THE ANATOMICAL RECORD 292:299–308 (2009) Effect of eda Loss of Function on Upper Jugal Tooth Morphology CYRIL CHARLES,1* SOPHIE PANTALACCI,2 RENATA PETERKOVA,3 PAUL TAFFOREAU,1,4 VINCENT LAUDET,2 AND LAURENT VIRIOT1* 1 iPHEP, CNRS UMR 6046, Faculté SFA, Université de Poitiers, 40 avenue du recteur Pineau, Poitiers Cedex, France 2 Molecular Zoology Team, Institut de Génomique Fonctionnelle de Lyon, Université de Lyon, CNRS, INRA, UCB Lyon 1, IFR128 Lyon Biosciences, Ecole Normale Supérieure de Lyon, 46 allée d’Italie, Lyon Cedex 07, France 3 Department of Teratology, Institute of Experimental Medicine, Academy of Sciences CR, Prague, Czech Republic 4 European Synchrotron Radiation Facility, 6 rue Jules Horowitz, Grenoble Cedex, France ABSTRACT The Tabby/eda mice, which bear a loss of function mutation for the eda (ectodysplasinA) gene, are known to display developmental anomalies in organs with an ectodermal origin. Although the lower jugal (cheek) teeth of Tabby/eda mice have been extensively studied, upper teeth have never been investigated in detail. However, this may help us to further understand the function of the eda gene in tooth development. In this work, the shape and size of both the crown and the radicular system were studied in the Tabby/eda mice upper jugal teeth. To deal with the high morphological variability, we deﬁned several morphotypes based on cusp numbers and position. Statistical tests were then performed within and between the different morphotypes to test the correlation between tooth size and morphology. Our analysis reveals that, as in lower teeth, eda is necessary to segment the dental lamina into three teeth with the characteristic size and proportions of the mouse. Nevertheless, since strong effects are observed in heterozygous upper teeth while lower are only mildly affected, it seems that the upper jaw is more sensitive than the lower jaw to the loss of eda function. Modiﬁcations in cusp number and the abnormal crown size of the teeth are clearly linked, and our results indicate a role of eda in cusp patterning. Moreover, we found that the Tabby mutation induces variations in the dental root pattern, sometimes associated with hypercementosis, suggesting a newly uncovered role played by eda in root patterning and formation. Anat Rec, 292:299– 308, 2009. Ó 2008 Wiley-Liss, Inc. Key words: Tabby/eda mouse; Mus musculus; dentition; jugal teeth; morphology Grant sponsor: ANR; Grant name: Quenottes, Ministry of Education, Youth, and Sports of the CR; Grant number: COST B23.002; Grant sponsor: Grant Agency of the CR; Grant number: 304/05/2665; Grant sponsors: Poitou-Charentes and Rhône-Alpes, The Fondation pour la Recherche Médicale, The Fondation Singer-Polignac), The COST B-23 Action (Brussels, EU). *Correspondence to: Cyril Charles, iPHEP, UMR CNRS 6046, 40 avenue du recteur Pineau, 86022 Poitiers Cedex, France. Tel: 1335-49-45-40-55. Fax: 133-5-49-45-40-17. E-mail: cyril.charles@univÓ 2008 WILEY-LISS, INC. poitiers.fr (or) Laurent Viriot, iPHEP, UMR CNRS 6046, 40 avenue du recteur Pineau, 86022 Poitiers Cedex, France. Tel: 133-5-49-4540-55. Fax: 133-5-49-45-40-17. E-mail: firstname.lastname@example.org Received 24 July 2008; Accepted 5 August 2008 DOI 10.1002/ar.20804 Published online 2 December 2008 in Wiley InterScience (www. interscience.wiley.com). 300 CHARLES ET AL. The mutant Tabby/eda mice display developmental anomalies in organs with an ectodermal origin such as hair, glands, teeth (Grüneberg, 1971), and palatal rugae (Charles et al., 2007). The mouse Tabby syndrome is homologous to the X-linked hypohidrotic ectodermal dysplasia syndrome in humans (Weeks and Blecher, 1983; Blecher, 1986; Srivastava et al., 1997). This syndrome is caused by a mutation in the eda gene on X-chromosome (Kere et al., 1996), which leads to a deﬁciency in the ectodysplasin A (EDA) protein (Srivastava et al., 1997). This protein initiates a signaling cascade, known as the EDA pathway, which ultimately regulates the activity of the NF-jB transcription factor (Courtney et al., 2005). During early development of ectodermal organs, the EDA pathway promotes placodal cell fate (Mustonen et al., 2004) and would thereby have an impact on the size of the forming organ. It has been demonstrated that overexpression of eda leads to the development of wider jugal teeth with an increased cusp number, and sometimes leads to a supernumerary tooth (Mustonen et al., 2003, Kangas et al., 2004). Tooth defects in Tabby/eda mice, including disruption in tooth shape, size, and number, were reported by several authors who all noticed a high phenotypic variability (Grüneberg, 1966; Sofaer, 1969a,b; Kristenova et al., 2002; Mustonen et al., 2003; Kangas et al., 2004; Peterkova et al., 2005). To deal with this variability, Kristenova et al. (2002) had to class the lower jugal teeth of the Tabby/eda mice into ﬁve morphotypes. Additionally, since the two to four lower jugal teeth of Tabby/eda mice cannot be homologized with certainty to the three molars of wild type mouse, they introduced a new nomenclature using T1, T2, T3, for lower jugal tooth 1, 2, and 3 (numbering from the most mesial jugal tooth). Further analysis of the ﬁve morphotypes during embryogenesis suggested that the phenotype could be explained by a defect in segmentation of the dental epithelium during early stages of odontogenesis (Peterkova et al., 2002). Only a few authors (Grüneberg, 1966; Sofaer, 1969a; Cermakova et al., 1998) have described the upper jugal teeth of the Tabby/eda mouse and with only general descriptions, mainly concerning size variation. We believe that studying in detail the upper tooth row of Tabby/eda mice might help to further illuminate the role of the eda gene in tooth development. We thus studied variations of the upper jugal dentition in homo/hemizygous and in heterozygous Tabby/eda specimens, both for the crown and the root apparatus. We deﬁned