The odontoblasts form the most peripheral celllayer of the dental papilla. Their terminal, func-tional, di�erentiation implies a withdrawal from thecell cycle, an elongation-polarization of the post-mitotic cells, and an up-regulation and/or initia-tion of the synthesis and polarized secretion ofpredentin-dentin components (for review, see ref. 1).In the Swiss mouse embryonic lower incisors, thisterminal di�erentiation is initiated at stage E-16 atthe anterior part of the teeth and progresses in aposterior direction. A gradient of di�erentiatingodontoblasts exists which includes, in antero-poster-ior direction, functional odontoblasts, polarizingodontoblasts and dividing preodontoblasts.

The ameloblasts derive from the inner dentalepithelium. Their terminal di�erentiation, whichalso implies withdrawal from the cell cycle,polarization, and polarized secretion of enamelcomponents (see ref. 2 and references therein), isrestricted to the labial aspect of the incisors and isinitiated in the presence of predentin. Consequentlywith regard to the odontoblasts, there exists anon-wedged, temporized gradient of ameloblastterminal di�erentiation.

Preodontoblasts-odontoblasts and preamelo-blasts-ameloblasts constitute a developmentalunit; reciprocal epigenetic signalling involving bothgrowth factors (including members of the TGFb

In vitro synchronizationof embryonic mouseincisor preodontoblastsand preameloblasts:repercussions on terminaldifferentiationSchmitt R, Ruch J-V. In vitro synchronization of embryonic mouse incisorpreodontoblasts and preameloblasts: repercussions on terminal di�erentiation.Eur J Oral Sci 2000; 108: 311±319. # Eur J Oral Sci, 2000

Preodontoblasts divide asynchronously and their terminal di�erentiation occursgradually. Experimental data suggested that the expression of competence bypreodontoblasts to respond to speci®c epigenetic signals, triggering their overtdi�erentiation, requires a minimal number of cell cycles. The intrinsic timingmechanism could imply division counting and preodontoblasts of juxtaposed cellgenerations might sequentially withdraw from the last physiological cycle. To testsuch an hypothesis, embryonic mouse lower incisors were cultured in vitro andtreated sequentiallywith nocodazole in order to induce a transitory synchronizationof the dividing preodontoblasts and preameloblasts. This synchronization led toa disorganization of the physiological gradual terminal di�erentiation of theodontoblasts, giving rise to three distinct domains comprising respectively:1) odontoblasts with altered polarization and predentin secretion; 2) odontoblastsdemonstrating equivalent polarization and predentin deposition; and3) preodontoblasts-odontoblasts involved in gradual terminal di�erentiation.These results strongly suggest that the gradient of odontoblast functionaldi�erentiation results from sequential withdrawal from asynchronous cell cyclesof competent cells able to overtly di�erentiate.

ReÂgine Schmitt, Jean-Victor Ruch

INSERM U424, Institut de Biologie MeÂdicale,Faculte de MeÂdecine de Strasbourg, France

ReÂgine Schmitt, INSERM U424, Institut deBiologie MeÂdicale, Faculte de MeÂdecine,11, rue Humann, 67085 Strasbourg Cedex,France

Telefax: z33±88±257817E-mail: Regine.Schmitt@medecine.u-strasbg.fr

Key words: odontoblast; in vitro;synchronization; nocodazole; differentiation

Accepted for publication April 2000

Eur J Oral Sci 2000; 108: 311±319Printed in UK. All rights reserved

superfamily) and matrix components trigger odon-toblast terminal di�erentiation ®rst, and later on,ameloblast functional di�erentiation (for reviews,see refs. 1, 2).

