TGF-β3-Induced Chondroitin Sulphate Proteoglycan Mediates Palatal Shelf Adhesion

TGF- ␤ 3-Induced Chond oi in Sulpha e P o eoglycan Media es Pala al Shel Adhesion A. Ga o,*M. L. Ma inez,†C. Tudela,†I. Alonso,*J. A. Mo o,* M. A. Fo moso,†M. W. J. Fe guson,‡and C. Ma ı´nez-A ´l a ez†,1 *Depa amen o de Ana omı´a Humana, Facul ad de Medicina, Uni e sidad de Valladolid, Valladolid 47005 Spain; †Depa amen o de Ciencias Mo olo´gicas I, Facul ad de Medicina, Uni e sidad Complu ense de Mad id, Mad id 28040, Spain; and ‡Cells, Immunology and De elopmen Di ision, School o Biological Sciences, Uni e si y o Manches e , Uni ed Kingdom In mammals, he adhesion and usion o he pala al shel es a e essen ial mechanisms in he de elopmen o he seconda y pala e. Failu e o any o hese p ocesses leads o he o ma ion o cle pala e. The mechanisms unde lying pala al shel adhesion a e poo ly unde s ood, al hough he p esence o ilopodia on he apical su aces o he supe icial medial edge epi helial (MEE) cells seems o play an impo an ole in he adhesion o he opposing MEE. We demons a e he e he appea ance o chond oi in sulpha e p o eoglycan (CSPG) on he apical su ace o MEE cells only immedia ely p io o con ac be ween he pala al shel es. This apical CSPG has a unc ional ole in pala al shel adhesion, as ei he he al e a ion o CSPG syn hesis by ␤ -D-Xyloside o i s speci ic diges ion by chond oi inase AC s ikingly al e s he in i o adhesion o pala al shel es. We also demons a e he absence o his apical CSPG in he cle ed pala es o ans o ming g ow h ac o be a 3 (TGF- ␤ 3) null mu an mice, and i s induc ion, oge he wi h pala al shel adhesion, when TGF- ␤ 3is added o TGF- ␤ 3 nullmu an pala alshel esincul u e.Whenchickpala alshel es( ha dono adhe ein i ono exp essTGF- ␤ 3,no CSPG in he MEE) a e cul u ed in i o, hey do no exp ess CSPG and pa ially adhe e, bu when TGF- ␤ 3is added o he media, hey exp ess CSPG and hei adhesion inc eases s ikingly. We he e o e conclude ha he exp ession o CSPG on he apical su ace o MEE cells is a key ac o in pala al shel adhesion and ha his exp ession is egula ed by TGF- ␤ 3.© 2002 Else ie Science (USA) Key Wo ds: pala e usion; chond oi in sulpha e p o eoglycan; cle pala e; TGF- ␤ 3null mu an mouse; chick. INTRODUCTION The seconda y pala e in mammals o ms by he union o he wo pala al shel es ha a ise om he medial aspec s o he maxilla y p ocesses o he i s b anchial a ch. Ini ially, hey g ow down bo h sides o he ongue [emb yonic day 13 (E13) in mice] and hen ele a e (E14) and app oach each o he , and he epi helium ha co e s hei ips, he medial edgeepi helium(MEE),con ac sin hemidlinea E14.5and adhe es, o ming he midline epi helial seam (Fe guson, 1988; Shule e al.,1991). Th ough p og ammed cell dea h (Mo i e al., 1994; Taniguchi e al., 1995; Ma ı´nez-A ´l a ez e al., 2000a), epi helial–mesenchymal ans o ma ion (Fi che and Hay, 1989; G i i h and Hay, 1992; Ma ı´nez- A ´l a ez e al., 2000a), and mig a ion o he o al and nasal aspec s o he pala e (Ca e e and Fe guson, 1992), he midline epi helial cells disappea , allowing con inui y o he mesenchyme in he usion zone a E15. E en ually, he mesenchyme o he an e io wo- hi ds o he pala e unde - goes in amemb anous ossi ica ion and becomes he ha d pala e. Du ing pala e usion in mammals, he adhesion o he opposing MEE is a c i ical e en whose al e a ion causes cle pala e(NewallandEdwa ds,1981a,b;P a e al.,1984; Abbo and P a , 1987). Howe e , his p ocess does no ake place in all species. Unlike in mammals, a ian pala al shel es app oach and con ac , bu do no adhe e in i o (G eene e al., 1983), esul ing in a na u al cle . Some ep iles show an a ian-like pa e n o pala e de elopmen , al hough in alliga o s and c ocodiles, pala al shel es adhe e a e hei con ac and he usion p og esses un il acon inu- 1To whom co espondence should be add essed. Fax: 34-91-394- 1374. E-mail: [email p o ec ed]. De elopmen al Biology 250, 393–405 (2002) doi:10.1006/dbio.2002.0792 0012-1606/02 $35.00 © 2002 Else ie Science (USA) All igh s ese ed. 393 ous pala e is o med ( e iewed in Fe guson, 1988). The e is a species and issue speci ici y in his adhesion p ocess, as i ails, o ins ance, when mouse MEE is placed in con ac wi h alliga o MEE, ongue, o supe icial maxilla y epi he- lium (Fe guson e al., 1984). Tempo al speci ici y in