ﬁve main morphotypes in heterozygous and two in homo/hemizygous mutant specimens. These various morphotypes revealed themselves to be linked with opposed length variations of the ﬁrst and second jugal teeth. We tested the relationship between tooth size and occlusal morphology to see if the crown size variations could be linked with modiﬁcations of cusp size or number. Taken together, our analyses support the idea that the eda gene is involved in the segmentation of the upper dental lamina and in the cusp number and arrangement, but also supports an uncovered role in root patterning and formation. MATERIALS AND METHODS Tabby Mice Breeding pairs of four different stocks of Tabby/eda mice corresponding to three independent alleles of the eda gene were purchased from the Jackson Laboratory (Bar Harbor, Maine, USA) from 1994 to 2001: C3H/HeJ-Ta5J/1 (stock number JR1232) C57BL/6J-Aw-J-Ta6J/1 (stock number JR0338) C57BL/6J-Aw-J-Ta1/1Tfm (stock number JR0569) B6CBACa-Aw-J/A-Ta/0 (stock number JR0314) The Ta allele is a null allele resulting from a 2 kb deletion in the 50 part of the eda gene. The Ta6J allele is also expected to be a null allele, as a base pair deletion results in a translation of an EDA protein truncated of its functional domains. The Ta5J allele is not yet characterized. The Tabby/eda mutant and wild type (WT) mice were generated by inbreeding (Kristenova et al., 2002; Kristenova-Cermakova et al., 2002) in the Institute of Experimental Medicine ASCR (Prague, Czech Republic). Mice phenotype/genotype was identiﬁed according to external morphological criteria, which made possible the distinction between heterozygous, homo/hemizygous Tabby/eda mice, and WT specimens. Those criteria are mainly the yellowish color of the coat and the bald spot behind ears for homozygous (2/2) and hemizygous Tabby/eda mice (2/Y or 2/0) and the striping of the coat for heterozygous specimens (1/2). Homozygous and hemizygous Tabby/eda mice exhibited the identical external traits and thus cannot be distinguished from one another. These animals were offspring of crosses between homo/hemizygous, hemizygous or heterozygous females with their hemizygous or WT male siblings. The WT mice were generated by further inbreeding of wild type (phenotypically normal, non-Tabby) animals derived from the same stock. The sample composed of 205 mice, 85 heterozygous females (Ta/1), 23 homo/hemizygous Tabby/eda females (Ta/Ta, Ta/0), 57 hemizygous Tabby/eda males (Ta/0), and 40 control (WT) females. For each specimen, left and right tooth rows have been studied independently. The age of the specimens ranged from 15 postnatal days to adults. Mice were killed by cervical dislocation and their heads ﬁxed in a series of graded ethanols 70%– 96%. The experimental protocol was designed in compliance with the recommendation of the European Economic Community (86/609/CEE) for the care and use of laboratory animals. All of the animals’ treatment satisﬁed the requirement of the Institutional Review Board. Observation and Imaging of Dental Rows After mandible dislocation, jugal teeth were examined using a Leica MZ 16 stereomicroscope. Based on the number and arrangement of the cusps, various morphotypes were deﬁned for upper tooth rows, and specimens were ranked according to these morphotypes. Among the entire sample, 14 upper jaws were selected as a representative panel covering the totality of observed morphologies to study in detail not only the shape of the crown but also the dental root number and position. The 28 selected upper tooth rows were imaged using X-raysynchrotron microtomography at the European Synchrotron Radiation Facility (ESRF, Grenoble, France), beamline ID19 and BM5, with a monochromatical beam at energy of 25 keV. X-ray synchrotron microtomography has been demonstrated to bring very high-quality results for accurate imaging of small teeth (Tafforeau UPPER TEETH MORPHOLOGY IN TABBY/eda MICE 301 et al., 2006) and allows nondestructive virtual extraction of teeth roots (Chaimanee et al., 2006). A cubic voxel of 7.46 mm was used. All 3D renderings and virtual slices were performed using VGStudiomax software. Statistics Various statistical tests were used to verify the signiﬁcance of differences in tooth size between the different morphotypes. Width and length of jugal teeth were measured from digitized pictures using Optimas software. A preliminary step was to test differences of morphotype occurrence between the various mouse strains. For that purpose, we used a v2 test to compare the morphotype observed repartition with the theoretical repartition to see if the four strains were similar. Then, to determine which morphotype displayed the same tooth size, we used ANOVA and the post-hoc Scheffé test (Winer et al., 1991). This test compares the tooth size mean for each morphotype and determines which morphotypes are statistically different from the others. The Scheffé test was preferred to the Tukey test because of the unequal number of specimens in the various morphotypes. The jugal tooth row length was measured for specimens exhibiting all jugal teeth fully erupted (65% of specimens). We also used ANOVA to test the homogeneity of jugal tooth row length among the different morphotypes. Finally, we tested the parallelism of the linear regression of the tooth size measurements among the various groups using a test of parallelism adapted from the Student’s t test. The t value was calculated as follows (Dagnelie, 1998): t 5 |b1 2 b2|/H(r2Y.x (1/(RSSx)1 1 1/(RSSc)2)), with b1 and b2 the slope coefﬁcient of the two considered linear regression, r2Y.x the common, pooled, estimate of the residual variance, and RSS the relative residuals sum of squares. RESULTS As for lower teeth (Kristenova et al., 2002), the uncertain homology of teeth between WT and Tabby/eda specimens led us to adopt a nomenclature using Tx, x designating the rank of the teeth (T1 for the