The temporo-spatial interplay of epigenetic con-trol mechanisms and the expression of speci®ccellular competence allowing for the establishmentof the gradients are, however, not well understood.The gradual ameloblast terminal di�erentiationappears to be (at least partially) a consequence ofthe odontoblast functional di�erentiation. Thesequential terminal odontoblast di�erentiationmight be explained in two ways. The epitheliallyderived epigenetic signal(s) reaches an operationalthreshold ®rst at the anterior part of the incisorand then progressively in a posterior direction. Inother words, there could be a spatial gradationof inducibility and an even distribution of com-petent preodontoblasts. Alternatively, the opera-tional epigenetic signal has an even distributionand the responsive, competent preodontoblastsemerge sequentially. Most of the current experi-mental data (reviewed in ref. 1) support the secondpossibility. The gradual emergence of competentpreodontoblasts might be related to cell kinetics:the crown size and morphology of a particulartooth is a function of the number of postmitoticodontoblasts and ameloblasts and of their spatialdistribution.

Histological investigations, combined with[3H]thymidine radioautography of in vitro culturedheterochronal enamel organ-dental papilla recom-binations (3±5), have suggested that the expressionof the competence of preodontoblasts to respond tospeci®c epigenetic signals triggering terminal di�er-entiation requires a minimum number of cell cycles.Terminal di�erentiation cannot be anticipated inheterochronal tissue recombinations. On the otherhand, in such experimental conditions supplemen-tary cell cycles do not hamper terminal di�eren-tiation and may facilitate regulatory events. Thelast division of preodontoblasts appears to beasymmetric; the spindle is oriented perpendicular tothe basement membrane, the two daughter cells aresuperimposed, and only the cell in contact with thebasement membrane will di�erentiate overtly (5, 6).The withdrawal from the cell cycle of odontoblastsfrom asynchronously dividing preodontoblasts willbe sequential (7). The speci®c competence expressedby postmitotic daughter cells may result either fromthe expression of a particular combination of sig-nalling receptors or from speci®c transductional-posttransductional steps. One way to test the cellkinetics-dependent hypothesis would be by experi-mental synchronization of dividing preodontoblasts.

A trial of in vitro synchronization using noco-dazole according to MACAULEY et al. (8) has

been performed. Notwithstanding the e�ects ofnocodazole on the polarization of odontoblastsand ameloblasts and the transitory synchroniza-tion of both preodontoblasts and preameloblasts,the observed disruption and/or transitory abroga-tion of the gradient of odontoblast overt di�er-entiation supports the hypothesis of their cellkinetics-dependent emergence.

Materials and methods

Tooth germs

Laboratory inbred Swiss mice were timed-matedand the day of ®nding a plug designated day zero.Embryonic lower incisors were removed on day 14(E-14) of gestation. The left and right cap-stageincisors were dissected together including the mostanterior part of Meckel's cartilage.

Materials

Nocodazole (Sigma-Aldrich, St. Louis, MO, USA)was used to block the passage through mitoses.Dimethylsulphoxide (DMSO; Braun, Boulogne,France) was used to solubilize nocodazole. Thestock solution of nocodazole was 0.5 mg noco-dazole in 2 ml DMSO. Cell proliferation wasinvestigated by mapping the S-phase cells afterincorporation of 5-bromo-2-deoxy-uridine (BrdU)using a cell proliferation kit (Amersham LifeScience, Les Ulis, France).

Organ culture

The E-14 lower incisor pairs (left and rightincisors) were cultured on 2 ml of semi-solidmedium per Petri dish (Nunc, Roskilde, Denmark;35610 mm). The medium consisted of BGJ-B(Fitton Jackson modi®ed; Gibco, Cergy-Pontoise,France) supplemented with ascorbic acid 0.18 mg/ml (Merck, Darmstadt, Germany), L-Glutamin2 mM (Seromed, Berlin, Germany), foetal calfserum 20% (Boehringer Bioproducts, Gagny,France), kanamycin 0.1 mg/ml (Gibco) and Difcoagar (0.5%). The teeth were incubated and grownat 37³C in a humidi®ed atmosphere of 5% CO2 inair. The medium was changed every 2 d.

