Abstract The role of major cellular serine/threonine-specific protein phosphatases, protein phosphatase 1 and2A, was investigated during chicken cartilage differenti-ation under in vitro conditions. Activity of protein phos-phatase 2A decreased parallel to differentiation of chon-drogenic cells, whereas activity of protein phosphatase 1remained unchanged as assayed in the supernatants ofthe homogenised chicken limb bud micromass cell cul-tures. When okadaic acid, a potent inhibitor of proteinphosphatase 1 and 2A was applied in 20 nM concentra-tion for 4 h during the second and third culturing days, itsignificantly increased the size of metachromatic carti-lage areas measured in 6-day-old colonies. Followingokadaic acid treatments, a significant inhibition in theactivity of protein phosphatase 2A was found, while theactivity of protein phosphatase 1 was unaffected as mea-sured an days 2 and 3. TRITC-phalloidin labelling dem-onstrated that okadaic acid disorganised actin filamentsand induced rounding of chondrogenic cells. This deteri-oration of actin filaments was reversible. Electron mi-croscopy and biochemical analysis of colonies revealedthat the ultrastructure and major components of cartilagematrix remained unchanged under the effect of okadaicacid. Okadaic acid-treatment applied to cultures contain-ing predominantly differentiated chondrocytes (after day 4) did not influence the cartilage formation. 3H-thymidine and bromodeoxyuridine incorporation-assays

demonstrated enhanced cell proliferation in the okadaicacid-treated colonies compared to that of the untreatedones. Our results indicate, for the first time, that pro-tein phosphatase 2A is involved in the regulation ofchondrogenesis. Inhibition of protein phosphatase 2Awith okadaic acid may result in increased chondrogene-sis via modulation of proliferation and cytoskeletal or-ganisation, as well as via alteration of protein kinase A-signaling pathway of the chondrogenic cells.

Keywords Cartilage differentiation · Cell proliferation ·Cyclic AMP-dependent protein kinase · Protein dephosphorylation · Protein phosphatase 1

Introduction

Nearly all cell functions are regulated by reversible pro-tein phosphorylation. Some examples include metabolicprocesses, gene regulation, cell cycle control, transportand secretory processes, the organisation of the cytoskel-eton and cell adhesion (Barford et al. 1998). Biologicalprocesses that depend an reversible phosphorylation re-quire not only protein kinases but also protein phospha-tases and the cellular concentrations of kinases are ap-proximately equal to that of phosphatases (Hunter 1995),suggesting, that targeted substrate proteins are specifical-ly phosphorylated by protein kinases and dephosphory-lated by protein phosphatases. Thus, modulation of ki-nase or phosphatase activities can alter the degree ofphosphorylation at a given cognate site.

Ser/Thr-phosphorylation of cellular proteins bycAMP-dependent protein kinase A (PKA; Leonard andNewman 1987; Zhang et al. 1996; Lee and Chuong1997), protein kinase C (PKC; Sonn and Solursh 1993)or tumor growth factor-β (TGF-β) superfamily receptors(Kulyk et al. 1989; Leonard et al. 1991) is an importantand well-studied event in the course of cartilage differen-tiation. However, the involvement and possible role ofdephosphorylation catalysed by Ser/Thr-specific proteinphosphatases (Ser/Thr PPs) are less understood.

R. Zákány (✉ ) · Sz. Felszeghy · K. Holló · L. MódisDepartment of Anatomy, Histology and Embryology, Medical and Health Science Center, University of Debrecen, Nagyerdei krt 98, Debrecen, 4012 Hungarye-mail: roza@chondron.anat.dote.huFax: +36–52–432290

É. Bakó · P. GergelyDepartment of Medical Chemistry, Medical and Health Science Center, University of Debrecen, Bem tér 18/B, Debrecen, 4026 Hungary

M. Balázs · H. BárdosDepartment of Preventive Medicine, Medical and Health Science Center, University of Debrecen, Nagyerdei krt. 98, Debrecen, 4012 Hungary

Anat Embryol (2001) 203:23–34 © Springer-Verlag 2001

O R I G I N A L A RT I C L E

Róza Zákány · Éva Bakó · Szabolcs Felszeghy Krisztina Holló · Margit Balázs · Helga Bárdos Pál Gergely · László Módis

Okadaic acid-induced inhibition of protein phosphatase 2A enhances chondrogenesis in chicken limb bud micromass cell cultures

Accepted: 26 June 2000

Ser/Thr PPs are classified as PP1, PP2A, PP2B andPP2C based an biochemical parameters. It is notablethat PP2A, more so than PP1, seems to play a majorrole in reversing all the steps of the protein kinase sig-nalling pathways (Brautigan 1997) and PPs also play acrucial role in regulating the cell cycle (Wera and Hemmings 1995). Inhibition of PPs with different PP-inhibitors is a relatively new approach to clarify therole of these enzymes in many cellular functions. Okadaic acid (OA) is a toxin that has been shown to bea potent tumour promoting substance. OA is able topenetrate into living cells, binds to and inhibits both ofthe major cellular phosphatases, PP1 and PP2A. Re-ported IC50 values for OA range from 0.04 to 1 nMagainst the PP2A catalytic subunit, and from 12 to 500 nM against the catalytic subunit of PP1 (Biajolanand Takai 1988; Cohen et al. 1989). The inhibition ofPPs leads to an increase in overall protein phosphoryla-tion in OA-treated cells (Haystead et al. 1989). Thetoxicity of OA and other PP inhibitory toxins affirmsthat phosphatases are essential for vital cellular func-tions (for references see Cohen et al. 1990; Wera andHemmings 1995; Brautigan 1997).

Since there is no account of the effect of OA on thedifferentiation of normal cells, the aim of our experi-ments was to investigate the alteration of chondrogenesisby OA in a well-studied differentiation model – chondri-fying high density cultures of chicken limb bud mesen-chymal cells (HD cultures).

