Development 120, 2723-2748 (1994) 2723 Printed in Great Britain © The Company of Biologists Limited 1994

Function of the retinoic acid receptors (RARs) during development (I) Craniofacial and skeletal abnormalities in RAR double mutants

David Lohnes*,†, Manuel Mark*, Cathy Mendelsohn*,‡, Pascal Dollé, Andrée Dierich, Philippe Gorry, Anne Gansmuller and Pierre Chambon¤ Laboratoire de Génétique Moléculaire des Eucaryotes du CNRS, Unité 184 de Biologie Moléculaire et de Génie Génétique de l’INSERM, Institut de Chimie Biologique, Faculté de Médecine, 11 rue Humann, 67085 Strasbourg Cedex, France *Should be considered as equal first authors †Present address: Institut de recherches cliniques de Montréal, Laboratoire de Biologie Cellulaire et Moléculaire, 110 Avenue des Pins-Ouest, Montréal, Québec H2W 1R7, Canada ‡Present address: Columbia University, Department of Physiology and Biophysics, 630 W 168th Street, New York, NY 10032, USA ¤Author for correspondence

SUMMARY

Numerous congenital malformations have been observed in abnormalities, which are reported in this and in the accom- fetuses of vitamin A-deficient (VAD) dams [Wilson, J. G., panying study. We describe here multiple eye abnormali- Roth, C. B., Warkany, J., (1953), Am. J. Anat. 92, 189-217]. ties which are found in various RAR double mutant fetuses Previous studies of retinoic acid receptor (RAR) mutant and are similar to those previously seen in VAD fetuses. mice have not revealed any of these malformations [Li, E., Interestingly, we found further abnormalities not previ- Sucov, H. M., Lee, K.-F., Evans, R. M., Jaenisch, R. (1993) ously reported in VAD fetuses. These abnormalities affect Proc. Natl. Acad. Sci. USA 90, 1590-1594; Lohnes, D., ocular glands, salivary glands and their associated ducts, Kastner, P., Dierich, A., Mark, M., LeMeur, M., Chambon, the axial and limb skeleton, and all skeletal elements P. (1993) Cell 73, 643-658; Lufkin, T., Lohnes, D., Mark, derived from the mesectoderm of the frontonasal mass and M., Dierich, A., Gorry, P., Gaub, M. P., LeMeur, M., of the second and third pharyngeal arches. RAR double Chambon, P. (1993) Proc. Natl. Acad. Sci. USA 90, 7225- mutants also exhibit supernumerary cranial skeletal 7229; Mendelsohn, C., Mark, M., Dollé, P., Dierich, A., elements that are present in the ancestral reptilian skull. Gaub, M.P., Krust, A., Lampron, C., Chambon, P. (1994a) The role of retinoic acid (RA) and of the RARs in the onto- Dev. Biol. in press], suggesting either that there is a con- genesis of the affected structures, particularly of those that siderable functional redundancy among members of the are derived from mesenchymal neural crest cells, is RAR family during ontogenesis or that the RARs are not discussed. essential transducers of the retinoid signal in vivo. In order to discriminate between these possibilities, we have generated a series of RAR compound null mutants. These Key words: retinoic acid, retinoic acid receptors, ontogenesis, neural RAR double mutants invariably died either in utero or crest, evolution, atavisms, mouse, homeotic transformations, eye, shortly after birth and presented a number of congenital skull, limb

INTRODUCTION retinol excess, causing many developmental abnormalities, the precise malformation depending largely on the time of admin- It has long been known that retinol (vitamin A) is crucial for istration (reviewed in Morriss-Kay, 1993; Nau et al., 1994; normal growth, vision, reproduction, maintenance of numerous Hofman and Eichele, 1994). The spectacular effects of topical tissues and overall survival (Wolbach and Howe, 1925; see RA application on limb development and regeneration popu- Sporn et al., 1994 and Blomhoff, 1994, for reviews and refer- larized the belief that RA could in fact be a morphogen (for ences). Retinol is also essential for normal development, as reviews, see Tabin, 1991; Hofman and Eichele, 1994). shown by the appearance of multiple congenital abnormalities The discovery of a nuclear receptor for RA, acting as a in fetuses from dams fed a vitamin A-deficient (VAD) diet (the ligand-inducible transcriptional regulator (RAR; Petkovich et fetal VAD syndrome, see Wilson et al., 1953 and references al., 1987; Giguère et al., 1987), greatly advanced our under- therein). Interestingly, with the exception of vision (Wald, standing of the molecular mechanisms underlying the 1968), retinoic acid (RA) appears to be the active derivative of pleiotropic effects of retinoids (synthetic and natural derivatives vitamin A, since its administration can prevent or reverse most of RA; reviewed in Leid et al., 1992; Kastner et al., 1994; Man- of the defects induced by postnatal VAD (Thompson et al., gelsdorf et al., 1994; Linney and LaMantia, 1994). Since this 1964). Furthermore, RA excess is much more teratogenic than initial finding, it has been shown that the RA signal can be trans- 2724 D. Lohnes and others duced in cultured cells through two families of retinoid Taken together, the above findings suggest either that there receptors. The RAR family (RARα, β and γ and their isoforms) is a high degree of functional redundancy among members of are activated by both all-tr a n s RA and 9-ci s RA, whereas the the RAR family, or that the RARs are not essential transduc- RXR family (RXR α, β and γ) are activated only by 9-ci s RA . ers of the retinoid signal in vivo. To discriminate between these The DNA-binding and ligand-binding regions (regions C and possibilities, we have generated and analyzed RAR compound E, respectively) of the three RAR types are highly similar, null mutants. The defects displayed by various double mutants whereas the C-terminal F region and the central D region recapitulate essentially all of the congenital malformations exhibit little, if any, conservation. The three RAR types also found in fetal VAD. These double mutants also exhibit a diverge in their N-terminal B regions and further diversifica t i o n number of abnormalities not previously described in VAD is generated for each receptor type by variant isoforms differing experiments (see also the accompanying study). We report here in their N-terminal-most A regions (RARα1 and α2, β1 to β4, a detailed analysis of the craniofacial and skeletal defects and γ1 and γ2), which originate from alternate splicing and dif- found in RAR double null mutants. ferential promoter usage (reviewed in Leid et al., 1992). Amino acid sequence comparisons have revealed that the interspecies conservation of a given RAR type and of each of its isoforms MATERIALS AND METHODS is greater than the similarity found between the three RARs within a given species (see Kastner et al., 1994 for review). Fur- Generation of RAR double mutants α α β γ thermore, the various RAR isoforms contain two transcriptional The generation of RAR 1, , 2 and single null mutants has been activation functions (AFs), located in the N-terminal A/B region described (Lohnes et al., 1993; Lufkin et al., 1993; Mendelsohn et al., 1994a). Initial intercrosses of these single mutants were performed to (AF-1) and C-terminal E region (AF-2) which act synergisti- derive double heterozygotes. With the exception of RARα+/−/γ+/− cally, and sometimes differentially, to activate various RA- offspring, second generation animals were obtained by mating double responsive promoters. Taken together with the distinct spa- heterozygotes with the appropriate RAR heterozygous or homozy- tiotemporal transcript distribution observed for each RAR and gous single mutants to test for viability and fertility of compound isoforms during mouse embryogenesis and in adult tissues, the mutants. This was performed in order to optimize subsequent gener- above interspecies sequence conservation and transcriptional ation of double null mutants. The RAR double mutants and the crosses used to generate them were as follows: RARα1−/−/β2−/− were derived activation specificities suggested that each RAR isoform may − − − − perform unique functions (for refs see Kastner et al., 1994; from RARα1+/ /β2 / intercrosses or crosses between RARα1+/ / β2−/− and RARα1+/−/β2+/− animals; RARα−/−/β2−/− and RARα−/−/ Chambon, 1994). Furthermore, the finding that RA-responsive β +/− α+/− β −/− promoters are likely controlled in cultured cells through RAR- 2 mutants were derived from crosses between RAR / 2 males and RARα+/−/β2+/− or RARα+/−/β2−/− females; RARα1−/−/γ−/− RXR heterodimers (reviewed in Kastner et al., 1994; Mangels- mutants were derived from crossing RARα1−/−/γ+/− males and dorf et al., 1994; Chambon, 1994) suggested that the diverse RARα1+/−/γ+/− females; RARα1−/−/γ−/−/α2+/− mutants were derived effects of retinoids may also reside in the control of various from crosses between RARα1−/−/γ+/− and RARα+/−/γ+/− animals; subsets of retinoid-responsive promoters by different combina- RARα−/−/γ−/− mutants were obtained from intercrosses between tions of RAR-RXR types (and isoforms). RARα+/−/γ+/− animals; RARβ2−/−/γ−/− mutants were obtained To evaluate the function of the various RARs (types and from crosses between RARβ2−/−/γ+/− males and RARβ2+/−/γ+/− or − − − isoforms) in vivo, we have created mice lacking several of RARβ2 / /γ+/ females. these receptors. Surprisingly, mice deficient for RARα1 (Li et Matings and genotyping of offspring al., 1993; Lufkin et al., 1993), RARβ2 (Mendelsohn et al., 1994a) or RARγ2 (Lohnes et al., 1993) isoforms were appar- Animals were mated overnight and females examined for a vaginal plug the following morning. Noon of the day of evidence for a vaginal ently unaffected. In contrast, mice deficient for RARα or γ α γ plug was considered 0.5 dpc. Embryos (10.5-14.5 dpc) or 18.5 dpc RAR receptors (all or isoforms disrupted) exhibited post- fetuses were obtained by Cesarean section and genotypes determined partum lethality and growth deficiency (Lohnes et al., 1993; by genomic Southern blotting using DNA isolated from the yolk sac Lufkin et al., 1993). Furthermore, RARα null mice presented or placenta. Probes, digests and other conditions for southern blotting a degeneration of the testicular germinal epithelium, which was have been detailed elsewhere (Lohnes et al., 1993; Lufkin et al., 1993; similar to that observed in male rats maintained on a VAD diet Mendelsohn et al., 1994a). (Howell et al., 1963). Male RARγ null mutants also exhibited VAD-like abnormalities, namely squamous metaplasia of the Histological and skeletal analysis For whole-mount skeletal analysis, fetuses were collected at 18.5 dpc seminal vesicles and prostate gland. Agenesis of the ocular − Harderian gland and homeotic transformations of the axial and stored at 20¡C. Skeletons were prepared as described (Lufkin et γ al., 1992). For histological analysis, embryos or skinned fetuses were skeleton were also observed in RAR null mutants, although fixed in Bouin’s solution. Paraffin sections, 7 µm thick, were stained both of these defects occurred with incomplete penetrance and with hematoxylin and eosin or Groat’s hematoxylin and Mallory’s expressivity. In no case, however, have these RAR null trichrome (Mark et al., 1993). mutants displayed any of the congenital malformations observed in fetal VAD studies (Wilson et al., 1953). Further- more, the phenotype of RARα and RARγ null mice was RESULTS confined only to a small subset of tissues expressing these receptor types. Thus, contrary to our expectations, the RAR (A) Viability types and isoforms disrupted to date do not apparently possess Animals lacking the RARα1 or β2 isoforms were apparently the unique functions that were predicted on the basis of their normal (Table 1A; see refs also Lufkin et al., 1993; Mendel- evolutionary conservation, expression pattern and in vitro tran- sohn et al., 1994a). Although exhibiting a high degree of scriptional regulatory characteristics. neonatal lethality, animals lacking all isoforms of either RARα RARs in ontogenesis 2725 or RARγ survived in isolation for at least 24 hours when (B) External features delivered by Cesarean section at 18.5 day postcoitum (dpc), Upon external inspection at 18.5 dpc, α1β2, αβ2 and β2γ being in this respect as viable as their control littermates (Table fetuses could not be distinguished from their littermates. In 1A; see Lohnes et al., 1993; Lufkin et al., 1993). With the contrast, αγ mutant fetuses could be readily identified by their exception of αγ double mutants, the distribution of RARα1β2, reduced size (compare Fig. 1a and c), the small size or apparent αβ 2, α1γ, α1γα 2+/ − and β2γ double mutant offspring at 18.5 absence of the eyes (asterisk, compare Fig. 1a with c and g-i) dpc indicated that loss of these receptors did not result in and the aspect of their mid-facial region (compare Fig. 1d and embryonic lethality [Table 1A; for the sake of simplicity f). Their snout was markedly foreshortened and divided by a mutants null for RARα1 and RARβ2, RARα (all isoforms) and sagittal median cleft (large arrow, Fig. 1f). The prolabium RA R β2, etc, are called hereafter and in the accompanying study (median third of the upper lip, dashed box in Fig. 1d) was α1β2, αβ 2, etc, mutants]. However, in contrast to the RAR absent (compare Fig. 1d and f). The maxillary processes which single mutants, these double mutants invariably died within at bear whiskers (double arrow, Fig. 1f) were located farther apart most 12 hours following Cesarean delivery at 18.5 dpc. than in WT animals. Paramedian swellings likely correspond- The frequency of RARαγ mutants found at 18.5 dpc was ing to the nasomedial processes (single small arrow, Fig. 1f lower than predicted from Mendelian distribution (Table 1B), and see below) were fused to the maxillary processes ventral indicating partial embryonic lethality. Although analysis of to the nostrils, which opened dorsally instead of rostrally earlier stages of development (10.5-13.5 dpc) yielded the (compare arrowheads, Fig. 1d and f). Additional external expected frequency, a large fraction were either dead or defects occasionally found in these mutants included: exteri- partially resorbed (Table 1B). The time of death appeared orized brain (exencephaly, Fig. 1h and i, and Table 1B), variable, as judged by the size and relative proportion of bilateral agenesis of the auricle (open arrows, compare Fig. 1a resorbed and dead mutant embryos (Table 1B, data not shown). with c and g), umbilical hernia (Fig. 1i, large black arrow), and The embryonic and postpartum lethality may be due to one or abnormal limbs (e.g. Fig. 1c,h and i, and see below). more malformations affecting the heart, aortic arches, kidney, A persistent opening of the rhombencephalic neural tube lung or trachea which were observed in RAR double mutant was observed in nine out of twenty living αγ mutant embryos fetuses (see the accompanying study). between 10.5 and 11.5 dpc (e.g. compare Fig. 2b and e) and in

Table 1. (A) Viability of 18.5 dpc RAR double null mutant fetuses Mutant Expected Mutant Total frequency frequency genotype offspring Mutants % % Viability RARα1 218 58 27 25 Viable RARα 137 39 28 25 Neonatal lethal† RARβ2 345 100 29 25 Viable RARγ 280 65 23 25 Neonatal lethal‡ RARα1β2 45 13 29 25 Viability* ≤12 hrs RARαβ2 170 15 9 12.5 Viability* ≤1 hr RARα1γ 243 29 12 12.5 Viability* ≤12 hrs RARα1γα2+/− 122 16 13 12.5 Viability* ≤1 hr RARαγ 355 11 3 6.25 Embryonic lethal Viability* ≤1 hr RARβ2γ 137 19 13 12.5 Viability* ≤12 hrs

*Fetuses were delivered by Cesarean section at 18.5 dpc and compared to littermates for postpartum (PP) viability in an isolated environment; under these conditions WT littermates survived for at least 24 hours. †,‡ Animals delivered by Cesarean section survived as well as WT littermates, but showed a high degree of mortality in the first 4 weeks of life following natural birth. †See Lufkin et al., 1993; ‡see Lohnes et al., 1993.