ﬁrst upper jugal tooth). We had four different tabby strains at our disposition, corresponding to three different mutations in the eda gene (see Materials and Methods section for details). However, we did not ﬁnd any statistical difference in morphotype occurrence between the various mouse strains when comparing observed and theoretical distributions (P > 0.1). We also tested the strain effect on tooth size using an ANOVA and found no statistically signiﬁcant effect (P > 0.7). The interstrain variability is thus negligible when compared with the intrastrain one. So, we considered all specimens together regardless of the mouse strain. This improved the number of studied Tabby/eda specimens, and therefore the reliability of our measurements and statistical values. Various Morphotypes of Upper Jugal Tooth Rows Wild type mice. The upper jugal dentition of wild type (WT) mice counts for three upper molars (M1, M2, and M3). The cusps of the M1 are linked by three trans- Fig. 1. Wild-type morphology, terminology used to name the main cusps of upper jugal teeth (terms from Miller, 1912). The red arrow indicates a supplementary cusp observed in one WT specimen. M, mesial; D, distal; L, lingual; V, vestibular. Scale bar: 1 mm. versal chevron-shaped crests (Fig. 1A). The most mesial chevron links together the three ﬁrst cusps t1 (lingual), t2 (central), and t3 (vestibular) according to Miller (1912). Similarly, the second crest composed of t4, t5, and t6 cusps, and the distal crest links only two cusps, t8 and t9 (Fig. 1A). The M2 shows the same general arrangement except that the mesial crest is limited to the t1 (absence of t2 and t3). M3 exhibits an isolated t1 and a cusp circle resulting from the two distal cusp rows interconnection. One of the spontaneous variations from the classical crown pattern was the occurrence of a tiny t3 cusp in 10% of the second molars. A rare second spontaneous variation was the presence in one specimen of a tiny supplementary cusp in the mesio-vestibular corner of the M1 (arrowed in Fig. 1B). The radicular part of the M1 (Fig. 1C) composed of three rounded roots (one mesial, one vestibular, and one disto-lingual). The M2 also exhibits three roots (one rounded mesio-lingual, one rounded disto-lingual, and one elliptic vestibular). The M3 variably exhibits two (one mesial and one distal) or three (two mesial and one distal) roots. Heterozygous Tabby/eda mice. Five major morphotypes, based on both the number and arrangement of cusps, were deﬁned in heterozygous specimens. We considered independently the right and the left tooth rows in each specimen and found that nearly half (46%) of the heterozygous mice do not exhibit the same morphotype on both sides (but the ﬁve morphotypes occurred in the same proportion on both sides). Nevertheless, tooth rows were always located at the same position at the left and right sides of the skull. Morphotype Heterozygous 0 (MT He0) was the most frequent (34%) among the heterozygous tooth rows. It is indistinguishable from WT (Fig. 2A). 302 CHARLES ET AL. Fig. 2. Morphotypes deﬁned on the heterozygous Tabby mice. A: WT; B: MT He1; C: MT He2; D: MT He3; E: MT He4. The red arrows indicate some speciﬁcs detailed in the text. M, mesial; D, distal; L, lingual; V, vestibular. Scale bar: 1 mm. Morphotype Heterozygous 1 (MT He1) was observed in 20% of upper tooth rows. In this morphotype, T1 displayed mild defects on the vestibular side when compared with the WT (Fig. 2B). The major systematic defect was the occurrence of a supplementary cusp distal to the t6 (t6b), which appeared either small or as large as a main cusp (arrowed in Fig. 2A). Its presence was not linked with the disappearance of the t9 as these two cusps could occur in a same tooth (data not shown). However, the t9 was often reduced or even sometimes absent. Occasionally, another supplementary minute cusp can occur between t3 and t6 cusps. The T2 crown was similar to the WT M2. The T3 showed a decrease in the cusp number associated with a reduction of tooth size. The T1 radicular system was similar to that of the WT M1. On the contrary, the root number of T2 and T3 differed from that observed for WT M2 and M3. Indeed, T2 can exhibit only one or two roots instead of three. The remaining roots were then not rounded, but displayed more complex shapes, suggesting that they might result from a fusion of roots (Fig. 2B). The T3 displayed either one or two roots. Morphotype Heterozygous 2 (MT He2) was found in 9% of tooth rows. The T1 of the MT He2 showed a vertical and rounded t2 cusp and a reduction of both the t3 and t9 cusps (arrowed in Fig. 2C). The T2 was grossly similar to the WT M2, except for a reduction or even disappearance of the t9 cusp. Moreover, in a few tooth rows (5% of He2), we found the supplementary tiny t3 cusp on T2 mesio-vestibular corner that is also occasionally found in WT M2. The T3 displayed either two or three cusps. The mesial root of the T1, located just below the t2 cusp, was more vertical than in WT M1 and was sometimes fused with the lingual root. As for MT He1, the number and shape of roots was variable in T2 and reduced to one in T3. Morphotype Heterozygous 3 (MT He3) was the least common morphotype with a frequency of 8.2%. It was characterized by a reduced T1 (Fig. 2D). This reduced tooth corresponds was considered as a supernumerary by Sofaer (1969b). The T1 notation, always used here to describe the ﬁrst jugal tooth, leads to homologize the small mesial tooth in type He3 with the T1 of other morphotypes. We, however, keep the T1 notation for this tooth as we saw intermediary sizes between the T1 of MT He3 and T1 of the other morphotypes, especially MT He4. Jugal tooth rows composed of either three or four teeth. The T1 occlusal side displayed only one to three distinct cusps. The T2 was the largest tooth of the row and exhibited a t3 cusp in 91% of specimens and always a reduced t9. The