pala al shel adhesion has also been epo ed, as p ecociously emo ed pala al shel es do no adhe e in cul u e (Pou ois, 1966; Humph ey, 1970; Smiley and Koch, 1971). The ini ial adhesion o pala al shel es seems o be co e- la ed wi h he appea ance o ce ain changes in he mos supe icial MEE cells ha occu jus p io o hei con ac . These cells bulge on he MEE su ace soon a e pala al shel ele a ion (Ma ı´nez-A ´l a ez e al., 2000b) and de elop mic o illi (DeAngelis and Nalbandian, 1968; Haywa d, 1969; Souchon, 1975), ilopodia and lamellipodia (Wa e - man e al., 1973; Melle e al., 1980; Schu¨pbach e al., 1983; Taya e al., 1999) on hei apical su ace, whose unc ion migh be o inc ease he adhesion a ea be ween opposing MEE(Ma ı´nez-A ´l a eze al., 2000a). Unlike he basal MEE cells, he supe icial MEE cells ha e an ex ensi e Golgi appa a us and nume ous cy oplasmic small coa ed esicles (DeAngelis and Nalbandian, 1968), all cha ac e is ics com- pa ible wi h apical sec e ion p ocesses. The p esence o hese ul as uc u al ea u es co ela es wi h he mos s iking change obse ed on he MEE su ace jus p io o he con ac o pala al shel es: he appea ance o a coa mos ly o med by glycoconjuga es (G eene and Kochha , 1974; Souchon, 1975; Melle and Ba on, 1978), whose expe imen al dis up ion in i o causes he inhibi ion o pala al shel adhesion (G eene and P a , 1977). By analyz- ing he lec in binding capaci y o his supe icial coa , Zscha¨bi z e al. (1994) demons a ed biochemical di e - ences be ween he ma ix ha co e s he MEE and o he pala al epi helia, and sugges ed i s impo ance in cell ec- ogni ion and adhesion du ing pala e usion. Glycop o ein molecules ha e also been obse ed in he in e cellula ma ix in e posed be ween he edges o he lens pla e (Van Ryb oeck and Olson, 1981; Yao e al., 1996) and he neu al p imo dium (Smi s Van P ooije e al., 1986; T asle and Mo iss-Kay, 1991) du ing hei apposi ion, and he e is expe imen al e idence showing ha dis up ion in he syn- hesis o enzyma ic deg ada ion o hese p o eoglycans se iously dis up s he usion p ocess (Mo iss-Kay and C u ch,1982;Mo iss-Kay and Tucke , 1989; Alonso e al., 1998). Among he di e en p o eoglycans ha could o m pa o hese glycoconjuga es and ha e a ole du ing pala al shel adhesion, ou ecen demons a ion o he p esence o chond oi in sulpha e p o eoglycan (CSPG) co e ing he supe icial bulging MEE cells (Ma ı´nez-A ´l a ez e al., 2000b), s ongly poin s o i s di ec pa icipa ion in his mechanism. Dis up ed pala al shel adhesion causes cle pala e in he ans o ming g ow h ac o be a 3 (TGF- ␤ 3) null mice (P oe zel e al., 1995; Kaa inen e al., 1995; Ma ı´nez- A ´l a ez e al., 2000a). TGF- ␤ 3is s ongly exp essed in he MEE cells o he s ill e ical pala al shel es, and his exp ession pe sis s un il he midline epi helial seam disap- pea s (Fi zpa ick e al., 1990; Pel on e al., 1990). TGF- ␤ 3is a c i ical molecule o pala al usion: i s inhibi ion by an ibodies o an isense oligodeoxynucleo ides p e en s in i o pala al shel adhesion in mice (B une e al., 1995) and mu a ion o he TGF- ␤ 3gene in bo h mice (P oe zel e al., 1995;Kaa inene al.,1995)and humans(Lid ale al.,1998; Mi chell e al., 2001) esul s in cle pala e. Pala al shel es om TGF- ␤ 3null mu an mice g ow, ele a e, and app oach no mally, bu hei adhesion ails due o dis up ion o an incomple ely cha ac e ized adhesion mechanism. A s ik- ing dec ease in he numbe o supe icial bulging cells and ilopodia p esen on he MEE su aces o TGF- ␤ 3null mu an mouse pala al shel es has been epo ed ecen ly (Taya e al., 1999; Ma ı´nez-A ´l a ez e al., 2000b). While he TGF- ␤ 3null pala al shel es appea ul as uc u ally o ha e a educed su ace coa (Taya e al., 1999), no in o ma- ion exis s on he s a us o he MEE glycoconjuga es. In o he sys ems, TGF- ␤ s egula e he syn hesis o ce ain glycoconjuga ed componen s o he ex acellula ma ix (Igno z and Massague´, 1986, 1987; Chimal-Mon oy and Dı´az de Leo´n, 1999), and TGF- ␤ s ha e been shown o egula e he exp ession o chond oi in sulpha e and de ma- an sulpha e p o eoglycans in di e en issues, epi helial and mesenchymal cells, and e en in he pala e (Bassols and Massague´, 1988; Sha pe and Fe guson, 1988; D’Angelo and G eene, 1991). To in es iga e whe he he p esence o CSPG on he mouse MEE su ace has a ole in pala al shel adhesion, we i s