(B) Embryonic lethality of RARαγ double mutant offspring from RARα+/−/γ+/− intercrosses RARαγ double mutants Mutants with neural tube defect‡ Mutants Relative to Days post- Total Total resorbed or Mutants Total living coitum offspring mutants* dead† living† mutants mutants 10.5 303 20 (6.6%) 6 (30%) 14 (70%) 6 43% 11.5 190 11 (5.8%) 5 (45%) 6 (55%) 3 50% 13.5 196 11 (5.6%) 4 (36%) 7 (64%) 1 14% 18.5 355 13 (3.7%) 2 (15%) 11 (85%) 5 45%

*Percentages are relative to the number of total offspring. †Percentages are relative to total mutants. ‡Neural tube defect in 10.5 to 13.5 dpc embryos was observed in the rhombencephalic region, whereas in 18.5 dpc fetuses this refers to exposure of the brain (exencephaly, see text). 2726 D. Lohnes and others

Fig. 1. Comparison of external features between 18.5 dpc (a,d) wild-type (WT) and (b,c,e-i) RAR double mutant fetuses. The mutant genotype is indicated in the lower right corner of each photograph. The asterisks indicate the region of the eye; the open arrows point to the region of the external ear; the arrowheads point to the nostrils; the large white arrow indicate midfacial clefts. (i) The large black arrow points to an umbilical hernia; (f) the single small arrow and the double arrow point to the nasomedian and maxillary processes of the αγ mutant fetus, respectively. (d) The dashed box indicates the prolabium. Same magnifications for a-c and for d-f. RARs in ontogenesis 2727

Fig. 2. Comparison of (a-f) whole mounts and (g-i) histological sections of facial and brain structures between (a,b,d,e,i) 10.5 dpc and (c,f-h) 11.5 dpc wild-type (WT, a-c,g) and αγ exencephalic mutant embryos (d-f,h,i). Histological sections were performed through the forebrain and frontonasal mesenchyme. Abbreviations: FC, median facial cleft; LV, lateral ventricles of the brain; MS, mesencephalon; NL; nasolateral process; NM, nasomedial process; NP, nasal placode; OP, olfactory pit; RH, rhombencephalon; TE, telencephalon (cerebral hemispheres). The arrowheads indicate the eye region; (d,e) the large black arrow indicates the scoliosis; (c,f) the broken line indicates the midline. (e) The bracket indicates the extent of the rhombencephalic opening;(i) the dashed box indicates an area of cell necrosis in the frontonasal mesenchyme. Magnifications: ×30 (g,h) and ×150 (i). The same magnifications were used for a,b,d,e and for c,f. 2728 D. Lohnes and others RARs in ontogenesis 2729 one of seven living αγ embryos at 13.5 dpc (Table 1B). In these and the incisive (or premaxillar, PX, Fig. 3a,k) and vomer (not embryos, the telencephalic hemispheres appeared markedly shown) bones (De Myer, 1975). These elements were grossly underdeveloped (TE, compare Fig. 2a,b with Fig. 2d, and Fig. deficient or absent in αγ mutant fetuses. The medial portions 2c with Fig. 2f). Some αγ mutant embryos also showed a of the frontal (F) and nasal (N) bones were lacking (compare severe degree of lateral deviation of the vertebral axis Fig. 3a with c,d). The nasal capsule was apparently reduced to (scoliosis, arrows in Fig. 2d,e). However, the only pathogno- laterocaudal rudiments (NC, compare Fig. 3a with c,d, and Fig. monic external feature of αγ mutant embryos between 11.5 and 3k with m). The rest of the nasal capsule, the nasal septum, the 13.5 dpc was the aspect of the frontonasal segment of the face: lamina cribriform and the vomer and incisive bones could not both the nasolateral (NL) and the nasomedial (NM) processes, be identified (NS and PX, compare Fig. 3a with c,d, and Fig. located on either side of the ofactory pit (OP), were present 3k with m), and were apparently replaced by aggregates of car- (compare Fig. 2c and f). The nasomedial processes were tilaginous and bony nodules or rods (asterisks in Figs 3m, 4c, normally fused with ipsilateral maxillary processes, but were 5c,d). In the cranial base, caudal to the ethmoid bone, similar never fused at the midline resulting in a median facial cleft aggregates replaced the presphenoid bone (data not shown). (FC; compare Fig. 2c with f, and Fig. 2g with h). This lack of The upper incisors, which are largely derived from the naso- fusion may result from cell deficiency caused by excessive cell medial processes (Lumsden and Buchanan, 1986), were death in the frontonasal mesenchyme, which was observed on lacking (not shown). The hypophyseal foramen of the serial histological sections in 10.5 dpc αγ embryos (dashed basisphenoid was never closed (arrowhead in Fig. 3m), thus box, Fig. 2i). the pituitary gland (HY, Fig. 4g) remained in contact with the 18.5 dpc α1γα2+/− fetuses were brachycephalic (compare pharynx. Fig. 1a and b) and occasionally showed a median cleft of the Many of the first pharyngeal arch-derived skeletal elements upper lip (compare Fig. 1d and e). In contrast, α1γ mutants did (Noden, 1988; Le Douarin et al., 1993 and references therein) not exhibit any abnormal external features. were also malformed (e.g. maxillary and palatine bones) or hypoplastic (e.g. alisphenoid) (compare X, AL and PL in Fig. (C) Craniofacial skeletal abnormalities 3k,m). Surprisingly, however, the mandibular (dentary) bone (1) Defects of the craniofacial skeleton and teeth (compare D in Fig. 3a,c,d), the temporomandibular joint (not Craniofacial skeletal deficiencies were not observed in 18.5 shown), the malleus middle ear ossicle (not shown) and the dpc αβ2, α1β2, or β2γ mutant fetuses. In contrast, most of the tympanic bone (T, compare Fig. 3k with m) appeared normal, neural crest cell (NCC)-derived craniofacial skeletal elements as did the patterning of the lower dentition and the shape were altered in αγ mutants. The most severe skeletal defects (cuspal pattern) of the first and second upper and lower molars were observed in the midfacial region and rostral cranial base, and of the lower incisors (compare UM and LM in Fig. 4a,c, consistent with the loss of midfacial structures described and data not shown). Second and third pharyngeal arch-derived above. The skeletal elements normally derived from the fron- skeletal elements were either absent (stapes, not shown) or tonasal mesectoderm are the frontal (F) and nasal (N) bones highly malformed (styloid and hyoid bones, see the accompa- (Fig. 3a), the cartilaginous template of the ethmoid bone [com- nying study and data not shown). prising the nasal septum (NS, Figs 3k and 4a), the nasal capsule In non-exencephalic αγ mutants, the skull vault caudal to the (NC, Figs 3a,k, 4a) and the lamina cribriform (LC, Fig. 4a)], frontal region was complete, although markedly underossified [compare the size of the parietal (P) and interparietal (IP) bones in Fig. 3a,d; also note the absence of interparietal (IP) and Fig. 3. Comparison of the craniofacial skeleton between 18.5 dpc supraoccipital (S) ossification centers in Fig. 3c]. In contrast, wild-type (WT) and RAR double mutant fetuses (genotypes as the entire cranial vault was absent in exencephalic mutants indicated on the photographs). (a-d) Lateral views of the skull. Note (Fig. 7k). that c and d represent different fetuses. (e,f) Dorsal views of the The cartilaginous otic capsule was always small and incom- cranial base; the orbitosphenoid bone has been removed in order to plete in αγ mutants (O, compare Fig. 3a,c,d), resulting in a show the pila prooptica (PP) and the pila metoptica (PM) more cystic protrusion of the epithelial inner ear within the braincase clearly. (g-j) Dorsal views of the WT incus (g) or incus plus (not shown). These malformations are likely to be secondary alisphenoid bone (h-j). (k-m) Ventral views of the cranial base; the to defects of the otocyst (discussed in Mark et al., 1993), which dentary bone has been removed from these whole mounts. αγ Abbreviations: AL, alisphenoid bone; BI, body of the incus; BO, was consistently hypoplastic in 10.5 dpc mutants (not basioccipital bone; BS, basisphenoid bone; D, mandibular (dentary) shown). That the epithelial inner ear is a RA-target organ bone; E, exoccipital bone; F, frontal bone; IF, incisive foramen; IP affected early during embryogenesis is supported by the interparietal bone; LI, long process of the incus; N, nasal bone; NC, absence of some of its derivatives (i.e. the spiral organ of Corti nasal capsule; NS, nasal septum; O, otic capsule; OB, orbitosphenoid and the spiral ganglion, data not shown) in 18.5 dpc αγ fetuses. bone; OF, optic foramen; P, parietal bone; PA, pila antotica; PL, Although less affected than αγ mutants, α1γα2+/− mutants palatine bone; PM, pila metoptica; PP, pila prooptica; PS, (but not α1γ mutants) also exhibited several defects of the presphenoid bone; PX, incisive (premaxillary) bone; Q, cranial skeleton. These included shortening of the frontal bone pterygoquadrate rod; S, supraoccipital bone; SI, short process of the (F, compare Fig. 3a and b), duplication of the cartilaginous incus; T, tympanic bone; X, maxillary bone; (m) asterisk marks cartilaginous nodules replacing the nasal capsule; (f) the arrows point nasal septum (NS, compare Fig. 3k and l), cleft palate, aplasia to a deficiency in the presphenoid bone; (c) the small arrows delimit of the presphenoid (PS, compare Fig. 3e and f), persistence of the median gap between the left and right frontal bones; (a,b) the the hypophyseal foramen (arrowhead Fig. 3f) and absence of broken line delimits the frontoparietal suture; (f,m) the arrowheads the incisive foramen (IF, compare Fig. 3k and l). The latter point to a persistent hypophyseal foramen. The same magnifications malformation was closely correlated with the absence or were used for a-d and k-m, for e and f, and for g-j. dysplasia of the upper incisors and may be secondary to this 2730 D. Lohnes and others RARs in ontogenesis 2731

Table 2. Fusion of the incus with a cartilaginous or osseous In mammals, the alisphenoid bone contributes to the base element in RAR double mutant fetuses and lateral walls of the skull between the optic and otic regions where it forms the lateral limit of the cavum epiptericum, Genotype and number Fusion of 18.5 dpc fetuses Unilateral Bilateral frequency whereas the incus (Fig. 3g) represents one of the three middle +/− +/− examined fusion fusion (%) ear ossicles. In the middle ears of α1β2, αβ2 , αβ2, α1γα2 αγ α γ β γ WT 16 0 0 0 and , but not 1 or 2 mutants, the medial aspect of the RARα1γ 10 0 0 0 body of the incus was continuous with a rostrally oriented car- RARα1γα2+/− 8 3 1 31 tilaginous or osseous rod (Q, Fig. 3h-j) which was frequently RARαγ 6 0 6 100 fused to the alisphenoid bone (Table 2). In a number of these RARα1β2 8 2 1 25 mutants, the short process of the incus (SI) was conspicuously RARαβ2+/− 9 3 2 39 RARαβ2 16 5 7 59 larger than its wild-type homologue (compare Fig. 3g and i). RARβ2γ 11 0 0 0 (D) Brain abnormalities The brains of 18.5 dpc α1β2, αβ2 and β2γ mutants appeared defect. Ectopic cartilaginous and bony nodules, formed from normal. In contrast, in 10.5 dpc αγ mutants, a wide persistent the meninges, were also found in 18.5 dpc αγ and α1γα2+/− opening of the rhombencephalic neural tube was frequently mutant fetuses. In particular, in αγ mutants, the falx cerebri observed (see above and Fig. 2e). The midbrain and forebrain was completely chondrified (FX, Fig. 4c). of these exencephalic embryos were invariably closed and covered by ectoderm (Fig. 2d,e). However, the neurectoderm (2) Supernumerary cranial skeletal elements of the telencephalic vesicles (TE) was abnormally folded and With the exception of the ectopic cartilaginous and bony the lateral ventricles (LV) were collapsed (compare Fig. 2g,h). deposits, the above skeletal abnormalities correspond to defi- Histological analysis of two 18.5 dpc exencephalic fetuses ciencies, including the duplicated nasal septum which arises showed a lack of hindbrain and cerebellar structures; further- by failure of coalescence of the nasomedial processes (De more the cerebral hemispheres were small and displayed Myer, 1975). Aside from these deficiencies and ectopias, two hemorragic foci (not shown). Taken together with the finding supernumerary skeletal elements were frequently detected. that persistent opening of the rhombencephalon at 10.5-11.5 In placental mammals, two cartilaginous pillars, the pila dpc and exencephaly at 18.5 dpc occurred with similar fre- prooptica and pila metoptica, connect the orbital (optic) region quencies (Table 1B), these results suggest that failure of of the fetal skull to the floor of the braincase (De Beer, 1985). closure of the rhombencephalic neural tube is the primary In normal mice at 18.5 dpc, these pilae (PP and PM in Fig. 3e) defect leading to exencephaly. This failure, which leaves the are located on either side of the optic foramen (OF) and are rhombencephalic neurectoderm exposed to the amniotic fluid, fused ventrally to the presphenoid bone (PS). All 18.5 dpc αγ , may result in rhombencephalon degeneration and impair the α1γα 2+/ − and α1γ mutant fetuses possessed a third, more caudal, accumulation of cerebrospinal fluid in the ventricular system. cartilaginous pillar which was fused ventrally to the basisphe- The absence of hydrostatic pressure would then lead to the noid bone to form a cartilaginous medial wall to the cavum abnormal folding of the cerebral hemispheres (Pexieder and epiptericum [PA, compare Fig. 3e with f, and Fig. 4e-g with 4d; Jelinek, 1970; Jacobson, 1981), thus altering the relationship the size of this additional pillar was greater in αγ and α1γα 2+/ − between the neurectoderm and overlying osteogenic cranial (not shown) mutants than in α1γ mutants]. The cavum epipter- mesectoderm and subsequently impairing the epithelial-mes- icum (bracketed in Fig. 4d), which corresponds to a normal enchymal skeletogenic interactions required for the formation extracranial space (i.e. located outside of the dura mater), is so of the bones of the skull vault (Hall, 1991). The improper inter- called as it is limited laterally by the alisphenoid bone (i.e. the actions between the folded neuroepithelium and presumptive mammalian homologue of the reptilian epipterygoid bone) (AL, dermis might also lead to the absence of skin covering the brain Figs 3e,k, 4d). This cavum lodges the trigeminal ganglion (G5, in 18.5 dpc exencephalic mutants. Fig. 4d-f) and is crossed by cranial nerves III, IV, V (N5, Fig. The brain of 18.5 dpc non-exencephalic αγ mutant fetuses 4g) and VI, and the internal jugular vein (JV, Fig. 4g). appeared considerably distorted (BR, Fig. 4c and data not shown), probably secondary to increased intracranial pressure caused by the shortening of the braincase and compression by Fig. 4. Comparison, on frontal histological sections, of craniofacial intracranial ectopic cartilaginous and bony nodules (e.g. FX in skeletal structures between 18.5 dpc wild-type (WT; a,d) and RAR double mutant fetuses (b,c,e-g). The genotype of the mutant fetuses Fig. 4c). In addition, failure of the rostral interhemispheric is indicated in the upper-right corner of each micrograph. (a-c) commissures (i.e. corpus callosum, hippocampal commissure Sections at the level of the first upper (UM) and lower (LM) molars. and anterior commissure) to cross the midline was consistently (d-g) Sections through the cavum epiptericum (brackets in d). observed, a condition that is also frequently encountered in Abbreviations: AL, alisphenoid bone; BR, brain; BS, basisphenoid humans with median cleft face syndrome (De Myer, 1975; bone; CP, cleft of the palate; EY, eye : FX, chondrification of the Cohen and Sulik, 1992). In the mutant hindbrain, the motor falx cerebri; G5, ganglion of the trigeminal nerve (5th cranial nerve); nucleus of the abducens nerve (derived from rhombomeres 5 HY, pituitary gland (hypophysis); JV, jugular vein; LC, lamina and 6, Lumsden et al., 1991) was not identifiable (not shown); cribriform of the ethmoid bone; LM, lower molar; N5, trigeminal whether this reflects a primary effect on the hindbrain or is nerve (5th cranial nerve); NC, nasal capsule; NS, nasal septum; OL, secondary to the abnormalities present in the ocular region (see olfactory epithelium; P, secondary palate; PA, pila antotica; PT, pterygoid bone; TO, tongue; UM, upper molar. (c) Asterisks indicate below) is unclear, since the target organ of the abducens nerve cartilaginous rods and nodules replacing the nasal septum and nasal is the external rectus muscle of the eye. capsule. Magnifications: ×16 (a,b); ×20 (c-f); ×32 (g). Distortions of the brain and absence of the interhemispheric 2732 D. Lohnes and others RARs in ontogenesis 2733