Materials and methods

Cell culturing

Chondrifying micromass cultures were prepared from distal limbbuds of Arbor Acres chicken embryos of Hamburger-Hamiltonstages 22–24 with minor modifications of the original protocol(Ahrens et al. 1977; Hadházy et al. 1982). Five 10 µl drops of cellsuspension containing 1.5×107 cells/ml were inoculated onto plas-tic coverslips (Nunc, Naperville, USA). Coverslips were placedinto plastic Petri dishes and cultured at 37°C in Ham’s F12 medi-um (Sigma, St. Louis, USA) supplemented with 10% fetal calf se-rum (FCS; Sebak, Aidenbach, Germany), antibiotics and antimy-cotics, in the presence of 5% CO2 and 95% humidity in a CO2 in-cubator. The medium was changed every second day.

OA treatment

OA (GIBCO, Gaithersburg, USA) was diluted (1:10) in DMSOand stored at –20°C. The stock solutions were further diluted withsterile bidistilled water (1:100) before treatments. OA was appliedto the colonies in different conditions: (1) Randomly selected cellcultures were treated with a medium containing 5, 10 or 20 nMOA for 4 h on the 2nd day. OA-application was repeated on the3rd day using the same protocol. (2) In some other experiments,50 and 100 nM OA were also applied for 4 h on the same days.After removal of the OA-containing medium, treated colonieswere fed with Ham’s F12 medium supplemented with 10% FCS.(3) Other colonies were fed continuously until the 6th day with 2,4, 6, 8, 10 or 20 nM OA containing culturing media starting on thefirst or second day. Some colonies were treated with DMSO inidentical dilutions to that of the OA-treated colonies. DMSO hadno effect on colonies.

Image analysis and light microscopical morphology

HD-cultures treated with 5, 10 or 20 nM OA on days 2 and 3, aswell as untreated controls were fixed in a 4:1 mixture of absoluteethanol and 40% formaldehyde for 30 min at day 6 of culturing.Cultures were stained with 0.1% dimethylmethylene blue(DMMB) (Aldrich, Germany) dissolved in 3% acetic acid for 5 min, washed in 3% acetic acid and mounted in gum arabic(Módis 1991). Due to its high proteoglycan content, the cartilagematrix appeared in purple metachromatic colour after DMMB-staining at low pH. Some cultures were stained with 0.1% aqueousDMMB for 5 min. The cartilaginous areas were measured by com-puter-assisted image analysis using a Hitachi HV-C20 video cam-era connected to a Leitz Diaplane microscope and IMAN 1.4 soft-ware (KFKI, Budapest, Hungary). The size of the metachromaticcartilage matrix was recorded from 6.8×106 µm2 rectangular areasof the central part of 25 different colonies of each experimentalgroup. Data were statistically analysed with an ambiguity probe(Bartels 1979; Módis 1991).

Labelling of actin microfilaments with TRITC-phalloidin

Colonies were washed twice in PBS, fixed in pre-cooled acetoneat –20°C for 20 min and repeatedly washed in PBS at room tem-perature. For labelling actin, 0.2 µg/ml phalloidin-TRITC (Sigma)in PBS was used for 1 h at room temperature in a wet-chamber.After repeated washing in PBS, colonies were mounted withMowiol 40-88 (Sigma-Aldrich, Steinheim, Germany). Labelledcells were examined with a Zeiss LSM 400 laser scanning confo-cal microscope system (Zeiss, Oberkochen, Germany) equippedwith a ×40 1.3 NA plan neofluar objective (Zeiss); 543 nm ofHeNe laser was used for the TRITC excitation, and the emissionfilter was LP 570 nm. The series of images were processed withthe LSM 3.8 base program and software package.

Electron microscopy

To investigate the effect of OA on cellular morphology, 6-day-oldcolonies were repeatedly washed in 0.02 M PBS and fixed over-night at 4°C in 2.5% glutaraldehyde (Merck, Darmstadt, Germa-ny) dissolved in the same buffer.

Proteoglycan (PG) staining was done as follows. Cell cultureswere rinsed in PBS, prefixed in 2.5% glutaraldehyde for 10 min atroom temperature. Colonies were fixed and stained in 0.05 M so-dium-acetate buffer pH 5.8 containing 2.5% glutaraldehyde, 0.5%cuprolinic blue (CB; BDH Chemicals, Poole, England) and 0.3 MMgCl2 overnight at 4°C (modified from van Kuppevelt et al.1985). The specimens were rinsed in the same buffer for 3×15 minand then transferred into 0.5% Na-tungstate in distilled water for 1 h. Colonies were removed from the coverslips by a razor bladeand the cartilage-containing central part of the colonies was dis-sected into approximately 1 mm×2 mm pieces, dehydrated andembedded in Durcupan ACM (Fluka, Buchs, Switzerland). Ultra-thin sections were collected an slot grids. Sections without CB-staining were counterstained with lead citrate and uranyl acetate.

Preparation of proteoglycan and collagen samples

After 8 days of culturing, 20 nM OA-treated and non-treated con-trol colonies were mechanically removed from the coverslips, ho-mogenised in liquid nitrogen and the soluble proteins extracted in4 M guanidine-hydrochloride (Reanal, Budapest, Hungary) in thepresence of 10 mM EDTA, 2 mM phenylinethylsulphonyl fluoride(PMSF), 2 mM iodoacetamide, 5 µg/ml pepstatin A, 5 µg/ml soy-bean trypsin inhibitor and 100 mM ε-aminocaproic acid at 4°C(Glant et al. 1986). After 48 h of extraction the samples were cen-trifuged, the supernatants were removed and extensively dialysedagainst distilled water and finally freeze-dried.