es ablished he iming o i s exp ession du ing pala e de elopmen and hen al e ed i s ac i i y h ough he in i o addi ion o ␤ -D-Xyloside o chond oi inase AC. To in es iga e he in luence o TGF- ␤ 3on he p oduc ion o CSPG by MEE cells, we analyzed i s exp ession in TGF- ␤ 3 null pala es and explo ed whe he he addi ion o TGF- ␤ 3 o ei he TGF- ␤ 3null mu an mouse o chick pala e cul u es s imula es bo h he exp ession o CSPG and pala al shel adhesion. Ou esul s demons a e he p esence o CSPG on he MEE su ace only immedia ely p io o pala al shel adhesionand ha his CSPG is equi ed o ha adhesion o occu . We also show ha TGF- ␤ 3null pala es lack CSPG on hei MEE su aces and ha addi ion o TGF- ␤ 3s imula es he p oduc ion o his p o eoglycan by he supe icial MEE cells, in bo h mice and chick, oge he wi h pala al shel adhesion. We he e o e conclude ha he cle pala e p o- duced by he absence o TGF- ␤ 3migh be caused, a leas in pa , by dis up ion o he syn hesis/sec e ion o CSPG by he supe icial MEE cells, hus impeding he app op ia e adhesion be ween pala al shel es. MATERIALS AND METHODS Animals Albino Swiss (CD1) o C57BL/6J TGF- ␤ 3he e ozygous mice (Jackson Labo a o ies) we e ma ed and he day o de ec ing he aginalplugwasdesigna ed day0. Time-ma edp egnan micewe e killed by an o e dose o chlo o o m, and he emb yos we e e- 394 Ga o e al. © 2002 Else ie Science (USA). All igh s ese ed. mo ed by Caesa ian sec ion, placed in s e ile cold 1/1 Dulbecco’s modi ied Eagle’s medium (DMEM/F12) (Sigma), and decapi a ed. Emb yos o be used o cul u e expe imen s we e emo ed unde s e ile condi ions. Once heads we e ob ained, he jaw and ongue we e emo ed. In C57BL/6J TGF ␤ 3he e ozygous mice, geno yping was pe o med as desc ibed in P oe zel e al. (1995). Fe ilized Whi e Legho n chick eggs (G anja San a Isabel, Co´ - doba, Spain) we e incuba ed a 37.5°C and 80% humidi y o 8.5 days [s age 35 (35HH) o Hambu ge and Hamil on, 1951]. Emb yos we e emo ed unde s e ile condi ions and decapi a ed, and he heads we e placed in s e ile cold DMEM/F12. Mouse and Chick Pala e Cul u es Pala al shel es om E13.5 CD1, TGF- ␤ 3⫹/⫹, and TGF- ␤ 3⫺/⫺ mouse o 35HH chick heads we e dissec ed unde s e ile condi- ions in DMEM/F12. Isola ed o pai ed mouse pala al shel es and isola ed chick pala al shel es we e hen placed on 2 ⫻2-mm Millipo e il e s (0.8- ␮ m po e size) and cul u ed in T owell’s issue cul u e, as desc ibed in B une e al. (1993). Cul u es we e incu- ba ed o 12 (isola ed shel es), 24, o 36 (pai ed shel es) h a 37°Cin a5%CO2incuba o . Chick pai ed pala al shel es we e cul u ed o 45 h in 0.5% aga gels, as desc ibed in Sun e al. (1998a), using DMEM/F12 ⫹1% asco bic acid as cul u e medium. Recombinan TGF- ␤ 3(10 ng/ml; Sigma) was added o he cul u e medium in hose expe imen s equi ing his condi ion. In all con ol and ea ed cul u es, medium was eplaced a e 24 h. In Vi o Inhibi ion o Chond oi in Sulpha e Ac i i y To inhibi in i o he biological ac i i y o CSPG, we used wo s a egies. Fi s , we dis up ed he syn hesis o CSPG by using he sulpha ed p o eoglycan syn hesis inhibi o P ni ophenyl ␤ -D- Xylopy anoside (Sigma) ( ␤ -D-Xyloside). Second, we speci ically di- ges ed CSPG by adding chond oi inase AC o he cul u e medium. E13.5 mouse pai ed pala al shel es we e cul u ed in DMEM/F12 in he p esence o ␤ -D-Xyloside (25 ␮ l/ml o 10.8 mg/ml o s e ile Hank’s; Sigma). Medium was eplaced a e 24 h. Cul u es we e main ained o 36 h o allow pala al shel usion. S e ile Hank’so a simila amoun o ␣ -D-Xyloside (ac i e anome e o ␤ -D-Xyloside) was added o he medium in con ol cul u es. All cul u es we e ixed in Ca noy ixa i e o 1 h and p ocessed o his ology. A o al o 20 ␮ l/ml o a solu ion o 9 U.I. Chond oi inase AC (Sigma) in 450 ␮ l o PBS was used o enzyma ically diges CSPG. We also used 10 U.I. o Hepa inase ype II (Sigma) in 100 ␮ lo PBS o speci ically diges hepa an sulpha e in some cul u e expe i- men s. Con ol cul u es had he same amoun o hea -inac i a ed enzymes added. Table 1 shows he numbe o cul u es s udied pe expe imen . CSPG Immunohis ochemis y TGF- ␤ 3⫹/⫹(and CD1) and TGF- ␤ 3⫺/⫺mouse emb yonic heads we e ca e ully washed wi h saline solu ion o a oid issue and blood con amina ion and hen ixed in Ca noy ixa i e o 2ha oom empe a u e. Mouse and chick isola ed pala al shel cul u es and