Involvement of Thyroid Hormone and Its & Receptor in Avian Neurulation · 2017-02-25 · neurulation a marked increase was observed in neurectoderm. A reporter cell line was then

DEVELOPMENTAL BIOLOGY 197, 1–11 (1998)ARTICLE NO. DB988872

Involvement of Thyroid Hormone andIts a Receptor in Avian Neurulation

Frederic Flamant1 and Jacques SamarutLaboratoire de Biologie Moleculaire et Cellulaire, l’Ecole Normale Superieure de LyonCNRS UMR49-INRA LA913, Allee d’Italie, 69364 Lyon Cedex 07, France

We have analyzed the expression pattern of c-erbAa and c-erbAb which encode the thyroid hormone receptors (T3Ra andT3Rb) during early chicken embryogenesis. Only c-erbAa expression was detected by RT-PCR and whole-mount in situhybridization. c-erbAa transcripts were found to be already present at low level in embryos before egg incubation. Duringneurulation a marked increase was observed in neurectoderm. A reporter cell line was then constructed and used todemonstrate the release of significant amount of thyroid hormone (T3) from egg yolk by area opaca cells before gastrulation.During gastrulation T3 was found to be enriched in the primitive streak and Hensen’s node. Introduction of excess T3frequently resulted in abnormal development of anterior structures, mainly neural tube defects and anencephalia. Theseobservations suggest that T3Ra, like the closely related retinoic acid receptors, fulfills functions which are important forembryonic development well before the onset of thyroid gland function. q 1998 Academic Press

Key Words: thyroid hormone; c-erbA; chick.

INTRODUCTION 1995). However, some response elements can mediate bothretinoids and T3 response (Umesono et al., 1988; Tini etal., 1994). DNA binding is not hormone dependent, andThyroxine is secreted by the thyroid gland of vertebratesboth T3R and RAR can regulate transcription in the absenceand deiodinated in several tissues to produce the more ac-of hormone. Although ligand binding usually induces ative hormone tri-iodo-thyronine (T3). T3 controls homeo-switch from transcription inhibition to transcription activa-stasis in adults and several aspects of development, includ-tion, some hormone response elements mediate hormone-ing amphibian metamorphosis and mammalian brain matu-induced repression (Horlein et al., 1995; Kurokawa et al.,ration (Chatterjee and Tata, 1992; Lebel et al., 1994; Nunez,1995). T3R, RAR, and RXR are involved in a large network1984). It acts on the in vitro differentiation of erythrocytesof protein interactions that constitute a molecular basis for(Gandrillon et al., 1994) myotubes (Cassar-Malek et al.,cross-talks with several other transcription regulation path-1994; Sachs et al., 1996) oligodendrocytes and neuronesways, including AP-1 (Pfahl, 1993).(Bernal and Nunez, 1995; Lezoualc’h et al., 1995) . Molecu-

In chicken, T3R are encoded by two loci: c-erbAa andlar cloning has revealed a close relationship between thyroidc-erbAb, which encode several isoforms of the receptor:hormone receptors (T3R) and retinoic acid receptors (RAR),T3Ra, T3Rb0, T3Rb2 (Sjoberg et al., 1992) which are nota surprising discovery considering the fact that the respec-functionally equivalent (Lezoualc’h et al., 1992; Zavacki ettive ligand molecules are not chemically related, and bindal., 1993; Safer et al., 1996). The precise development stageonly to their own receptor. The two types of receptors be-at which thyroid hormone starts to regulate gene expressionlong to the same subfamily of nuclear hormone receptorsis not known. Thyroid gland is functional 9.5 days after theand share a number of properties. They can act as homodi-beginning of egg incubation but because both T3 and itsmers or as heterodimers with 9-cis-retinoic acid receptorspoorly active precursor thyroxine are present in egg yolk(RXR) (Mangelsdorf et al., 1995). Each monomer can recog-(Prati et al., 1991) these hormones may be accessible earliernize the same core motif, and binding specificity of dimersto embryonic tissues. c-erbAb transcripts could not be iden-on DNA depends on the configuration of core motifs withintified by Northern blotting before embryonic day 4, wherethe hormone response element (Mangelsdorf and Evans,it is abundant in retina, whereas c-erbAa mRNA is presentat embryonic day 3 (Forrest et al., 1990). There is no reportedexperiment addressing the possibility of earlier expression.1 To whom correspondence should be addressed. Fax: (00 33) 04

72 72 86 86. E-mail: [email protected]. These incomplete observations prompted us to explore an

1

0012-1606/98 $25.00Copyright q 1998 by Academic PressAll rights of reproduction in any form reserved.