To melt the collagens the remaining pellets were washed with0.5 M acetic acid and digested with pepsin (Sigma) for 48 h. The

24

supernatants were further purified an a DEAE-Sepharose CL-6Bcolumn (Pharmacia, Uppsala, Sweden) for further purification. Af-ter dialysis against 0.1 M acetic acid the supernatants were freeze-dried, and protein as well as uronic acid contents were measuredin each sample.

Agarose-polyacrylamide composite gels, SDS polyacrylamide gels,Western blotting and immunostaining

Cartilage extracts were analysed on large-pore-size agarose-poly-acrylamide composite slab gels (Cs-Szabó et al. 1995). The gelswere stained with toluidine blue in order to visualise the proteo-glycans. The pepsin-digested samples were separated on 10%polyacrylamide minigels under reducing conditions (Laemmli1970). The proteins were electrophoretically transferred to a nitro-cellulose membrane (Millipore, Bedford, USA) in TRIS/glycinebuffer pH 8.6 at 80 V for 1 h. The nitrocellulose membranes wereblocked with 5% BSA in 20 mM TRIS, 137 mM NaCl, pH 7.6 at room temperature for 2 h. Polyclonal antibodies raised against chicken-collagen types I and II, respectively (Biodesign,Kennebunk, USA) were added in 1:100 dilution overnight. Thetwo antibodies were highly specific for the given type of collagenwithout any cross reactivity.

The membranes were extensively washed with TWB-TWEENand labelled with horseradish peroxidase conjugated anti-rabbit Ig(DAKO, Glostrup, Denmark) at room temperature for 2 h. Thecollagen specific bands were visualised with 3,3’-diaminobensi-dine (Sigma) chromogen reaction.

The reason for demonstration of the presence of type I colla-gen in the matrix of colonies is that the cultures still contain (usu-ally on the periphery) mesenchymal cells/fibroblasts, and thesecells are producing type I collagen. Thus, we wanted to demon-strate that not only the chondrocytes (synthesizing type II colla-gen), but also the mesenchymal cells/fibroblasts produce normalmatrix molecules – in this case collagen type I – and their proteinproduction is not altered by the okadaic acid treatment.

Preparation of cell extracts and measuring protein phosphatase activity

On days 2 and 3, immediately after the OA-treatments, as well as ondays 4 and 6, colonies were removed from coverslips, washed twicein 10 vol of physiological NaCl solution to remove traces of OA-containing medium and sedimentated with centrifugation (1200 g, at4°C for 5 min). Cell pellets (10 colonies per tube) were resuspendedin 100 µl of homogenization buffer containing 50 mM TRIS-HCl(pH 7.5), 1 mM EGTA, 1 mM PMSF, 1 mM benzamidine and 50 mM β-mercaptoethanol. After storing them at –70°C, suspen-sions were sonicated by pulsing burst for 2 min by 60 cycle (Branson Sonifier, Danbury, USA). After centrifugation at 10,000 gfor 15 min at 4°C, the soluble fraction was used for a phosphataseassay performed with 32P-phosphorylase (10 µM). One unit of pro-tein phosphatase activity releases 1 µmol of Pi from the phospho-substrate per min. The details of phosphatase and protein assayswere the same as previously described by Murányi et al. (1998).

Investigation of cell proliferation with 3H-thymidine-labelling

Ten µl droplets of cell suspension were inoculated into wells of 96-well microtiter plates. On day 3, following the second 20 nM OA-treatment, medium containing 1 µCi/ml 3H-thymidine(diluted from methyl-3H-thymidine solution, 37 Mbq, 1 mCi/ml at185 GBq, Amersham, Little Chalfont, UK) was added to each wellfor 16 h. Wells were washed twice with PBS and the proteins wereprecipitated with ice-cold 5% trichloroacetic acid for 20 min. Af-ter repeated washing with PBS, cells were digested with 0.25%trypsin (1:250, Sigma)-0.15% EDTA in calcium/magnesium freephosphate-buffered saline (CMF-PBS) at 37°C for 10 min andharvested with a semiautomatic cell-harvester (Skatron, Lier, Nor-

way). Harvested colonies were dried on scintillation filters and ra-dioactivity was counted by a beta-counter (Pharmacia, Uppsala,Sweden). The same procedures were carried out on days 4 and 6.Thymidine-incorporation of OA-treated colonies was compared tothat of untreated ones of the same age. Data were statisticallyanalysed with the F-test.

Investigation of cell proliferation with 5-bromodeoxyuridine incorporation

The proliferation of cells was also investigated using 5-bromode-oxyuridine (BrdUrd) incorporation assay according to Balázs et al.(1991). Randomly selected 3-, 4- and 6-day-old 20 nM OA-treatedand untreated control cultures were incubated with 20 µM BrdUrd(Sigma) for 1 h at 37°C, then colonies were washed with CMF-PBS. To obtain single cell suspension cell cultures were di-gested with 0.25% trypsin, 0.15% EDTA for 10 min at 37°C. Fol-lowing the trypsin treatment cells were further digested with5×106 U/mg pronase (Calbiochem, La Jolla, USA) for 55 min at37°C. Cells were suspended, washed with CMF-PBS, centrifuged,resuspended and fixed in 70% ethanol (diluted in CMF-PBS).Fixed cells were placed onto microscopic slides using Cytospin-3centrifuge (Shandon, England), and airdried. Cells were denaturedin 2.5 M HCl/0.5% Triton-X 100 for 1 h, then washed in PBS(2×5 min). Samples were pre-blocked with 1.5% Carnation drymilk/0.1% Triton-X 100/PBS for 10 min and then incubated withmonoclonal anti-BrdUrd antibody (Boehringer Mannheim, India-napolis, USA; diluted 1:50) for 1 h at room temperature. Afterwashing twice with PBS for 10 min, cells were incubated with an-ti-mouse-FITC (Sigma; diluted 1:500) for 30 min at room temper-ature, washed with PBS, dried and stained with propidium-iodide(Vector Laboratories, Burlingame, USA).