Table 3. Main abnormalities of the eye of RAR double null mutants RAR mutant genotype and number of 18.5 dpc fetuses examined α1γ α1γα2+/− αγ αβ2+/− α1β2 αβ2 β2γ VAD 5 5 6 4 3 7 3 Coloboma of the retina 0 0 12/12 0 0 0 1/6 + Coloboma of the optic nerve 0 0 12/12 0 0 0 6/6 + Fibrous retrolenticular membrane 0 0 12/12 0 4/6 14/14 6/6* + Unfused eyelids 0 4/10 12/12 0 0 0 0 + Small conjunctival sac 0 0 12/12 0 0 0 6/6 + Corneal - lenticular stalk 0 4/10 4/12 0 0 0 0 NR Abnormal corneal stroma 0 0 8/12 0 0 0 6/6 + Absence of the anterior chamber 0 0 12/12 0 0 0 6/6 + Abnormal lens fibers 0 0 4/12 0 0 0 0 + Agenesis of the cornea and conjunctiva 0 0 4/12 0 0 0 0 NR Agenesis of the lens 0 0 2/12† 0 0 0 0 NR Shorter ventral retina 0 0 0 0 0 0 0 +

The number of eyes showing the abnormality versus eyes analysed on serial sections is given; *chondrified. †Observation made on a 12.5 dpc embryo (see text). Note that all abnormalities were found bilaterally, with the exception of coloboma of the retina in β2γ mutants. NR =not reported. commissures was also found in two out of five 18.5 dpc ventral gap permitted communication between periocular mes- α1γα2+/− mutants (BR, Fig. 4b). In these mutants, which enchyme and persistent retrolenticular mesenchyme (F, Fig. display nearly normal eyes, the abducens nucleus was always 5c), which occupied the space normally taken by the vitreous present. The brains of α1γ mutants, which did not exhibit gross body (V, Fig. 5a). The two ventral gaps were joined caudally. cranial skeletal malformations, appeared normal. In this plane of section, however, they were separated by a portion of the retina with a duplicated internal leaf. This aspect (E) Eye defects is characteristic of an eversion of the retina (ER, Fig. 5c), The eye of a normal 18.5 dpc fetus is covered by fused eyelids which is thought to arise by metaplastic transformation of (Y, Fig. 5a,e). A wide conjunctival sac (J, Fig. 5a,e) is present pigmented epithelium into neural epithelium (Coulombre and between the lids and the cornea (C). There are two cell-free Coulombre, 1977). The cleft in the ventral portion of the retina, spaces, one between the cornea and the lens (the anterior the penetration of the optic cup by mesenchymal tissue and the chamber; A, Fig. 5a,e), and the other between the lens and the eversion of the retina in the cleft region are characteristic of retina (the vitreous body; V, Fig. 5a,e). The retina, comprising the typical complete coloboma of the retina (Mann, 1937). The the internal (neural) and external (pigmented) epithelia (IR and developmental fault underlying this defect is a complete per- OR, Fig. 5a,e), encloses the lens and the vitreous body and is sistence of the optic fissure (also called fetal or choroid fissure, continuous with the iris (I, Fig. 5a,e) laterally. which normally closes completely by 14.0 dpc) through its RARαγ, α1γα2+/−, α1β2, αβ2 and β2γ mutants exhibited a entire length from the region of the optic disc (the optic nerve number of ocular defects (Table 3). Microphtalmia, coloboma exit point) to the iris. Additional defects in this eye included: of the retina and abnormalities of the cornea, eyelids and con- absence of fusion of the eyelids (compare Y, in Fig. 5a and c), junctiva, were constant features of 18.5 dpc αγ fetuses. As agenesis of the upper (dorsal) conjunctival sac (compare J, in shown in Fig. 5c, the eye was conspicuously smaller than Fig. 5a and c), persistent corneal-lenticular stalk (see below), normal (compare Fig. 5a and c). The dorsal margin of the retina absence of differentiation of the corneal stroma (C, in Fig. 5a,e) extended well past the equator of the lens (L, Fig. 5c), where and absence of the anterior chamber of the eye (A, in Fig. 5a,e). it formed a rudimentary iris (I, Fig. 5c). Ventrally, the retina Another αγ mutant eye is displayed in Fig. 5d showing a cleft barely reached the equator (possibly secondary to the restricted to the medial portion of the ventral retina (i.e. typical coloboma), so that the ventral lens region was in direct contact ‘partial’ coloboma of the retina). The corneal stroma was with periocular mesenchyme (PO, Fig. 5c). Medially, a second present (C, in Fig. 5d,f), but was fused with that of the iris (I, Fig. 5e,f), resulting in an absence of the anterior chamber; the corneal epithelium was locally hyperplastic and keratinized Fig. 5. Comparison of frontal sections through the left eye between (EP, compare Fig. 5e and f). Additional ocular abnormalities 18.5 dpc (a,e) wild-type (WT) and (b-d,f-i) RAR double mutant in two 18.5 dpc exencephalic αγ mutants included agenesis of fetuses; the genotype of the mutant fetuses is indicated in the upper the conjunctiva and cornea, and abnormal lens fibers (data not right corner of each micrograph. Abbreviations: A, anterior chamber shown). Primary aphakia (i.e. failure of lens formation) was of the eye; C, corneal stroma; EP, corneal epithelium; ER, everted observed bilaterally in one non-exencephalic 12.5 dpc αγ retina; F, fibrous retrolenticular membrane; I, iris, IR, internal leaf of mutant (data not shown; see also Table 3). the retina; J, conjunctival sac; L, lens; ON, optic nerve; OR, external (outer or pigmented) leaf of the retina; PO, periocular mesenchyme; Interestingly, the only ocular malformations found in two of α γα +/− V, vitreous body; Y, eyelids; asterisks, cartilaginous nodules five 1 2 mutants were unfused eyelids (not shown) and replacing the ethmoid bone. (g) The large black arrow points to the a corneal-lenticular stalk. This latter abnormality was charac- persistent corneal-lenticular stalk. (i) The arrow points to a terized by a persistent continuity of the corneal and lens cartilaginous nodule developed from the fibrous retrolenticular epithelia at the center of the cornea (large arrow, Fig. 5g), membrane. Magnifications: ×46 (a-d,h); ×76 (e,f,i); ×229 (g). which most probably results from the failure of the lens cup to 2734 D. Lohnes and others pinch off from its parental surface ectoderm (Coulombre and The parenchyma of the submandibular and sublingual Coulombre, 1977). The eyes of α1γ mutants were unaffected salivary glands develop from downgrowth of the buccal (Fig. 6b; Table 3). ectoderm and from the adjacent mesenchyme, whereas the A partially chondrified fibrous retrolenticular membrane (F main ducts of these two glands are formed by closure, in a and arrow, Fig. 5h,i) was a specific feature of β2γ mutants. It rostral direction, of gutter-like grooves in the oral ectoderm was always continuous with a thick strand of mesenchyme, (Hamilton et al., 1945). In 18.5 dpc wild-type fetuses, the two which was embedded in the optic nerve (ON, Fig. 5h) and was ducts (MAD and LID, Fig. 6e) open in the mouth cavity at the carried along with it into the eye through a coloboma of the sublingual caruncle, a median mucosal fold located at the level optic nerve. The retina appeared normal, with the exception of of the rostral third of the lower incisors (LI, Fig. 6e). Both ducts a unilateral coloboma of the ventral retina in one out of three were shortened in all α1γ fetuses (Table 4): the submandibu- mutants (see ER, Fig. 5h). All β2γ eyes also exhibited a small lar duct (MAD, Fig. 6f) opened at the caudal end of the lower conjunctival sac (J, compare Fig. 5a and h, and Fig. 5e and i), incisor (LI, Fig. 6f), whereas the sublingual duct opened even poorly differentiated corneal stroma and absence of the anterior more caudally at the level of the 2nd lower molar (not shown). chamber of the eye (Fig. 5h,i; see also Table 3). The sublingual caruncle was always absent. In contrast, only A fibrous retrolenticular membrane (F, Fig. 5b) was the only the sublingual duct was shortened in β2γ fetuses and the sub- abnormality present in the eyes of all αβ2 and of two (out of lingual caruncle was always present (Table 4). Dysplasia of the three) α1β2 18.5 dpc fetuses (Table 3). It is noteworthy that, sublingual gland, consisting of cystic epithelial formations with the exception of the fibrous retrolenticular membrane, the within the parenchyma (compare LIG, Fig. 6g and h), was fre- ocular defects were confined to animals disrupted for RARγ quently observed bilaterally in α1γ fetuses (Table 4), whereas plus either α or β2 (Table 3). the submandibular glands appeared normal (compare MAG, Fig. 6g with h; Table 4). (F) Glandular defects As expected, glandular defects similar those described above The intraorbital, submandibular and sublingual glands and in α1γ mutants were found in all α1γα2+/− and αγ 18.5 dpc their ducts were normal in 18.5 dpc α1β2 and αβ2 mutants fetuses (Table 4), with agenesis of both sublingual and sub- (Table 4). In contrast, in all 18.5 dpc α1γ (Fig. 6b) and β2γ mandibular glands being occasionally observed (Table 4). (not illustrated) mutant fetuses, the epithelial rudiments of all intraorbital glands [which include the lachrymal and Harderian (G) Abnormalities of the axial skeleton glands (HG, Fig. 6a)] were missing bilaterally (Table 4). RARγ null mice exhibit various vertebral abnormalities which However, the neural crest-derived, melanocyte-containing include homeotic transformations and affect primarily the stroma of the Harderian glands was always present (data not cervical region (Lohnes et al., 1993). These abnormalities shown; see Lohnes et al., 1993). The nasolachrymal duct, occurred with variable penetrance and bilateral expressivity, which was consistently missing in α1γ fetuses (compare NLD although the majority (89%) of RARγ null offspring exhibited in Fig. 6c and d), was sometimes present in β2γ fetuses, indi- one or more vertebral malformations. Although vertebral cating that its development was independent from that of the defects were not initially found in RARα null mutants (Lufkin lachrymal glandular epithelium. et al. 1993), analysis of RARα null offspring inbred in a 129

Table 4. Abnormalities of the Harderian, sublingual and submandibular glands and their associated excretory ducts in RAR double null mutants RAR mutant genotype and number of 18.5 dpc fetuses examined† α1γ α1γα2+/- αγ β2γ 5 5 5 3 Agenesis of the Harderian glands* 5 (B) 5 (B) 5 (B) 3 (B) 1 (B) Agenesis of the nasolachrymal duct 5 (B) 5 (B) 5 (B) 2 (U) 1 (B) Cystic dysplasia of the sublingual gland 4 (B) NA 0 3 (U) 2 (B) Cystic dysplasia of the submandibular gland 0 2 (B) 0 1 (U) 1 (B) Agenesis of the sublingual gland and duct 0 5 (B) 0 3 (U) 2 (B) Agenesis of the submandibular gland and duct 0 0 0 1 (U) 1 (B) Shortening of the sublingual duct 5 (B) NA 3 (B) 3 (U) 2 (B) Shortening of the submandibular duct 5 (B) 5 (B) 0 1 (U) Absence of the sublingual caruncle 5 5 5 0