AID DB 8872 / 6x3c$$$141 04-03-98 08:01:56 dbal

2 Flamant and Samarut

calcium phosphate coprecipitation and selected with G418 (500 mg/eventual involvement of thyroid hormone and its receptorsml) 2 days later. Forty cell clones were picked 10 days later andat early stages of chicken development. We report that T3tested for b-galactosidase activity by Xgal staining after 2 days inand its a receptor are present at the blastula stage, and thatthe presence of 1007 M of T3. The responsive clones were furtherstriking analogies exist between retinoic acid and T3 actiontested for the absence of b-galactosidase activity in the absence ofduring early morphogenesis.T3. One of the clones, called GG1, was used as a reporter cell linein all experiments. GG1 cells were grown in Dulbecco’s modifiedEagle medium supplemented with penicillin, streptomycin, gluta-

MATERIALS AND METHODS mine, and 10% fetal calf serum depleted of T3 by ion exchange onAG 1-X8 beads (Bio-Rad). GG1 cells were used for only seven pas-sages, during which no major variation in T3 response was ob-RT–PCRserved. b-Galactosidase activity in cell lysates were measured in

Total RNA was prepared using Trizol (Eurobio) in the presence triplicate using the Galactolight kit (Tropix).of 10 mg/ml of carrier glycogen (Boehringer, Mannheim). Polyade-nylated RNA was then purified using oligo(dT)-loaded magneticbeads (Dynal). Reverse transcription was initiated from random Cultures of Embryonic Cellsprimers, using AMV reverse transcriptase (Promega). cDNA were

Whole embryos or area opaca were mechanically dissociated andamplified by PCR with Goldstar Taq polymerase (Eurogentec).grown on gelatin-coated dishes in the same medium. Conditionedc-erbAa cDNA was amplified by 35 temperatures cycles at 957Cmedium was harvested 48 h later and cell debris was removed by(30 s), 687C (15 s), and 727C (5 s) using the oligonucleotides primers:centrifugation and stored frozen before assaying for the presence5*-CTATCTTATCCACGCAGATCAGCCCCGTCGG-3* and 5*-of T3 on GG1 cells.CACGCCGGCCATCACCCGCGTGG-3*.

c-erbAb (all isoforms) cDNA was amplified by 40 temperaturescycles at 957C (30 s), 567C (15 s), and 727C (5 s) using the oligonucle- Cocultivation of Embryonic Cells with GG1otides primers: 5*-GGTGTTGGATGACAGCAAGAGG-3* and 5*-

Reporter CellsTTCCAGTGGCTTCCTTGTGC-3*.Glycerol-3- phosphate dehydrogenase cDNA was amplified by GG1 cells were seeded 1 day before embryo dissection at 2 1

25 cycles at 957C (30 s), 567C (15 s), and 727C (5 s) using the oligonu- 105 cells per 35-mm dish containing 1 ml of the medium describedcleotides primers: 5*-CTGTCCATGCCATCACAGCCACAC-3* above. Embryos were dissected in phosphate-buffered saline usingand 5*-TCAGCAGCAGCCTTCACTACCCTC-3*. tungsten needles and identical fragments of five different embryos

were added in one dish. In the case of area opaca cells, which canquickly overgrow GG1 cells, only two samples were added in one

In Situ Hybridization coculture dish.

Whole-mount in situ hybridization was performed with digoxi-genin-labeled RNA probe (Boehringer, Mannheim) as described T3 Injection in Egg Yolk(Riddle et al., 1994). Probes were synthesized with T7 polymerasefrom linearized plasmids derived from pSG5 (Stratagene) carrying Solutions of 1002 and 1003 M thyroid hormone (Sigma) were pre-the entire coding sequence of chicken c-erbAa in sense (control pared in 10 mM NaOH, and diluted in calcium-free Tyrode’s solu-probe) or antisense orientation. Antisense probe specificity was tion supplemented with 0.1mg/ml of Fast Green (Sigma) to lightlycontrolled on fibroblasts infected by a retrovirus vector carrying stain the solution. Controls were performed with 10 mM NaOHthe c-erbAa coding sequence. As this probe may cross-react with diluted in the same solution. Eggs were windowed, and 1 to 2 mlc-erbAb mRNA, we used it only at stages when c-erbAb mRNA is was injected in the subgerminal cavity with a glass capillary pipettenot detectable by RT–PCR. (Over et al., 1980). Eggs were returned to incubation after sealing

with tape.

PlasmidsRESULTSpGCnlslacZ was constructed by inserting the nlslacZ gene which

encodes a nuclear targeted b-galactosidase in pGC, a plasmid con-Expression of c-erbAa Starts before Gastrulationtaining (5* to 3*) 4 response elements for the yeast Gal4 transcrip-

tion factor, a minimum HSV thymidine kinase promoter, a multi-To determine at which time c-erbAa and c-erbAb tran-ple cloning site, a b-globin intron, and an SV40 polyadenylation

scription starts during chicken development, polyadenyl-signal (Braselman et al., 1993). pGC-Gal4-cerbAa was constructedated RNA were purified from pools of embryos and theirin two steps: First, the XbaI of the rat c-erbAa cDNA was intro-extraembryonic annexes after various time of incubationduced into pMFH2 (Witzgall et al., 1994), creating a reading frameand analyzed by RT–PCR. c-erbAa transcripts were foundcovering the Gal4 DNA binding domain (amino acids 1–147), the

hinge and ligand binding domains of T3Ra. The SalI fragment of in all samples from stage 1 (unincubated blastoderm (Ham-the resulting plasmid was then transferred into pGC. burger and Hamilton, 1951)) to stage 10 (35 h of incubation,

10 somites). By contrast c-erbAb transcripts were only justdetectable after extensive cycling in late-stage samples only

Reporter Cells (Fig. 1A). This confirmed that c-erbAb transcription startsat relatively late stages (Forrest et al., 1990). To address aSwiss 3T3 cells (ATCC) were cotransfected with pGCnlslacZ,

pGC-Gal4-cerbAa, and pKoNeo (Maniatis et al., 1982) by DNA/ possible restriction of c-erbAa pattern expression, embryos

Copyright q 1998 by Academic Press. All rights of reproduction in any form reserved.