Several concentrations of nocodazole were testedafter 1, 2 or 3 d of preincubation of the teeth.Knowing that the average cell cycle duration ofpreodontoblasts and preameloblasts in vitro isabout 20 h, nocodazole was applied for 20 h. Thebest results, highest rate of arrested cells, and rapidrelease from nocodazole was observed with0.25 mg/ml of nocodazole. This concentration wasused in all subsequent experiments. The ®nal

312 Schmitt & Ruch

concentration of DMSO alone had non-dicernablee�ects as judged by light microscopy. BrdU wasused at 3 mg/ml of culture medium.Finally, the precise modalities of the performed

cultures were: E-14 explants were ®rst culturedfor 2 or 3 d in normal medium, followed by 20 hof culture in the presence of nocodazole (test) orcontrol medium (controls), respectively. Specimensto be processed by histology only were then furthercultured in normal medium for 1, 2, 3 or 4 d.Specimens intended for BrdU labelling were treatedin the same way, except that after the 20 hnocodazole or control medium incubation, theywere further cultured for 8, 24 or 48 h, followedby a 2-h BrdU pulse-labelling. For each particularculture condition, at least 3 pairs of incisorswere used.

Histology and immunohistochemistry

The specimens were ®xed in Bouin-Hollande ¯uid,embedded in para�n wax, and cut serially in 5-mm-thick sagittal sections. Histological staining wasperformed with Mallory's Alun hematoxylin.BrdU incorporated into DNA was detected on

the de-waxed sections with a speci®c mouse mono-clonal antibody and immunoperoxidase labellingfollowing the manufactor's instructions (AmershamLife Science). After immunostaining the sectionswere counterstained with eosin.

Cell counting

After 20 h of nocodazole treatment, the ratio ofcells in mitotic arrest was evaluated by cellcounting. All preameloblasts, preodontoblasts andthe respective mitotic arrests were counted, indistinct areas of 55 mm640 mm using a 640objective, on each third section. All counts wererepeatedon serial sections of 3 incisors. The standarddeviation was calculated according to:

� �����������������������p�100ÿ n�

n

rwhere p represents the experimental proportion inpercentage and n the number of examined cells.

Results

Effects of nocodazole after 2 d preincubation

Taking into consideration the particularities ofrodent incisors, we focused our attention on thelabial aspect of the teeth, where both odonto-blasts and ameloblasts di�erentiate. All the E-14incisors were not at the same developmental stage.

Statistically after 2 d of culture, functional odonto-blasts were not yet present. However the ®rstpostmitotic, polarizing odontoblasts could beobserved at the anterior end of some teeth.

E-14z2 dz20 h: In the controls, polarizingodontoblasts were present (Fig. 1A, C). After20 h of nocodazole treatment, 52.5¡1.9% of thepreameloblasts and 33.4¡1.6% of the preodonto-blasts demonstrated mitotic arrest (Fig. 1B, D).

E-14z2 dz20 hz2 d: In the controls, the gradi-ent of polarizing odontoblasts was fairly evident(Fig. 1E). On the other hand, after 2 d of releasefrom the nocodazole treatment, the odontoblastlayer appeared to be subdivided into three moreor less distinct areas: a) the most anterior onecomprised odontoblasts demonstrating a disturbedpolarization; b) an intermediate area includedrather similarly polarized odontoblasts; c) the mostposterior area included some postmitotic odonto-blasts and dividing preodontoblasts (Fig. 1F).The inner dental epithelium comprised dividingpreameloblasts (Fig. 1E, F).

E-14z2 dz20 hz3 d or 4 d: In the controls, thegradient of di�erentiating odontoblasts extendedinto the posterior direction including (in antero-posterior direction) functional, polarizing, post-mitotic odontoblasts and dividing preodontoblasts.The antero-posterior gradient of polarizing amelo-blasts, superimposed on the gradient of predentinaccumulation, was obvious (Fig. 1G, I).

After 3±4 d of release from nocodazole, theodontoblast layer was clearly subdivided into threeconsecutive antero-posterior domains: a) odonto-blasts demonstrating disturbed polarization andpredentin secretion; b) equivalent functional odon-toblasts; and c) odontoblasts exhibiting the normalgradient of functional di�erentiation (Fig. 1F, H, J).

