STABLE, POSITION-RELATED f!,ESPONSES TO ...

III Vitro CeU. Dev. BioI. 30A:181-186. March 1994 it) 1994 TISsue Cullure A~~oc", uoll

1071·2690/ 94

$0 1.50+0.00

STABLE, POSITION-RELATED f!,ESPONSES TO RETINOIC ACID BY CHICK

LIMB-BUD MESENCHYMAL CELLS IN SERUM-FREE CULTURES

DOUGLAS F" PAULSEN, MICHAEL SOLURSH, ROBERT M" LANGILLE, LEI PANG,

AND

WEI-DANG CHEN

Department ofAnaJomy, Morehouse School qJ"Medicine, 720 We~tview Drive, S.W. Atlanta, Georgia, 30310 (D. P., L P., W.-D . C.);

Depar'menl qJ"Biological Sciences, University of Iowa, Iowa, City, Iowa,. 52242 (M. S.); and Department (J./·Ana,tomy a.nd Neurobiology,

University of Ottawa, Ottawa,. OnLaria, KI H 8M5 (R. L.)

(Reeeived 28 April 1993 ; accepted 19 .Tuly 1993)

SUMMARY

Retinoic acid (RA) has dramatic effects on ~mb-skeletal patterning in vivo and may weU playa pivotal role in normal limb morphogenesis. RA 's effects on the elCpl"ession of pall ern-related genes in the developing limb are probably mediated by cytoplasmic RA-binding proteins and nnclear RA-receptors. Liltle is known, however, about how R,.A modifies specmc cellular behaviors required for skeletal morphogenesis. Earlier studies supported a role for regional differences in RA concentralion in generating the region-specific cell behaviors that lead to pattem fon-nation. The present study explores the possibility that position-related, cell-autonomous diffe re nces in the way limb mesenchymal ceUs respond to RA mighl have a role in genemting pauern-related cell behavior. Mesenchymal cells from different proximodistal regions of stage 21 - 22 and 23-24 chick wing-buds were growlI in chemically defined medium and exposed to 5 or 50 ng/ ml of RA for 4 days in high-density micro titer cultures. The etlects of RA on chondrogenesis in thcse cultures clearly differed depending on the limb region from which the cells were isolated. Regional differences in RA's effects on growth over 4 days in these cultures were less striking. Thc region-dependent responses of these cells to RA proved relatively stable in cullure despite ongoing cytodifferentiation. This scrum-free culture modcl will be useful in exploring the mechanisms underlying Lhe region-dependent responsiveuess of these cells to RA..

Key words: limb bud; reLinoic acid; pattern formation; chondrogenesis; cell culture; serum-free medium. and away from the complex and rapidly changing environment of the developing limb. Our previous studies have shown that limb mesencLyme cells respond directly to RA (29). In serum-containing medinm, RA causes a dose--dependent inhibition of chondrogenesis. In serum­ free medium , however. RA at physiologic doses promotes, and al higher (pharmacologic) doses inhibits chondrogenesis. Similar re­ sponses to RA have been noted in cultures of chick mandibular meseuchyme (18) where RA's effects are also stage--dependent (33). Thus, the availability of a defined medium that allows isolated chick limb mesenchymal eells to proliferale, differenliate, and re­ spond to RA in vitro more as they do in vivo than when grown in standard serum-containing medinm (29,30) provides a useful model for eontinuing such studies. One view of RA's role in limb development suggests that the mesenchymal cells derive positional information from regional dif­ ferences in the RA concentration to which they are exposed (9,36). From this perspective. cells from differenl limb rcgions should re­ spond similarly to identical RA concentrations. The present results indicate that al RA conccntrat ions in the physiologic range, the region of Oligin significant ly influenees Lhe way limb mesenchymal eells respond to a parti cular concentration of RA in vitro.. Further­ more, Ihese cell-autonomous, stage- and region-related differences in response are relatively stable in culture, indicating that further