AID DB 8872 / 6x3c$$$141 04-03-98 08:01:56 dbal

3Thyroid Hormone in Avian Neurulation

experiments performed with the sense probe. With the anti-sense probe, staining was observed before neurulation. Atstage 4, expression was low but ubiquitous. At stage 6, gen-eralized expression was still observed, but the newly formedneural fold was more strongly stained. At later stages thisstrong signal proceeded to the entire neural tube. Observa-tion of serial sections of stage 8 embryos confirmed thatc-erbAa expression is found in ectoderm, mesoderm, andendoderm, and high in neural tube.

A Reporter Cell Line to Detect Small Amountsof Thyroid Hormone

The above data raise the possibility that T3Ra activatesits target genes during the very early stages of chicken devel-opment. Alternatively, unliganded T3Ra may repress targetgene expression. To look for the possible presence of thyroidhormone in the embryonic cells at these stages, we designeda reporter cell line able to react to femtomolar quantitiesof T3. This was performed by transfecting mouse fibroblastswith two plasmids: a first plasmid carrying the nlslacZ geneencoding a nuclear b-galactosidase driven by a minimumpromoter preceded by several response elements for theyeast Gal4 transcription factor (Gal4RE), and a second plas-mid encoding a Gal4-c-erbAa protein which is able to bindFIG. 1. RT–PCR detection of c-erbAa mRNA during earlyto Gal4RE. This binding entails a repression of transcriptionchicken development. Polyadenylated RNA was extracted fromin the absence of T3, and a strong activation when T3 isembryos before incubation or after 1 or 2 days of incubation. In theadded (Katz and Lazar, 1993). The second plasmid was alsolatter case, area pellucida with the embryo proper was separated

from area opaca. cDNA was synthesized using random DNA prim- driven by the same Gal4-responsive promoter. In this situa-ers from 1 mg of poly(A) RNA for all samples, except for unincubated tion, the addition of T3 to the culture medium produces anembryos where only 0.5 mg was used. cDNA was then used for increase in the expression of the hybrid receptor, concomi-PCR amplification. (A) c-erbAa mRNA is present from stage 1 and tant to the activation of the reporter nlslacZ gene, thereforeexpressed in both area pellucida (AP) and area opaca (AO). Low ensuring a high specificity and sensitivity of the systemlevels of c-erbAb mRNA is only detected at late stage in area pellu- (Fig. 3). In the clone called GG1, the frequency of cells withcida. Glycerol-3-phosphate dehydrogenase (GAPDH) mRNA is used

a visible nuclear b-galactosidase activity (lac/ cells) wasas an internal control. (B) c-erbAa mRNA is widely expressed invery low in the absence of T3 (around 1004 after 3 h of Xgalembryonic tissues. mRNA were prepared from three fragments ofstaining at 377C). A small but significant increase in theindividual embryos. Somites were used as landmarks to cut areafrequency of lac/ cells was observed in the presence ofpellucida in three parts. The upper fragment covers the region that

is rostral to the first somite. The middle fragment is the somitic 10011 M of T3, and the majority of the cells were lac/ afterregion. The lower fragment is caudal to the last somite. Variations 48 h of culture with more than 1008 M of T3. To confirmof amplification were observed among the different embryos and the specificity of GG1 cells response, we did also verify thatfragments, but analysis of six other embryos failed to reveal any the frequency of lac/ cells is not significantly affected bysignificant spatial or temporal restriction in c-erbAa expression. retinoic acid (1006 M).

Early Embryonic Cells Release Thyroid Hormonewere dissected, and individual samples of polyadenylatedRNA were prepared from upper, middle, and lower part of Unincubated blastoderm embryos were isolated and cellsstage 8 to 10 embryos. The c-erbAa transcripts were present were cocultivated for 2 days with GG1 reporter cells (seein these three fragments at a similar level, indicating that Materials and Methods). Cells were then stained with Xgalno tight spatial regulation of c-erbAa transcription occurs to detect nuclear b-galactosidase. Although some embry-in early chicken embryos (Fig. 1B). onic cells had endogenous cytoplasmic enzymatic activity,

the fact that GG1 cells b-galactosidase is nuclear preventedWidespread Expression of c-erbAa: High Level any confusion. To correct for leaky expression of nlslacZin Neurectoderm the number of lac/ cells was compared to the one observed

in the control GG1 cells that were included in each experi-To precisely define the expression pattern of c-erbAa dur-ing the first stages of chicken development, whole-mount in ment. A twofold increase in the number of lac/ cells was

observed in the cocultures, demonstrating that blastodermsitu hybridization was performed with digoxigenin-labeledRNA probes (Fig. 2). No staining was observed in control cells produce thyroid hormone. Quantifying b-galactosidase