In most cultured incisors, the polarization of theameloblasts, facing the abnormal odontoblasts, wasseverely a�ected, but their polarization appearedquite normal in front of the uniformly functionalodontoblasts (Fig. 1G±J).

Effects of nocodazole after 3 d of preincubation

E-14z3 dz20 h: In the controls, functional odon-toblasts kept close to polarized and polarizingodontoblasts. The gradient of odontoblast terminaldi�erentiation was obvious. The ameloblasts facingpredentin demonstrated polarization (Fig. 2A).

The e�ects of the 20-h nocodazole treatment areshown in Fig. 2B. Polarized odontoblasts werepresent at the anterior end of the incisor. In a moreposterior direction, the polarization of postmitoticodontoblasts was disrupted and dividing preodon-toblasts demonstrated mitotic arrest. The mitoticarrest of preameloblasts was obvious (Fig. 2B).

Synchronization of preodontoblasts and preameloblasts in vitro 313

314 Schmitt & Ruch

The ratio of mitotic arrest was respectively52¡1.5% and 32.4¡1.9% for preameloblasts andpreodontoblasts.E-14z3 dz20 hz1 d, 2 d or 3 d: In the controls,

the progressive, gradual, posterior extension offunctional odontoblasts was evident, as well as thedelayed posterior extension of ameloblast polariza-tion (Fig. 2C, E). The ®rst functional ameloblastswere seen at stage E-14z3 dz20 hz3 d (Fig. 2G).The e�ects of nocodazole after 1, 2 or 3 d of

release are shown in Fig. 2D, F, H, respectively.At each stage, a disrupted gradient of odontoblastdi�erentiation was observed. The odontoblastlayer was always subdivided into a most anteriordomain, comprising rather normal functionalodontoblasts (Fig. 2D), followed in a posteriordirection ®rst by a) abnormal, functional, odonto-blasts secreting an irregular layer of predentin, andthen by b) equally polarized and/or functionalodontoblasts, and ®nally by c) a normal lookinggradient of polarizing odontoblasts. The stage-related di�erences (1, 2 or 3 d of release fromnocodazole) were re¯ected in the amount ofpredentin and the posterior extension of thegradually polarizing odontoblasts.As far as the ameloblasts are concerned, some

polarized ameloblasts, albeit without enameldeposition, were observed at the anterior end. Infront of the irregular predentin layer polarizingameloblasts were present, and facing polarizingodontoblasts and preodontoblasts the preamelo-blasts had a normal histological appearance(Fig. 2D, F, H).

BrdU incorporation

A 2-h pulse labelling with BrdU was performedafter 8, 24 and 48 h following the nocodazoletreatment and in corresponding control cultures.Eight h after the nocodazole treatment only a veryfew cells were labelled (not shown). Twenty-four h

after nocodazole treatment, adjoining preodonto-blasts and preameloblasts were labelled (Fig. 3A).The labelling was discontinous and less cells werelabelled in corresponding controls (Fig. 3B). Forty-eight h after nocodazole treatment, more cellswere labelled than in corresponding controls.However, this labelling occurred in scattered cells(Fig. 3C, D).

Discussion

The gradual odontoblast and ameloblast terminaldi�erentiation can be observed easily at the labialaspect of sagittal sections of developing incisors.