Results

Morphology

It is well known, that the majority of the cells growing inHD-cultures are chondrogenic, but some myoid and epi-thelial cells are also present. Metachromatic cartilagematrix can be observed in the colonies from the end ofday 3. Numerous elongated mesenchymal cells are pres-ent at the periphery of the colonies.

Our results obtained by measurement of metachro-matic cartilage areas demonstrated that the cartilage for-mation was stimulated by OA in a dose-dependent man-ner in HD colonies. Repeated treatments on days 2 and 3with 5 nM OA resulted in a 20% but not significant in-crease, while 10 nM and 20 nM OA treatments resultedin 34% and 115% significant increases of the metachro-matic cartilage areas, respectively (Fig. 1a–d; Table 1).Due to the cytotoxic effect of OA, a degeneration of col-onies was attained in the case of application of higherconcentrations of OA (50 or 100 nM) at any stage of culturing. Lower concentrations of OA (2, 4, 6, 8 and 10 nM) started an day 2 and applied until the end of cul-turing, resulted in a moderate, dose-dependent increaseof cartilage formation but colonies detached during thestaining procedure (data are not shown). Two nM OAdid not have any visible influence on cultures.

The age of the culture selected for OA-treatment wascritical. All cells were lost if the treatment was carriedout at the time of inoculation of the cells. When the cul-

25

tures were treated on day 1, the majority of the cells de-tached. Only a very moderate increase of cartilage areawas detected when the first OA-treatment was applied onor after the fourth culture day (data are not shown).

The microfilament structure of elongated mesenchy-mal-chondrogenic cells located at the periphery of colo-nies deteriorated following 20 nM OA treatment and ac-

26

Fig. 1a–d Microphotographs from 6-day-old HD cultures afterstaining with DMMB at pH 3. Colonies shown in b and d weretreated with 20 nM OA for 4 h on the 2nd and 3rd days. The size

of the metachromatic cartilaginous areas is considerably increasedin 20 nM OA-treated colonies (b,d) as compared to the untreatedcontrols (a,c). Bars 100 µm

Fig. 2 Laser confocal microscopic pictures of TRITC-phalloidinlabelled chondrogenic cells from untreated (a) and 20 nM OA-treated (b) 3-day-old colonies. OA caused disorganisation of actinmicrofilaments and condensation of phalloidin-labelled materialunder the cell membrane. Numerous round cells were visible fol-lowing OA treatments (b). Bars 25 µm

27

Fig. 3 Electron micrographstaken from untreated (a) and 20 nM OA-treated (b) 6-day-old colonies after contrastingthe sections with uranyl acetateand lead citrate. Chondrocytesshow high similarity in cellularmorphology. Both the untreatedand OA-treated cells have en-larged endoplasmic reticulumcysternae (*) demonstratinghigh biosynthetic activity. Notethe delicate filamentous colla-gen network in ECM surround-ing the cells (arrowheads). Fol-lowing en bloc staining withCB size, shape and distributionof CB-stained proteoglycans(arrowheads) are almost identi-cal in ECM of either untreated(c) or OA-treated (d) cartilagenodules. Distribution of proteo-glycan particles in cartilagematrix including pericellularmatrix is demonstrated in un-treated (e) and OA-treated col-onies (f). Bars 1 µm (a,b,e,f);500 nm (c,d)

Table 1 Average areas of the metachromatic cartilage as recordedfrom 6.8×106 µm2 segments of the 6-day-old control and OA-treated micromass colonies by computer-assisted image analysisafter staining with DMMB at pH 3. For statistical analysis, theambiguity probe was used to compare data recorded from control

and OA-treated groups shown in histograms. Using this probe,ambiguity values (A) are calculated which may range between 1.0(no difference) and 0 (total separation). If A< 0.7, the differencebetween the sets of data is significant. Cartilaginous areas are in-creased in OA-treated cultures in a dose-dependent manner

Experimental groups Average area SD Relative average area(n=25 in each group) (µm2) (%)

Control 1,664,927 382,933 (23%) 1005 nM OA 2,006,462a 280,904 (14%) 12010 nM OA 2,239,856b 425,572 (19%) 13420 nM OA 3,579,721c 680,146 (19%) 215

a A=0.790 (minor difference) b A=0.624 (considerable difference) c A=0.08 (total separation)

tin condensed under the cell membrane (Fig. 2a,b) as itwas detected by confocal microscopy on day 3, afterTRITC-phalloidin labelling. Some cells remained elon-gated after OA-treatment and preserved their organisedmicrofilament structure, but the number of these cellssubstantially decreased (Fig. 2b). Disorganisation of ac-tin was reversible; when OA was removed, the normalmicrofilament structure recovered in cells.

As revealed by electronmicroscopy, ultrastructure ofOA-treated chondrocytes remained almost identical tothat of untreated control ones, but enlarged rough endo-plasmic reticulum cisternae were visible in a highernumber inside the treated cells (Fig. 3a,b).

Ultrastructural and biochemical analysis of cartilage ECM

In electron microscopic observations a delicate networkof thin collagen filaments with similar density was seenin the ECM of both 20 nM OA-treated and untreatedcontrol cartilage nodules (Fig. 3a,b). As revealed by CB-staining, the size and distribution of PG molecules in theECM of cartilage nodules were identical in OA-treatedand control groups (Fig. 3c,d). The distribution of pro-teoglycan particles in cartilage matrix, including thepericellular area, is identical in untreated and OA-treatedcolonies, as shown in Fig. 3e,f.

28

Fig. 3c,d

The wet masses of the samples treated with 20 nMOA were significantly higher than that of non-treatedcontrol ones. The uronic acid and protein contents of aunit (1 mg) freeze-dried material were approximately thesame in colonies of the above mentioned two experimen-tal groups (data are not shown). On the basis of the com-posite gel patterns (Fig. 4a), the size and the ratio of thelarge and small proteoglycans were identical in the OA-treated and non-treated samples. The same major PGcomponent with a molecular mass characteristic for ag-grecan was isolated from both OA-treated and untreatedcolonies. Collagen patterns (Fig. 4b) were identical inthe OA-treated and non-treated samples. In the pepsin-

solubilised samples only a single band with 100 kDa mo-lecular mass representing the α chains of collagen type Iand II was detected.