TGF- ␤ 3⫹/⫹(and CD1) and TGF- ␤ 3⫺/⫺mouse pala e cul u es om di e en expe imen al condi ions we e ixed in Ca noy ixa- i e o 1 h. No less han ou di e en specimens pe expe imen al condi ion we e analyzed. Ca noy ixa i e was used because i main ains he ca bohyd a e epi ope ecognized by he CS-56 an i- body used. Specimens we e dehyd a ed in a g aded e hanol se ies and embedded in pa a in ollowing s anda d p ocedu es. Sec ions (8- ␮ m- hick) we e p eincuba ed in 1% PBS–BSA o 30 min and incuba ed o e nigh wi h he an i-chond oi in sulpha e monoclo- nal an ibody CS-56 (Sigma), which ecognizes he glycosaminogly- FIG. 1. E14.5 mouse pala al shel es immunolabeled wi h he an i-CSPG monoclonal an ibody CS-56. (A) When pala al shel es a e s ill a om each o he , he e is no an i-CSPG immunos aining on he MEE su ace. Inse shows an adjacen sec ion s ained wi h hema oxylin and eosin. (B) As pala al shel es app oach each o he , an an i-CSPG posi i e ma e ial appea s on he MEE su ace, and he supe icial bulging cells a e su ounded by abundan CSPG (a ow). The boxed a ea is shown in he inse a highe magni ica ion. No e he p esence o an i-CSPG posi i e ma e ial on he MEE su ace (a owheads). (C) This apical CSPG (a ows) g ea ly inc eases when con ac be ween pala al shel es becomes imminen . (D) Abundan CSPG con aining ma e ial accumula es a he poin o con ac be ween pala al shel es (a ows). (E) Howe e , when pala al shel adhesion has occu ed, no CSPG is obse ed in he midline epi helial seam. The inse shows an adjacen sec ion s ained wi h hema oxylin and eosin. (F) Islands o o ally an i-CSPG nega i e epi helium emain in he midline when pala al mesenchyme becomes con luen (a ow). No ice ha CSPG is ne e localized be ween mouse MEE cells. The inse shows an adjacen sec ion s ained wi h hema oxylin and eosin. As e isk, medial edge epi helium. M, mesenchyme. TABLE 1 Pala al Shel Adhesion in Mouse and Chick Pala e Cul u es Type o pala e cul u e No. cul u es s udied A e age leng h o adhe ed/ used MEE Con ols 33 Hank’s10 ␣ -D-Xyloside ea ed 10 2530 ⫾327 ␮ m (100%) Inac i a ed enzymes ea ed 13 ␤ -D-Xyloside ea ed 38 943 ⫾427 ␮ m (37%)a Chond oi inase AC ea ed 10 143 ⫾105 ␮ m (5%)a Hepa inase II ea ed 7 1600 ⫾812 ␮ m (63%)a TGF- ␤ 3 ⫹/⫹33 2801 ⫾395 ␮ m (100%) TGF- ␤ 3 ⫺/⫺12 970 ⫾559 ␮ m (34%)a TGF- ␤ 3 ea ed, TGF- ␤ 3 ⫺/⫺10 2133 ⫾379 ␮ m (77%)a Chick (con ols) 17 400 ⫾413 ␮ m (100%) TGF- ␤ 3 ea ed, Chick 17 1500 ⫾968 ␮ m (375%)a No e. Measu emen s a e mean ⫾s anda d e o . In all cases P⬍ 0.001 by wo- ailed S uden ’s es . aPe cen age o adhesion/ usion in each condi ion ela ed o he adhesion/ usion ob ained in con ol o TGF- ␤ 3 ⫹/⫹pala e cul u es (conside ed 100%). 395Chond oi in Sulpha e and Pala al Fusion © 2002 Else ie Science (USA). All igh s ese ed. 396 Ga o e al. © 2002 Else ie Science (USA). All igh s ese ed. can moie y o na i e chond oi in sul a e p o eoglycan (A nu and Geige , 1984). Sec ions we e hen eincuba ed in a luo escein- conjuga ed goa an i-mouse IgM (Vec o ) o 30 min. Some sec ions we e coun e s ained wi h p opidium iodide (Molecula P obes) (1/12,000 dilu ion) o 30 min and all we e moun ed in Aquamoun (GURR). Samples we e analyzed by using ei he an a gon (488/514 ␭ )o helium/neon (543 ␭ ) lase ZEISS LSM T-300 Scan Con ocal Mic o- scope. Digi alized images we e eco ded and o e lapped when equi ed. His ological P epa a ion Emb yonic heads and pala e cul u es om di e en expe imen- al condi ions we e ixed in 4% pa a o maldehyde o Ca noy ixa i e, dehyd a ed in a g aded e hanol se ies, and embedded in pa a in. T ans e sal sec ions (8- ␮ m- hick) we e hema oxylin and eosin-s ained ollowing s anda d p ocedu es. Sec ions we e s udied by using a Nikon Op ipho ligh mic oscope and pho og aphed wi haNikon FX35A came a.To measu e heleng h o he adhe ed opposing MEE in all pala e cul u es, a measu ing g id inse ed in a 20⫻ocula lens was used. The leng h o he adhe ed MEE in e e y 10 (mouse) o 20 (chick) sec ions aken om he middle 100 (mouse) o 200 (chick) sec ions o each pala e cul u e was added. The a e age leng h o adhe ed MEE o each g oup was hen calcula ed.The inal aluesa e exp essed as hea i hme ic mean ⫾ s anda d e o . Fo compa ison o he a e age measu emen s be- ween expe imen al and con ol samples, a wo- ailed S uden ’s es o independen samples was applied