AID DB 8872 / 6x3c$$$141 04-03-98 08:01:56 dbal


FIG. 2. Whole-mount in situ hybridization of c-erbAa mRNA. Embryos were hybridized to a RNA antisense probe corresponding to the3* of c-erbAa. (A) Primitive streak embryo stage 4. (B) Stage 6. (C) Stage 8. (D) Stage 8. (E) Stage 9. (F) 7 mm transversal section after paraffinembedding of stage 8 embryo mid trunk region. (G) Stage 8 embryo hybridized to the control plus strand probe. hn, Hensen’s node; ps,primitive streak; nf, neural fold; ect, ectoderm; nt, neural tube; mes, mesoderm; end, endoderm; so, somites. Horinzontal bars correspondto approximatively 1 mm.


AID DB 8872 / 6x3c$$8872 04-03-98 08:01:56 dbal


AID DB 8872 / 6x3c$$8872 04-03-98 08:01:56 dbal


FIG. 4. Quantitation of T3 release by embryonic cells cultivated in vitro. (A) Calibration of GG1 response at high T3 concentration: b-Galactosidase activity was measured in GG1 cell lysate after 48 h of culture in a medium supplemented with increasing amounts of T3.Three cell dishes were used for each T3 concentration, and the measurement of b-galactosidase activity in each lysate was triplicated.(B) Quantitation of T3 release by embryonic cells in vitro: Cell medium was replaced 2 days before lysis by control medium (3 dishes),medium with low concentration of T3 (3 dishes for each T3 concentration), or supernatant of embryonic cell culture (48 h of culture incontrol medium, 1 ml/embryo).

enzymatic activity in the coculture lysate would be mis- chemoluminescent assay. A small but significant increasein b-galactosidase enzymatic activity was observed (Fig. 4),leading because it would not distinguish GG1 cells activity

from embryonic cell endogenous activity. To quantify em- corresponding to a production exceeding 10 fmol T3/em-bryo at stage 1 and 100 fmol T3/embryo at stage 4. Thisbryonic cell hormone production, pools of 10 blastoderms

were cultivated alone for 48 h and the presence of thyroid increase in hormone production correlates with the rapidincrease in blastoderm size: if one roughly estimates thehormone in the supernatant was addressed by transferring

the medium to GG1 cells for another 48 h. b-Galactosidase embryonic volume to be less than 10 ml at stage 1 and 100ml at stage 4, the average concentration of free hormone inactivity was then quantified in the GG1 cell lysate using a

FIG. 3. GG1 cells response to thyroid hormone. GG1 cells have been designed to allow a highly sensitive and specific detection of T3in culture medium. (A) Plasmids integrated in 3T3 cells genome. In the absence of T3, pGC-Gal4-cerbAa1 is weakly expressed, and thefew Gal4-T3Ra hybrid receptors which are possibly synthesized bind to the Gal4 response elements present in the two transcriptionpromoters of pGC-Gal4-cerbAa1 and pGCnlslacZ and repress transcription. When T3 is present in the medium Gal4-T3Ra activatestranscription. This results in a positive feedback loop, causing the accumulation of both hybrid receptors and nuclear-targeted b-galactosi-dase. (B) GG1 cells cultivated for 48 h in various concentrations of T3 for 48 h and stained for 3 h with Xgal at 377C.


AID DB 8872 / 6x3c$$$141 04-03-98 08:01:56 dbal


FIG. 5. A cocultivation assay to reveal unequal distribution of T3 within early embryonic tissues. Embryos were explanted and fiveidentical fragments of embryonic tissues (shaded area) were pooled and cocultivated with GG1 reporter cells for 48 h, in 1 ml of medium.Lac/ GG1 reporter cells were counted 2 days later, after Xgal staining. Due to leaky expression of nlslacZ cells in the absence of T3, theratio between the number of lac/ cells in each dish and the number of lac/ cells in the negative control dish was calculated in eachcase. The dotted line correspond to a ratio of 1, expected in the complete absence of T3. The first two columns report the calibrationperformed with standart T3 solutions. Error bars represent to extreme values observed in at least four independent experiments, not tothe statistical calculation of standard deviation. The fact that activation is superior for Hensen’s node that for entire stage 4–6 areapellucida suggests an uptake or degradation of T3 by some other embryonic cells.

an embryo might be constantly superior to 1009 M, which maintained, whereas a small release of T3 from the areapellucida was observed (Fig. 5). The area pellucida was fur-is more than sufficient for receptor activation.ther dissected and four small fragments were cocultivatedwith GG1 reporter cells: The first fragment includedArea Opaca Cells Release T3 before GastrulationHensen’s node, the second the caudal part of the primitive

