EU Research

Developmental biology and building a toothIrma Thesleff, DDS, DrOdont'

During the lasf 15 years, we have started to understand tooth deveiopment at the gene level. The list ofgenes known to reguiale tfie position, shape, or number of teeth Is lengthening rapidly. Interestingly, so farall these genes have important ¡unctions in the mediation ot cell communication, which is generaliy con-sidered the most important mechanism driving embryonic deveiopment,The communication is mediatedby smail signal molecules that are sent to nearby cells, thereby affecting their behavior and advancing dif-ferentiation,There are dozens of different signals and fheir receptors and target genes, which togetherform complicated signaling networks. The defects in several human conditions affecting tooth developmenthave been identified recently, and these genes have turned out to be necessary components of signalingnetworks. Experimental studies using transgenic mice as models for human syndromes such as ectoder-mal and oleidocranial dysplasia have pinpointed the exact roles of the disease genes and indicated waysfor possible new therapies. It is also possible that by combining the knowledge of molecuiar regulation oftooth development with the recent breakthroughs in stem cell research, dreams of building new teeth indental practice may come true in the future, (Quintessence int 2003:34:613-620)

Deveiopmental biology is at present one of tbemost rapidly progressing fields in biology and bio-

medicine. Tbe advances of gene technology have ledto a rapid explosion in the understanding of the mole-cular mechanisms regulating embryonic development.New genes and their functions are continuousiy beingdiscovered in experimental studies using animal em-bryos, and molecular genetic studies in humans areunraveling gene mutations causing congenital defects.Today animal models can be created for human dis-eases and new possibilities are opening up for preven-tion, diagnosis, and treatment of congenital defects. Inaddition, tbere is now great hope that the knowledgeon the molecules driving tissue and organ develop-ment and eell differentiation will lead to tools for tis-sue regeneration and stem cell therapies in the future.It may even be possible to grow whole new organs,such as teeth.

'Director, Developmentai Sioiogy Prograrrîme, institute of Biotechnology,University oí i-lelsinki, Helsinki, Finiand.

Reprint requests: Dr Irma Thesieff, Institute of Biotechnology, POB56,University of Heisinki. 00014 Helsinlii, Finlanfl. E-maii. irma.thesleff®helsinki.fi

This artide was originaiiy pubiished as "How to buiid a fooift?Developmentai biology is reveaiing the instructions." In: Scfiou L (ed).Nordic Dentistry 2003 Yeartooit. Copenhagen: Quintessence.2003:117-128.

I entered the exciting field of developmental biol-ogy 30 years ago as an undergraduate dentai sfudentand bave been fortunate to be part of its dramaticprogress, wbicb realiy can be called a revolution. Intbe 1980s, it became possible to study deveiopment attbe level of genes, and witb excellent students in mygroup, we combined gene technology witb classicalmethods of experimental embryology. We dissecteddental tissues ftom mouse embryos and cultured tbemin various conditions, and studied the expression ofgenes. Over the years, we analyzed the functions ofmany different genes by various approaches. Recently,we used transgenic mouse technology to examine thefunctions of some genes, in particular genes that affectthe number of teetb in humans (see following sectionson cieidocranial dysplasia and ectodermal dysplasia).

The fascinating concept of a common moleculartoolkit regulating development evolved during the1980s and 1990s, It was realized that same genes reg-ulate developmental processes in all animals and in alldifferent organs and tissues. Our group contributed tothis concept by unraveling the roles that many impor-tant molecules belonging to the conserved develop-mental toolkit have in teeth,'-^ This indicated that theteeth are no exception of the rule; on the contrary, it isnoteworthy that no "tooth-specific" regulatory mole-cules have so far been identified (besides some thatare involved in the formation of dentin and enamel).

Quintessenceinternational 613

• Thesleff

Initiation

Morphogenesis

Dentallamina

AmeioblastOdontoblast

Cytodifferentiation

Matrix secretion"'

Fig 1 Morphogenesisn th

cf a too hphogeneInteractions between the epithelialand mesenchymal tissues regulateadvancing development.