Preodontoblasts and preameloblasts divide asyn-chronously. These asynchronous cell divisions leadto the periodic antero-posterior distribution ofmitoses as shown in Fig. 4A. According to ourworking hypothesis, the sequential withdrawalfrom the cell cycle could explain the gradualemergence of overtly di�erentiating odontoblasts(Fig. 4B). The timing of our experiments wasadapted to try both to synchronize the preodonto-blasts in the absence or in the presence of alreadypostmitotic, polarizing or functional odontoblastsand to follow the temporal-spatial behavior of theprogressively emerging postmitotic odontoblastswhich were synchronized either during their lastcell cycle or during earlier ones. The nocodazoletreatment of intact incisors a�ected both thepreodontoblasts and the preameloblasts. Theoreti-cally to override this inconvenience, the dentaltissues (dental papillae and enamel organs) could beisolated and synchronized individually and thenrecombined in vitro (i.e. synchronized dentalpapillae recombined with normal enamel organsand vice-versa). However, the control of preodon-toblast and preameloblast cell kinetics is depen-dent on heterotopic cell interactions (9), and thecell kinetics of isolated dental tissues become

Fig. 1. Sagittal sections of E-14 incisors cultured for 2 d on control medium, then for 20 h (B, D) in presence of nocodazole and®nally for 2, 3 or 4 d on control medium (F, H, J). A, C, E, G, I corresponding controls. The anterior part of the teeth is consistentlyoriented to the right. (A, C) Low and high magni®cation of an E-14z2 dz20 h control incisor: the gradient of di�erentiatingodontoblasts is initiated. (B, D) Low and high magni®cation illustrating the e�ect of 20 h nocodazole treatment. The mitotic arrest isobvious. Control cultures (E, G, I) demonstrate the progressive posterior extension of the gradient of odontoblast terminaldi�erentiation. The gradual polarization of the ameloblasts occurs in the presence of predentin (G, I). After nocodazole treatment thedisorganization of the gradients is obvious (F, H, J). With time, three distinct areas of odontoblast di�erentiation may be identi®edmore and more distinctly:

a) Disturbed polarization and predentin secretionb) Odontoblasts equivalent as far as their polarization and predentin secretion is concernedc) Gradual cytological and functional di�erentiation.

Lab, labial; Ling, lingual; pd, predentin; PO, preodontoblasts; pO, polarizing odontoblasts; O, odontoblasts; PA, preameloblasts;pA, polarizing ameloblasts; DP, dental papilla. Scale bar, 100 mm.

Synchronization of preodontoblasts and preameloblasts in vitro 315

disturbed (3, 4). Furthermore, to perform reallyisotopic recombinations appears to be quiteimpossible.

In vitro the mean value of the cell cycle duration(TC) for preodontoblasts and preameloblasts has

been shown to be about 20 h (10). According tothese data, the nocodazole treatment lasted 20 h totry to arrest most of the cycling cells. The growthfractions for preodontoblasts and preameloblastshave been evaluated to be respectively about 40%

316 Schmitt & Ruch

and 50% (11). The rates of observed mitotic arrestswere compatible with these values.Obviously the nocodazole treatment had a

double e�ect. The cytological polarization of post-mitotic odontoblasts and ameloblasts as well as

the polarized secretion of predentin were disturbedand dividing cells demonstrated mitotic arrest. Thee�ects on polarization were foreseen, knowing theaction of colcemid (and cytochalasin B) on dentalcytodi�erentiation (12).

Fig. 2. Sagittal sections of E-14 incisors cultured for 3 d on control medium, then for 20 h (B) in presence of nocodazole and ®nallyfor 1, 2 or 3 d on control medium (D, F, H). A, C, E, G corresponding controls. The anterior part of the teeth is oriented to the right.(A) The antero-posterior gradient of di�erentiating odontoblasts is obvious in this control incisor. The ameloblast polarization isinitiated in front of predentin. (B) The mitotic arrest after 20 h nocodazole treatment is obvious. The most anterior odontoblasts werealready postmitotic at the onset of nocodazole treatment. In the control cultures (C, E, G), the progressive continuous posteriorextension of the gradient of odontoblast terminal di�erentiation is obvious. The gradient of polarizing ameloblasts is superimposedto the gradient of predentin secreted by functional odontoblasts. The ®rst functional ameloblasts secreting enamel components (En)are located at the anterior part of the incisor (G). After nocodazole treatment, a disorganization of the gradient of odontoblastterminal di�erentiation is observed (D, F, H). Three distinct areas coexist:

a) The most anterior one: functional odontoblasts demonstrate disturbed polarization and irregular predentin secretion. With time,this area has a more posterior localization.

b) The more posterior shorter intermediate area comprises odontoblasts equivalent as far as their polarization (D) and laterpredentin secretion (F, H) are concerned.

c) The most posterior area comprises preodontoblasts and odontoblasts progressively demonstrating the physiological gradient ofterminal di�erentiation.