T3 was smaller than the T2 and exhibited ﬁve cusps. Of note, T2 and T3 in MT He3 looks like T1 and T2, respectively, in MT He2 (Fig. 2D). When present, a bicuspidate T4 was smaller than the T3. The T1 was single-rooted (with a root inclined toward the mesial direction), whereas the T2 displayed two roots similar to the MT He2 T2. The T3 had three roots and the T4 only one. Morphotype Heterozygous 4 (MT He4) corresponded to 25% of observed tooth rows (Fig. 2E). It was characterized by a disruption of the t4-t5 interconnection in T1, and the appearance of a new mesio-distal connection between the t2 and t4 cusps by a lingual crest. This later connection was sometimes prolonged until the t8 cusp, resulting in a crest which occupied the entire lingual border of the tooth. Like T1, the T2 can exhibit a disruption of the t4-t5 interconnection. In 47% of MT He4 specimens, the T2 also showed a crest connecting t1 and t3 cusps. The T3 displayed three cusps. The radicular system of the T1 generally composed of three roots. However, the vestibular root tended to split into two different roots (Fig. 2E). The T2 displayed two roots, the elliptic mesial one resulted from the fusion of two rounded roots. The T3 exhibited three rounded roots. Other rare morphologies represented only 4% of the observed tooth rows (Fig. 3). The ﬁrst case showed the appearance of both a t1bis cusp associated to an extra root on the T1 (arrowed in red in Fig. 3A). A vestibular extra cusp (arrowed in white) occurred between t3 and t6 cusps. The mesial cusps occupied abnormal positions: t1 was distally displaced while both the t1bis and t3 were closer to the t2, resulting in an acute triangular conﬁguration (Fig. 3A). Another specimen exhibited a T1 morphologically close to the T1 of the MT He4 with a duplication of t5-t6 enclosing an extralingual cusp (arrowed in Fig. 3B). A third exceptional case displayed a tooth row similar to that of MT He3 with fused T1 and T2 crowns and roots (Fig. 3C). The radicular complex of both the T1 and the ﬁrst crest of the T2 were grouped into a unique root. Of note, whatever the morphotype occurring in the two sides of heterozygous asymmetric specimens, the UPPER TEETH MORPHOLOGY IN TABBY/eda MICE 303 Fig. 4. Morphotypes deﬁned on the homo/hemizygous Tabby mice. A: MT Ho1; B: MT Ho2; C and D: Radicular patterns observed in homozygous Tabby mice. The red arrows indicate the occlusal surface speciﬁcs detailed in the text. M, mesial; D, distal; L, lingual; V, vestibular. Scale bar: 1 mm. cusps. Based on these characters, MT Ho1 and MT Ho2 were very close to the MT He4 deﬁned for heterozygous Tabby/eda mice. The MT Ho1 and MT Ho2 sometimes occur on the same specimen, resulting in an asymmetric conﬁguration. Apart from these occlusal traits, two different root patterns were observed for T1 in both Ho1 and Ho2. In the ﬁrst pattern (Fig. 4C), mesial and lingual roots were fused and not connected to the vestibular one. In the second pattern (Fig. 4D), the three roots were fused and met in the center of the tooth. In Tabby/eda mice, the radicular complex sometimes displayed hypercementosis (e.g., Fig. 2B–D). The amount of cementum that covered the root dentine was highly variable from one specimen to another. Fig. 3. Rare morphologies of heterozygous Tabby mice. Arrows referred to supplementary cusps, roots, and abnormal connection indicated in the text. M, mesial; D, distal; L, lingual; V, vestibular. Scale bar: 1 mm. mesial and distal extremities of the jugal tooth rows are located at the same mesio-distal position on the skull. Homo/hemizygous Tabby/eda mice. In all homo/hemizygous specimens, the T1 was characterized by a reduction of vestibular cusps, a disruption of the t4t5 interconnection and the appearance of a new mesiodistal connection between t2 and t4 cusps by a lingual crest. Two major morphotypes were deﬁned, mainly differing by the T1 morphology (Fig. 4A,B). In the MT Ho1 (69% of tooth rows), T1 exhibited only one central cusp in the mesial area (t2) linked to the t4 by a long semicircular lingual crest (Fig. 4A). In the second morphotype MT Ho2, two cusps (t1 and t4 arrowed on Fig. 4B) were still present in the lingual area and attached to a lingual semicircular crest connected to the t2. In both morphotypes, T2 systematically lacked the t9 cusp, and a lingual crest similar to the one observed on the T1 was sometimes present. The T3 displayed either two or three Classes of Morphotypes To further characterize the different morphotypes, we used measurements of the total jugal tooth row and of the ﬁrst jugal tooth. Measurements of the tooth row length revealed two groups (Fig. 5). The ﬁrst group composed of WT and a part of the heterozygotes (MT He0 and He1), which have a normal tooth row length and are morphologically very close to wild type. The second included the other heterozygous (MT He2, He3, and He4) and all of the homo/hemizygous Tabby/eda tooth rows, which all have a reduced tooth row length. These associations were statistically signiﬁcant and we found no difference between the measurements of jugal tooth row length in WT, MT He0, and MT He1 mice (ANOVA, F 5 2.83, P 5 0.07) or between MT He2, MT He3, MT He4, MT Ho1, and MT Ho2 (ANOVA, F 5 1.08, P 5 0.37). Measurements of the ﬁrst jugal tooth (Fig. 6) conﬁrmed homogeneity of the ﬁrst group containing WT, MT He0, and MT He1 (Scheffé test, threshold value 0.05). Tooth rows in this group displayed the largest T1, with mild or no cusp defects. Measurements of the ﬁrst jugal tooth also allowed the division of the second group into three subgroups. (1) MT He2 group: The T1 was shorter than M1 of WT in relation with a more vertical 304 CHARLES ET AL. Fig. 5. A: Length of the upper jugal tooth rows for the different morphotypes. Black lines indicate the range of variation of each morphotype. B: Proportion of the various jugal teeth of the tooth row (mean values). Black: ﬁrst jugal tooth; dark gray: second jugal tooth; clear gray: third jugal tooth; white: fourth jugal tooth. Tooth Size Among the Jugal Tooth Row Fig. 6. First jugal tooth morphologies and ﬁrst jugal tooth size modiﬁcations. Each morphotype is represented by the mean value and the standard deviation of the mean. The ellipses regroup morphotypes exhibiting close morphologies for the ﬁrst upper cheek tooth. mesial area. The T2 was also shorter than wild type M2 (see Fig. 5b); (2) MT He4, MT Ho1, and MT Ho2 group: In this homogeneous group (Scheffé test, threshold value 0.05), the T1 was still shorter, displaying strongly reduced mesial and vestibular areas and a lingual crest. The T2 is of normal or even increased size as compared with M2 (Fig. 5b); (3) MT He3 group: The T1 consisted of a minute tooth composed of one, two, or three cusps. The T2 is clearly longer than M2 (Fig. 5b). Kavanagh et al. (2007) proposed a model to explain size relationships within the tooth row (the ‘‘inhibitory cascade model’’). In this model, sequential development along with cumulative inhibition explain the relative size of the mouse molars: the ﬁrst molar develops ﬁrst and inhibits the second molar, which as a consequence is smaller, and the third molar, which cumulates inhibition of both the ﬁrst and the second, is even smaller. To further evaluate changes in the relative size of the molars in Tabby mice (Fig. 5), we used a graph that represents the cumulative nature of the ‘‘inhibitory cascade model’’ (Fig. 7). We represented the size of a tooth as a function of the cumulated size of the previous teeth. On this graph, each wild-type tooth row is represented by two points: one linking M2 with M1 size, and the second linking M3 with M11M2 size. On the same graph, we can incorporate the Tabby 3-teeth row, but also the 4-teeth row, with three possible relationships: T2 and T1, T3 and T11T2, and lastly, T4 and T11T21T3. Importantly, in this last case, the three points were found aligned, showing that the same cumulative relationship as the one linking T1 and T2 size to T3 size can predict T4 size, and thus supported this representation. We can consider three different regressions: WT (Fig. 7, line ), MT He 0-1 (Fig. 7, line ), and MT He2-3-4 plus MT Ho1-2 Tabby (Fig. 7, line ). The position of these lines on the graph is mainly linked to the T1/T2 ratio. Line 2 is slightly shifted and line 3 is strongly shifted toward the left. This reﬂected the unusual T1/T2 ratio in Tabby morphotypes (especially in the more affected morphotypes, He2-3-4 and Ho1-2), as already shown. Then, the slope of these lines is related to the T3 size relative to the T2 size (for a 3-teeth row, the slope coefﬁcient would be: size of T3/size of T2-1). First, we noted that lines 1, 2, and 3 had very close slope values. UPPER TEETH MORPHOLOGY IN TABBY/eda MICE Fig. 7. Relationship between the size of a tooth and the total size of the more mesial jugal teeth. The black line represents the linear regressions: for WT specimens, for MT He0 and MT He 1 specimens, for MT He2, MT He 3, MT He 4, and homo/hemizygous Tabby specimens. Measurements indicate similar tooth proportion on the entire jugal row among MT He2-3-4 and homo/hemizygous Tabby specimens and similar T2, T3 and when present T4 proportions among all of the Tabby mice (parallelism of the linear regression and ). We used a test of nonparallelism to see if differences were statistically signiﬁcant. We found a statistically relevant nonparallelism (Pnonparallelism > 0.05) between the regression lines 1 and 2 (WT versus He0-1), and 1 and 3 (WT versus He2-3-4-Ho1-2), but not between 2 and 3 (light tabby phenotype He0-1 versus strong tabby phenotype He2-3-4-Ho1-2; Pnonparallelism < 0.05). Thus, between the two groups of Tabby morphotypes, the T2 2 T3 proportions are not signiﬁcantly different. In conclusion, the T1/T2 proportions strongly differ between homo/hemizygous and MT He2-3-4 specimens on the one hand, and the MT He0-1 specimens on the other hand, yet the T2 2 T3 proportions are similar among all of the Tabby specimens and only slightly differ from the WT condition (slope values are close even if statistically different). Correlation Between Tooth Morphology and Size The tooth measurements showed a higher variability in heterozygous than in homo/hemizygous Tabby/eda mice. The Coefﬁcient of variation of the T1 length was 2.9 for the WT (interval of conﬁdence 2.5–3.4), 16.6 (IC 14.8–18.8) for heterozygous, and 4.5 (IC 3.9–5.4) for homo/hemizygous Tabby/eda specimens. This size variability accompanied the morphological diversity among the heterozygous specimens, suggesting that there is a relation between tooth size and tooth shape. To test this relation, the next step consisted of comparing the variations of cusp number with tooth size among the heterozygous sample because this genotype exhibits the most important morphological variations. Results presented in the Fig. 8 showed that for each tooth, the size clearly increased with the number of cusps. 305 Fig. 8. Relationships between teeth size and cusps number in heterozygous Tabby mice in red: ﬁrst jugal tooth (T1), in blue: second jugal tooth (T2), in black: third jugal tooth (T3). The verticals lines represent the range of size variation for each number of cusps. For each tooth, the black line represents the linear regression. DISCUSSION Natural Variability of Upper Jugal Teeth in WT Mice Among the WT mouse sample, two main variations have been found: (1) a supplementary cusp mesial to the ﬁrst chevron of the M1; this has already been described in Mus setulosus but occurs less frequently in Mus musculus (Misonne, 1969); and (2) a mesial cingulum in 10% of the M2. This has also been reported by Cermakova et al. (1998) in WT and ICR mice. Morphological Diversity of the Upper Jugal Tooth Row in Tabby/eda Mutant Mice Previous studies on Tabby/eda mouse dentition mainly documented