*The lachrymal glands are not individualized at this stage. †No abnormalities were seen in α1β2 and αβ2 double null mutants. NA, not applicable; B, bilateral defect; U, unilateral defect. RARs in ontogenesis 2735

Fig. 6. Comparison of the Harderian glands, nasolachrymal duct, and sublingual and submandibular glands and ducts between (a,c,e,g) day 18.5 wild-type (WT) and (b,d,f,h) α1γ fetuses. (a,b) Frontal sections at the level of the eye and upper molar (UM). (c,d) Frontal sections at the level of the upper incisors (UI). (e,f) Frontal sections at the level of the rostral third (e) or caudal end (f) of the lower incisors (LI). (g,h) Sections through the submandibular (MAG) and sublingual (LIG) glands. Abbreviations: CY, epithelial cyst in the sublingual gland; HG, Harderian gland; L, Lens; LI, lower incisor; LID, main duct of the sublingual gland; LIG, sublingual gland; MAD, main duct of the submandibular gland; MAG, submandibular gland; N5, maxillary branch of the trigeminal nerve (5th cranial nerve); NLD, nasolachrymal duct; NS, nasal septum; NT, nasal cavity; PX, incisive (premaxillary) bone; R, retina; TO, tongue; UI, upper incisor; UM, first upper molar. The large arrows in (c) and (d) point towards the midline. Magnifications: ×30 (a,b,e,f); ×75 (c,d,g,h). 2736 D. Lohnes and others

SV background (third generation of inbreeding) revealed a low α1γα2+/− offspring that cannot be identified due to the lack of frequency of malformations affecting the second (C2) and third morphological landmarks (see Kessel and Gruss, 1991). Since (C3) cervical vertebrae (Table 5). the C2 to C1, C6 to C5 and C7 to C6 anterior transformations were found simultaneously in two α1γ and in six α1γα2+/− (1) Homeotic transformations mutants (data not shown), the entire cervical region may have With the exception of fusion of the basioccipital bone with the undergone an anterior transformation in these fetuses. anterior arch of the atlas, which likely represents a posterior RARα1γ, α1γα2+/− and αγ (but not γ) mutants also exhibited homeotic transformation (Lohnes et al., 1993), α1γ and a posterior homeotic transformation characterized by an α1γα2+/− mutants showed increases in the frequency of all extensive rib anlage on C7, which in some cases fused homeotic transformations previously observed in RARγ null ventrally with the first thoracic rib (data not shown; Table 5). fetuses (Table 5). Anterior transformation of C2 to a first This transformation was usually unilateral and the ectopic C7 cervical identity, evidenced by an increase in the thickness of rib never contacted the sternum. Ectopic cervical ribs (CR) the neural arches and the appearance of an ectopic anterior arch were found unilaterally on both C7 and C6 in two αγ mutants, appeared two to four times more frequently in α1γ and with the C6 rib joining the cervical rib projecting from C7 α1γα2+/− mutants, respectively, when compared to γ null (compare Fig. 7i with d; Table 5). These additional ribs did not fetuses (this transformation was, however, incomplete since alter the total number of presacral vertebrae, thus this likely the axis dens was unaffected). Although an ectopic anterior represents homeotic transformations of C6 and C7 to thoracic arch of atlas was occasionally observed in αγ skeletons (e.g. vertebral identities. (‘T1’ and ‘T2’ in Fig. 7i). AAA* in Fig. 7i and l), the severe malformations of the Relative to RARγ null offspring, αβ2 mutant skeletons cervical vertebrae in these mutants usually precluded evalua- exhibited an increase only in the frequency of anterior trans- tion of homeosis. formations of C6 to C5, and C7 to C6, which were bilateral Anterior transformation of C6 or C7 to a C5 or C6 identity, transformations in 50% of the cases (Table 5 and data not respectively, was found twice as frequently in α1γ null mutants shown), whereas RARβ2γ mutants did not exhibit any increase and approximately five times more frequently in α1γα2+/− in the frequency of the homeotic transformations described for mutants than in RARγ null fetuses (Table 5 and data not RARγ null mutants (Table 5). No skeletal abnormalities were shown). C7 to C6 transformation was evidenced by the appear- found in α1β2 mutant fetuses. ance of tuberculi anterior (TA) on the ventral aspect and of foramina transversaria on the lateral processes of C7 (see (2) Malformations of the axial skeleton Lohnes et al., 1993 and Fig. 7d). C6 to C5 transformation was In RARα1γ, α1γα2+/− and β2γ mutants, bifidus of C1 and/or inferred from loss of the C6-specific tuberculi anterior. Inter- fusion with C2 occurred eight to ten times more frequently than estingly, in 50% of the affected α1γα2+/− mutants, the C6 to in RARγ null mutants (Table 5 and data not shown). A similar C5 and C7 to C6 transformations were bilateral (data not increase was also observed for fusion of the neural arches of shown), whereas these transformations were essentially uni- C2 and C3 (compare Fig. 7d and e, white arrow; Table 5 and lateral in RARγ and α1γ null mutants. Additional anterior data not shown). RARα1γ, α1γα2+/− and β2γ mutants also transformations of cervical vertebrae may exist in α1γ and exhibited malformations not previously observed in RARγ null

Table 5. Axial skeletal malformations in RAR double mutants Genotype and number of 18.5 dpc mutant featuses examined RARα RARγ RARα1γ RARα1γα2+/− RARαγ RARβ2γ RARαβ2 21 29 16 11 6 9 10 Number of mutants with abnormal skeletons 4(19%) 25(86%) 16(100%) 11(100%) 6(100%) 7(78%) 10(100%) ABNORMALITIES Homeotic Transformations Basioccipital AAA fusion 0 8(28%) 3(19%) 2(18%) NA 2(22%) 0 Anterior transformations of C2 to C1 1(5%) 5(17%) 5(31%) 7(64%) NA 1(11%) 2(20%) C6 to C5 0 4(14%) 4(25%) 8(73%) NA 0 8(80%) C7 to C6 0 4(14%) 4(25%) 8(73%) NA 0 8(80%) Posterior transformations of C6 to T1 0 0 0 0 2(33%) 0 0 C7 to T1 or T2 0 0 6(38%) 3(27%) 4(67%) 0 0 Malformations C1 malformed 0 2(7%) 10(63%) 9(82%) 6(100%) 5(56%) 8(80%) C2 malformed 3(14%) 3(10%) 4(25%) 5(45%) 6(100%) 5(56%) 7(70%) Fusions of cervical neural arches 1(5%) 5(17%) 5(31%) 11(100%) 6(100%) 6(66%) 4(40%) Agenesis of cervical neural arches 0 0 0 0 6(100%) 0 0 Dyssymphysis of cervical neural arches 3(14%) 0 10(63%) 10(91%) 6(100%) 2(22%) 4(40%) Ectopic bone in cervical region 0 0 0 0 2(33%) 0 0 Rib fusions 0 4(14%) 3(19%) 2(18%) 2(33%) 0 0 Basioccipital Exoccipital fusion 0 0 0 0 1(17%) 0 7(70%) Sternum malformations 0 0 6(38%) 3(27%) 6(100%) 0 0

Skeletal malformations were not observed in RARα1 or RARβ2 single mutants, nor in RARα1β2 double mutants. NA, not applicable; AAA, anterior arch of atlas. RARs in ontogenesis 2737

Fig. 7. Malformations of the axial skeleton in RAR double mutant mice. (a-c) Dorsal views of the cervical and upper thoracic region of wild-type (a) and αβ 2 mutant (b,c) skeletons. (b) The long arrow indicates bifidis of the neural arch of the first cervical vertebra, the short arrow indicates fusion between the neural arches of the second and third cervical vertebrae. (c) The asterisk denotes dyssymphysis of the first cervical vertebra. C1 to C7, first to seventh cervical vertebrae. (d,e,f,i) Lateral views of the cervical and upper thoracic region of wild-type (d) α1γα 2+/ − mutant (e) and αγ (f,i) mutant skeletons. (e,f) Asterisks indicate dyssymphysis of the first cervical vertebrae. (e) Arrows indicate fusions between the neural arches of the second and third and third and fourth cervical vertebrae. (f) Arrow indicates fusion of the neural arches of the third, fourth and fift h cervical vertebrae. (i) AAA* indicates an ectopic anterior arch ventral to the second cervical vertebra, CR indicates cervical ribs and ‘T1’ and ‘T2’ denote posterior transformation of the sixth and seventh cervical vertebrae to a first and second thoracic identity, respectively. (f,i) Note the agenesis of the neural arch of the second cervical vertebrae. (g,h) Ventral views of the cranial base of wild-type (g) and αβ 2 mutant (h) skeletons. (h) The asterisk indicates an ossified fusion between the basioccipital (BO) and exoccipital (E) bones. (j-l) Lateral views of the cranial to lumbar regions of a wild-type (j) and αγ mutant (k,l) skeletons. (k) EC indicates an ectopic ossified structure in the cervical region. Note also the complete lack of the cranial vault of this specimen. (l) Asterisk denotes dyssymphysis of the first cervical vertebra and AAA* indicates an ectopic anterior arch. Note also the agenesis of the neural arch of the second cervical vertebra. T1, first thoracic vertebra; L1, first lumbar vertebra. 2738 D. Lohnes and others