Fig 2 Secreted signal molecules are the most important media-tors of cell communication during development. Binding oí thesignal fo a specific receptor ai the surlace of the responding cellleads tc cfianges in ifs behavioi

In other words, all currently known genes that affectthe position, shape, or number of teeth also have de-velopmental regulatory functions in other tissues.

SIGNALING MOLECULES:KEY REGULATORS OF TOOTH DEVELOPMENT

The special aim of our research has heen to elucidatethe mechanisms whereby cells communicate during

development. It has been known for more than a cen-tury that cells signal to each other and thereby directceils to new pathways, thus controlling advancementof development. This has heen called embryonic in-duction, and today this is believed to be the singlemost important mechanism of developmental regula-tion. Many embryologists showed already in the 1950sand 1960s that intercellular signaling is an importantfeature of tooth development and that in teeth, the sig-naling events take place mainly between the epithelialand mesenchymal tissue components (Fig I).""-*

The developmental signal molecules are small pro-teins that usually act by binding on specific receptorsat the surface oí the responding cells (Fig 2). A multi-step intracellular cascade leads to regulation of geneexpression in the nucleus, and the cell then changes itsbehavior. Our work has pinpointed the roles of severalsignal molecules and their targets during the earlysteps of tooth development when teeth are initiatedand their shapes are determined. We have proposed amodel on how conserved signaling molecules mediatesequential and reciprocal interactions hetween dentalepithelium and mesenchyme and thereby regulate ad-vancing tooth morphogenesis (Fig 3), The model isbased on results from many laboratories, including ourown.̂ The most studied signals belong to four differentfamilies including fihroblast growth factor (FGF),hone morphogenetic proteins (BMP), hedgehog, andWnt, In addition to signals, the model in Fig 3 also in-cludes several genes, which are regulated by the sig-nals in the responding tissues. Mutations in many ofthese genes already have been shown to cause dentaldefects in mice as well as in humans,^

614 Volume 34, Number 8, 2003

• Theslefl

Fig 3 Model of the molecular regulation of tooth development trom initiation to crown morphogene-sis, interaotions between epitheiiai (biue) and mesenchymal tissues (yeiiow) are mediated by signalmolecuies (BMP = bone morphogenetic proteins, FGF = tibrobiast growth tactors; Shh = sonichedgehog. TNF = tumor necrosis factor). These signals operate throughout development and regu-iate the expression of genes in the responding tissues {shown in the boxes). Signaling centers (reä)appear m the epithelium reiteratively and secrete iocaiiy more than 10 different signals that regulatemorphogenesis and tooth shape. The green-colored genes have been shown to be necessary fornormal tcoth development.

A breakthrougb in our researcb was the discoverythe BMP is an important signal during the initiation oftooth development. Bone morphogenetic proteins aresynthesized by tbe early dental epithelium, and tbey in-duce in the mesenchyme the expression of many genes,including the transcription factor Msxl.' For this study,an organ culture method was designed in which thelocal effects of signal molecules were analyzed by in-troducing them with small beads on developing dentaltissues (Fig 4). Tooth germs were dissected from mouseembryos and epithelial and mesenchymal tissues wereseparated and placed on metal grids covered with cul-ture medium. The beads were incubated in a high con-centration of BMP protein and placed on dental mes-enchyme, and the expiant was cultured for 1 day in anincubator. Thereafter, the expression of candidate tar-get genes was examined by in situ bybridization fromwhole mounts (Fig 5). Later, it was sbown by otbersthat Msxl is, in fact, a necessary gene for tooth devel-opment both in mice and humans. No teeth developedin so called knockout mice lacking the function ofMsxl. Their tooth development was arrested at the budstage.» In bumans, mutations in tbe Msxl gene causesevere oligodontia (eight or more missing teeth).^

ENAMEL KNOTS: SIGNALING CENTERSGOVERNING TOOTH SHAPE

An important leap forviiard in our studies was the dis-covery of signaling centers, called enamel knots in the