INTRODUCTION

The vitamin A agonist, retiuoic acid (RA). is a well-known skele­ tal teratogen (16, 17). Depending on th e dose and mode of delivery, exogenous RA has profound effects on limb skeletal pattern ranging from distal deletions to skeletal pattern dupJications (16, 17,39,43, 44). Together wiLh evidence that developing limbs contain endoge­ nous RA (6,42), these effects support a role for RA as a morphogen during normal limb development (9,36). On the otJler hand , evi­ dence is mounting that RA is ouly one component of a complex cascade of morphogenic effectors in the developing limb (27,47). Nevertheless, the patterns of expression of nuclear RA-receptors and cytoplasmic RA-binding proteins in developing limbs suggest Lhat, whatever its position in the cascade, RA's role in normal skele­ tal patterning is pivotal and involves regional differences in ligand­ receptor iuteraetions (8 and references Lherein; 37). Inasmuch as limb morphogenesis involves the spatiotemporal co­ ordination of gwwth and differentiation, regional difi'erences in the effects of RA on these behaviors in limb cells deserve further study. Of related interest is the question of whether RA's morphogenic and teratogenic effects arc due to direc t effects on the cells of the limb mesoblast or whcther RA aets indirec tly by altering the fu~ction of the overlying ectoderm, specifieally, the apical ectodermal li.dge (19,38,45). It is thus of particular interes t to lest, the responses of the limb mesenchymal cells to RA in isola lion from the ectoderm,

181

182

PAULSEN ET AL.

Stcl]lI 21-22

Stoge 23-24

,)

Proximo I

Dletal

CHONDROGENESIS

0.3

{

0.2

PROUFERAnON

Wing-Bud Region

D.'

• Di,tal Tip BSubdistol

0 .3

Dproximol

8

:e

0.2

{ ;! 0

[ 0.1 0.1

0.0

21-22 23-24

SloglI ~f Danor

21-22

23- 2'"

0.0

FIG. 1. Above, sehe matic representation of the wing-bud reglons from wh.ich Ihe mesenchyme was isolated in these s tudies. Below, defined me­ dium (OM) eonlrol c ultures. Quantitative estimates of chondrogenesis (t\:.oo/weU) a nd prolife ration (p-g DNA/weD) in 4-day, se ru m-free, micro­ cultures of mesenchyme from different regions of s tage 21 - 22 and slage 23-24 chicK wing bnds. Values are mean ± standard error.

studies of factors ullderlying regional differences ill lhe way these cells respond to RA will be possible using Ibis expetimenlal model. MATER.IAl.B AND METHODS

Twue i!olotion.. White leghorn chick embryos we re coUected in cold Tyrode's solution (Sigma, SI. Louis, MO) and staged according to Ham­ burger and Hamilton (1 3). Stage 23- 24 wing-buds were divided into three regions (Fig. 1); the proximal region, eontaining a visible precartilage con­ densation ; the distal 2S0-,um tip, containing rapidly dividing (40), undiffer­ entiated mesenchyme known to be mainly chondrogenic and free of myo­ genic ceUs (1 ,25,26,28); and the intervening s ubdis tal region containing a mixture of rapidly dividing, undifferentiated mesenchymal precursors of both chondrogeruc and myogenic cells. In s tage 21 - 22 wing buds, proximal precartilagp, condensations are not yet visible and thus cannot serve as a landmark to separate prox.imal and subdistAl regions. These buds were divided inlO only the distAl 250 -~m tips and tue remaining proximal por­ tions (Fig. 1). Microtiter cultures. High de nsity microcuhures of mesenchyme fr om each region were established as described previously (30) with some modifi­ cations. The defined medium (DM) consisted of Oulbecco's modified Ea­ gle's mediuru (DMEM)/F12 (1 :1, MediaTech, Herndon, VA) snpple­ mented ...nth 5 ~g/ml bovine insulin (CoUaborative Research, Waltham, MA), 5 ~g/ml chicken transferrin (conalbumin), 100 nM hydrocortisone, 50 p.g/ml L-ascorbic acid (aU Sigma), and antibiotics (GlECQ, Grand ls­ land, NY). The eultures were grown in half-well microtiter cnlture plates (Corning, Corning. NY), which were inoculated with 2.5 X lOs cells in 10 ~l of OM . After a I-h attachment period, the cultures were each carefully flood ed w"ilh an additional 290 ~l of the appropriate medium, followed by daily medium changes. Crystalline all-trans retin oic aeid (Sigma) was pre­ pared as a 0.5-mg/rnl freezer s tock in absolute ethanol and stored as aliquots unde r nitrogen gas at -BO° C. On the day the cultures were seeded. a n atlquol was diluted with OM to 50 and 500 ng RA/ ml (l OX stocks). and added at appeopliate limes and at each s ubsequent medium