and P alues below 0.001 we e in e p e ed as indica ing s a is ical signi icance. RESULTS Exp ession Pa e n o CSPG in he Medial Edge Epi helium Immunolabeling o E14 (no shown) and E14.5 mouse pala es wi h he an i-CSPG CS-56 monoclonal an ibody FIG. 2. Thi y-six-hou pai ed pala e cul u es un ea ed (A, B, and I) and ea ed wi h ␤ -D-Xyloside (C and D), chond oi inase AC (E, F, and J) o hepa inase II (G and H). (B, D, F, and H) His ological sec ions aken om he cul u es shown in (A, C, E, and G), espec i ely, a he le el indica ed by he line. (I and J) Con ocal images o an i-CSPG immunolabeled sec ions. (A and B) Pai ed pala al shel es cul u ed o 36 h always adhe e/ use. A ows in (B) indica e he a ea o usion. (C and D) Howe e , when ␤ -D-Xyloside is added o he cul u e medium, pala al shel adhesion is g ea ly diminished (a ea be ween a ows in D), i no o ally impeded (as e isk in C). (E and F) Simila esul s a e ob ained when chon- d oi inase AC is added o he cul u e medium, al hough he adhe ed a ea is equen ly smalle (a ow in F). (G and H) When hepa inase ype II is added o he cul u e medium, pala al shel adhesionissomewha be e and anepi helial seam o ms (a owin H). (I and J) Pala e cul u es ea ed wi h chond oi inase AC show a g ea dec ease in he an i-CSPG immunolabeling in bo h he MEE (as e isk in J) and mesenchyme (M) when compa ed wi h con ols (I). Opposing MEE (as e isk) ha e pa ially adhe ed in (I) and a e sepa a ed in (J). The boxed a eas in (D) and (F) a e shown in he op igh co ne a highe magni ica ion. 397Chond oi in Sulpha e and Pala al Fusion © 2002 Else ie Science (USA). All igh s ese ed. showed posi i e s aining o he pala al mesenchyme in all emb yos s udied (Figs. 1A–1F). A E14, pala al shel es we e ei he e ical alongside he ongue o had ele a ed and we e posi ioned ho izon al abo e he ongue. No an i- CSPG-posi i e ma e ial was obse ed among MEE cells no on heMEEsu ace a his imepoin (no shown).A E14.5, he pala al shel es app oach each o he and ha e con ac ed in he middle hi d o he pala e in mos cases, o ming he midline epi helial seam. No an i-CSPG-posi i e s aining was obse ed on he MEE su ace when pala al shel es a e s ill a om each o he (Fig. 1A). Howe e , an i-CSPG- posi i e ma e ial co e ed mos o he MEE su ace when pala al shel es became close (Figs. 1B and 1C), and CSPG- posi i e s aining g ea ly inc eased when con ac be ween pala al shel es was imminen (Fig. 1D). Posi i e s aining was ne e no iced be ween basal MEE cells, bu was ob- se ed su ounding he bulging supe icial MEE cells (Fig. 1B). In he zones whe e con ac be ween pala al shel es had occu ed, he midline epi helial seam showed o ally nega- i e an i-CSPG s aining (Fig. 1E). Howe e , as he midline epi helialseam dis up ed,an i-CSPG-posi i emesenchyme occupied he midline, lea ing only small islands o o ally nega i e an i-CSPG labeling (Fig. 1F). These esul sindica e ha aCSPG-con ainingma e ialis loca ed ansien ly on he MEE su ace only p io o pala al shel adhesion, and ha his ma e ial is absen among basal MEE cells. In Vi o Al e ed Chond oi in Sulpha e Exp ession In e e es Pala al Shel Adhesion Pala al shel es cul u ed o 36 h in con ol condi ions always use (Figs. 2A and 2B). Howe e , he addi ion o he sulpha ed p o eoglycan syn hesis inhibi o P ni ophenyl ␤ -D-Xylopy anoside ( ␤ -D-Xyloside) o he cul u e medium esul ed in a comple e absence o usion o he pala al shel es, which emained ei he sepa a ed o minimally adhe ed a e 36 h o cul u e (Fig. 2C). His ological sec ions aken om hese cul u es showed ei he no adhesion o minimalcon ac be weenopposing MEE(Fig.2D,andTable 1). Simila esul s we e ob ained when he CSPG-deg ading enzyme chond oi inase AC was added o he cul u e me- dium, al hough i seemed o al e pala al shel adhesion mo e s ikingly (Figs. 2E and 2F, and Table 1). Pala al shel adhesion and usion was somewha al e ed when he hepa- an sulpha e-deg ading enzyme hepa inase ype II was added (Figs. 2G and 2H, and Table 1). The wo- ailed S uden ’s es applied o he leng h o adhe ed/ used MEE in all specimens o each g oup showed P⬍0.001 in all cases, hus indica ing i s s a is ical signi icance. Immunolabeling o 36-h cul u ed pala es wi h he an i- CSPG CS-56 monoclonal an ibody showed a g ea ly e- duced exp ession o CSPG in bo h he mesenchyme and he MEE su