Protein phosphatase activities in the extracts of micromass cultures

Phosphorylase phosphatase activity measured in thepresence of different concentrations of OA showed a sig-nificant decrease up to 2 nM concentration of OA repre-senting the inhibition of PP2A. Higher concentrations ofOA caused the total inhibition of PP1 (Fig. 5). PP2A and

29

Fig. 3e,f

PP1 were found to be the only enzymes in the superna-tants of the homogenised micromass cell cultures as-sayed with phosphorylase substrate, since 10 µM of OAsuppressed totally the enzymatic activity.

PP2A-activity of the colonies can be measured by theselective inhibition of PP1 by inhibitor-2. OA decreasedthe activity of PP2A in a dose-dependent manner. 20 nMof OA significantly reduced the PP2A-activity when it

was applied on the 2nd or the 3rd days of culturing (Fig. 6).

Phosphatase activities, measured in the supernatants ofthe homogenised control colonies demonstrated that PP1-activity remained almost unchanged, while the PP2A-activity decreased parallel to the progress of the differenti-ation of chondrogenic cells (Fig. 7A). The greatest reduc-tion of PP2A activity was found between the 2nd and the

30

Fig. 4 Collagen (a) and PG (b)content of 8-day-old cultureswith or without OA-treatment.Collagens were detected by im-munoblotting, and PGs werestained with toluidine blue.Pepsin solubilised samplesshow type I and type II colla-gen production of HD colonies(a). Single bands detected byimmunostaining represent the αchain of the collagen moleculeswith approximately 100 kDamolecular mass. The 4 M gu-anidinium-hydrochloride ex-tracts were loaded onto a poly-acrylamide-agarose compositegel for analysis of PG-content(b). The two lanes show a verysimilar pattern. The upper bandrepresents the hyaluronan-aggrecan aggregate, while thelower band is the aggrecanmonomer

3rd culture-day, at the time of differentiation of the chon-drocytes in micromass cultures (Solursh et al. 1978). Avery significant suppression of PP2A-activity without no-table change in activity of PP1 was detected when colo-nies were treated with 20 nM OA (Fig. 7B).

Cell proliferation assays

Although identical experiments led to similar results, av-eraging data would produce high standard errors. Conse-

quently, we selected the most representative data to pres-ent the proliferation pattern (Fig. 8). We measured in-creased 3H-thymidine-incorporation of OA-treated colo-nies, compared to untreated ones at any time during cul-turing. The highest difference was found on day 3, afterthe second OA-treatment (Fig. 8). A similar effect of OAwas observed when BrdUrd-incorporation of cells wasinvestigated (Table 2).

31

Fig. 5 Inhibition of phosphorylase phosphatase activity by ok-adaic acid. Phosphorylase phosphatase activities were measured inthe supernatants of 1-day-old (● ), 2-day-old (■ ) and 6-day-old(▲) untreated control cultures. Phosphatase activities measured inthe absence of okadaic acid were taken as 100%

Fig. 6 Protein phosphatase 2A (PP2A) activity upon okadaic acidtreatment. Colonies after treatment with 5, 10, 20 nM OA on the2nd and after repeated treatment on the 3rd days of cultures, aswell as the 2- and 3-day-old untreated control colonies (columnslabeled by 0) were assayed in the presence of 100 nM inhibitor-2for PP2A activity. Total phosphatase activity of 2-day-old untreat-ed control colonies was taken as 100% (*P<0.05)

Fig. 7 Phosphorylase phosphatase activity upon aging of untreat-ed (A) and 20 nM OA-treated (B) chondrifying micromass cul-tures. Protein phosphatase activities were measured from the su-pernatants of disrupted cells as given in Materials and methods.PP1 activity was assayed in the presence of 2 nM OA (open columns), PP2A activity was measured in the presence of 100 nMinhibitor-2 (closed columns). In the untreated cultures (A) PP1 ac-tivities were rather similar (P>0.65), while PP2A activitiesshowed significant decreases by Student’s t-test. Following treat-ments with 20 nM OA, unaffected PP1 and suppressed PP2A ac-tivities did not show further significant changes (P >0.65) mea-sured in colonies in different stages of culturing (B)

Discussion

Measurement of the activities of phosphatases in the su-pernatants of homogenates of HD cultures demonstratedthat parallel to differentiation of chondrogenic cells theactivity of PP2A significantly decreased, while that ofPP1 remained almost unchanged. In our experiments weinhibited PP2A with 20 nM OA on days 2 and 3, whichis a crucial time period in the differentiation of chondro-genic cells (Solursh et al. 1978). As a marker of en-hanced chondrogenesis, the metachromatic areas of car-tilage matrix became twice as large in 6-day-old 20 nMOA-treated colonies compared to the untreated controls.OA is frequently referred as a tumour-promoting agent(Cohen et al. 1990); however, normal morphology ofcells and extracellular matrix (ECM) in the cartilage up-on OA-treatment were found in our experiments. Thecomposition of proteoglycans and collagens in the ECMof treated colonies were also identical compared to thematrix of untreated colonies.