a reduction of the jugal tooth size and a decrease in the cusp number (Grüneberg, 1966; Sofaer, 1969a; Miller, 1978; Cermakova et al., 1998). Our study also revealed these two general trends. However, among heterozygotes, we observed very few specimens with larger crown size and higher cusp number. These latter correspond to the ‘‘Incomplete twinning’’ specimens described by Grüneberg (1966). At least one of these teeth (Fig. 3C) can clearly be considered as the result of an abnormal fusion between the ﬁrst two teeth of the MT He3. Among heterozygous Tabby/eda mice, 54% of the tooth rows display a normal length and a morphology similar (MT He0) or close (MT He1) to the WT one. It corresponds to tooth rows in which the effect of the Tabby mutation is very small. However, the mild modiﬁcations observed between the crown shapes of these two morphotypes might reﬂect an effect of the eda gene on cusp number and arrangement, which could be compared with the changes that can be observed between two related taxa in nature (Misonne, 1969). The rest of the heterozygous and homozygous tooth rows all have a reduced length, with an opposed 306 CHARLES ET AL. 1 2 variation in tooth length between T and T . MT He3 is characterized by a T1, which is smaller than the T2. This is a reversed situation as compared with the wild type. MT He2 may be interpreted as a MT He3 with an aborted development of the most mesial tooth. Such a situation in front of the future T1 was previously observed during the development of the lower tabby tooth row (Peterkova et al., 2002). Here, it could have provoked the vertical disposition of the t2 cusp in the T1 of MT He2 and the change in the orientation of its underlying mesial root (Fig. 2). As for the lower jaw, both the aborted tooth of MT He2 and the T1 of MT He3 might be interpreted as a ‘‘revival’’ of tooth rudiments found in wild type mouse (Viriot et al., 2002; Peterkova et al., 2005; Klein et al., 2006). However, developmental data will be necessary to test this possibility. Finally, the last heterozygous morphotype (MT He4), corresponding to about 25% of the heterozygous specimens displays similar features to those of the two homozygous morphotypes (MT Ho1 and MT Ho2), i.e., the reduction of vestibular cusps and the occurrence of a lingual cingulum. Grüneberg (1966), Sofaer (1969a), and Miller (1978) considered that this mesial cingulum is a typical feature of Tabby/eda mice. However, it has been observed in WT and ICR mice. In the latter, the mesial cingulum is tiny and usually smaller than the mesial crest observed in Tabby/eda mice. We thus state that the occurrence of a mesial cingulum is not characteristic of the Tabby mutants, but the bigger size of this cingulum is a speciﬁcity of Tabby/eda mice. Of note, the reduction of the total tooth row in these three morphotypes is mainly attributed to the reduction of the T1, which is nevertheless still meanly 25% longer than the T2. The Tabby Mutation Putatively Modiﬁes the Segmentation of the Upper Dental Lamina The various morphotypes deﬁned for lower jugal tooth rows have been related to differences in the segmentation of dental epithelium during determination of tooth boundaries; although the total antero-posterior length of the dental epithelium is similar, the length of the particular tooth primordia differs between the jaws, even in the same embryonic head (Peterkova et al., 2002, 2005). Moreover, Kristenova et al. (2002) described an opposed size variation between the ﬁrst and second lower jugal teeth associated with various morphotypes. In our study of Tabby/eda upper tooth rows, we also found that the limits between the different teeth are located at variable mesio-distal positions despite similar row length, revealing a similar modiﬁcation in the segmentation of the upper dental lamina. We also found that the size of a tooth depends on the size of the more mesial jugal teeth (Fig. 7). The occurrence of a fourth tooth in some Tabby heterozygous specimens is then a consequence of the reduced size of the three mesial ones, teeth continuing to develop according to the observed relationship until the tooth size falls under the minimal size (Fig. 7). Interestingly, this size relationship might be explained by the recent ‘‘inhibitory cascade model’’ proposed by Kavanagh et al. (2007). These authors proposed that the developing ﬁrst molar exerts an inhibitory effect on the second molar, and that this inhibitory effect will remain as strong as the ﬁrst molar development continues. This model could explain the tooth size relationships observed in Tabby heterozygous mice as well as the number of teeth. The observed proportions of the various teeth among the jugal tooth row could thus directly result from the variability of size of the ﬁrst tooth (T1) among heterozygous Tabby mice. This T1 size variability might itself be the consequence of a variable amount of EDA protein in heterozygous Tabby specimens. Relationship Between Tooth Size and Morphology Measurements emphasized a link between cusp number and crown size in T1, T2, and T3 (Fig. 8). According to the ‘‘patterning cascade mode of cusp development’’ published by Jernvall (2000), the succession of the signaling centers, and consequently the number and position of cusps, is linked to the crown size. This relationship had previously been observed for the lower teeth of Tabby mice (Kangas et al., 2004) and holds true here for upper Tabby jugal teeth (Fig. 8). As the number of cusp is clearly linked to the tooth size, one could think that the morphological modiﬁcations could only be the result of a primary effect on the tooth size, which consecutively provokes modiﬁcations of the cusp pattern. This is clearly happening in Tabby mutants. However, we do not think this is the only explanation, essentially for two reasons: ﬁrst, eda gene is not only expressed in early stages of tooth formation but also