Before egg incubation, the blastoderm can be divided into streak, the third one contained a lateral portion of the areatwo concentric areas: the area opaca and the area pellucida. pellucida, and the fourth fragment was taken in the regionThe area opaca contain cells that will participate only in that is anterior to the primitive streak. The two fragmentsforming extraembryonic membranes: These can be easily containing the primitive streak were observed to secreterecognized by their large size and the presence of numerous the highest amount of T3 (Fig. 5). Each Hensen’s node regioncytoplasmic inclusions which reflects a very active uptake can produce approximately 10 fmol of T3. From stage 8 toof yolk droplets. The central region, area pellucida, contains 14, T3 was detected only in the caudal half of the embryosmall cells which are totipotent progenitors of the embryo (data not shown)proper (Pain et al., 1996) and are separated from egg yolk bythe subgerminal cavity. These two regions were manually T3 Regulates Area Opaca Cell Growth in Vitrodissected and cocultivated separately with GG1 reporter

We took advantage of the fact that area opaca cells growcells. Because area opaca cells grow very rapidly in culture,

very actively in culture as a relatively homogenous cellonly the central part of area pellucida was kept to avoid

population to address a possible regulation of their prolifera-any contamination. This experiment showed that only area

tion by T3 (1008 or 1007 M). T3 was found to exert a signifi-opaca can secrete a detectable amount of T3 at this stage

cant negative influence on the growth of primary area opaca(Fig. 5). The egg yolk itself did not activate nlslacZ expres-

cell cultures (Fig. 6). The morphology of the cells and thesion in GG1 cells, whereas, as already noted (Prati et al.,

appearance of characteristic multilayered structures pre-1991), methanol treatment of egg yolk was sufficient to

viously described as ‘‘dome-like structures’’ (Mitrani andextract significant amounts of hormone (data not shown).

Eyal-Giladi, 1982) remained unchanged.

Thyroid Hormone Is Enriched in Primitive Streak Exogenous T3 Is a Potent TeratogenThe concomitant presence of T3Ra and its ligand in earlyWhen cocultivation experiments were performed with

stage 4–6 embryos, the release of T3 by the area opaca was chicken embryos suggests that precise regulation of target


AID DB 8872 / 6x3c$$$141 04-03-98 08:01:56 dbal


ing early chicken development. (a) c-erbAa expression, butnot c-erbAb expression, starts before the blastula stage andis high in neural plate and neural tube. (b) Area opaca cellssecrete T3 continuously from the blastula stage. Duringgastrulation and neurulation Hensen’s node, which is theembryo organizer, and primitive streak, produce T3. (c) Ex-ogenous T3 disturbs anterior neural tube morphogenesis.Thus T3 can act at a short distance during early chickendevelopment, well before the onset of blood circulation andthyroid function; these features characterize a new modefor thyroid hormone action.

Distribution of Thyroid Hormone in Early Embryosand Possible Functions

Our observations strongly suggest that area opaca cells re-lease T3 that is bound to yolk proteins. Another nonexclusivepossibility would be that these cells deiodinate yolk thyrox-ine to synthesize T3. The released hormone could then diffuseand reach area pellucida cells at a stage where they alreadyexpress the c-erbAa at low level. The fact that the in vitroproliferation of area opaca cells can be modulated by T3 maybe an indication for a function in the control of early embry-

FIG. 6. Inhibition of area opaca cells growth in vitro by T3. Area onic disk expansion. Disk expansion does not seem to beopaca cells dissected from unincubated stage 1 embryos were modified by in ovo T3 injection, but in this case endogenousseeded on gelatine coated dishes (1 area opaca/3 cm2) and grown in level of T3 may mask exogenous hormone effect.the presence (1008 or 1007 M) or absence of T3 for 3 days. At this After primitive streak formation, the unequal distribu-time, density was superior to the seeding density in all dishes.

tion of the receptor and its ligand in the embryo properCells were then separated by trypsine treatment and counted. Thisshould entail marked regional differences in target genesdemonstrated a dose-dependent effect of T3 on area opaca cellregulation by T3 and provides a molecular basis for a posi-growth Although this might reflect a limitation in cell prolifera-tional information. In particular, neural plate cells expresstion, a moderate effect on the cell death rate cannot be ruled out.c-erbAa at a high level and are located near the two majorsites of T3 release: Hensen’s node and primitive streak. Apossible involvement of thyroid hormone in neurulation is

gene by T3 is important for embryonic development. If this supported by the observation that injection of excess thy-is true, we can predict that adding an excess of hormone will roid hormone frequently interferes with normal neural tubedisturb embryonic development. To test the hypothesis, we morphogenesis. Previous experiments have shown that T3injected T3 in the subgerminal cavity at stage 1 and 3, and can disturb in a similar way the early development of Xeno-compared 2 or 3 days later these embryos with control em- pus embryo, provided that T3Ra mRNA is injected in oo-bryos that had been injected with saline solution. T3 treat- cyte (Banker and Eisenman, 1993; Old et al., 1992).ment did not apparently impair area opaca expansion, asexpected from the in vitro experiments described above. T3