Fig. 8. Limb defects in αγ double mutants. (a,b) External ventral views of wild-type (a) and αγ mutant (b) right forelimbs; I* indicates a supernumerary preaxial digit; I-V, first to fifth digits. (c,d) Frontal external views of a wild-type (c) and αγ double mutant (d) left forelimbs; note the abnormal aspect and fusions of the digits. (e,f) Dorsal external views of the right forefeet of a wild-type (e) and an αγ mutant (f); note the abnormal aspect of the fifth digit. (g,h) Lateral views of the right hindlimbs of a wild-type (g) and an αγ mutant (h); open arrowhead indicates delayed ossification and abnormal appearance of the phalanges; F, fibula; T, tibia. (i-l) Whole-mount skeletal stainings of the forelimbs of wild-type and αγ double mutants. (i) Small arrow indicates supernumerary preaxial digit. Arrows (j-l) indicate missing first digit. Asterisks (i-l) indicate malformations of the scapulae. Note that the αγ mutant limbs (i,j) were from the same animal. H, humerus; R, radius; S, scapula; U, ulna. RARs in ontogenesis 2739 offspring, including dyssymphysis of the neural arch of C1 The entire forelimb was shorter when compared to 18.5 dpc (asterisk in Fig. 7e compare to 7d) or C2 (not shown), and control littermates (Fig. 8i-l). This may be related to the fusion of C3 with C4 (white arrow in Fig. 7e compare to 7d). general growth deficiency of these mutants (e.g. compare Fig. RARαγ mutant fetuses had severe defects of all cervical 1a and c). The scapula was always malformed in αγ mutants. vertebrae, usually making detailed analysis impossible, save In one fetus, the two scapulae were partially agenic with a for cervical ribs on C6 or C7 (discussed above). Bifidus (not greatly reduced shaft diameter and vertebral (medial) region shown) and dyssymphysis of C1 (asterisks in Fig. 7f,i,l, (asterisk in Fig. 8l). In another fetus, the median portion of one compare to Fig. 7d,j) and agenesis or fusions of the neural scapula appeared bifurcated (Fig. 8k). In a third fetus, one arches of C2 to C5 (large arrow in Fig. 7f, compare to 7d; note scapula was partially agenic, whereas the central portion of the the absence of the neural arch of C2 in Figs. 7f,i,l) were contralateral scapula was bifurcated (data not shown). The observed in all specimens. In two cases an ectopic ossified other scapula defects corresponded to mild aplasia of the structure between C1 and C4 or C5 was observed dorsal to the superior or inferior margins of the vertebral region (e.g. Fig. cervical region (EC in Fig. 7k). 8i and j, asterisks). Most interestingly, with the exception of rib fusions, which The humerus, although smaller than controls, appeared occurred in α1γ, α1γα2+/− and αγ mutants with a frequency otherwise normal (e.g. Fig. 8i-l). In contrast, the radius and similar to that in RARγ null mutants (arrow in Fig. 7l; Table ulna, which were also reduced in size, were abnormal in all αγ 5), vertebrae caudal to the cervical region appeared unaffected mutants. In four of six fetuses, the radius of either the left or in all double mutants analyzed (e.g. compare thoracic (T) and right forelimb was missing unilaterally (e.g. Fig. 8l; Table 6). lumber (L) vertebrae in Fig. 7j and k; note the complete loss The ulna was always formed, but exhibited an abnormal of the cranial vault of the specimen in Fig. 7k). The only addi- curvature as did the radius when present (e.g. Fig. 8i-l). tional malformations found in these animals involved occipital With the exception of the pyramidal (PY) and pisiform (PI) bones (see below) and the sternum. In some 18.5 dpc α1γ, bones [the nomenclature of Milaire (1978) is used for the α1γα2+/− and αγ mutant fetuses, the sternum was distorted and, carpal bones], which were well defined (e.g. compare Fig. 9e in one case, appeared incompletely closed (arrowhead, Fig. 7l) with f and g with h), all other carpal bones were malformed. and possessed four instead of five sternebrae. The central bone (C) was missing in all but one case (compare The cervical region of αβ2 mutants was also malformed, Fig 9e with f, g with h; asterisks denote the missing central with a high frequency of dyssymphysis and bifidus of C1, and fusions between C1 and C2 or C2 and C3 (Fig.7a compare to b and c; arrow in Fig. 7b denotes C1 bifidus and arrowhead Table 6. Forelimb malformations in RARαγ double null indicates C2-C3 fusion; asterisk in Fig. 7c denotes dyssymph- mutants αβ αγ ysis of C1). Furthermore, the majority of 2 and 1 (of 6) 18.5 dpc mutant fetuses mutant fetuses exhibited an osseous fusion of the basioccipital (BO) and exoccipital (E) bones (asterisk in Fig. 7h, compare 1 2 3 4 5 6 Total to g; Table 5), a malformation not observed in RARγ mutants. Malformation or partial L + + ND + + + 5 agenesis of the scapula R + + ND + + + 5 (H) Malformations of the appendicular skeleton Agenesis of the radius L + + 2 The limbs of all double mutants were essentially normal with R + + 2 the exception of RARαγ mutants. Malformation of the L + + + + + + 6 scapholunatum R + + + + + + 6 (1) External features Ectopic distal carpal bone L R + 1 Although forelimbs were affected in all 18.5 dpc αγ mutants Agenesis of the D1 L + + + 3 examined, the nature of the limb defects showed considerable carpal bone R + + 2 variation. Syndactyly was frequently observed (compare Fig. Agenesis of the central L + + + + + 5 8e and f). The precise digits fused varied among animals and carpal bone R + + + + + + 6 fusion was at the level of soft tissues only (see below). Digits Prepollex agenic or L + + + + + 5 frequently looked abnormal (compare Fig. 8c and d, e and f) rudimentary R + + + + 4 and polydactyly (with 6 digits; compare Fig. 8a and b), or 6 Digits L 0 ectrodactyly (3 or 4 digits; not illustrated) was often apparent. R +* 1 RARαγ mutant hindlimbs exhibited a consistent aberrant 5 Digits L + + 2 external aspect with the hindfoot twisted such that the footplate R + + 2 faced the lateral abdominal wall (compare Fig. 1a and c). 4 Digits L +† +¤ +† 3 However, no cases of syndactyly or other digit malformations R +† +† +† 3 were found in the hindlimbs upon external examination. 3 Digits L +‡ 1 R 0 (2) Skeletal analysis Hypoplastic first digit L + 1 Whole-mount skeletal preparations from six 18.5 dpc αγ R 0 mutants were analyzed in detail. All six exhibited a number of ND; not determined; *additional presumptive first digit; †loss of malformations affecting diverse skeletal elements of the presumptive first digit; ‡loss of presumptive first and second digit; §loss of forelimb. However, the severity and frequency of most of these presumptive second digit. L and R, left and right forelimb, respectively. 1 to defects varied both among animals and between contralateral 6, correspond to the different fetuses which were examined. Total, total forelimbs from the same animal (Table 6). number of limbs that exhibited a given malformation. 2740 D. Lohnes and others bone). The scapholunatum (SL), which was always present tently hypoplastic (compare Fig. 9e and f; g and h). The regardless of the absence or presence of the radius (e.g. see SL prepollex (PX) was frequently rudimentary or absent (compare in Fig. 9m), was always misshapen and small (compare SL in Fig. 9i and j). Fig 9e and f, g and h), and in many cases was partially cleft In 6 (of 12) mutant forelimbs, only 4 digits were present, but (arrow in Fig. 9f). All distal carpals (D1-D4/5) were consis- this defect was bilateral in only one instance (Table 6 and see RARs in ontogenesis 2741 below). In five cases, the missing digit was presumably digit 1 comparison to 18.5 dpc controls of only slightly larger size (e.g. compare Fig. 9c and d), as determined by phalangeal (e.g. compare Fig. 9a and b) indicates that this delay may be a count (digit one has two phalanges, whereas digits 2 to 5 have direct effect of the mutations. In this respect, note a complete three) and the position of the remaining digits with respect to lack of metacarpal ossification centers in the mutant forelimb the distal carpal bones (using D4/5 as a landmark). The loss of shown in Fig. 9m (although hypertrophic chondroblasts are digit 1 was correlated with the absence of the carpal D1 in four evident), even though ossification of metacarpals 3 and 4 cases (Table 6 and data not shown). In one other specimen, the usually commences as early as 15.5 dpc (Dollé et al., 1993; first digit was present, but markedly smaller (Table 6 and data Kaufman, 1992). The phalanges of αγ mutants were also not shown). In one fetus, which was missing the first digit on abnormal, appearing bulbous with the outline of the synovial the right forelimb, the second digit was absent on the con- joints between the phalanges difficult to discern (compare Fig. tralateral forelimb, with only a small cartilaginous anlage 9a to b, c to d, k to l). This latter malformation occurred irre- found in its place (arrow II in Fig. 9h). Interestingly, the first spective of other forelimb malformations. digit in this forelimb resembled a normal second digit rather Examination of the hindlimbs revealed (in five of six cases) than a thumb, as judged by its length and the advanced ossifi- a dramatic bilateral reduction in the length and increase in the cation of the metacarpal bone, yet only two phalanges were diameter of the ossification of the fibula (F), with a concomi- present (not shown). Unilateral ectrodactyly with three digits tant bending of the tibia (T; compare Fig. 8g and h; note that a was observed in one animal, where digits 1 and 2 were missing connection exists between the head of the tibia and the proximal and carpal D1 was agenic (Table 6 and data not shown). end of the fibula in the specimen in Fig. 8h). The dispropor- Polydactyly with six digits was found in one mutant tionate growth between the fibula and tibia is the likely cause (compare Fig. 9a and b, e and f); the additional preaxial of the bending of the tibia and of the abnormal external aspect element was associated with an ectopic distal carpal bone, of the hindlimbs. In one αγ mutant fetus, the ossification of the whereas the prepollex was lacking (data not shown). The fibula was intermediate between controls and severely affected diameter of the putative metacarpal bone was reduced and only mutants (data not shown). The only additional defects found in one phalange was apparent. Interestingly, in this specimen the the hindlimbs of RARαγ skeletons were a slight retardation in normal first digit may have undergone a transformation to a os s i fi cation of the metatarsals and phalanges, and bulbous second digit, as supported by the increase in its length and phalanges with indistinct synovial joints as described above for diameter, advanced ossification of the metacarpal bone and the forelimbs (compare Fig. 8h and g). No loss or gain of larger size of the phalanges (compare digit I in Fig. 9a and b, hindlimb digits or tarsal elements was observed. e and f; note that the third phalange was broken during prepa- The forelimbs of α1γα2+/− fetuses were unaffected. ration of this specimen). Remarkably, the contralateral limb of However, of 11 RARα1γα2+/− fetuses examined, 1 exhibited a this mutant possessed only four digits (Fig. 9d; Table 6). defect of the fibula identical to that shown in Fig. 8h and two Additional malformations observed in αγ mutant forelimbs other specimens exhibited a similar, but milder malformation included a delay in the ossification of the metacarpals and/or (data not shown). phalanges (e.g. compare Fig. 9c and d). Although this may result from the general growth deficiency of these animals, DISCUSSION αγ Fig. 9. Skeletal analysis of the malformations of the forefeet in The present results show that RARs are essential for normal mutants. (a,b) Dorsal views of skeletons of the right forefeet of a wild-type (a) and an αγ mutant (b); I* indicates supernumerary development of many structures in the mouse. However, with preaxial digit; note that the adjacent digit (digit I) of this foot the exception of malformations of the axial skeleton and originally possessed three phalanges; I-V, first to fifth digits; R, agenesis of the Harderian glands (which have been previously radius; U, ulna. (c,d) Dorsal views of the skeletons of the left forefeet described in RARγ null offspring; Lohnes et al., 1993), con- of a wild-type (c) and an αγ mutant (d); arrow indicates missing first genital defects are observed only in RAR double mutants. digit. (e,f) Higher magnification of the carpal regions of the Offspring from dams fed a vitamin A-deficient (VAD) diet also specimens shown in a and b, respectively; asterisk denotes missing exhibit a number of developmental abnormalities (Wilson et central (C) carpal bone, arrow indicates split in scapholunatum (SL) al., 1953). The recapitulation of most of these VAD-associated and I* indicates supernumerary preaxial digit. Note also the small congenital malformations (see also Mendelsohn et al., 1994b) size of the distal carpals (D1-D4/5). PI, pisiform carpal bone; PY, indicate that retinoic acids, the known ligands of the RARs, pyramidal carpal bone. (g,h) Dorsal views of the carpal skeleton of the left forefeet of a wild-type (g) and an αγ mutant (h); asterisk most probably represent the active retinoid developmental indicates missing central carpal bone and arrow indicates signal. With the exceptions of eye defects (Warkany and cartilaginous rudiment in place of the second digit. Note also the Schraffenberger, 1946) and cleft palate (Hale, 1933), the mal- increased size and the advanced ossification of the first digit in h formations described here in RAR double mutants have not compared to g. (i,j) Lateral views of skeletons of the left forefeet of a been reported in VAD studies, probably because extreme wild-type (i) and an αγ mutant (j). Note the reduction of the deprivation of vitamin A results in embryonic lethality (Mason, prepollex (PX) in j. (k,l) Dorsal views of the phalanges of the forefeet 1935; Wilson and Barch, 1949). Complete inactivation (by full of a wild-type (k) and an αγ mutant (l); arrows indicate the bulbous dietary deprivation of RA) of all RARs is likely to result in appearance of the phalanges and the indistinct synovial joint between early embryonic death and resorption, as reflected here by the the second and third phalanges. (m) Dorsal view of the left forefoot embryonic death of approximately 50% of the RARαγ double of an αγ mutant. Short arrow indicates hypertrophic chondroblasts and absence of ossification of the metacarpals and the long arrow mutants. These observations indicate that some RA-dependent indicates missing first digit. Note also the severe malformation of the developmental processes are more sensitive than others to RA scapholunatum and the absence of the radius. deficiency (e.g. eye development; see below) and that only 2742 D. Lohnes and others abnormalities related to these events are observed in VAD distinct stages of eye development. The complete absence of studies. lens tissue observed in one 12.5 dpc RARαγ mutant embryo RARβ isoform transcripts and amount of RARβ2 protein likely results from a fault in the process of lens induction were not affected in 13.5 or 14.5 dpc αγ double mutant occurring before 9.5 dpc (reviewed in Grainger et al., 1992). embryos, respectively, nor was the pattern of expression of The presence of a corneal lenticular stalk implies an arrest at RARβ transcripts apparently altered in 10.5, 11.5 or 13.5 dpc 11.5 dpc in the separation of the lens from the parental αγ mutant embryos as judged by in situ hybridization data (our ectoderm. These processes depend on interactions between the unpublished results). Thus the extreme malformations specific lens placode and the optic vesicle. Interestingly, the embryonic to αγ double mutants are unlikely to reflect an additional defect retina is capable of synthesizing RA (McCaffery et al., 1992) in RARβ expression. Moreover, since RARβ2 is believed to and RA-reporter mice suggest that the eye contains RA as early be transcriptionally modulated by RA in vivo, its expression as 9.5 dpc, with later synthesis (12.5 dpc) in the neural retina in RARαγ double null mutants may be controlled by the other (Balkan et al., 1992; Rossant et al., 1991). These data suggest RARβ isoforms (RARβ1, β3 and β4), through an autoregu- that the lack of a RA-dependent inductive signal from the optic latory mechanism (i.e. by RARβ2 itself), or by other nuclear vesicle may be the underlying basis for these lens defects. receptors, such as the RXRs. The fusion of the two lips of the optic fissure, which starts with an inversion of the retinal pigmented epithelium, occurs (A) RA is required at several stages during eye first between undifferentiated cells at the junction of the neural morphogenesis retina and the retinal pigmented epithelium, and involves the The eye is the most sensitive organ to retinol deprivation and disintegration of the basement membrane between the retina in less severely affected VAD fetuses, it is often the only site and the mesodermal tissue (Geeraets, 1976; Suzuki et al., 1988; of malformation (Warkany and Schraffenberger, 1946). The Hero, 1989, 1990). Partial or complete persistence of the optic spectrum of VAD-induced ocular malformations is largely fissure (coloboma) in 18.5 dpc RARαγ and β2γ mutants might recapitulated in αγ (with the exception of a shorter ventral β γ then be caused by: (i) overgrowth of the inner layer of the optic retina) and to a lesser extent in 2 , double mutants (see Table cup relative to its external counterpart, thus preventing the 3). normal inversion of the latter along the line of the optic fissure, Eye formation involves the coordinated development of (ii) precocious differentiation of the retinal cells located at the forebrain neuroectoderm (which gives rise to the retina, optic junction between the two layers or, (iii) maintenance of the nerve and epithelial portion of the iris), surface ectoderm basement membrane preventing the fusion of the two lips of (which gives rise to the lens and epithelia of the cornea and the optic fissure. Further studies at the electron microscopic conjunctiva) and cranial neural crest-derived mesenchyme level are required to distinguish between these possibilities. (which forms the choroid, sclera, stroma of the cornea and iris, Although the RA-dependent events leading to coloboma are anterior chamber, and vitreous body; Pei and Rhodin, 1970; unknown, this defect appears to be determined shortly before Johnston et al., 1979; Le Douarin et al., 1993; and references therein). In WT mice at 9.5 dpc, the optic vesicle, which has the begining of the closure of the optic fissure (12.5 dpc), as evaginated from the forebrain neuroectoderm, comes into administration of vitamin A to VAD rat embryos can reduce contact with the lens placode, which immediately invaginates the incidence of coloboma if given before the equivalent of and, at 11.5 dpc, pinches off from its parental surface ectoderm 11.5 dpc of mouse gestation (Wilson et al., 1953). The persis- to become a hollow lens vesicle, leaving behind the presump- tence of the retrolenticular mesenchyme appears to correspond tive corneal epithelium at the surface. At 10.5 dpc, the optic to a later arrest, since its occurrence could be likewise vesicle invaginates from its ventral side to form the optic cup prevented in VAD embryos by vitamin A administration before (i.e. the anlage of the retina and of the epithelial portion of the 13.5 dpc (Wilson et al., 1953). The additional malformations iris; Kaufman, 1992; Pei and Rhodin, 1970). The area of in the eyes of both VAD offspring and RAR double mutants invagination represents the optic fissure. Mesoderm extends (including isolated persistent retrolenticular mesenchyme and from this fissure into the cup forming the primary vitreous malformations of the eyelids, conjunctival sac, cornea and body. The two lips of the optic fissure come into contact and anterior chamber) may be due to defects in mesenchymal NCC, fuse at 12.5 dpc, except in the regions of the iris and of the as all of these structures are derived from NCC originating optic disc, where closure is delayed until 14.0 dpc. By 14.5 from the forebrain and/or midbrain of the developing embryo dpc, the mesenchymal cells of the vitreous body (i.e. the retro- (see below; Serbedzija et al., 1992; Le Douarin et al., 1993 and lenticular mesenchyme) disappear. references therein). The ocular defects in RAR double mutants correspond The major ocular defects were found in RARαγ and mostly to structures arrested in ontogenesis and partly to RARβ2γ double null mutants, suggesting that RARγ is abnormal embryonic formations. Absence of the lens, cornea, essential (in the absence of RARα1 and RARα2 or RARβ2) conjunctiva or anterior chamber, persistence of a corneal lentic- for normal eye development. Furthermore, in contrast to − ular stalk or a fibrous retrolenticular membrane, coloboma of RARαγ mutants, RARα1γα2+/ double mutants had near the retina or optic nerve, poor differentiation of the corneal normal eyes, suggesting that (in the absence of RARγ) one stroma or lens fibres, hypoplasia of the conjunctival sac and lack copy of RARα2 suffices for most of the events needed for eye of fusion of the eyelids are developmental arrests. However, the development. The eye defects observed in RAR mutants are in ch o n d r i fi cation of the persistent retrolenticular tissue in good agreement with the presence of RA in the eye (see above) RA R β2γ mice and the keratinization of the corneal epithelium and the expression pattern of the RARs during the period of have no equivalent at earlier developmental stages. eye development known to be sensitive to vitamin A depriva- RARs are clearly required for morphogenetic processes at tion (8.5 to 13.5 dpc in the mouse; Dollé et al., 1990; Ruberte RARs in ontogenesis 2743 et al., 1990, 1991). RARα is expressed ubiquitously, whereas observed in αγ double mutants may occur when the expression RARγ and RARβ are expressed in the periocular mesenchyme of several of these Hox genes are concommitantly altered. throughout this period of development. Furthermore, RARβ2 In situ hybridization studies have shown that RARγ (and α) (but not RARγ) transcripts are also detected at 12.5 to 13.5 dpc are expressed posterior to the caudal neuropore in the late gas- in the retrolenticular mesenchyme (our unpublished results), trulating mouse embryo in all three germ layers prior to somite the abnormal persistence of which is the only ocular abnor- formation. RARγ expression then apparently disappears con- mality found in RARα1β2 and αβ2 double mutants. Thus, in comitant with the appearance of somites (Ruberte et al., 1990), these mutants, this abnormality may be a primary defect suggesting that RARγ and α may control Hox gene expression whereas, in RARαγ and β2γ mutants, it may be secondary to during somite formation and specification. This also coincides the coloboma. with the embryonic stages when RA excess affects both Hox gene expression and vertebral identities (Kessel and Gruss, (B) RARs and specification of the axial skeleton 1991). However, there is also evidence indicating that vertebral The homeotic transformations and other axial malformations transformations can occur at later stages (10.5-11.5 dpc) that occur in RAR double mutants are strikingly confined to through mechanisms not involving Hox genes (Kessel, 1992). cervical vertebrae. The present results indicate that RARs may At these later stages, RARα and γ transcripts are found in scle- be functionally redundant for specification of these vertebrae, rotomes (Ruberte et al., 1990, 1991; Dollé et al., 1990), sug- as the penetrance and expressivity (bilateral versus unilateral gesting that RA could also be involved in the maintenance of defects) of cervical anterior transformations previously vertebral identities through a Hox gene-independent observed in RARγ null offspring (Lohnes et al., 1993) mechanism. Thus, the vertebral malformations observed here increased in a graded manner with subsequent loss of RARα1 could reflect a RA requirement at these later stages. RARβ and RARα2 isoforms from the RARγ−/− background. Further- transcripts were not detected in presomitic mesoderm, somites more, RARβ2 (in the absence of RARα) also appears to play or sclerotomes in mouse embryos, while present in the neural a role in axial specification, as RARαβ2 double mutants (but tube (Ruberte et al., 1991; Dollé et al., 1990; and our unpub- not RARα mutants) displayed a high frequency of anterior lished results). This suggests that the defects seen in RARαβ2 homeotic transformations, particularly of the sixth and seventh and RARβ2γ mutants may reflect an indirect effect of RARβ cervical vertebrae. That the posteriorization of the sixth and/or in vertebral morphogenesis, involving RA-dependent signals seventh cervical vertebrae was observed only in α1γ, α1γα2+/− emanating from the neural tube, since perturbations of this and αγ double null mutants suggests that RARα and RARγ are structure can result in vertebral malformations (Hall, 1977). redundant with regard to events leading to these posterioriza- However, isolated transformations of C6 and C7 and a restric- tions and that RARβ2 has no role in this particular event. tion of malformations to the cervical region have not been Gain-of-function (Kessel et al., 1990: Lufkin et al., 1992) reported in such studies. and loss-of-function (Le Mouellic et al., 1992; Ramirez-Solis et al., 1993; Jeannotte et al., 1993; Condie and Capecchi, 1993) (C) RARs and patterning of the limb studies have shown that some Hox genes specify the identity We previously speculated that the lack of limb malformations of somites. Although there are notable exceptions (Pollock et in RARα or γ null mutants may be due to functional redun- al., 1992; Jegalian and De Robertis, 1992), Hox gain-of- dancy between these receptors (Lohnes et al., 1993; Lufkin et function mutations usually lead to posteriorization, while Hox al., 1993), since each is uniformly expressed throughout the loss-of-function mutations lead to anteriorization of vertebral mesenchyme of the limb bud at 9.5-11.5 dpc (Dollé et al., identities. RAR double mutants and some Hox null mutants 1989). This is clearly the case, as the limbs from RARαγ exhibit similar cervical vertebral transformations, suggesting double mutants consistently exhibited malformations. Strik- that RA may affect vertebral patterning by controlling Hox ingly, forelimbs from RARα1γα2+/− mutants were normal, gene expression. The most notable similarities are with Hoxb- showing that (as for the eye) a single copy of RARα2 suffices 4 null mice, which exhibit anterior transformation of C2 to a for limb morphogenesis in the absence of RARγ and RARα1. C1 identitiy (Ramirez-Solis et al., 1993), and with Hoxa-5 null The anteroposterior axis of the limb is patterned by the zone mice, which exhibit anterior transformation of C6 to a C5 of polarizing activity (ZPA), whereas limb outgrowth requires identity and posterior transformation of the C7 to a T1 identity a functional apical ectodermal ridge (AER; see Tabin, 1991 for (Jeannotte et al., 1993). That Hox gene expression may be con- review). The effect of the ZPA on the specification of the trolled by RA during development has been suggested by the anteroposterior axis of the limb can be mimicked by topical observation that some Hox gene transcripts accumulate in application of RA, and both RA and the ZPA can trigger the cultured embryonal carcinoma (EC) cells exposed to RA expression of some Hox genes believed to be critical for limb (Simeone et al., 1990, 1991; Mavilio, 1993 and refs therein) specification (for reviews, see Duboule, 1992; Dollé and and by the presence of functional RAREs in the promoter Duboule, 1993). A role for RA in maintaining AER activity is region of some of these RA-responsive genes (e.g. Hoxa-1, suggested by the finding that RA, in combination with FGF-4, Langston and Gudas, 1992; and Hoxd-4, Pöpperl and Feather- can fulfill most AER functions (Niswander et al., 1993). stone, 1993). That only cervical vertebrae were transformed in However, the AER of RARαγ mutants appeared histologically RAR double mutants may reflect the greater sensitivity (at least normal, although its formation may be slightly delayed (our in EC cells) of 3′ Hox paralogues to RA (Mavilio, 1993). Some unpublished results). It is currently believed that RA may indi- of these 3′ genes may be preferentially affected in RAR double rectly influence limb patterning by generating (or maintaining) mutants, leading to selective vertebral transformation in the a functional ZPA (Wanek et al., 1993), possibly through the cervical region. The severe malformations of cervical vertebrae regulation of a secreted protein, sonic hedgehog (Riddle et al., 2744 D. Lohnes and others