Following OA treatments, we found enhanced cellproliferation detected by the incorporation of either 3H-thymidine or BrdUrd. The highest increase of cell prolif-eration was observed on day 3, which is the day of thebeginning of cartilage differentiation in HD cultures.Proliferation and condensation of chondrogenic cells are

essential steps in the formation of cartilage (George et al.1983; Hall and Miyake 1995). Commitment of chondro-genic cells happens on days 1 and 2 in HD cultures, par-allel to rapid proliferation and condensation of chondro-genic elements into small nodules (Solursh et al. 1978).If cell density is proper, direct cell-cell contacts andparacrine signals (e.g. prostaglandins) enhance the com-mitment of the chondrogenic cells. Committed cells dif-ferentiate into chondroblasts and in a further step intochondrocytes most probably on day 3 of culturing. Stim-ulation of cell proliferation can promote cartilage forma-tion via establishment of a proper cell density for suc-cessful paracrine and direct cell-cell interactions ofchondrogenic cells, resulting in stronger commitment ofthem. Moreover, OA was also reported to arrest cells inmitosis in a reversible manner (Zheng et al. 1991; Linand Arndt 1995) and it is known that chondrogenic cellsfollowing short and reversible inhibition of mitosis be-came committed to cartilage cells more effectively(Solursh and Reiter 1975). Taken together, OA influenc-es the cell cycle in two opposite but not incompatibleways: stimulation of proliferation and arresting in mito-sis. Both effects may act in the same direction, enhanc-ing commitment of chondrogenic cells.

We found that OA caused disorganisation of actin fi-laments and condensation of actin under the cell mem-brane in chondrogenic cells of HD cultures. It had beenclearly demonstrated that disruption of the actin micro-filament structure increases cartilage formation (Zanettiand Solursh 1984; Daniels and Solursh 1991). It is alsoknown that inhibition of PPs by OA modifies the struc-ture of different cytoskeletal components, probably de-pending on cell type. The architecture of intermediate fi-laments of hepatocytes (Blankson et al. 1995) and neu-rons (Giasson et al. 1996), microtubules and actin micro-filaments of lung tumour-cells (Maier et al. 1995), actinorganization of fungi (Lin and Arndt 1995), green alga(Menzel et al. 1995) or platelets (Erdődi et al. 1995) can all be modified by OA-induced inhibition of PP-activities. Disruption of actin microfilaments influences

32

Table 2 Average BrdUrd-labelling indexes (BLI) calculated fromdata counted in control and 20 nM OA-treated HD-colonies fol-lowing incorporation of 20 µM BrdUrd for 1 h. Note the consider-able increase of BLI on days 3 and 4. Data are averages of threedifferent experiments. Each value was calculated as the average of10 measurements

BLI (%)

Control OA-treated

Day 3 7.2 15.0Day 4 4 6.6Day 6 2.8 2.9

Fig. 8 Effect of OA on the 3H-thymidine-incorporation ofchondrifying micromass cul-tures. Open columns representthe radioactivity of coloniestreated with 20 nM OA, whileclosed columns show the radio-activity of control colonies. Ra-dioactivities were counted ondays 2, 3 and 6. OA-treatedcolonies showed significant elevation of DNA-synthesis ondays 3 and 6 according to theF-probe (* P< 0.05)

cartilage formation presumably in various ways. Whenintegrins, anchoring the cells to fibronectin were down-regulated, chondrogenic cells became round (Danielsand Solursh 1991). Either ligand binding of ECM recep-tors or loss of this binding may cause modulation of fo-cal adhesion sites, changing the organisation of actin mi-crofilaments and influencing gene expression (Aplin andJuliano 1999). It has been also reported that OA treat-ment modifies the expression of some proteins that regu-late gene expression, like c-fos or jun family (Schonthalet al. 1991a,b) including AP-1 transcription factor com-plex (Miller et al. 1998). These proteins are involved in the regulation of chondrogenesis (Wang et al. 1992;Beier et al. 1999).

The level of the phosphorylation of different targetproteins is determined by the balance of respective pro-tein kinases and phosphatases. The inhibition of PP2Aor/and PP1 by OA shifts the balance of reversible cellu-lar protein phosphorylation toward increased phosphory-lation of target proteins (Schönthal 1998). Some Ser/Thrspecific protein kinases such as PKA (Smales and Biddulph 1985; Zhang et al. 1996), PKC (Chang et al.1998) and receptor kinases of TGF-β-superfamily (Kulyk et al. 1989; Leonard et al. 1991) have been de-scribed as modulators of chondrogenesis. Signal trans-duction processes of PKA (Gonzalez and Montminy1989), PKC (Xie and Rothstein 1995) and TGF-β path-ways (Potchinsky et al. 1997) can induce the phosphory-lation of cyclic AMP response element binding protein(CREB) on Ser-133. The phosphorylation of CREB is in-volved in both bone morphogenetic protein-2 (a memberof TGF-β-superfamily) and cAMP-induced chondrogen-esis (Lee and Chuong 1997). On the other hand, PP1 andPP2A can dephosphorylate CREB (Hagiwara et al. 1992;Wadzinski et al. 1993; Alberts et al. 1994; Wheat et al.1994), and PP2A is also involved in dephosphorylationof the catalytic subunit of PKA itself (Liauw and Steinberg 1996). Thus, either the inhibition of CREB-dephosphorylating effect of PP2A by OA or the pro-longed activity of PKA, if PP2A fails to dephosphorylateits catalytic subunit, may result in the prolonged phos-phorylation of CREB, which thereby can induce chon-drogenesis more effectively.

Furthermore, PKA is involved in chondrogenesis dueto the phosphorylation of a nuclear basic 35 kDa proteindesignated as p35 (Leonard and Newman 1987). The de-phosphorylation of this protein may be blocked by OAtreatment. Since the PKA II holoenzyme can be bound tocytoskeletal elements via its regulatory subunit (Zhang et al. 1996), the alteration of the actin microfilaments by OA can modify the cAMP signalling to the nucleus(Feliciello et al. 1997) further facilitating cartilage dif-ferentiation. According to our preliminary experimentsthe OA-stimulated chondrogenesis was prevented by theselective inhibition of PKA with the cell permeable H89during cell culturing.