Thyroid Hormone and Retinoids in Hensen’s Nodeinjection did increase the frequency, but not the severity,of large embryo malformations. This effect was dose depen- Hensen’s node is a structure equivalent to Xenopus Spem-dent and not observed with reverse T3, which is a poorly ann’s organizer. It is made of cells that migrate before gas-active isomere of T3 (Table 1). The observed abnormalitiesincluded anencephaly and defect in anterior neural tubeclosure (Fig. 7). Spina bifida was observed once. No alter-

TABLE 1ations were seen in nonneural tissues, like heart, somites,Frequency of Embryo Malformations after T3 Injectionand extraembryonic membranes, although subtle defects

may not be easily visible. 10 ngO 1 ng T3 10 ng T3 reverse T3

DISCUSSION Stage 1 6/61 (10%) 10/31 (30%) 17/35 (48%) 2/14 (14%)Stage 3 2/29 (7%) 5/18 (28%) 4/12 (33%) NDTotal 8/90 (9%) 15/49 (30%) 21/47 (44%) 2/14 (14%)A Nonconventional Role for Thyroid Hormone

We present here several observations that support the idea Note. T3 has a significant effect on the frequency of malforma-tions (x2 Å 26, 8 omitting the reverse T3 data P õ 0.001).that thyroid hormone plays a hitherto unsuspected role dur-


AID DB 8872 / 6x3c$$$141 04-03-98 08:01:56 dbal


FIG. 7. Teratogenic effect of exogenous T3. 1 ng of T3 was injected in the subgerminal cavity before incubation. Embryo were dissected48 h later and observed under a dissection microscope (A) or, after dehydration and gold shadowing, under a scanning electron microscope.Note the defect in neural tube closure (A to D) extending from the forebrain sometimes to the somitic region (A). (E) Complete anencephaly.fb, forebrain; nt, neural tube; np. anterior neuropore; hb, hindbrain.

trulation from a caudal structure of the marginal zone with abnormal cephalic structure (Chen and Solursh, 1992).Finally, RA inhibits area opaca cell growth in vitro (Mitraniknown as Koller’s sickle (Izpisua-Belmonte et al., 1993), and

it gives rise to notochord and floor plate (Catala et al., 1996). and Shimoni, 1989; Modak et al., 1993).Although these data clearly establish the importance ofIn ectopic grafting experiments, it can induce embryonic

secondary axis (Waddington, 1936), or, at later stage, digits RA in early development, the underlying molecular mecha-nisms remain poorly documented. RA activates Sonicduplication (Hogan et al., 1992). Retinoic acid (RA) is en-

riched in Hensen’s node (Chen et al., 1992; Hogan et al., hedgehog expression, at least in limb bud (Riddle et al.,1994). Because the implication of Sonic hedgehog in a num-1992). Immunolocalization confirmed the presence of RA

in Hensen’s node and indicated that it is also present in the ber of inductive interactions is well established (Bumcrotand Mac Mahon, 1995; Riddle et al., 1994; Stolow and Shi,posterior primitive streak (Twal et al., 1995).

A proper regulation of RA level is crucial during early 1995), it has been proposed that Hensen’s node-inducingproperty is due to RA synthesis and transduced via Sonicdevelopment: early injection of a monoclonal antibody di-

rected against RA causes cardiovascular and head abnormal- hedgehog, which is expressed in the Hensen’s node. Thissimple view does not take into account the fact that thereities (Twal et al., 1995). Vitamin A deficiency prevents the

synthesis of all retinoids and is responsible, in quail, for the is only a partial spatial overlap between Sonic hedgehogexpression and retinoic acid distribution at primitive streakabsence of posterior hindbrain (Maden et al., 1996, 1997).

Excess of retinoids also induces a number of malformations stage (Levin et al., 1995). More importantly, it has beenshown that Hensen’s node-inducing activity and its abilityin chicken, mice, frog, and zebra fish (Alles and Sulik, 1990;

Holder and Hill, 1991; Papalopulu et al., 1991). RA applied to synthesize retinoic acid are separate properties (Chen etal., 1996; Selleck et al., 1996). Retinoic acid may thereforeto primitive streak stage chicken embryos induces defi-

ciency in anterior neural structures, open neural tubes, and be only one of several diffusible factors involved in neuralinduction, a process that appears to be sequential and com-situs inversus in heart (Chen and Solursh, 1992). At the

same stage, local release induces secondary axis formation plex (Streit et al., 1997).