Dissectedtooth 'g=i^:^?tanläge

Signal' molecule

EpitheliumMesenchyme

Culture on filter in Trowell-type organ culture

Fig 4 In vitro method developed for ihe analysis of local effectsof signal molecules.

tooth germ epithelium.'" Although the enamel knotshad been described already a century earlier in mor-pbologic studies as clusters of epithelial cells in thecap-stage tooth organs, their function was not known.The first signal discovered in the enamel knot wasFGF-4, and today, more than 10 different signals be-longing to the four families mentioned above bavebeen localized in enamel knots {Figs 2 and 6). Jemvall

Quintesseoce International 615

• Thesleff

has continued studies on enamel knots and shownthat the enamel knots instruct the patterning of thetooth crowns.'^ They determine the location andheight of tooth cusps hy inducing new, secondaryenamel knots at the sites of future cusps. His recentresults'- indicate that slight changes in enamel knotsignaling can explain the differences hetween theshapes of molars in various mammalian species andthat such changes can also account for the evolution-ary changes in dental morphology that are well knownfrom the fossil record."'- In Fig 6, the localized geneexpression in the primary and secondary enamel knotsprefigures the formation of cusps in a mouse mandibu-lar molar.

GENE EXPRESSION IN TEETH: A DATABASE

During the years we and others have reported the ex-pression patterns of numerous genes during tooth de-velopment. Most data were obtained by in situ hy-bridization, a method detecting the sites of mRNAexpression in tissues. As the information of gene ex-pression patterns was accumulating with increasingspeed, Nieminen'^ in our group constructed in 1995an internet database to help ourselves as well as thewhole scientific community to keep track of genes ex-pressed in teeth. This graphic database shows gene ex-pression patterns at various stages of tooth develop-ment, and in addition, contains supplementary dataon important genes. At present, the database includesinformation on approximately 250 genes. As an exam-ple, the page on Msxl expression is shown in Fig 7.

Fig S EffeclE of BMP signals ondental tissues cultured as shown inFig 4. f/eflj Light microscopic view ofan expiant of combined epilhelium(e) and mesenchyme (rn¡ and abead releasing BMP signal, (righijWhole mount in si t j hybridizalionanai/sis of a simiiar expiant as in tinee!t image, induction of Msxl ex-pression around the bead (b) and in(he mesenchyme adjacent to epithe-lium (ej

Fig 6 (leit) The primary enameiknoi IS visuaii2ed in a fronlai sectionthrough ihe cap slage tooth germ(expression of Edar, the receptor torectodysplasin). (rigiilj Secondaryenamei knots of the beil stage firstmoiar ¡Ml) prefigure c jsps. Thesecond moiar (M2) is at cap stageand the primary enamei knot isseen.

The database is now widely used internationally, andmembers of our laboratory keep it updated.

Altbough the information on the patterns of geneexpression does not directly tell about gene function,tbey give important insights. For example, the coex-pression of several signal molecule genes in theenamel knots actually led to the unraveling of thefunction of the enamel knot as a signaling center ofteeth. Today, our gene expression database is used as atool to detect coexpression of genes and to discovermolecular interactions between genes.

CLEIDOCRANIAL DYSPLASIA GENE RUNX2 ANDTHE INDUCTION OF NEW TEETH

Cleidocranial dysplasia is a syndrome affecting boneand tooth development and is caused by mutations inRUNX2 (earlier called CBFAl), a gene encoding a tran-scription factor." The symptoms include hypoplasticbone, in particular deficient formation of caivarialbones and clavicles. The tooth phenotype is particularlyinteresting as the patients have supernumerary teeth,and sometimes an almost complete third dentition de-velops.'= This indicates, remarkably, that all humansbave the potential to develop a tbird dentition and thatthis is normally inhibited by tbe RUNX2 gene.