ace in hose cul u es ea ed ei he wi h chon- d oi inase AC (Fig. 2J) o ␤ -D-Xyloside (no shown) when compa ed wi h con ol (Fig. 2I) o hepa inase- ea ed (no shown) pala e cul u es. These obse a ions indica e ha CSPG is equi ed on he opposing MEE o achie e pala al shel adhesion in i o. TGF- ␤ 3Induces CSPG Exp ession by MEE Cells and Pala al Shel Adhesion in Vi o We immunolabeled E14.5 TGF- ␤ 3 ⫹/⫹and TGF- ␤ 3 ⫺/⫺ mouse pala e sec ions wi h he an i-CSPG CS-56 monoclo- nal an ibody and s ained cellula nuclei wi h p opidium iodide o allow he obse a ion o he en i e MEE (Figs. 3A and 3B). E14.5 TGF- ␤ 3 ⫺/⫺almos con ac ing MEE showed a s iking educ ion o CSPG on hei su aces (Fig. 3B) when compa ed wi h con ols (Fig. 3A). In o de o de e mine he in luence o TGF- ␤ 3on he p oduc ion o CSPG by MEE cells, we cul u ed E13.5 TGF- ␤ 3 ⫺/⫺isola ed pala al shel es o 12 h in he p esence o absence o TGF- ␤ 3. Since he MEE disappea s when mouse isola ed pala al shel es a e cul u ed o 24 h o mo e (unpublished obse a ion), we selec ed 12 h as he cul u e pe iod o his expe imen . To analyze di e ences in pala al shel adhesion, pai ed pala al shel eswe e cul u ed o 24 h, since his pe iod is su icien o con ol pala al shel es o adhe e and use. As expec ed, CSPG was ne e de ec ed on he MEE su ace in he TGF- ␤ 3 ⫺/⫺cul u es(Fig. 3C). The addi ion o TGF- ␤ 3 obo h 12-h TGF- ␤ 3 ⫺/⫺isola ed pala al shel cul u es and o 24-h TGF- ␤ 3 ⫺/⫺pala ecul u es esul ed in an inc ease inbo h he exp ession o CSPG on he MEE su ace and in pala al shel adhesion ha ai ly mimicked he esul s ob ained in he TGF- ␤ 3 ⫹/⫹pala es (Figs. 3D–3J, and Table 1). TGF- ␤ 3 ⫺/⫺ isola ed pala al shel es cul u ed o 12 h in a medium con aining 10 ng/ml TGF- ␤ 3clea ly showed an an i-CSPG- posi i e ma e ial co e ing he MEE su ace (Fig. 3D). In cul u es o TGF- ␤ 3 ⫹/⫹pai ed pala al shel es, opposing MEE had almos comple ely adhe ed a e 24 h (Fig. 3E), and as expec ed, he epi helial seam was o ally an i-CSPG- nega i e (no shown). To analyze he p esence o CSPG on he MEE, we selec ed hose sec ions whe e adhesion be- ween opposing MEE had pa ially occu ed and no iced he p esence o CSPG on he MEE su ace (Fig. 3F). TGF- ␤ 3 ⫺/⫺ pai ed pala al shel es cul u ed o 24 h had ailed o adhe e adequa ely(Fig.3G,and Table1)andshowedno an i-CSPG- posi i e s aining on he MEE su ace (Fig. 3H). Howe e , when TGF- ␤ 3was added o 24-h TGF- ␤ 3 ⫺/⫺pala e cul u es, an inc ease in pala al shel adhesion occu ed (Fig. 3I, and Table 1) oge he wi h he p esence o CSPG on he MEE su ace (Fig. 3J). We inally in es iga ed whe he TGF- ␤ 3is able o s imu- la e he in i o adhesion o chick pala al shel es and/o induce CSPG p oduc ion by MEE cells. When cul u ed in aga gels o 45 h, pai ed chick pala al shel es adhe ed pa ially in mos cases (Fig. 4A). The addi ion o TGF- ␤ 3 o he cul u e medium s ikingly inc eased pala al shel adhe- sion, leading some imes o pala al shel usion (Fig. 4B, and Table 1). To de e mine whe he his e ec o TGF- ␤ 3on chickpala al shel adhesionco ela eswi h he induc iono 398 Ga o e al. © 2002 Else ie Science (USA). All igh s ese ed. he syn hesis o CSPG by MEE cells, we cul u ed isola ed chick pala al shel es o 12 h wi h o wi hou he addi ion o TGF- ␤ 3 o he cul u e medium, and hen immunolabeled he sec ions wi h he an i-CSPG CS-56 monoclonal an i- body. Con ocal isualiza ion o he con ol cul u es showed li le o no an i-CSPG s aining in he MEE (Fig. 4C). Howe e , inc eased exp ession o CSPG bo h among (Fig. 4D) and o e he MEE su ace (Fig. 4E) was p esen in he TGF- ␤ 3- ea ed cul u es. The wo- ailed S uden ’s es applied o he leng h o adhe ed/ used MEE in all specimens o each g oup showed P⬍0.001 in all cases, hus indica ing i s s a is ical signi i- cance. Taken oge he , hese esul sindica e ha TGF- ␤ 3s imu- la es pala al shel adhesion in bo h mouse and chick em- b yos while inducing he p oduc ion o CSPG by MEE cells. DISCUSSION The CSPG-Con aining Ma e ial Tha Co e s he MEE P io o Con ac be ween Pala al Shel es Is Necessa y o Pala al Shel Adhesion Ou wo k demons a es ha he p esence o CSPG co - e ing he MEE apical su ace p io o and du ing ini ial mouse pala al shel adhesion is essen ial o pala al shel adhesion. The p esence o a glycoconjuga e- ich cell coa , likely syn hesized by MEE cells (P a and Hassell, 