The ameloblasts demonstrate more or less disturbed polarization in presence of the irregular predentin. The preameloblasts±ameloblasts superimposed to the areas b and c have a normal histological aspect. Lab, labial; Ling, lingual; pd, predentin; d, dentin;En, enamel; pO, polarizing odontoblasts; O, odontoblasts; pA, polarizing ameloblasts; A, ameloblasts; DP, dental papilla. Scale bar,100 mm.

Fig. 3. Sagittal sections illustrating BrdU incorporation in E-14 incisors cultured in vitro in absence or presence of nocodazole. Theanterior part of the teeth is oriented to the right. (A, C) control incisors, pulse-labelled, respectively after 2 dz20 hz1 d (A) or3 dz20 hz2 d (C) on control medium. Scattered BrdU-labelled preodontoblasts and preameloblasts exist. These cycling cells arelocated in a more posterior position in the older incisor (C). (B, D) Nocodazole-treated incisors pulse-labelled respectively after2 dz20 h nocodazolez1 d (B) or 3 dz20 h nocodazolez2 d (D). (B) After 1 d of release from nocodazole, the BrdU pulse revealslabelling of rather adjacent preodontoblasts and preameloblasts. More cells are labelled than in the corresponding controls (A). (D)After 2 d of release from nocodazole, the BrdU pulse leads to scattered labelled cells. Again more cells are labelled than in thecorresponding control (C). Lab, labial; Ling, lingual; PO, preodontoblasts; PA, preameloblasts. Scale bar, 100 mm.

Synchronization of preodontoblasts and preameloblasts in vitro 317

During release from nocodazole mitotic arrest thecycling cells should achieve mitoses, and thestill cycling cells should proceed through the nextcycle synchronously. The BrdU incorporationexperiments indicated that the re-entering to theS-phase occurred later than after 8 h of release;after 24 h, most of the cycling cells were still inS-phase. After 48 h the BrdU labelled cells werescattered again. Most probably, the synchroniza-tion was operational for one cycle only. Thistransitory synchronization a�ected cycling pre-odontoblasts and preameloblasts. The synchronizedpreodontoblasts next to the more anterior post-mitotic, polarizing odontoblasts accomplished theirlast physiological mitosis and, most interestingly,the competent postmitotic daughter cells overtlydi�erentiated synchronously, whilst the more pos-terior preodontoblasts involved again in hetero-chronous cycling gave rise to gradually polarizingodontoblasts.

Consequently, our results suggest that thetransitory synchronization of the preodontoblastsleads to a transitory suppression of their normal

gradual terminal di�erentiation. However, sincein our experiments the transitory synchronizationa�ected both the preameloblasts and the pre-odontoblasts, the particular behavior of theodontoblasts might be a consequence also of thesynchronization of the preameloblasts. Neverthe-less, considering that gradual terminal di�erentia-tion of odontoblasts can be triggered in vitro in theabsence of the preameloblasts but in presence ofimmobilized, uniformly distributed active growthfactors including TGFb-1 or 3, BMP-2 or 4(13, 14), we strongly believe that the synchroniza-tion of the preodontoblasts undergoing their lastcycle leads to the synchronous emergence ofcompetent postmitotic odontoblasts able to overtlydi�erentiate in presence of an evenly distributedmultifactorial epigenetic control.

It will be of further importance to investigatewhether the synchronization of preameloblasts, inphysiological conditions involved in the control ofodontoblast terminal di�erentiation, might a�ecttheir transcriptional-translational and secretoryactivities.