at the time of cusp formation (Laurikkala et al., 2001). Second, in MT He4, Ho1, and Ho2, T2 is about the normal size of WT mice M2, but the cusp pattern is nevertheless strongly affected. Rather, we think that the effect on tooth morphology is a consequence of both a primary effect on tooth size and a direct effect on cusp formation. Finally, Jernvall and Thesleff (2000) explained the Tabby/eda jugal tooth morphology by closer locations of cusp tips. Our results show that the whole morphological variability observed among Tabby/eda jugal dentition is not only related to closer cusp positions but also to new cusp arrangements and occurrence of neoformed cusps. Comparison of the Tabby/eda Phenotype Between Lower and Upper Tooth Rows For lower tooth rows, Cermakova et al. (1998) found only very minor defects in heterozygotes and still a small percentage of wild type tooth rows among homozygotes. This situation contrasts with the situation in upper tooth row, where 46% of heterozygous and 100% of homozygous tooth rows were abnormal. It is interesting to further compare the morphotypes we described in both hetero- and homozygotes with the one described for the homozygous lower tooth row (Kristenova et al., 2002; Peterkova et al., 2002). Class I lower tooth rows have T1 shorter than T2. Thus, they may be analogous to MT He3 upper tooth rows. In Class IIa, the ﬁrst tooth aborts, and adult T1 is longer than T2. Class IIa might thus correspond to MT He2 for upper tooth row. Finally, in Class IIb, T1 develops ﬁrst and is longer than T2. Class IIb might then correspond to MT He4, Ho1, and Ho2 for upper tooth row. Taken all together, this comparison suggests that the effect of the eda mutation on the upper and lower tooth row is rather similar, but that the upper tooth row development is more sensitive to eda loss of function. Differences in the developmental UPPER TEETH MORPHOLOGY IN TABBY/eda MICE control between upper and lower jugal teeth have already been suggested by Shimizu et al. (2004) after a study of size variation in ﬁrst upper and lower molars of SMXA recombinant inbred strains of mice. Several mutations lead to a complete loss of molars on one jaw while molars on the opposite jaw develop normally (see dlx12/2,dlx22/2 double mutant, Qiu et al., 1997 or activinbA2/2 mutant, Ferguson et al., 1998). The case of eda mutants provides a novel and more subtle example of differences in the genetic control of upper and lower tooth row development. Associated Modiﬁcations of the Radicular Complex We observed modiﬁcations of the root pattern in Tabby/eda mice. Since root formation is consecutive to crown formation, an abnormal crown size speciﬁcation could very logically disrupt the root patterning program. Except for some rare specimens showing an increase in both crown size and number of roots (Fig. 3), the major tendency in Tabby/eda is a decrease in the crown size associated with a decrease in the number of roots. Notably, Grüneberg (1966) described a unique root for the second upper jugal tooth, whereas we observed specimens displaying one, two, or three roots. However, the position of these roots and the enlarged shape of the unique root of other specimens indicate that the unique root results from the fusion of the three normal ones (Fig. 4C,D). Importantly, however, we also often observed variations in the number of roots for similar crown size, suggesting a more direct perturbation of root patterning. Moreover, the observation of Tabby/eda specimens with a higher amount of cementum than the normal condition could indicate a role of eda not only in root patterning but also in the rhizogenesis itself. Very interestingly, another gene implicated in the EDA pathway, edar, was very recently found in a subtractive hybridization screen for genes speciﬁcally involved in the molar multiroot rhizogenesis versus the incisive single root pseudo-rhizogenesis (Xing et al., 2007). Together with our results, these data strongly suggest that the EDA pathway plays a role in root patterning and formation. CONCLUSION We show that, as in the lower jaw, eda function is necessary during development to obtain a correct tooth row length and a correct segmentation of this tooth row. However, the upper jaw seems more sensitive to Tabby/ eda heterozygoty than the lower jaw. Moreover, we have indications that, apart from determining tooth crown size, which is linked to cusp and root number, eda may play a more direct role in cusp and root patterning and formation. ACKNOWLEDGMENTS The authors thank J.-J. Jaeger and V. Lazzari for comments concerning this study. They also thank J. Baruchel and ID 19 staff of the European Synchrotron Radiation Facility of Grenoble. They acknowledge P. Kristenova for collecting the material; L. Foley-Ducrocq for language revision; the reviewers for their comments; C. Noël and G. Florent for administrative guidance. 307 LITERATURE CITED Blecher S. 1986. Anhidrosis and absence of sweat glands in mice hemizygous for the Tabby gene: supportive evidence for the hypothesis of homology between Tabby and human. J Investig Dermatol 87:720–722. Cermakova P, Peterka M, Capkova J, Turečková J, Ruch J-V, Lesot H, Peterkova R. 1998. Comparison of the tooth shape and size in tabby and non-tabby mice. Acta Vet Brno 67:3–14. Chaimanee Y, Yamee C, Tian P, Khoawiset K, Marandat B, Tafforeau P, Nemoz C, Jaeger J-J. 2006. Khoratpithecus piriyai, a late Miocene hominoid of Thailand. Am J Phys Anthropol 131: 311–323. Charles C, Pantalacci S, Peterkova R, Peterka M, Laudet V, Viriot L. 2007. Disruption of the palatal rugae pattern in Tabby (eda) mutant mice. Eur J Oral Sci 114:1–8. Courtney J-M, Blackburn J, Sharpe PT. 2005. The ectodysplasin and NFkB signalling pathways in odontogenesis. Arch Oral Biol 50:159–163. Dagnelie P. 1998. Statistique théorique et appliquée. Tome 2: inférence statistique à une et à deux dimensions. Bruxelles: De Boeck, 659 