1993). The malformations in RARαγ mutants do not appear to and Alberch, 1986; and refs therein). Since RARα and RARγ result from an early ZPA defect, since the limbs displayed a are both expressed in the perichondrial region of blastemal clear anteroposterior asymmetry. This does not exclude a role condensations (Dollé et al., 1989), the loss of these receptors for RA in anteroposterior limb patterning, since RARβ tran- could conceivably alter the functional integrity of the pha- scripts (which appear unaffected in the limbs of αγ mutants) langeal perichondrium, thus leading to their altered morphol- are expressed in the flanking mesenchyme and proximal ogy. regions of the early limb bud in a region that overlaps with the It is noteworthy that preaxial malformations of the limbs ZPA (Dollé et al., 1989; Mendelsohn et al., 1991; our unpub- exhibited by RARαγ double mutants are restricted to the lished results). In addition, the expression of some molecular forelimbs (with the exception of bending of the tibia, which is markers of ZPA and AER activity, including Hoxd-9 and likely secondary to the severe malformation of the fibu l a ) . Hoxd-13 (Dollé and Duboule, 1993), MSX-1 (Robert et al., These observations suggest that either RA plays different roles 1989; Hill et al., 1989), BMP-2 (Lyons et al., 1990), FGF-4 in forelimbs and hindlimb development, or that events related (Niswander et al., 1993) and sonic hedgehog appeared unaf- to differences in time of development of the two limbs allow fected as judged from in situ hybridization studies from 10.5 phenotypic rescue of the preaxial derivatives of the hindlimb. and 11.5 dpc RARαγ mutants (P. Dollé and D. Décimo, unpub- Note in this respect that several mouse mutants, such as Po lished data). (postaxial; Nakamura et al., 1963) and Px (postaxial hemimelia; Although two RARαγ mutants exhibited putative digit trans- Searle, 1964) also exhibit defects confined largely to the formations, firm conclusions cannot be drawn concerning the forelimbs. However malformations affecting the anteroposte- role of RA in specification of limb axes. It is clear, however, rior axis of the limb usually affect homologous structures in that RA is essential for the generation of some skeletal both forelimbs and hindlimbs (e.g. Xt , Batchelor et al., 1966). elements of the forelimb. In tetrapods, the limb skeletal (D) Neural crest and RARs elements are established through a conserved series of branching and segmentation events from prechondrogenic Malformations of most of the structures derived from cranial blastemas arising during limb outgrowth (reviewed in Shubin and cardiac mesenchymal neural crest cells (NCC) were and Alberch, 1986; Shubin, 1991). Branching of the humeral observed in RAR double mutants (this and the accompanying blastema gives rise to those of the radius and ulna. The study of Mendelsohn et al., 1994b). These structures originate proximal and central carpals, as well as the digital arch, arise from osteogenic NCC (i.e. membrane bones of the skull and face), from chondrogenic NCC [i.e. endochondral bones of the from segmentation and branching events initiating from the ulnar condensation, while distal carpal bones and subsequent skull and face and second (hyoid) and third arch skeletal elements], from odontogenic NCC (i.e. upper incisors), from digit formation generally proceed in a posterior-to-anterior smooth muscle NCC precursors (i.e. the tunica media of the direction in the mouse. All of the elements consistently aortic arches and the aorticopulmonary septum), from dermal affected in RARαγ mutants are derived from the preaxial NCC precursors (contributing to the pinna, eyelids and (anterior) portion of the limb bud (i.e. radius, central and D1 prolabium) and from periocular NCC (i.e. corneal stroma and carpals, digits 1 and 2 and the prepollex) and arise from the retrolenticular mesenchyme). Mesenchymal NCC also con- last branching event occurring in a given region of the limb. tributes to the stroma of the glands whose development and/or Loss of radius, first digit, central carpal bone and prepollex migration are altered in RAR double mutants (i.e. Harderian, may be due to the absence of the final specific branching events lachrymal, thyroid, thymus and parathyroid glands) (for review giving rise to these structures (i.e. branching of the radial con- and references see Noden, 1988; Le Douarin et al., 1993). densation from the humeral condensation, branching of the Since NCC appeared with the emergence of vertebrates (Gans central carpal condensation from the pyramidal condensation and Northcutt, 1983) and RARs have been found only in ver- and final branching of the digital arch to yield the prepollex or tebrates (Kastner et al., 1994; Linney and LaMantia, 1994; and first digit). These defects may reflect a requirement of RA to references therein), our observations raise the interesting pos- generate the proper amount of limb mesenchyme, since a sibility that these receptors may have evolved to fulfill deficit in this mesenchyme leads to a preferential loss of functions necessary for the development of mesenchymal anterior skeletal elements (Alberch and Gale, 1983). Such a NCC-derived structures. deficit is also suggested by the generalized size reduction of In αγ (and to a lesser extent in α1γα2+/−) mutant mice, all the carpals in RARαγ double mutants (note also in this respect the structures derived from mesenchymal NCC originating that the entire forelimb of RARαγ mutants was often shorter). from the forebrain and the rostral midbrain (i.e. prolabium, Smaller mesenchymal condensations may also provide the frontal, nasal, premaxillary, ethmoid, presphenoid and extra space needed for generating supernumerary condensa- sphenoid bones, and upper incisors; Le Douarin et al., 1993 tions, occasionally resulting in polydactyly. and references therein) were either agenic or severely The phalanges of both the forelimbs and hindlimbs of 18.5 malformed. Ocular structures derived from fore- and rostral dpc RARαγ double mutants were consistently malformed, midbrain mesenchymal NCC were also affected in αγ and appearing bulbous with poorly defined boundaries. This mal- other RAR mutants (see above). All elements derived from formation does not appear to correspond to a developmental more caudal mesenchymal NCC emanating from the level of delay, since the phalanges of WT 15.5 to 17.5 dpc fetuses are rhombomeres 4 and 6 (R4 and R6) and from the unsegmented always well defined (our unpublished observation; Kaufman, caudal portion of the rhombencephalon, and populating the 1992). The perichondrial cells, surrounding the emerging second, third, fourth and sixth arches, (Lumsden et al., 1991; blastemae, are believed to be important in directional growth Noden, 1988; Le Douarin et al., 1993; Serbedzija et al., 1992; of the blastemae and in determining their final shape (Shubin Sechrist et al., 1993), were also likely malformed or ectopi- RARs in ontogenesis 2745 cally localized in RARαγ and other double mutants. These of RAR inactivation on mesenchymal NCC are unknown, two structures included the hyoid and styloid bones and stapes (i.e. phenomena have been observed: (i) abnormal cell death in the second and third pharyngeal arch derivatives), aortic arches, mesectoderm of the frontonasal process, which precedes the aorticopulmonary septum and thymus, thyroid and parathyroid aplasia of midfacial structures in RARαγ mutants, and (ii) glands (Mendelsohn et al., 1994b). In contrast, the first pha- abnormal specification of the fate of some NCC populations, ryngeal arch skeletal elements, which are derived from caudal as evidenced by chondrification of the meninges and the per- midbrain and rostral hindbrain (i.e. R1 and R2) levels, were all sistence and occasional chondrification of the retrolenticular identifiable in RARαγ double mutant fetuses, although some mesenchyme. Additional cartilaginous ectopias found in the were abnormal. Most of the misshapen first pharyngeal arch- diaphragm, the peritoneum and the semilunar cusps of the heart derived skeletal elements (e.g. maxillary and palatine bones) (see Mendelsohn et al., 1994b), may also be of NCC origin, are derived from the maxillary (rostral) process of this arch, here reflecting either an abnormal specification of trunk NCC which is populated mainly by mesenchymal NCC from the (which normally have no chondrogenic potential; Hall, 1991 caudal mesencephalon. Note that the alisphenoid and incus and references therein) or abnormal migration or specification bones, which are also derived from the maxillary process, were of cranial NCC. fused, but otherwise not grossly affected (see below). In It is remarkable that, with the exception of the lack of the contrast, most of the unaffected skeletal elements (e.g. dentary spiral ganglion (acoustic portion of the VIIIth cranial nerve) in bone, Meckel’s cartilage, malleus and tympanic bone) are RARαγ mutants (which, however, may be secondary to abnor- derived from the mandibular (caudal) process of the first pha- malities of the otocyst; see results section), we have not ryngeal arch, which is essentially populated by mesenchymal observed primary malformations of neurogenic NCC-derived NCC originating from R1 and R2 (Lumsden et al., 1991). structures. Thus, whether the effect of RA excess on many of This lack of effect of RARαγ mutation on first pharyngeal these structures (reviewed in Armstrong et al., 1994) is a phar- arch mesectoderm is unlikely due to a compensation by RARβ, macological phenomenon awaits examination of RARβ (all since RARβ transcripts are expressed at a much lower level in isoforms) null mice. In this respect, RA has been proposed to the first arch than in the frontonasal region, nor can it be regulate directly transcription of the Hoxa-1 gene (Langston et explained by the absence of expression of RARα and γ, since al., 1992; Boylan et al., 1993). However, with the exception of their transcripts are equally abundant in the first arch and fron- the lack of structures derived from the otocyst and lack of the tonasal region (Dollé et al., 1990; Ruberte et al., 1990, 1991). abducens nerve, RAR double mutants do not exhibit any of the Interestingly, first arch NCC appear to be embodied with a defects found in Hoxa-1 null mice (Lufkin et al., 1991; Chisaka ground state morphogenetic program, which is also present in and Capecchi, 1991; Mark et al., 1993). the second pharyngeal arch mesectoderm where it is respeci- Finally, the phenotype of αγ mutants (see also Mendelsohn fied by expression of Hoxa-2 (Rijli et al., 1993). Frontonasal et al., 1994b) present some striking similarities with a human mesectodermal cells are also embodied with a similar program, neurocristopathy called the CHARGE syndrome (Pagon et al., since they can generate first arch skeletal elements when 1981; Siebert et al., 1985): Coloboma of the retina, Heart grafted to the level of either the presumptive first or second disease (consisting of aorticopulmonary septal defects and arch (Noden, 1983). Thus, our data suggest that the realization abnormalities of aortic arch-derived great arteries), Atresia of of at least part of the ground state program present in the first the choana (likely resulting from a defect in the formation of arch does not require RA, whereas modification of this the ethmoid bone), Retardation of physical and mental devel- program both in the frontonasal and second arch mesectoderm opment, Genital hypoplasia in males and Ear abnormalities involves RA-dependent processes. Since the lack of expression (i.e. malformation of the pinna) and/or deafness. Additional of Hoxa-2 results in homeotic transformation of second arch defects in these patients include hypoplasia or agenesis of the to first arch skeletal elements (Rijli et al.,1993), the second arch thymus and parathyroid glands. Although it is unlikely that the defects seen here cannot simply reflect a requirement of RA CHARGE association is due to inactivation of both RARα and for Hoxa-2 expression. RARγ, it could result from the mutation of a RA-dependent Substantial evidence suggests that RA excess can affect pre- gene that is mis-expressed in αγ mutant mice (potential candi- migratory NCC (Morriss-Kay, 1993 and references therein). dates are discussed in the conclusion section of Mendelsohn et However, our results suggest a RA requirement for events al., 1994b). occurring during and/or after NCC migration, since RARγ has not been detected in the presumptive forebrain or midbrain, nor (E) RAR double mutants exhibit atavistic changes in presumptive NCC progenitors in the rhombencephalon Two types of supernumerary skeletal structures were often (Ruberte et al., 1990). In contrast, all three RARs are highly detected in the skulls of RAR double mutant fetuses, the first expressed in the frontonasal process, whereas RARα and forming a cartilaginous medial wall to the cavum epiptericum, RARγ, and to a lesser extent RARβ2, are expressed in the pha- the second linking the incus and alisphenoid. Comparative ryngeal arches after (or possibly during) NCC migration into anatomical data strongly suggest that these two skeletal these structures (Dollé et al., 1990; Ruberte et al., 1990, 1991; elements correspond to atavistic structures (Allin, 1975; our unpublished results). Furthermore, VAD-induced aortic Presley, 1989; De Beer, 1985). arch and aorticopulmonary septal defects can be prevented by In reptiles and monotremes (lower mammals), the orbital vitamin A administration up to mouse 9.5 or 10.5 dpc, respec- and the basal regions of the skull are connected by three pillars, tively, whereas all NCC contributing to these structures have the pilae prooptica, metoptica and antotica (cranial to caudal migrated by 9.0 dpc (Wilson et al., 1953; Le Douarin et al., order) all of which arise by local chondrification of the 1993; and references therein; see also Mendelsohn et al., primitive cranial wall (i.e. the dura mater). The caudalmost 1994b). Although the molecular defects underlying the effect reptilian pillar (the pila antotica) forms a solid medial wall to 2746 D. Lohnes and others the cavum epiptericum, thus physically separating it from the REFERENCES braincase. The ability to form a pila antoptica from the dura mater has been lost in placental mammals, thus permitting the Alberch, P. and Gale, E. A. (1983). Size-dependence during the development expansion of the braincase. The resulting gap caused by this of the amphibian foot: colchicine induced digital loss or reduction. J. Embryol. Exp. Morphol. 76, 177-197. loss became filled in a more lateral plane by a first pharyngeal Allin, E. F. (1975). Evolution of the mammalian middle ear. J. Morph. 147, arch-derived skeletal element, the alisphenoid bone. In the 403-438. RARγ−/− genetic background, disruption of the RARα1 Armstrong, R. B., Ashenfelter, K. O., Eckhoff, C., Levin, A. A. and isoform resulted in the bilateral appearance of a small super- Shapiro, S. S. (1994). In The Retinoids (second edition) (eds. M. B. Sporn, numerary chondrification center at the base of the medial wall A. B. Roberts and D. S. Goodman), pp. 545-572. Raven Press, New York. Balkan, W., Colbert, M. C., Bock, C. and Linney, E. (1992). Transgenic of the cavum epiptericum. With subsequent loss of only one indicator mice for studying activated retinoic acid receptors during copy of the α2 isoform, a complete supernumerary pillar was development. Proc. Natl. Acad. Sci. USA 89, 3347-3351. formed which, based on its anatomical relationships, likely cor- Batchelor, A. L., Phillips, R. J. S. and Searle, A. G. (1966). A comparison of responds to the reptilian pila antotica. Thus RA appears to the mutagenic effectiveness of chronic neutron- and γ-irradiation of mouse modify the ancestral reptilian program, suppressing the spermatogonia. Mutat. Res. 3, 218-229. Blomhoff, R. B. (1994). In Vitamin A in health and disease. (ed. Blomhoff, R.), formation of this pillar in mammals. Marcel Dekker Inc, New York. In reptiles, the quadrate bone (the dorsal element of the jaw- Boylan, J. F., Lohnes, D., Taneja, R., Chambon, P. and Gudas, L. J. (1993). joint) and the epipterygoid bone (which forms the lateral limit Loss of retinoic acid receptor γ function in F9 cells by gene disruption results of the cavum epiptericum) develop from a single element, the in aberrant Hoxa-1 expression and differentiation upon retinoic acid treatment. Proc. Natl. Acad. Sci. USA 90, 9601-9605. pterygoquadrate (upper jaw) cartilage. During the emergence Chambon, P. (1994). The retinoid signaling pathway: molecular and genetic of mammal-like reptiles (therapsids), the quadrate bone, while analyses. Sem. in Cell Biol. 5, 115-125. losing its connection with the epipterygoid bone, underwent Chisaka, O. and Capecchi, M. R. (1991). Regionally restricted developmental size reduction, with its otic process becoming the short process defects resulting from targeted disruption of the mouse homeobox gene Hox- of the incus middle ear bone [the long or stapedial process of 1.5. Nature 350, 473-479. Cohen, M. M. Jr and Sulik, K. K. (1992). Perspectives on holoprosencephaly: the incus likely represents a mammalian neomorph (Presley, Part II. Central nervous system, craniofacial anatomy, syndrome 1989)]. It is also widely held that the mammalian pterygoid commentary, diagnostic approach, and experimental studies. J. Craniofac. and alisphenoid bones evolved from the reptilian epipterygoid Genet. Dev. Biol. 12, 196-244. bone which develops from the rostral portion of the pterygo- Condie, B. G. and Capecchi, M. R. (1993). Mice homozygous for a targeted quadrate cartilage. A number of RAR double mutant fetuses disruption of Hoxd-3 (Hox-4.1) exhibit anterior transformations of the first and second cervical vertebrae, the atlas and the axis. Development 119, 579- exhibited abnormal skeletal structures in which the body of the 595. incus was continuous with a rostrally oriented cartilaginous or Coulombre, A. J. and Coulombre, J. L. (1977). Abnormal organogenesis in osseous rod, which was often fused to the alisphenoid (Table the eye. In Handbook of Teratology. (eds J. G. Wilson and F. Clarke Fraser), 2), while the short process of the incus was larger than in pp. 329-354. Plenum Press, New York. de Beer, G. R. (1985). The Development of the Vertebrate Skull. The normal fetuses. It is likely that this is an atavistic structure cor- University of Chicago Press, Chicago. responding to the therapsid evolutionary state in which the DeMyer, W. (1975). Median facial malformations and their implications for incus had appeared, but was still linked to the newly derived brain malformations. In Morphogenesis and Malformation of Face and alisphenoid bone through quadrate remnants. Interestingly, the Brain. (eds D. Bergsma, J. Langman and N. W. Paul), pp. 155-181. New present day ground state morphogenetic program of the first York: A. R. Liss, Inc. Dollé, P. and Duboule, D. (1993). Structural and functional aspects of arch appears to encode a similar atavistic structure (Rijli et al., mammalian Hox genes. In Adv. Dev. Biochem. vol. 2, (ed. P. Wassarman), 1993). It is tempting to conclude that an RA-dependent process pp. 57-109. JAI Press, Inc. has been recruited during the course of evolution to modify this Dollé, P., Dierich, A., LeMeur, M., Schimmang, T., Schuhbaur, B., program. Chambon, P. and Duboule, D. (1993). Disruption of the Hoxd-13 gene The re-emergence of ancestral skeletal features in RAR induces localized heterochrony leading to mice with neotenic limbs. Cell 75, 431-441. double mutants not only indicates that the underlying mesec- Dollé, P., Ruberte, E., Kastner, P., Petkovich, M., Stoner, C. M., Gudas, L. todermal developmental programs are still present in J. and Chambon, P. (1989). Differential expression of genes encoding α, β mammals, but also that RA-dependent mechanisms have been and γ retinoic acid receptors and CRABP in the developing limbs of the recruited during the reptilian-mammalian transition to modify mouse. Nature 342, 702-705. some features of the reptilian skull. Dollé, P., Ruberte, E., Leroy, P., Morriss-Kay, G. and Chambon, P. (1990). Retinoic acid receptors and cellular retinoid binding proteins. I. A systematic study of their differential pattern of transcription during mouse We would like to thank Dr T. Pexieder for a critical reading of the organogenesis. Development 110, 1133-1151. manuscript, Dr M. LeMeur for her collaboration and the members of Duboule, D. (1992). The vertebrate limb, a model system to study the the retinoid group for useful discussions; B. Weber, C. Fischer and Hox/HOM gene network during development and evolution. BioEssays 14, V. Giroult and the technical staff of the animal facility for excellent 375-384. help; B. Boulay, J. M. Lafontaine and C. Werlé for the illustrations Gans, C. and Northcutt, R. G. (1983). Neural crest and the origin of and the secretarial staff for assembling the manuscript. D. L. was a vertebrates: A new head. Science 220, 268-274. recipient of a fellowship from the MRC Canada and C. M. was Geeraets, R. (1976). An electron microscopic study of the closure of the optic supported by a fellowship from the NIH (5F32 GM13597-03) and fissure in the golden hamster. Am. J. Anat. 145, 411-432. from the ARC. This work was supported by funds from the Institut Giguère, V., Ong, E. S., Segui, P. and Evans, R. M. (1987). Identification of a receptor for the morphogen retinoic acid. Nature 330, 624-629. National de la Santé et de la Recherche Médicale, the Centre National Grainger, R. M., Henry, J. J., Saha, M. S. and Servetnick, M. (1992) Recent de la Recherche Scientifique, the Centre Hospitalier Universitaire progress on the mechanisms of embryonic lens formation. Eye 6, 117-122. Régional, the Association pour la Recherche sur le Cancer, the Human Hale, F. (1933). Pigs born without eyeballs. J. Hered. 24, 105-127 Frontier Science Program and the Fondation pour la Recherche Hall, B. K. (1977). Chondrogenesis of the somitic mesoderm. Adv. Anat. Médicale. Embryol. Cell Biol. 53-54. RARs in ontogenesis 2747