AID DB 8872 / 6x3c$$$141 04-03-98 08:01:56 dbal


Cassar-Malek, I., Marchal, S., Altabef, M., Wrutniak, C., Samarut,Our data reveal a striking parallel between RA and T3. BothJ., and Cabello, G. (1994). v-erbA stimulates quail myoblast dif-are enriched in Hensen’s node, which is able to release approx-ferentiation in a T3 independent, cell-specific manner. Oncogeneimately 10 fmol of T3, and 3 fmol of active retinoids (Chen9, 2197–2206.et al., 1992; Hogan et al., 1992). At later stages, we have found

Catala, M., Teillet, M. A., De Robertis, E. M., and Le Douarin,that T3 is secreted only by the caudal half of the embryo.N. M. (1996). A spinal cord fate map in the avian embryo: While

Studies that used indicator transgenic mice led to similar con- regressing, Hensen’s node lays down the notochord and floorclusions for RA (Tsou et al., 1994; Zimmer and Zimmer, plate thus joining the spinal cord lateral walls. Development 122,1992). The correlation between T3 and RA distribution may 2599–2610.therefore be maintained after neurulation. Transthyretin, Chatterjee, V. K. K., and Tata, J. R. (1992). Thyroid hormone recep-which is a plasma protein recognized by a specific membrane tors and their role in development. Cancer Surv. 147–166.

Chen, Y., Huang, L., Russo, A. F., and Solursh, M. (1992). Retinoicreceptor during oocyte growth (Vieira et al., 1995) is able toacid is enriched in Hensen’s node and is developmentally regu-transport in egg yolk T3, RA, and their respective precursors,lated in the early embryo. Proc. Natl. Acad. Sci. USA 89, 10056–thyroxine and retinol (Ingenbleek and Young, 1994). If these10059.molecules are released from egg yolk by area opaca cells, they

Chen, Y. P., D., D., Kostetskii, I., and Zile, M. H. (1996). Hensen’smay be accumulated by presumptive Hensen’s node cells be-node from vitamin A-deficient quail embryo induces chick limbfore their migration from the Koller sickle. The apparent corre-bud duplication and retains its normal asymmetric expression of

lation between RA and T3 distribution would therefore result Sonic hedgehog (Shh). Dev. Biol. 173, 256–264.from a common transport pathway. Based on these observa- Chen, Y. P., and Solursh, M. (1992). Comparison of Hensen’s nodetions, and the fact that excess of T3 or RA lead to similar and retinoic acid in secondary axis induction in the early chickanterior neural tube defects, we propose that T3 could be embryo. Dev. Dyn. 195, 142–151.another inducer molecule released by Hensen’s node, and, Forrest, D., Sjoberg, M., and Vennstrom, B. (1990). Contrasting de-

velopmental and tissue-specific expression of a and b thyroidbecause some DNA response elements can transduce both T3hormone receptor genes. EMBO J. 9, 1519–1528.and RA signal, that its function might partially be redundant

Gandrillon, O., Ferrand, N., Michaille, J. J., Roze, L., Zile, M. H.,with retinoids functions. This hypothesis would explain whyand Samarut, J. (1994). c-erbA a/T3R and RARs control commit-in Xenopus, v-erbA, a viral dominant negative version ofment of hematopoietic self-renewing progenitor cells to apop-chicken T3Ra, can reduce the teratogenic effect of RA (Schuhtosis or differentiation and are antagonized by the v-erbA onco-et al., 1993). Similar cross-talks between T3 and RA pathwaysgene. Oncogene 9, 749–58.have also been shown in mouse embryonic stem cells (Lee

Hamburger, V., and Hamilton, H. L. (1951). A series of normal stageset al., 1994). The late divergence during chordate evolution in development of the chick embryo. J. Morphol. 88, 49–92.between retinoic acid and thyroid hormone receptors may Hogan, B., Thaller, C., and Eichele, G. (1992). Evidence thattherefore reflect similarities in the functions of their ligands Hensen’s node is a site of retinoic acid synthesis. Nature 359,at early stages of development. 237–238.

Holder, N., and Hill, J. (1991). Retinoic acid modifies developmentof the midbrain–hindbrain border and affects cranial ganglion

ACKNOWLEDGMENTS formation in zebrafish embryos. Development 113, 1159–1170.Horlein, A. J., Naar, A. M., Heinzel,T., Torchia, J., Gloss, B., Kuro-

kawa, R., Ryan, A., Kamei, Y., Soderstrom,M., Glass, C. K., andWe thank M. Busslinger and J. V. Bonventre for the kind gift ofRosenfeld, M. G. (1995). Ligand-independent repression by theplasmids, and G. Joly for scanning electron microscopy. This workthyroid hormone receptor mediated by a nuclear co-repressor.was supported by the Centre National de la Recherche Scientifique,Nature 377, 397–403.Institut National de Recherche Agronomique, Association pour la

Ingenbleek, Y., and Young, V. (1994). Transthyrenin (prealbumin)Recherche contre le Cancer, and Ligue Departementale de l’Yonnein health and disease: Nutritional implications. Annu. Rev. Nutr.contre le Cancer.14, 495–533.

Izpisua-Belmonte, J. C., De Robertis, E. M., Storey, K. G., and Stern,C. D. (1993). The homeobox gene goosecoid and the origin ofREFERENCESorganizer cells in the early chick blastoderm. Cell 74, 645–659.