Fig. 4. (A) Sagittal sections of an E-17 in vivo incisor. The periodic distribution of mitoses of asynchronously dividing preameloblastsand preodontoblasts is shown (PA, PO). (B) Schematic drawing of the sequential withdrawal of odontoblasts (O) and ameloblasts(A) from the asynchronous cell cycles and the possible cell kinetics-related establishment of the gradients. Scale bar, 100 mm.

318 Schmitt & Ruch

Acknowledgements ± We wish to thank Dr. A. Gritli-Linde whosuggested such an experimental approach, Pr. A. Linde forcritical reading of this manuscript, Dr. B. Senger for statisticalhelp and Mr. A. Ackermann for technical help. This researchwas partially ®nanced by the International Human frontierScience Program (grant TG-558/95 M) and by the FondationDentaire de France (UB/SS 500144±98002598).

References

1. RUCH JV. Odontoblast commitment and di�erentiation.Biochem Cell Biol 1998; 76: 923±938.

2. COIN R, HAIKEL Y, RUCH JV. E�ects of apatite, transform-ing growth factor bone morphogenetic protein-2 andinterleukin-7 on ameloblast di�erentiation in vitro. Eur JOral Sci 1999; 107: 487±495.

3. RUCH JV, KARCHER-DJURICIC V. Mise en e vidence d'un roà lespe ci®que de l'epithelium adamentin dans la di�e rencia-tion des odontoblastes. Ann Embryol Morphogen 1971; 4:359±366.

4. RUCH JV, KARCHER-DJURICIC V, THIEBOLD J. Cell divisionand cytodi�erentiation of odontoblasts. Di�erentiation 1976;5: 165±169.

5. RUCH JV, LESOT H, KARCHER-DJURICIC V, MEYER JM,OLIVE M. Facts and hypotheses concerning the controlof odontoblast di�erentiation. Di�erentiation 1982; 21: 7±12.

6. RUCH JV. Determinisms of odontogenesis. Cell Biol Rev1987; 14: 1±112.

7. RUCH JV. Patterned distribution of di�erentiating dentalcells: facts and hypotheses. J Biol Buccale 1990; 18: 91±98.

8. MACAULEY A, WERB Z, MIRKES PE. Characterization ofthe unusually rapid cell cycles during rat gastrulation.Development 1993; 117: 873±883.

9. OLIVE M, RUCH JV. Does the basement membrane controlthe mitotic activity of the inner dental epithelium of theembryonic mouse ®rst lower molar. Dev Biol 1982; 93:301±307.

10. AHMAD N, RUCH JV. Comparison of growth and cellproliferation kinetics during mouse molar odontogenesisin vivo and in vitro. Cell Tissue Kinet 1987; 20: 319±329.

11. OSMAN A, RUCH JV. Contribution aÁ l'e tude des parameÁ tresdu cycle cellulaire au cours de l'odontogeneÁ se chez la souris.J Biol Buccale 1978; 6: 43±54.

12. RUCH JV, KARCHER-DJURICIC V, STAUBLI A, FABRE M.E�ects of cytochalasin B and colchicine on dentalcytodi�erentiation in vitro. Arch Anat Microsc MorpholExp 1975; 64: 113±134.

13. BEGUE-KIRN C, SMITH AJ, RUCH JV, WOZNEY JM, PURCHIO

A, HARTMANN D, LESOT H. E�ects of dentin proteins,transforming growth factor beta 1 (TGF beta 1) and bonemorphogenetic protein 2 (BMP2) on the di�erentiation ofodontoblast in vitro. Int J Dev Biol 1992; 36: 491±503.

14. BEGUE-KIRN C, SMITH AJ, LORIOT M, KUPFERLE C,RUCH JV, LESOT H. Comparative analysis of TGFbs,BMPs, IGF1, msxs, ®bronectin, osteonectin and bonesialoprotein gene expression during normal and in vitro-induced odontoblast di�erentiation. Int J Dev Biol 1994; 38:405±420.

Synchronization of preodontoblasts and preameloblasts in vitro 319