.. -- - --_._ - - - - - - - ­

change to final eoncentrations of 5 or 50 ng/mI. These concentrations were selected based on pre\'ious dose-response studies using mesenehyme from whoLe limb buds (29,30). Unde r conditions like those in the prese nt study, RA at 5 ng/ mi maximally s limulates. and at 50 nglml inhibits, in vitro ehondrogenesis. The euhures we re grown for 0 , ] 2, 24, or 48 h before adding the RA. the n harvested after 4 days. Control c ultures we re grown in DM alone. ClwndrogeneslJ. Cartilage differentialion in RA-treated And eontrol cultures was assessed eoiorimenicaUy by a modifieation of the method of HasseU and Horigan (14) as described previously (29,30). Briefly, the net absorbanee al 600 nm (~) per well was measnred in 4 M guanidinium Hel (Sigma) extracts of alcian blne stained eartilage matrix from fixed 4-day cultures. Alcian blue (I.CN. Cleveland, OH) binds quantit.ati\'ely to sulfate groups in the cartilage matrix (21) and the bound and extracled stain in mieromass c ultures of limb-bud mesenchym"e is proportionate to aecu­ mulated s ulfated proteoglycan (2,20) . Growth. DNA accumulation was measured in eompanion 4-day c ul­ tures seeded from the same ceU suspe nsions used for the chondrogenesis s tndies. DNA content in a sonicate of the cell layer (rom eaeh c ulture well was quantitated using the Hoeehst 33258 (Polysc ie nees, Warrington, PAl fluorescence enhaneement method of Brunk et al. (4), as in pre\'ious studies (29,30). DNA aceumulation in defined medium is known to be similar la, but slightly less than, that in s tandard serum-containing medium (30). The moderate baseline growth rate is expected because the cultures were essen­ tially conflu enl al seeding. Data a.na.lysi.s. Each data point (A600/ well and /J.g DNA/well) in Figs. 1, 2 and 4 represe nts the mean and st.andard error for four to six separate experiments, including a minimum of two and typically tUree separate cul­ tures in eaeh experiment, for each condition . Each data point thus repre­ sents a minimum of 10, and more typically 12 10 16 euhures. To coUect s uffieie nt tissue from each s ubregion (tue distal tip being the s mallest) a minimum of 15 dozen e mbryos were routinely used for each experime nt. Inasmuch as two s tages were included in each age group, there was some

.

~

250

A

250

o

Control

Is ng RA/ml

200

8 50 nlil RA/rnl

J 150 "Ii D

u

0

200 150

0

50

50

250

0 250

B

200

200

150

1

8

.

100

"Ii 100

~

•

i

50~

.3

~

.0- -'"

- ~

«~

24 H

0

.2 .....

o

C

C ~

'z E ~-

12 H

0: '"

"'~ Cf­

"'