p. Ferguson CA, Tucker AS, Christensen L, Lau AL, Matzuk MM, Sharpe PT. 1998. Activin is an essential early mesenchymal signal in tooth development that is required for patterning of the murine dentition. Genes Dev 12:2636–2649. Grüneberg H. 1966. The molars of the tabby mouse, and a test of the ‘single-active X-chromosome’ hypothesis. J Embryol Exp Morph 15:223–244. Grüneberg H. 1971. The Tabby syndrome in the mouse. Proc R Soc Lond B Biol Sci 179:139–156. Jernvall J. 2000. Linking development with generation of novelty in mammalian teeth. Proc Natl Acad Sci USA 97:2641–2645. Jernvall J, Thesleff I. 2000. Reiterative signaling and patterning during mammalian tooth morphogenesis. Mech Dev 92:19–29. Kangas A, Evans AR, Thesleff I, Jernvall J. 2004. Nonindependence of mammalian dental characters. Nature 432:211–214. Kavanagh K, Evans AR, Jernvall J. 2007. Predicting evolutionary patterns of mammalian teeth from development. Nature 449: 427–432. Kere J, Srivastava AK, Montonen O, Zonana J, Thomas N, Ferguson B, Munoz F, Morgan D, Clarke A, Baybayan P, Chen EY, Ezer S, Saarialho-Kere U, Chapelle ADL, Schlessinger D. 1996. X-linked anhidrotic (hypohidrotic) ectodermal dysplasia is caused by mutation in a novel transmembrane protein. Nat Genet 13:409–416. Klein OD, Minowada G, Peterková R, Kangas A, Yu BD, Lesot H, Peterka M, Jernvall J, Martin GR. 2006. Sprouty genes control diastema tooth development via bidirectional antagonism of epithelial-mesenchymal FGF signaling. Dev Cell 11:181–190. Kristenova-Cermakova P, Peterka M, Lisi S, Lesot H, Peterkova R. 2002. Postnatal lower jaw dentition in different phenotypes of tabby mice. Connect Tissue Res 43:283–288. Kristenova P, Peterka M, Lisi S, Gendrault J-L, Lesot H, Peterkova R. 2002. Different morphotypes of functional dentition in the lower molar region of tabby (EDA) mice. Orthod Craniofacial Res 5:205–214. Laurikkala J, Mikkola M, Mustonen T, Åberg T, Koppinen P, Pispa J, Nieminen P, Galceran J, Grosschedl R, Thesleff I. 2001. TNF signaling via the ligand–receptor pair ectodysplasin and Edar controls the function of epithelial signaling centers and is regulated by Wnt and Activin during tooth organogenesis. Dev Biol 229:443–455. Miller GS. 1912. Catalogue of the mammals of Western Europe (exclusive of Russia) in the collection of the British Museum. London: British Museum (Natural History). Miller WA. 1978. The dentition of Tabby and Crinkled mice (an upset in mesodermal:ectodermal interaction). In: Butler PM, Joysey KA, editors. Development, function and evolution of teeth. London: Academic Press. p 99–109. Misonne X. 1969. African and Indo-Australian Muridae, evolutionary trends. Ann Musée Roy Afr Centr Ser 8 Sci Zool 172:1–219. 308 CHARLES ET AL. Mustonen T, Ilmonen M, Pummila M, Kangas AT, Laurikkala J, Jaatinen R, Pispa J, Gaide O, Schneider P, Thesleff I, Mikkola ML. 2004. Ectodysplasin A1 promotes placodal cell fate during early morphogenesis of ectodermal appendages. Development 131:4907–4919. Mustonen T, Pispa J, Mikkola ML, Pummila M, Kangas AT, Pakkasjärvi L, Jaatinen R, Thesleff I. 2003. Stimulation of ectodermal organ development by Ectodysplasin-A1. Dev Biol 259: 123–136. Peterková R, Kristenová P, Lesot H, Lisi S, Vonesch J-L, Gendrault J-L, Peterka M. 2002. Different morphotypes of the tabby (EDA) dentition in the mouse mandible result from a defect in the mesio-distal segmentation of dental epithelium. Orthod Craniofacial Res 5:215–226. Peterková R, Lesot H, Viriot L, Peterka M. 2005. The supernumerary cheek tooth in tabby/EDA mice—a reminiscence of the premolar in mouse ancestors. Arch Oral Biol 50:219–225. Qiu M, Bulfone A, Ghattas I, Meneses JJ, Christensen L, Sharpe PT, Presley R, Pederson RA, Rubenstein JL. 1997. Role of the Dlx homeobox genes in proximodistal patterning of the branchial arches: mutations of Dlx-1, Dlx-2, and Dlx-1 and -2 alter morphogenesis of proximal skeletal and soft tissue structures derived from the ﬁrst and second arches. Dev Biol 185:165–184. Shimizu T, Oikawa H, Han J, Kurose E, Maeda T. 2004. Genetic analysis of crown size in the ﬁrst molars using SMXA recombinant inbred mouse strains. J Dent Res 83:45–49. Sofaer JA. 1969a. Aspects of the tabby-crinckled-downless syndrome I. The development of Tabby teeth. J Embryol Exp Morph 22: 181–205. Sofaer JA. 1969b. Aspects of the tabby-crinckled-downless syndrome II. Observations on the reaction to changes of genetic background. J Embryol Exp Morph 22:207–227. Srivastava AK, Pispa J, Hartung AJ, Du Y, Ezer S, Jenks T, Shimada T, Pekkanen M, Mikkola ML, Ko MSH, Thesleff I, Kere J, Schlessinger D. 1997. The Tabby phenotype is caused by mutation in a mouse homologue of the EDA gene that reveals novel mouse and human exons and encodes a protein (ectodysplasin-A) with collagenous domains. Proc Natl Acad Sci USA 94:13069–13074. Tafforeau P, Boistel R, Boller E, Bravin A, Brunet M, Chaimanee Y, Cloetens P, Feist M, Hoszowska J, Jaeger J-J, Kay RF, Lazzari V, Marivaux L, Nel A, Nemoz C, Thibault X, Vignaud P, Zabler S. 2006. Applications of X-ray synchrotron microtomography for non-destructive 3D studies of paleontological specimens. Appl Phys A Mater Sci Process 83:195–202. Viriot L, Peterková R, Peterka M, Lesot H. 2002. Evolutionary implications of the occurence of two vestigia tooth germs during early odontogenesis in the mouse lower jaw. Connect Tissue Res 43:139–133. Weeks N, Blecher S. 1983. Evidence from thiol histochemistry for homology between the Tabby-crinkled syndrome in mice and human ectodermal dysplasia. J Histochem Cytochem 31:1407– 1411. Winer BJ, Brown R, Michels KM. 1991. Statistical principles in experimental design. 3rd ed. New York: McGraw-Hill. Xing X, Deng Z, Yang F, Watanabe S, Wen L, Jin Y. 2007. Determination of genes involved in the early process of molar root development initiation in rat by modiﬁed substractive hybridization. Biochem Biophys Res Commun 363:994–1000.