Hall, B. K. (1991). Cellular interactions during cartilage and bone migration of neural crest cells in the hindbrain region of the chick embryo. development. J. Craniofac. Genet. Dev. Biol. 11, 238-250. Development 113, 1281-1291. Hamilton W. J., Boyd, J. D. and Mossman, H. W. (1945). Human Lumsden, A. G. S. and Buchanan, J. A. G. (1986). An experimental study of Embryology. Cambridge: W. Heffer & Sons Ltd. timing and topography of early tooth development in the mouse embryo with Hero, I. (1989). The optic fissure in the normal and microphthalmic mouse. an analysis of the role of innervation. Arch. Oral. Biol. 31, 301-311. Exp. Eye Res. 49, 229-239. Lyons, K. M., Pelton, R. W. and Hogan, B. L. (1990). Organogenesis and Hero, I. (1990). Optic fissure closure in the normal cinnamon mouse. pattern formation in the mouse: RNA distribution patterns suggest a role for Investigative Ophthalmology & Visual Science 31, 197-209. bone morphogenetic protein-2A (BMP-2). Development 109, 833-844. Hill, R. E., Jones, P. F., Rees, A. R., Sime, C. M., Justice, M. J., Copeland, Mangelsdorf, D. J., Umesono, K. and Evans, R. M. (1994). The Retinoid N. G., Jenkins, N. A., Graham, E. and Davidson, D. R. (1989). A new Receptors. In The Retinoids (second edition). (eds. M. B. Sporn, A. B. family of mouse homeo box-containing genes: molecular structure, Roberts and D. S. Goodman), pp. 319-350. New York: Raven Press. chromosomal location, and developmental expression of Hox-7.1. Genes Mann, I. (1937). Developmental abnormalities of the eye. Cambridge Dev. 3, 26-37. university Press. Hofman, C. and Eichele, G. (1994). Retinoids in Development. In The Mark, M., Lufkin, T., Vonesch, J.-L., Ruberte, E., Olivo, J.-C., Dollé, P., Retinoids (second edition) (Eds. M. B. Sporn, A. B. Roberts and D. S. Gorry, P., Lumsden, A. and Chambon, P. (1993). Two rhombomeres are Goodman), pp. 387-442. Raven Press, New York. altered in Hoxa-1 null mutant mice. Development 119, 319-338. Howell, J. Mc. C., Thompson, J. N. and Pitt, G. A. (1963). Histology of the Mason, K. E. (1935). Foetal death, prolonged gestation, and difficult lesions produced in the reproductive tract of animals fed a diet deficient in parturition in the rat as a result of vitamin A-deficiency. Am. J. Anat. 57, 303- vitamin A alcohol but containing vitamin A acid. I. The male rat. J. Reprod. 349. Fertil. 5, 159-167. Mavilio, F. (1993). Regulation of vertebrate homeobox-containing genes by Jacobson, A. G. (1981). Morphogenesis of the neural plate and tube. In morphogens. Eur. J. Biochem. 212, 273-288. Morphogenesis and Pattern Formation. (eds. T. G. Connelly et al.), pp. 233- McCaffery, P., Lee, M. O., Wagner, M. A., Sladek, N. E. and Dräger, U. C. 263. New York: Raven Press. (1992). Asymmetrical retinoic acid synthesis in the dorsoventral axis of the Jeannotte, L., Lemieux, M., Charron, J., Poirier, F. and Robertson, E. J. retina. Development 115, 371-382. (1993). Specification of axial identity in the mouse : role of the Hoxa-5 Mendelsohn, C., Mark, M., Dollé, P., Dierich, A., Gaub, M.-P., Krust, A., (Hox1.3) gene. Genes Dev. 7, 2085-2096. Lampron, C. and Chambon, P. (1994a). Retinoic acid receptor β2 Jegalian, B. G. and De Robertis, E. M. (1992). Homeotic transformations in (RARβ2) null mutant mice appear normal. Developmental Biology, in press. the mouse induced by overexpression of a human Hox3.3 transgene. Cell 71, Mendelsohn, C., Lohnes, D., Décimo, D., Lufkin, T., LeMeur, M., 901-910. Chambon, P. and Mark, M. (1994b). Function of the retinoic acid receptors Johnston, M. C., Noden, D. M., Hazelton, R. D., Coulombre, J. L. and (RARs) during development. Development 120, 2749-2771. Coulombre, A. J. (1979). Origins of avian ocular and periocular tissues. Mendelsohn, C., Ruberte, E., LeMeur, M., Morriss-Kay, G. and Chambon, Exp. Eye Res. 29, 27-43. P. (1991). Developmental analysis of the retinoic acid-inducible RAR-β2 Kastner, P., Leid, M. and Chambon, P. (1994). The role of nuclear retinoic promoter in transgenic animals. Development 113, 723-734. acid receptors in the regulation of gene expression. In Vitamin A in Health Milaire, J. (1978). Etude morphologique, histochimique et autoradiographique and Disease (ed. Blomhoff, R.), pp. 189-238. New York: Marcel Dekker Inc. du développement du squelette des membres chez l’embryon de souris. I. Kaufman, M. H. (1992). The Atlas of Mouse Development. London: Academic Membres antérieurs. Arch. Biol. 89, 169-216. Press. Morriss-Kay, G. (1993). Retinoic acid and craniofacial development: Kessel, M. (1992). Respecification of vertebral identities by retinoic acid. molecules and morphogenesis. BioEssays 15, 9-15. Development 115, 487-501. Nakamura, A., Sakamoto, H. and Moriwaki, K. (1963). Genetical studies of Kessel, M. and Gruss, P. (1991). Homeotic transformations of murine post-axial polydactyly in the house mouse. Ann. Rep. Nat. Inst. Genet. Jap. vertebrae and concomitant alteration of Hox codes induced by retinoic acid. 13, 31. Cell 67, 89-104. Nau, H., Chahoud, I., Dencker, L., Lammer, E. J. and Scott, W. J. (1994). Kessel, M., Balling, R. and Gruss, P. (1990). Variations of cervical vertebrae Teratogenicity of Vitamin A and Retinoids. In Vitamin A in Health and after expression of a Hox-1.1 transgene in mice. Cell 61, 301-308. Disease (ed. Blomhoff, R.), pp. 615-664. New York: Marcel Dekker Inc. Langston, A. W. and Gudas, L. J. (1992). Identification of a retinoic acid Niswander, L., Tickle, C., Vogel, A., Booth, I. and Martin, G. R. (1993). responsive enhancer 3′ of the murine homeobox gene Hox-1.6. Mech. Dev. FGF-4 replaces the apical ectodermal ridge and directs outgrowth and 38, 217-228. patterning of the limb. Cell 75, 579-587. Le Douarin, N., Ziller, C. and Couly, G. (1993). Patterns of neural crest Noden, D. M. (1983). The role of the neural crest in patterning of avian cranial derivatives in the avian embryos: in vivo and in vitro studies. Dev. Biol. 159, skeletal, connective, and muscle tissues. Dev. Biol. 96, 144-165. 24-49. Noden, D. M. (1988). Interactions and fates of avian craniofacial mesenchyme. Le Mouellic, H., Lallemand, Y. and Brûlet, P. (1992). Homeosis in the mouse Development 103Supplement, 121-140. induced by a null mutation in the Hox-3.1 gene. Cell 69, 251-264. Pagon, R. A., Graham, J. M. Jr., Zonana, J. and Yong, S. L. (1981). Leid, M., Kastner, P. and Chambon, P. (1992). Multiplicity generates Coloboma, congenital heart disease, and choanal atresia with multiple diversity in the retinoic acid signalling pathways. Trends Bioch. Sci. 17, 427- anomalies. J. Pediat. 99, 223-227. 433. Pei, Y. F. and Rhodin, J. A. G. (1970) The prenatal development of the mouse Li, E., Sucov, H. M., Lee, K.-F., Evans, R. M. and Jaenisch, R. (1993). eye. Anat. Rec. 168, 105-126. Normal development and growth of mice carrying a targeted disruption of Petkovich, M., Brand, N. J., Krust, A. and Chambon, P. (1987). A human the α1 retinoic acid receptor gene. Proc. Natl. Acad. Sci. USA 90, 1590-1594. retinoic acid receptor which belongs to the family of nuclear receptors. Linney, E. and LaMantia, A-S. (1994). Retinoid signalling in mouse Nature 330, 444-450. embryos. Advances in Developmental Biology 3, in press. Pexieder, T. and Jelinek, R. (1970). Pressure of the CSF and the Lohnes, D., Kastner, P., Dierich, A., Mark, M., LeMeur, M. and Chambon, Morphogenesis of the CNS II. Pressure necessary for normal development of P. (1993). Function of retinoic acid receptor γ (RARγ) in the mouse. Cell 73, brain vesicles. Morphologica 18, 181-192. 643-658. Pollock, R. A., Jay, G. and Bieberich, C. J. (1992). Altering the boundaries Lufkin, T., Dierich, A., LeMeur, M., Mark, M. and Chambon, P. (1991). of Hox3.1 expression: evidence for antipodal gene regulation. Cell 71, 911- Disruption of the Hox-1.6 homeobox gene results in defects in a region 923. corresponding to its rostral domain of expression. Cell 66, 1105-1119. Pöpperl, H. and Featherstone, M. S. (1993). Identification of a retinoic acid Lufkin, T., Lohnes, D., Mark, M., Dierich, A., Gorry, P., Gaub, M. P., response element upstream of the murine Hox-4.2 gene. Mol. Cell. Biol. 13, LeMeur, M. and Chambon, P. (1993). High postnatal lethality and testis 257-265. degeneration in retinoic acid receptor α (RARα) mutant mice. Proc. Natl. Presley, R. (1989). Ontogeny and the evolution of the mammalian jaw Acad. Sci. USA 90, 7225-7229. complex. In Complex Organismal Functions : Integration and Evolution in Lufkin, T., Mark, M., Hart, C. P., Dollé, P., Le Meur, M. and Chambon, P. Vertebrates (eds. D. B. Wake and G. Roth), pp. 53-61. New York: John (1992). Homeotic transformation of the occipital bones of the skull by Wiley & Sons Ltd. ectopic expression of a homeobox gene. Nature 359, 835-841. Ramirez-Solis, R., Zheng, H., Whiting, J., Krumlauf, R. and Bradley, A. Lumsden, A., Sprawson, N. and Graham, A. (1991). Segmental origin and (1993). Hox-B4 (Hox-2.6) mutant mice show homeotic transformation of 2748 D. Lohnes and others