OH

183

mal cultures (Fig. 1). At stage 23-24, chondrogenesis from proxi­ mal mesenchyme was still less than that from distal or subdistal mesenchyme, which exhibited similar levels of chondrogenesis. Chondl'ogenesis was greater in proximal stage 21 - 22 cultures than in proximal stage 23-24 cultures (P < 0.002), which exhibited the least chondrogenesis of the regions tested. The levels of chondro­ genesis in control cultures of mesenchyme from distal and subdistal stage 23 - 24 and distal stage 21-22 wing-bud regions were similar. Growth. In 4-day control cultures of stage 21 - 22 wing mesen­ chyme, DNA accumulation was slightly less (P < 0.01) in the distal tip cells than in the proximal cultures (Fig. 1). For stage 23-2 4 limb mesenchyme, distal and subdistal cultures exhibited similar levels of growth accumulation. At Ihis lat er stage, growth was slightly greater in subdistal than in proximal cultures (P < 0 .05), whereas that in the distal cultures did not differ significantly from the other regions (Fig. 1). In comparing the stages, it was evid ent that growth was grcater in distal and subdistal stage 23- 24 cuhures than in distal .stage 2 1-22 cultures (P < 0.05), whereas th ai in the proximal cultures from both stages was similar. Region-Dependent Differences in the Effects of Ret.inoic Acid

-.'E., ..,

OH ')007'"

~

'0 > «
184

PAULSEN ET AL. CHONDROGENESIS

300

J ]

A

DISTAl

200

~

8

PROXIIMI..

B ::;

DISTAl

200

~ RA 5

RA 5 ng/m\

ng/ml RA 50 ng/ml

RA 50 l1Q/rnl

'5

300

STAGE 21-22

PROXIWJ....

c

< ;, •

PROLIFERATION

STAGE 21-22

100

0

8

'5

100

~

C

•~ •

"­

0 300

} ~

..c

II C

a 300

STAGE 2.3-24

Dr.nAl.

SUBDISTAL

PROXIIML

200

0

u

'5

is in chick Limb mesenchyme in vi tro derived from distal limb bnd tips: changes in cyclic AMP and in prostaglandin responsiveness. J. CeU. Phys;. L 13608 1-87; 1988. 3. Boylan, J. F.; Gudas, L. 1. The level of CRABP-l expression inUuences the amounts and types of all·trans reti nOlC acid metabolites in F9 teratocarcinoma stem cells. 1. BioI. Chern. 267:21 486 - 21491; 1992. 4. Brunk, C. F.; Jones, K. c.; James, T. W. Assay for nanogram quanti­ ties of DNA in cellular homogenates. Anal. Biochem. 92:497-500;

1979. 5. Caplan, A. I. Effects of the nicotinamidc-scnsiLlve teratogen a-aeetyl­ pyridine on chick limb-bud cells in culture. Exp. Cell Res. 62:341­ 355; 1970. 6. Chen. Y. P.; Huang, L.; Russo, A. F., et al. ReLinoic acid is enriched iII Hensen's node and is developmentally regulat cd in the early chicken embryo. Proc. Nat!. Acad. Sci. USA 89: 10056- 10059; 1992. 7. Coelho. C. N. D. ; Snmoy, L. ; Kosher, R. A. , et al. GHox-7: a chicken horneobox-containing gene expressed in a fash ion consistent with a role in pauerning events during embryonic chick development. Dif­ ferentiation 49:85-92; 1992. 8. DoUe. P.; Ruherte. E.; Leroy, P., et al . Re1.inoie aeid receptors and cellular retinoid binding proteins. I. A systematic study of their different ial pattern of transcription dmi ng mouse orga nogeuesis. Developmen t llO:1l33-1151; 1990. 9. Eichele. G. Retinoids and vertebrate limb pattern form ari on. Trends Gene t. 5:246- 251; 1989. 10. Evans, R. M. The steroid and thyroid hormone receptor snperfamily. Scicnee 240:889-895: 1988. 11 . Fiorella, P. D.; Napon, J. L. Expression of cellular ret,inoic acid bind­ in g protein (CRABP) in Escherichi,a coho Characterization a.nd evi­ dence thal holo-CRABP is a substrate in retinoic acid metabolism. J. BioI. Chern . 266: 16572-16579: 1991.

i

L

186

PAULSEN ET AL.