1975), was epo ed in ea lie in es iga ions o co e he MEE apical su aces. This p esence was demons a ed by using u henium ed, a poly alen ca ion ha binds ai ly speci i- cally o polyanions such as glycosaminoglycans (GAGs) (G eene and Kochha , 1974; Souchon, 1975), o by he p e e en ial adhe ence o concana alin-A o he MEE su - ace (P a and Gibson, 1973). This glycoconjuga e su ace coa was also shown o be equi ed o he ini ial adhe ence o pala al MEE, as he addi ion o pai ed pala al shel cul u es o Diazo-oxo-no leucine, a glu amine analog ha inhibi s glu amine ans e eac ions and he eby blocks glycosaminoglycans and glycop o ein syn hesis, esul ed in a s iking inhibi ion o pala al shel adhesion (G eene and P a , 1977). Which glycoconjuga es a e esponsible o pala al shel adhesion was ne e de e mined. We demons a e he e he appea ance o CSPG on he mouse MEE apical su ace only immedia ely p io o pala al shel adhesion and ha his exp ession inc eases as he con ac be ween pala al shel es becomes imminen . This empo al pa e n o exp ession is in con as wi h he obse a ion o he appea ance o he ca bohyd a e coa on he MEE su ace 48 h p io o pala al shel adhesion (G eene and Kochha , 1974), al hough a hicke coa was obse ed sho ly be o e usion (Souchon, 1975). The CSPG de ec ed he e could be pa o a mo e complex ex acellula ma ix co e ing he MEE su ace, o med by di e en glycoconjuga es (B inkley e al., 1992). The appea ance o CSPG only a E14.5 sugges s ha he molecule con aining i is sec e ed jus p io o pala al adhesion and ha his molecule is essen ial o such adhe- sion. We also demons a e ha CSPG is absen among nonapi- cal MEE cells du ing mouse pala e de elopmen . This migh be a cha ac e is ic o he mouse MEE. When he exp ession o CSPG was induced in TGF- ␤ 3null mouse pala e cul u es h ough he addi ion o TGF- ␤ 3, CSPG exp ession was only obse ed on he MEE su ace and ne e among MEE cells. On he con a y, a g ea exp ession o CSPG was no iced among MEE cells ( oge he wi h CSPG exp ession on he MEE su ace) when his expe imen was pe o med by using chick pala es. This is in keeping wi h ea lie in es iga ions showing heabsenceo u henium ed s aining in in e cellula a eas in he mouse MEE (G eene and Kochha , 1974; Souchon, 1975) bu i s p esence in humans (Melle and Ba on, 1978). We ha e in es iga ed whe he CSPG is equi ed o pala al shel adhesion in mice. We speci ically al e ed he syn hesis o sulpha ed p o eoglycans h ough he addi ion o he inhibi o P ni ophenyl ␤ -D-Xylopi anoside ( ␤ -D- Xyloside) o pai ed pala e cul u es. This ␤ -D-Xyloside ac s as an a i icial chain ini ia o o sulpha ed p o eoglycans biosyn hesis ha s imula es he syn hesis o sho co e- p o ein- ee unde sulpha ed chond oi in sulpha e chains (Gibson e al., 1979) and o hepa an sulpha e o a lesse ex en (Sobue e al., 1987). ␤ -D-Xylosides ha e become a popula ool o s udying he ole o sulpha ed p o eogly- cans in a wide a ie y o biological p ocesses. By using chond oi inase AC in sepa a e expe imen s, we also speci i- cally deg aded CSPG in mouse pai ed pala e cul u es. A e bo h ea men s,immunolabelingwi h he CS-56 monoclo- nal an ibody showed he absence o CSPG on he MEE su ace and a s iking educ ion o he in i o pala al shel adhesion. In addi ion, ou esul s indica e ha ea men o pala e cul u es wi h hepa inase ype II al e s pala al shel adhesion less s ikingly. This means ha Hepa an sulpha e could also play a ole in his p ocess. These da a con i m ha pala al shel adhesion depends on mechanisms local- ized in he apical su ace o he MEE cells (Schu¨pbach e al., 1983; Taya e al., 1999) and ha he syn hesis and sec e ion o CSPG by he supe icial MEE cells plays a c ucial ole in his adhesion. This s a emen has addi ional suppo in he simul aneous expe imen al s imula ion by TGF- ␤ 3o he exp ession o he apical CSPG and pala al shel adhesion in he chick, whose pala al shel es do no adhe e in i o (despi e hei con ac du ing pala ogenesis) and lack he apical CSPG. Mo eo e , abundan CSPG su ounds he supe icial bulging MEE cells, which ha e been shown o be p ima ily in ol ed in pala al shel adhesion (Ma ı´nez- A ´l a ez e al., 2000a,b). CSPG has been shown o be equi ed in he adhesion o o he epi helial edges, as du ing he usion o he neu al olds (Mo is-Kay and Tucke , 1989; Alonso e al., 1998) o in he