cervical vertebra and defective closure of the sternal rudiments. Cell 73, 279- features of the charge association: support for involvement of the neural 294. crest. Teratology 31, 331-336. Riddle, R. D., Johnson, R. L., Laufer, Ed. and Tabin, C. (1993). Sonic Simeone, A., Acampora, D., Arcioni, L. Andrews, P. W., Boncinelli, E. and hedgehog mediates the polarizing activity of the ZPA. Cell 75, 1401-1416. Mavilio, F. (1990). Sequential activation of HOX2 homeobox genes by Rijli, F., Mark, M., Lakkaraju, S., Dierich, A., Dollé, P. and Chambon, P. retinoic acid in human embryonal carcinoma cells. Nature 346, 763-766. (1993). A homeotic transformation is generated in the rostral branchial Simeone, A., Acampora, D., Nigro, V., Faiella, A., D’Esposito, M., region of the head by disruption of Hoxa-2, which acts as a selector gene. Cell Stornaiuolo, A., Mavilio, F. and Boncinelli, E. (1991). Differential 75, 1333-1349. regulation by retinoic acid of the homeobox genes of the four HOX loci in Robert, B., Sassoon, D., Jacq, B., Gehring, W. and Buckingham, M. (1989). human embryonal carcinoma cells. Mech. Dev. 33, 215-228. Hox-7, a mouse homeobox gene with a novel pattern of expression during Sporn, M. B., Roberts, A. B. and Goodman, D. S. (1994). In The Retinoids embryogenesis. EMBO J. 8, 91-100. (second edition) (eds. M. B. Sporn, A. B. Roberts and D. S. Goodman), pp. Rossant, J., Zirngibl, R., Cado, D., Shago, M. and Giguère, V. (1991). 319-350. New York: Raven Press. Expression of a retinoic acid response element-hspβ-gal transgene defines Suzuki, T., Shirai, S. and Majima, A. (1988) Morphological study on the specific domains of transcription activity during mouse embryogenesis. mechanism of closure of the embryonic fissure. Acta. Soc. Ophthalmol. Genes Dev. 5, 1333-1344. Japan 92, 238-242. Ruberte, E., Dollé, P., Chambon, P. and Morriss-Kay, G. (1991). Retinoic Tabin, C. J. (1991). Retinoids, homeoboxes, and growth factors: toward acid receptors and cellular retinoid binding proteins. II. Their differential molecular models for limb development. Cell 66, 199-217. pattern of transcription during early morphogenesis in mouse embryos. Thompson, J. N., Howell, J. McC. and Pitt, G. A. J. (1964). Vitamin A and Development 111, 45-60. reproduction in rats. Proc. Royal Soc. 159, 510-535. Ruberte, E., Dollé, P., Krust, A., Zelent, A., Morriss-Kay, G. and Wald, G. (1968). Molecular basis of visual excitations. Science 162, 230-239. Chambon, P. (1990). Specific spatial and temporal distribution of retinoic Wanek, N., Gardiner, D. M., Mineska, K. and Bryant, S. V. (1991). acid receptor gamma transcripts during mouse embryogenesis. Development Conversion by retinoic acid of anterior cells into ZPA cells in the chick wing 108, 213-222. bud. Nature 350, 81-83. Searle, A. G. (1964). The genetics and morphology of two ‘luxoid’ mutants in Warkany, J. and Schraffenberger, S. (1946) Congenital malformations the house mouse. Genet. Res. 5, 171-197. induce in rats by maternal vitamin A deficiency. I. Defects of the eye. Arch. Sechrist, J., Serbedzija, G. N., Scherson, T., Fraser, S. E. and Bronner- Ophth. 35, 150-169. Fraser, M. (1993). Segmental migration of the hindbrain neural crest does Wilson, J. G. and Barch, S. (1949). Fetal death and maldevelopment resulting not arise from its segmental generation. Development 118, 691-703. from maternal vitamin A deficiency in the rat. Proc. Soc. Exp. Biol and Med. Serbedzija, G. N., Bronner-Fraser, M. and Fraser, S. E. (1992). Vital dye 72, 687-693. analysis of cranial neural crest cell migration in the mouse embryo. Wilson, J. G., Roth, C. B. and Warkany, J. (1953). An analysis of the Development 116, 297-307. syndrome of malformations induced by maternal vitamin A deficiency. Shubin, N. H. (1991). The implications of the bauplan for development and Effects of restoration of vitamin A at various times during gestation. Am. J. evolution of the tetrapod limb. In Developmental Patterning of the Anat. 92, 189-217. Vertebrate Limb(Eds. J. R. Hinchliffe, J. M. Hurle and D. Summerbell). New Wolbach, S. B. and Howe, P. R. (1925). Tissue changes following deprivation York: Plenum Publishing Corp. of fat-soluble A vitamin. J. Exp. Med. 42, 753-777. Shubin, N. H. and Alberch, P. (1986). A morphogenetic approach to the origin and basic organization of the tetrapod limb. Evol. Biol. 20, 319-387. Siebert, J. R., Graham, J. M. Jr. and MacDonald, C. (1985). Pathologic (Accepted 15 July 1994)