RUNX2 is a "master gene" of bone developmentand is needed for osteoblast differentiation." RUNX2knockout mice have no bone, and their skeleton iscomposed solely of cartilage.'^ We studied tbe toothphenotype of these mice and quite unexpectedly ob-served that their teeth failed to develop {Fig 8). Teeth

616 Volume 34, Number 6, 2003

• TheslefJ

Expression ofMsxI in mouse tooth

dial epiULcbuDí, cltnm! rpuhfhim

i, dcitlal fpuhrfctmi

: denial pjpiHa, dfCiEBliPC

Eiprtsaioa; doiCal p^nHfl. dEnlaJ liNo apttsiimt: DEJI cpiChcbon, QuEc

, ecitmeL Vjmi. ncllaic

! ¡ho J flm Mtirl n nccf «atv For i^oih J

l Î4S-356

an J G . and, Setdman.C E f t»W^ Abl

Fig 7 Page descfibing the expression pattern of the Msx1 gene m the internet database"Gene expression in teeth" at www.bite-it.heisjniti.fi.

Quintessence International 617

• Thesleff

Fig 8 The BunKÍ gene is neoessary for tooih deveiop-ment. The deietion of its function m knockout mice(Run«2 -/-¡ arrests tooth development at the bud stage.In normal mice (vuiid type. WT) at the same age, tootfigerms have reached the oap stage

were initiated, but their development was arrestedafter the bud stage, indicating that RUNX2 function isnecessary for bud to cap stage transition."

Cleidocraniai dysplasia is caused by reduced func-tion of RUNX2, Though supernumerary teetb developin these patients whereas the complete loss of RUNX2function in mice inhibits tootb formation appears con-troversial, it is a challenging problem. Unfortunately,tbe mice do not normally develop a secondary denti-tion at all and therefore cannot be used as model ani-mals for studies on secondary tootb formation.However, we hope that by clarifying the function ofRUJNX2 in primary tooth development in mice, wecan shed light on the mechanisms whereby the sec-ondary dentition develops in humans, and that thiswill also explain why the third dentition develops inthe human cieidocranial dysplasia patients.

We are now analyzing RUNX2 knockout mice andaim to clarify tbe cause of arrested tooth development.We are searching for genes, which are regulated byRUNX2 by comparing gene expression between themutants and wildtype mice. We use mieroarray tech-nology and DNA chips, wbich allow tbe simultaneousanalysis of tbousands of genes, including presently un-known genes. Could this information be used for in-ducing a new set of teetb in the future?

ECTODYSPLASIN:THE MOLECULELACKING IN ECTODERMAL DYSPLASIA.A STIMULATOR OF TOOTH FORMATION

Ectodermal dysplasia syndromes are characterized byabsence or hypopiastic development of several ecto-dermal organs. In addition to teeth, these organs in-clude hairs, nails, and a variety of exocrine glands,such as sweat, salivary, and mammary glands. The de-velopment of all these organs are regulated by epithe-lial mesenebymal interactions, which are mediated bythe same signals as tooth development. The most com-

mon form is the hypohidrotic or anhidrotic ectodermaidysplasia (HED), The tooth phenotype of these pa-tients includes severe hypodontia, and sometimes eom-plete anodontia (Fig 9), Several gene mutations caus-ing HED bave been identified recently by positionalcloning of the human and corresponding mouse mu-tants,' Interestingly, these genes function in the tumornecrosis factor (TNF) signal pathway. X-chromosomalHED is caused by mutations in tbe actual TNF signal,called ectodysplasin, and otber forms result from mu-tations in the ectodysplasin receptor (Edar) and othercomponents in the same signal pathway.

We have shown that ectodysplasin regulates thefunctions of the enamel knots and hair placodes (whichare signaling centers for bair formation),'^'^^ We alsobave used several mouse mutants in tbese studies, in-cluding tbe naturally occurring ectodysplasin knockoutcalled the Tabby mouse. In addition, we have producedtransgenic mice overexpressing either ectodysplasin orits receptor Edar in the ectoderm. Interestingly, whilethe third molars are mostly missing in the Tabby mice,the ectodysplasin overexpressing mice have extra teeth(Fig 10),̂ ' Hence, ectodysplasin is an important signalregulating tooth number and sbape. Since ectodysplasinis a secreted protein, it is tempting to imagine that itcouid be used in the future to induce new teetb andhairs in ectodermal dysplasia patients. Could it be usedto grow new teetb in other patients too?