12. Greene, S.; Chamboo. P. Nuclear receptors e nha nce our understand­ ing of transcription regulatio n. Trends Genel. 4:309- 314; 1988. 13. Ha mburger. V.; Hamilton. H. L. A series oC normal stages in the development of the chick embryo. 1. Morphol. 88:49-92; 1951 . 14. Hassell, J. R.; Horigan, E. A. Chondrogenesis: a model developmental system for measuring the teratogenic poten tial of compounds. Tera­

togen. Carcirlogen.

Mu~e n .

2:325- 331; 1982.

15. Ide, H.; Aono. H. Retinoie acid promotes proliferation a lld chondro­ genesis in the distal mesodermal cells of chiek limb bud. Dev . BioI.

130,767-773; 1988. 16. Koehhar, D. M. Teratogenic activi ty of retinoic acid. Aeta Pathol. Mi­

crobial Scand. 70:398-404; 1967. 17. Kochhar, D. M. Limb development in mouse embryos. 1. Analysis of teratogenic effects of retinoie acid. Te ratology 7:289-29B; 1973. 18. Langille, R. M.; Paulsen, D. F. ; Solurs h, M. Differen tial effeets of physiological coneentrations of re tinoic acid in vitro on ehondrogen­ esis and myogenesis in ehiek c raniofacial mesenchyme. Differentia­ tion 40:84-92; 1989, 19, Lee, J.; Tiekle, C. Retinoie aeid and pattern formation in the develop­ illg c hick wing: SEM and qnanlitativc studies of early effects on the apical ectodermal ridge and bud ontgrowth. J. Embryo!. Exp. Mor­

pho!. 90,139-169; 1985. 20. Leonard. C. M, ; Bergman, M.; Frenz, D. A., e t a!. Abnormal ambie nt glueose levels inhibit proleoglycan core protein gene expression and reduee proteoglycan aceumulation during chondrogenesis: poss ible mechanism for tel'3.togenic effects of maternal diabetes. Proe. Nail.

Aoad. Sc;. USA 86,10113 -1 0 11 7; 1989. 21. Lev. R. ; Spicer, S, S. Speeific staining of snIfate groups I"jlh alcian blue allow pH. J. Histochem. Cytochem. 12:309; 1964. 22. Maden, M.; Ong, D. E.; Summerbell, D" ct al. Spatial distribution of eellular protein binding to retinoic aeid in thechiek limb bud, Nature

335,733- 735; 1988. 23, Mangelsdorf, D. 1.; Ong, E. S.; Duek, 1. A., et al Nuc\earreceptor tha t identifies a novel relinoic acid response pathway. Na tnre 345:224­ 229; 1990. 24. Napoli, 1. L.; Posch, K. P.; Fiorella, p , D. , ct at Physiological occm ­ renee, biosynthesis and metabolism of retinoie acid: evide nce for roles of ceUular rctinol-bindi ng protein (CRBP) and cellula r retinoie acid·bi nding protein (CRABP) in the pathway of retinoic acid homeo­ stasis. Biomed. Pharmacother. 45:131 - 143; 1991. 25. Newman, S. A. Lineage and pattem in the developing wing bud. In: Ede, D. A, ; Hinchliffe, J. R; Balls, M., cds. Vertebrate limb a nd somite morphogenesis. Cambridge: Cambridge University Press;

19n 181-200. 26. Newman, S. A.; Pa uloll, M,.P.; Ki eny, M. The distal boundary of myogenic primordia in chimeric avian limb buds and its relation to an accessible population of cartilage progenitor ceUs. Dev. Biol.

84,440-448; 198 1. 27. Noji, S.; Nohno, T.; Koyama, E., e t aI. Re1inoic acid indnces polarizing activity but is unlikely to be a morphogen in the chick limb bud. Natnre 350:83- 86; 1991. 28. Paulsen, D. F,; Parker, C. L.; Finch, R. A. Region-dependent capacilY for limb chondrogenesis: patterns of chondrogenesis in enilures from different regions of the developing c hick wi-ng. Differentiation

lH59- 165; 1979.