o ma ion o he lens (Alonsoe al., 1996; Ga o e al.,2001).In e es ingly, as wi h he CSPG immunos aining, he u henium ed ma e ial obse ed on he MEE su ace in ea lie s udies is absen when pala al shel adhesion is accomplished (Souchon, 399Chond oi in Sulpha e and Pala al Fusion © 2002 Else ie Science (USA). All igh s ese ed. FIG. 3. Co ela ion be ween pala al shel adhesion and p esence o apical CSPG and TGF- ␤ 3in he mouse. (A–D, F, H, and J) Con ocal images o an i-CSPG immunolabeled sec ions ha ha e been s ained wi h p opidium iodide o be e isualize nuclei. (E, G, and I) Hema oxylin and eosin-s ained sec ions o pala e cul u es simila o hose obse ed in (F, H, and J), espec i ely. (C and D) Twel e-hou cul u es o E13.5 TGF- ␤ 3 ⫺/⫺(C) and TGF- ␤ 3supplemen ed (S) TGF- ␤ 3 ⫺/⫺(D) isola ed mouse pala al shel es. (E–J) Twen y- ou -hou TGF- ␤ 3 ⫹/⫹(E and F), TGF- ␤ 3 ⫺/⫺(G and H), and TGF- ␤ 3supplemen ed TGF- ␤ 3 ⫺/⫺(I and J) pai ed pala e cul u es. (A) App oaching E14.5, TGF- ␤ 3 ⫹/⫹pala al shel es ha e a laye o CSPG-con aining ma e ial (a ows) co e ing he MEE su ace. (B) The e is no CSPG on he MEE su ace o almos con ac ing E14.5 TGF- ␤ 3 ⫺/⫺pala al shel es. (C) Twel e-hou E13.5 TGF- ␤ 3 ⫺/⫺isola ed pala al shel cul u es do no show an i-CSPG-posi i e immunos aining on he MEE su ace. (D) When TGF- ␤ 3is added o TGF- ␤ 3 ⫺/⫺pala e cul u es, an an i-CSPG-posi i e ma e ial is clea ly obse ed on he MEE su ace (a ow). (E and F) Opposing MEE o 24-h TGF- ␤ 3 ⫹/⫹pai ed pala e cul u es adhe e, and an epi helial seam is obse ed (a ow in E). In hose sec ions whe e pala al shel adhesion is no comple e, an an i-CSPG-posi i e ma e ial is loca eda he placeo con ac be ween opposingMEE (a ow inF). (G and H)Howe e , TGF- ␤ 3 ⫺/⫺pala alshel es a e mos lysepa a ed when cul u ed in pai s o 24 h (G), and no an i-CSPG-posi i e immunos aining is p esen on he MEE su ace in hese cul u es (H). (I and J) The addi ion o TGF- ␤ 3 o TGF- ␤ 3 ⫺/⫺pala e cul u es escues pala al shel adhesion (a ow in I) and induces he p esence o CSPG on he MEE su ace (a ows in J). As e isk, MEE. M, mesenchyme. The boxed a eas in (A, C, D, and J) a e shown in he lowe le co ne s a highe magni ica ion. 400 Ga o e al. © 2002 Else ie Science (USA). All igh s ese ed. 1975). This sugges s ha CSPG is needed mainly o he ecogni ion and es ablishmen o he ini ial adhesion be- ween opposing MEE. The monoclonal an ibody CS-56 used he e is di ec ed agains an epi ope p esen in he glycidic moie y o CSPG and i is common o mos CSPG ypes (A nu and Geige , 1984). This implies ha we can de ec a b oad numbe o molecules ha ha e CSPG among hei componen s. The esolu ion capaci y o ou s udy does no allow us o dis inguish whe he he CSPG de ec ed on he MEE apical su ace is pa o a comple ely ex acellula ma ix o belongs o he ex acellula domain o molecules wi h ansmemb ane componen s. Howe e , since his CSPG plays a ole in pala al shel adhesion, i migh belong o o be ela ed wi h a cell adhesion molecule. Di e en ypes o CSPG a e in ol ed in he main enance o cell adhesion in o he sys ems h ough hei in e ac ion wi h cadhe ins, CAMs, o in eg ins (Re zel e al., 1996; Iida e al., 1998; S an o d e al., 1999; Li e al., 2000). Recen in es iga ions ound mu a ions o PVRL1, encoding he FIG. 4. Co ela ion be ween pala al shel adhesion and p esence o apical CSPG and TGF- ␤ 3in he chick. (A and B) His ological sec ions aken om un ea ed (A) and TGF- ␤ 3- ea ed (B) 35HH chick pai ed pala al shel es cul u ed o 45 h. (C–E) Con ocal images o an i-CSPG immunolabeled sec ions aken om 12-h 35HH chick isola ed pala al shel cul u es, un ea ed (C) o TGF- ␤ 3- ea ed (D and E). (A) Chick opposing MEE adhesion is pa ial (a owheads) in mos con ol cul u es. (B) Opposing MEE adhesion inc eases when TGF- ␤ 3is added o he cul u e medium (a owheads) and e en pala e usion occu s (a ows). (C) The e is almos no an i-CSPG-posi i e immunos aining among MEE cells no on he MEE su ace in un ea ed chick pala al shel cul u es. (D) Howe e , when TGF- ␤ 3is added o he cul u e medium, a g ea amoun o CSPG is obse ed amongs MEE cells. (E) A di e en con ocal sec ion om he specimen shown in (D) allows he obse a ion o CSPG on he MEE su ace. As e isk, MEE. M, mesenchyme. 401Chond oi in Sulpha e and Pala al Fusion © 2002 Else ie Science (USA). All igh s ese ed.