PROSPECTS FOR GROWING NEW TEETH:DENTAL STEM CELLS AND SIGNAL MOLECULES

A frequent question asked after my lectures and semi-nars over the years bas been "Wben can you grownew teeth for us?" Earlier, I used to say "never," butmore recently, due to the fantastic progress in thefields of developmental biology and stem cell biology,I have started to be more optimistic. This is first of alldue to the rapid increase in our understanding of the

618 Volume 34. Number 8, 2003

• Thesleff •

Fig 9 Oligodontia in a patient wtth hypohidrotic (an-htdrotic) ectodermal dysplasia (HED].

molecular regulation of tooth development. Of partic-ular importance here is the identification of the signaimolecules, such as ectodysplasin and many othersguiding tooth development at different stages (see Fig3), Secondly, the recent scientific breakthroughs instem cell research have indicated that adult cells mayhe much more plastic in their behavior than previ-ously thought."'^^ In other words, cells may changethe directions of differentiation according to environ-mental signals. The interest in tissue regeneration ingeneral has increased tremendously and data are accu-mulating rapidly on the competence of various celltypes to be programmed to different fates and on thesignals which induce their differentiation.

Stem cells also have heen identified in teeth.Mesenchymal stem cells, which formed dentin whentransplanted to muscle, were found in the adult dentalpulp.-" These cells could perhaps provide the mes-enchymal tissue for bioengineering teeth. We haveidentified epithelial stem cells in the cervical loop ofthe continuously growing mouse incisors and shownthat their maintenance and differentiation depends onFGF signals (Fig 11)." Unfortunately human teeth donot grow continuously, and they do not contain a poolof differentiating epithelial celis that could be used forgrowing new teeth. However, the characterization ofthe stem cells in mice may lead to the discovery ofsuitable human cells that could be triggered to takepart in tooth development.

It is ohvious that it will take a long time heforeteeth can be grown in dental practice. It will still re-quire more knowledge on the molecules regulatingtooth development. This is the specific area where weare continuing research. Also, more research is neededin order to find ways to identify and harvest potentialstem cells and to induce their differentiation to spe-cific directions. In addition, the practical designs fortransplantation and culture of cells and tissues presentgreat challenges. However, dreams ahout tooth regen-eration may afterall come trtie in the future.

Fig 10 Ectodysplasin stimulates tooth formation, (a)Normal mice have three molars; (b) Tabby mouse mu-tants, which lack ectodysplasin in gene function, do nothave the ihird molar, and (o) mioe with increased functionof ectodysplasin gene have an extra tooth in front of fhetirst molar

Quintessence fnternational 619

Stem celts

Fig 11 Stem cells are present in the cen/ical loop ot the continu-ously growing mouse incisor. Their maintenance and differentia-tion depends on FGF-10 signal secreted by the dental mes-enchyme.

REFERENCES

1. [ernvall J, Thesleff I. Reiterative signalling and patterningduring mammahan tooth morphogenesis. Mech Dev 2000;92:19-29.

2. Thesleff I. Genetic basis of tooth developmental and dentaldefects. Acta Odontol Scand 2000;58:191-194.

3. Thesleff I, Mikkola M. The role of growth factors in toothdevelopment. Int Rev Cytol 2002;217:93-135.

4. Kollar E], Paird GR. The influence of the dental papilla onthe development of tooth shape in emhryonic mouse toothgerms. J Embryol Exp Morphol 1969;21:131-148.

5. Ruch JV, Lesot H, Karcher-Djuriric V. Meyer JM, Mark M.Epithelial-mesenchymal interactions in tooth germs: Mecha-nisms of differentiation. J Biol Buccale 1983;11:173-183.