Acknowledgements This work was supported by grants OTKAT022621 and OTKA T026541 from the Hungarian Science Re-search Fund, and MKM FKFP 0784/1977 from the HungarianMinistry of Education. We thank Mrs. Júlia Bárány, Mrs. Magdolna Szanitter, Ms. Erzsébet Bak, Mrs. Ilona Rónai and Mrs.Ágnes Miklós-Durkó for expert technical assistance. The JEOL1010 TEM that we used for electron microscopy had been provid-ed by the Japanese International Cooperation Agency (JICA) forthe Hungarian-Japanese Electron Microscopy Center of our Uni-versity.

References

Ahrens PB, Solursh M, Reiter RS (1977) Stage-related capacityfor limb chondrogenesis in cell culture. Dev Biol 60:69–82

Alberts AS, Montminy M, Shenolikar S, Feramisco JR (1994) Ex-pression of a peptide inhibitor of protein phosphatase 1 in-creases phosphorylation and activity of CREB in NIH 3T3 fi-broblasts. Mol Cell Biol 14:4398–4407

Aplin AE, Juliano RL (1999) Integrin and cytoskeletal regulationof growth factor signaling to the MAP kinase pathway. J CellSci 112:695–706

Balázs M, Mayall BH, Waldman FM (1991) Simultaneous analy-sis of chromosomal aneusomy and 5-bromodeoxyuridine in-corporation in MCF-7 breast tumor cells. Cancer Genet Cyto-genet 57:93–102

Barford D, Das AK, Egloff M-P (1998) The structure and mecha-nism of protein phosphatases. Annu Rev Biophys BiomolStruct 27:133–164

Bartels PH (1979) Numerical evaluation of cytologic data. III. Se-lection of features for discrimination. Anal Quant Cytol1:153–159

Beier F, Leask TA, Haque S, Chow C, Taylor AC, Lee RJ, PestellRG, Ballock RT, LuValle P (1999) Cell cycle genes in chon-drocyte proliferation and differentiation. Matrix Biol 18:109–120

Biajolan C, Takai A (1988) Inhibitory effect of a marine-spongetoxin, okadaic acid, on protein phosphatases. Specificity andkinetics. Biochem J 256:283–290

Blankson H, Holen I, Seglen PO (1995) Disruption of the cytoker-atin cytoskeleton and inhibition of hepatocytic autophagy byokadaic acid. Exp Cell Res 218:522–530

Brautigan DL (1997) Phosphatases as partners in signalling net-works. Adv Second Messenger Phosphorylation Res 31:113–124

Chang S-H, Oh Ch-D, Yang M-S, Kang S-S, Lee Y-S, Sonn J-K,Chun J-S (1998) Protein kinase C regulates chondrogenesis ofmesenchymes via mitogen-activated protein kinase signalling.J Biol Chem 273:19213–19219

Cohen P, Klump S, Schelling DL (1989) An improved procedurefor identifying and quantitating protein phosphatases in mam-malian tissues. FEBS Lett 250:596–600

Cohen P, Holmes CFB, Tsukitani Y (1990) Okadaic acid: a newprobe for the study of cellular regulation. Trends Biochem Sci15:98–102

Cs-Szabó G, Roughley PJ, Plaas AHK, Glant TT (1995) Large andsmall proteoglycans of osteoarthritic and rheumatoid articularcartilage. Arthritis Rheum 38:660–668

Daniels K, Solursh M (1991) Modulation of chondrogenesis bythe cytoskeleton and extracellular matrix. J Cell Sci 100:249–254

Erdo′′di F, Tóth B, Hirano M, Hartshorne DJ, Gergely P (1995) En-dothall thioanhydride inhibits protein phosphatases-1 and -2Ain vivo. Am J Physiol 269:1176–1184

Feliciello A, Li Y, Avvedimento EV, Gottesman ME, Rubin CS(1997) A-kinase anchor protein 75 increases the rate and mag-nitude of cAMP signaling to the nucleus. Curr Biol 7:1011–1014

George M, Chepenik KP, Schneiderman MH (1983) Proliferationof cells undergoing chondrogenesis in vitro. Differentiation24:245–249

33

Giasson BI, Cromlish JA, Athlan ES, Mushynski WE (1996) Acti-vation of cyclic AMP-dependent protein kinase in okadaic acid-treated neurons potentiates neurofilament fragmentationand stimulates phosphorylation of Ser2 in the low-molecular-mass neurofilament subunit. J Neurochem 66:1207–1213

Glant TT, Mikecz K, Poole AR (1986) Monoclonal antibodies todifferent protein-related epitopes of human articular cartilageproteoglycans. Biochem J 234:31–41

Gonzalez GA, Montminy MR (1989) Cyclic AMP stimulates so-matostatin gene transcription by phosphorylation of CREB atSer-133. Cell 59:675–680

Hadházy Cs, László MB, Kostenszky KS (1982) Cartilage differ-entiation in micro-mass cultures of chicken limb buds. ActaMorphol Acad Sci Hung 30:65–78

Hagiwara M, Alberts A, Brindle P, Meinkoth J, Feramisco J, DengT, Karin M, Shenolikar S, Montminy M (1992) Transcriptionalattenuation following cAMP induction requires PP-1 mediateddephosphorylation of CREB. Cell 70:105–133

Hall BK, Miyake T (1995) Divide, accumulate, differentiate: cellcondensation in skeletal development revisited. Int J Dev Biol39:881–893

Haystead TAJ, Sim ATR, Carling D, Honner RC, Tsukitani Y, Co-hen P, Hardie DG (1989) Effects of the tumour promoter ok-adaic acid on intracellular protein phosphorylation and metab-olism. Nature 337:78–81

Hunter T (1995) Protein kinases and phosphatases: The yin andyang of protein phosphorylation and signaling. Cell 80:225–236

Kulyk WM, Rodgers BJ, Greer K, Kosher RA (1989) Promotionof embryonic chick limb bud cartilage differentiation by trans-forming growth factor-β. Dev Biol 135:424–430

