molecule involved, our results suggest that two distinct functional domains on the PTH molecule can
exert different hormonal effects on collagen II and X metabolism by epiphyseal chondrocytes, depending on the differentiation stage of the cells.
We thank Dr. G. Gross (GBF Braunschweig) for helpful discussions and Dr. L. Sorokin for reading the manuscript. Initial Ca2+-imaging experiments were performed in the laboratory of Professor Dr. D.
Swandulla (Department of Pharmacology and Toxicology, University of Erlangen). We are grateful to Professor Dr. J.R. Kalden (Department of Internal Medicine III, University of Erlangen) for generous support.
This work was supported by the German Ministry of Research and Technology (BMBF No. 01VM8702) and by the Deutsche Forschungsgemeinschaft (Heisenberg grant to Wolfgang Müller, grant Br 1497/1-1 to Peter Bruckner, and SFB263 grant C3 to H. Burkhardt).
Received for publication 9 July 1996 and in revised form 29 August 1996.