29. Paulsen, D. F.; Langille, R. M, ; Dress, V., et al. Selective s timulation of in vitro limb-bud chondroge nesis by retinoic acid. Differentiation 39, 123-130; J988. 30, Panlsen. D. F.; Solursh, M. Microtiter micromass cnllnres of limb-bud mesenchymal cells. In Vitro Cell. Dev. Bio1. 24: 138-14 7; 1988. 3 1. Pa ulsen, D. F.; Pang, L. Positiou-related eft'ects of re tinoic acid (RA) on chick limb-bnd chondrogenesis in serum-free microeulture . Anal. Rec, 226:78A; 1990 , 32. Paulsen, D. F,; Chen. W.-D.; Pang, L., ela!. Unique responsiveness of dis tal antcrior chick wing-bud mesenc hyme to retilloi e acid a nd serum in vilro. AI13t. Rec. 229:68A; 1991. 33. Sakai, A. ; LangiUe, R, M. Differe ntial and s tage dependent effects of retinoie acid on chondrogcnesis a nd syn thesis of extraceUnla r ma1rix macromolecules in chick craniofacial mesellchyme in vitro. Differ­ e ntia tion 52:19- 32; 1992. 34. Saunders, J. W., .Ir, The proximo-distal seqnellee of the oligin of thc parts of the chick wing a ud Ine role of the eetoderm. J. Exp. Zool. 108,3 63-404 ; J 948. 35, Sc hofield, J. N.; Rowe, A.; Brickell, P. M. Position-dependence of retmoic acid receptor-.B gene expression in the ehiek limb bud, Dev.

Bio!. 152,344- 353; 1992. 36. Slack, J. M, W. We have a morphogen! NatrJre 32 7:553-554; 1987. 37. Smith, S. M.; Eiehele, G, Temporal and regional differences in the ex.press ion pattern of distinct retinoic acid receptor-,B lra nscripts in the chick embryo, Developme nt 111:245-252; 1991. 38. SnLik, K. K.: Dehart, D. B, Retinoic acid induced limb malformations resulting from apical ectode rmal eell death. Tera tology 37:527­

537; 1988. 39. Sllmmerbeli . D. The effect of local application of Jetinoic acid lo the anterior margin of the developing ehick Limb. J. Embryo!. Exp.

Mm·phoL 78,269-289; 1984. 40. SummerbeU . D.; Wolpert, L. Cell density a nd cell division in the early morphogenesis of the chick wing. Nalure 239:24- 26; 1972. 41. SUZIlki, H. R. ; Padanilam, B, J.; Vitale, E. , e t a1. Repeating develop­ men tal expression ofG-Hox 7, a novel homeobox containing gene in the chicken, Dev, Biol. 148:375- 388; 1991. Eichele, G. Ide ntifieation and spatial dis tribution of reti­ 42. Thalle r, noids in the developing ch ick limb . Na1ure 327:625- 628; 1987. Alberts, B.; Wolpert, L., el al. Loeal application of retinoic 43 . Tickle, aeid to the limb buds mimics the action of the polarizing regi on. Natme 296:564-566; 1982. 44. Tickle, C. ; Lee. J.: Eichele, G. A qnantita tive analysis of the effect of all-trans retinoic acid on the paUe rn of chiek wing developme nt. Dev. Biol 109:82- 95; 1985. 45. Tickl e, c.; Crawley, A.; Farrar, J. Retinoic acid application to ehiek wing buds leads to a dose-dependent reorganization of the apical ectodermal ridge that is mediated by the mesenchyme. Development

c.; c.;

106,691-705; 1989. 46. Uma nsky, R. The effect of eel! population density on the developm..:n­ tal fate of reaggregaling mou se limb-blld mesenc hyme. Dev. BioI. 13,31-56; J 966. 47. Wanek, N.; Gardine r. D. M. ; Muneoka, K., e t al. Conversion by reti­ noic acid of ante rior cells to ZPA cells in the cruck wing bud. Nature 3 50,81-83; J 991. 48, Yok ouchi , Y.; Sasaki, H.; Knroiwa, A. Homeobox gene expression eorrelated with the bifurcation process of limb cartilage develop­ ment. NalUre 353:443-445; 1991.