Kuppevelt THMS van, Cremers FPM, Domen JGW, BeuningenHM van, Brule AJC van den, Kuyper CMA (1985) Ultrastruc-tural localization and characterization of proteoglycans in hu-man lung alveoli. Eur J Cell Biol 36:74–80

Laemmli UK (1970) Cleavage of stuctural proteins during the as-sembly of the head of bacteriophage T4. Nature 227:680–685

Lee YS, Chuong CM (1997) Activation of protein kinase A is apivotal step involved in both BMP-2- and cyclic AMP-inducedchondrogenesis. J Cell Physiol 170:153–165

Leonard CM, Newman SA (1987) Nuclear events during earlychondrogenesis: phosphorylation of the precartilage 35.5 kDdomain-specific chromatin protein and its regulation by cyclicAMP. Dev Biol 120:92–100

Leonard C, Fuld HM, Frenz DA, Downie SA, Massague J, Newman SA (1991) Role of transforming growth factor β inchondrogenic pattern formation in the embryonic chick limb.Dev Biol 145:99–109

Liauw S, Steinberg RA (1996) Dephosphorylation of catalyticsubunit of cAMP-dependent protein-kinase at Thr-197 by acellular protein phosphatase and by purified protein phospha-tase-2A. J Biol Chem 271:258–263

Lin FC, Arndt KT (1995) The role of Saccharomyces cerevisaetype 2A phosphatase in the actin cytoskeleton and in entry intomitosis. EMBO J 14:2745–2759

Maier GD, Wright MV, Lozano Y, Djordjevic A, Matthews JP,Young MR (1995) Regulation of cytoskeletal organization intumor cells by protein phosphatases-1 and -2A. Int J Cancer61:54–61

Menzel D, Vugrek O, Frank S, Elsner-Menzel C (1995) Proteinphosphatase 2A, a potential regulator of actin dynamics andactin-based organelle motility in the green alga Acetabularia.Eur J Cell Biol 67:179–187

Miller C, Zhang M, He Y, Zhao J, Pelletier J-P, Martel-Pelletier J,Di Battista JA (1998) Transcriptional induction of cylooxyge-

nase-2 gene by okadaic acid inhibition of phosphatase activityin human chondrocytes: co-stimulation of AP-1 and CRE nu-clear binding proteins. J Cell Biochem 69:392–413

Módis L (1991) Organization of the extracellular matrix. A polar-ization microscopic approach. CRC, Boca Raton

Murányi A, Erdődi F, Ito M, Gergely P, Hartshorne DJ (1998)Identification and localization of myosin phosphatase in hu-man platelets. Biochem J 330:225–231

Potchinsky MB, Weston WM, Lloyd MR, Greene RM (1997)TGF-beta signaling in murine embryonic palate cells involvesphosphorylation of the CREB transcription factor. Exp CellRes 231:96–103

Schonthal A, Tsukitani Y, Feramisco JR (1991a) Transcriptionaland post-translational regulation of c-fos expression by the tu-mor promoter okadaic acid. Oncogene 6:423–430

Schonthal A, Alberts AS, Frost JA, Feramisco JR (1991b) Differ-ential regulation of jun family gene expression by the tumorpromoter okadaic acid. New Biol 3:977–986

Schönthal AH (1998) Role of PP2A in intracellular signal trans-duction pathways. Front Biosci 3: d1262–1273

Smales WP, Biddulph DM (1985) Limb development in chick em-bryos: cyclic AMP-dependent protein kinase activity, cyclicAMP, and prostaglandin concentrations during cytodifferentia-tion and morphogenesis. J Cell Physiol 122:259–265

Solursh M, Reiter R (1975) Determination of limb bud chondro-cytes during a transient block of the cell cycle. Cell Differ4:131–137

Solursh M, Ahrens PB, Reiter RS (1978) A tissue culture analysisof the steps in limb chondrogenesis. In Vitro Cell Dev BiolAnim 14:51–61

Sonn JK, Solursh M (1993) Activity of protein kinase C duringthe differentiation of chick limb bud mesenchymal cells. Dif-ferentiation 53:155–162

Wadzinski BE, Wheat WH, Jaspers S, Peruski LF Jr, Lickteig RL,Johnson GL, Klemm DJ (1993) Nuclear protein phosphatase2A dephosphorylates protein kinase A-phosphorylated CREBand regulates CREB transcriptional stimulation. Mol Cell Biol13:2822–2834

Wang Z-Q, Ovitt C, Grigoriadis AE, Moehle-Steinlein U, RuetherU, Wagner EF (1992) Bone and haemopoietic defects in micelacking c-fos. Nature 360:741–745

Wera S, Hemmings BA (1995) Serine/threonine phosphatases.Biochem J 311:17–29

Wheat WH, Roesler WJ, Klemm DJ (1994) Simian virus 40 smalltumor antigen inhibits dephosphorylation of protein kinase A-phosphorylated CREB and regulates CREB transcriptionalstimulation. Mol Cell Biol 14:5881–5890

Xie H, Rothstein TL (1995) Protein kinase C mediates activationof nuclear cAMP response element binding protein (CREB) inB lymphocytes stimulated through surface Ig. J Immunol154:1717–1723

Zanetti NC, Solursh M (1984) Induction of chondrogenesis inlimb mesenchymal cultures by disruption of the actin cyto-skeleton. J Cell Biol 99:115–123

Zhang Q, Carr DW, Lerea KM, Scott JD, Newman SA (1996) Nu-clear localization of type II cAMP-dependent protein kinaseduring limb cartilage differentiation is associated with a noveldevelopmentally regulated A-kinase anchoring protein. DevBiol 76:51–61

Zheng B, Woo CF, Kuo JF (1991) Mitotic arrest and enhanced nu-clear protein phosphorylation in human leukemia K562 cellsby okadaic acid, a potent protein phosphatase inhibitor and tu-mor promoter. J Biol Chem 266:10031–10034

34