6. Thesleff I, Nieminen P. Tooth induction. In: Encyclopedia ofLife Sciences. London: Nature Publishing Groups, 2001.

7 Vainio S, Karavanova I, Jowett A, Thesleff I. Identificationof BMP-4 signal mediating secondary induction betweenepithehal and mesenchymai tissue during early tooth devel-opment. Cell 1993;75:45-58.

8. Satokata I, Maas R. MSXl deficient mice exhihit cleft palateand abnormalities of craniofacial and tooth development.Nat Genet 1994;6:348-356.

9. Vastardis H, Karimhux N, Guthua SW, Seidman JG,Seidman CE. A human MSXl homeodomain missense mu-tation causes selective tooth agenesis. Nat Genet 1996;13:417-42 L

10. [emvall I, Kettunen P, Karavanova I, Martin LB, Thesleff I.Evidence for the role of the enamel knot as a control centerin mammalian tooth cusp formation. Non-diving cells ex-press growth stimulating Fgf-4 gene. Int | Dev Biol 1994;58:463-469.

11. [emvall |, Kcranen SV, Thesleff 1. Evolutionary modificationof development in mammalian teeth: Quantifying gene ex-pression patterns and topography. Proc Nat Acad Sei USA20G0;97:14444-14448.

12. S a la zar-Ciudad 1, |ernvall |. A gene network model accouQt-ing for development and evolution of mammalian teeth.Proc Nat Acad Sei USA2002;99:8116-8120.

13. Nieminen P, Pekkanen M, Âberg T, Thesleff 1. A graphicalwww-database on gene expression in tooth. Eur J Oral Sei1998;106:7-ll.

14. Karsenty G. Minireview: Transcriptional control of os-teohlast differentiation. Endocrinology 2001;142:2731-2733.

15. fensen BL, Kreihorg S. Development of the dentition in clei-docranial dysplasia. J Oral Pathoi Med 1990;19:89-93.

16. Otto P. Thornell AP, Crompton T, et al. Cbfal, a candidategene for cleidocranial dysplasia syndrome, is essential forosteobiast differentiation and bone development. Cell1997;89:765-771.

17 D'Souza RN, Âberg T, Gaikwad J, Cavender A, Owen M,Karsenty G, Thesleff I. Chfa 1 is required for epithelial-mes-enchymal interations regulating tooth development in mice.Development 1999;126:2911-2920.

18. Pispa J, Jung HS, Jernvall ], Kettunen P, Mustonen T, TabataM|, Kere J, Thesleff I. Cusp patterning defect in Tabbymouse teeth and its partial rescue hy FGF. Dev Biol 1999;216:521-534.

19. Laurikkala |, Mikkola M, Mustonen T, et al. TNF signallingvia the lig-and-receptor pair ectodysplasin and edar controlsthe function of epithelial signalling centers and is regulatedhy Wnt and activin during tooth organogénesis. Dev Biol2001,229:443-455.

20. Laurikkala ], Pispa J, Jung HS, et al. Development and dis-ease. Regulation of hair follicle development by the TNFsignal ectodysplasin and its receptor Edar. Development2002;129:2541-2553.

21. Mustonen T, Pispa J, Mikkola M, et al. Stimulation of ecto-dermal organ development by Ectodysplasin-Al. Dev Biol(in press).

22. Donovan PJ, Gearhart J. The end of the beginning forpiuripotent stem cells. Nature 2001;414:92-97.

23. Jiang Y, Jahagirdar BN, Reinhardt RL, et al. Pluripotency ofmcsenchyraal stem cells derived from adult marrow. Nature2002;418:41-49.

24. Gronthos S, Mankani M, Brahim J, Robey PG, Shi S.Postnatal human dental pulp stem cells (DPSCs) in vitroand in vivo. Proc Nat Acad Sei USA 2000;97:13625-13630.

25. Harada H, Kettunen P, Jung HS, Mustonen T, Wang YA,Thesleff L Localization of putative stem cells in dental ep-ithelium and their association with Notch and FGF sig-nalhng. J Cell Biol 1999;147:: 05-120.

620 Volume 34, Number 8, 2003