Katz, D., and Lazar, M. A. (1993). Dominant negative activity ofAlles, A. J., and Sulik, K. K. (1990). Retinoic acid induced spinaan endogenous thyroid hormone receptor variant (alpha-2) is duebifida: Evidence for a pathogenic mechanism. Development 108,to competition for binding sites on target genes. J. Biol. Chem.73–81.266, 20904–20910.Banker, D. E., and Eisenman, R. N. (1993). Thyroid hormone recep-

Kurokawa, R., Soderstrom, M., Horlein, A. J., Halachmi, S., Brown,tor can modulate retinoic acid-mediated axis formation in frogM., Rosenfeld, M. G., and Glass, C. K. (1995). Polarity-specificembryogenesis. Mol. Cell. Biol. 13, 7540–7552.activities of retinoic acid receptors determined by a co-repressor.Bernal, J., and Nunez, J. (1995). Thyroid hormones and brain devel-Nature 377, 451–453.opment. Eur. J. Endocrinol. 133, 390–398.

Lebel, J. M., Dussault, J. H., and Puymirat, J. (1994). OverexpressionBraselman, S., Graninger, P., and Busslinger, M. (1993). A selectiveof the beta 1 thyroid receptor induces differentiation in neuro-transcriptional induction system for mammalian cells based on2a cells. Proc. Natl. Acad. Sci. USA 91, 2644–2648.gal4–estrogen receptor fusion proteins. Proc. Natl. Acad. Sci.

Lee, L. R., Mortensen, R. M., Larson, C. A., and Brent, G. A. (1994).USA 90, 1657–1661.Thyroid hormone receptor-a inhibits retinoic acid-responsiveBumcrot, D. A., and Mac Mahon, . P. (1995). Sonic signals somites.

Curr. Biol. 5, 612–614. gene expression and modulates retinoic acid-stimulated neural


AID DB 8872 / 6x3c$$$141 04-03-98 08:01:56 dbal


differentiation in mouse embryonic stem cells. Mol. Endocrinol. Hedgehog mediates the polarizing activity of ZPA. Cell 75, 1401–1416.8, 746–756.

Sachs, L., de Luze, A., Lebrun, J. J., Kelly, P. A., and Demeneix,Levin, M., Johnson, R. L., Stern, C., Kuehn, M., and Tabin, C.B. A. (1996). Use of heterologous DNA-based gene transfer to(1995). A molecular pathway determining left–right asymmetryfollow physiological, T3-dependent regulation of myosin heavyin chick embryogenesis. Cell 82, 803–814.chain genes in Xenopus tadpoles. Endocrinology 137, 2191–2194.Lezoualc’h, F., Hassan, A. H. S., Giraud, P., Loeffler, J. P., Lee, S. L.,

Safer, J. D., Langlois, M. F., Cohen, R., Monden, T., John-Hope,and Demeneix, B. A.(1992) Assignment of the b-thyroid hormoneD., Madurra, J., Hollenberg, A. N., and Wondisford, F. E. (1996).receptor to 3,5,3*-triiodothyronine-dependent inhibition of tran-Isoform variable action among thyroid hormone receptor mu-scription from the thyrotropin-releasing hormone promoter intants provides insight into pituitary resistance to thyroid hor-chick hypothalamic neurons. Mol. Endocrinol. 6, 1797–1804.mone. Mol. Endocrinol. 11, 16–26.Lezoualc’h, F., Seugnet, I., Monnier, A. L., Ghysdael, J., Behr, J. P.,

Schuh, T. J., Hall, B. L., Kraft, J. C., Privalsky, M. L., and Kimelman,and Demeneix, B. A. (1995). Inhibition of neurogenic precursorD. (1993). v-erbA and citral reduce the teratogenic effects of all-proliferation by antisense a thyroid hormone receptor oligonucle-trans retinoic acid and retinol, respectively, in Xenopus em-otides. J. Biol. Chem. 270, 12100–12108.bryogenesis. Development 119, 785–798.Maden, M., Gale, E., Kostetskii, I., and Zile, M. (1996). Vitamin-A

Selleck, M. A., Hayamizu, T. F., Ohsugi, K., Bronner-Fraser, M.,deficient quail embryos have half a hindbrain and other neuraland Bryant, S. V. (1996). Digit duplication by Hensen’s node anddefect. Curr. Biol. 6, 417–426.notochord involves the expression of shh but not RAR-b2. Dev.Maden, M., Graham, A., Gale, E., Rollinson, C., and Zile, M. (1997).Biol. 173, 318–326.Positional apoptosis during vertebrate CNS development in the

Sjoberg, M., Vennstrom, B., and Forrest, D. (1992). Thyroid hor-absence of endogenous retinoids. Development 124, 2799–2805.mone receptors in chick retinal development: differential expres-Mangelsdorf, D. J., and Evans, R. (1995). The RXR heterodimerssion of mRNA for a and N-terminal variant receptors.and orphan receptors. Cell 83, 841–850.

Stolow, M. A., and Shi, Y. B. (1995). Xenopus sonic hedgehog as aMangelsdorf, D. J., Thummel, C., Beato, M., Herrlich, P., Schutz,potential morphogen during embryogenesis and thyroid-hor-G., Umesono,K. , Blumberg, B., Kastner, P., Mark, M., Chambon,mone dependent metamorphosis. Nucleic Acids Res. 23, 2555–P., and Evans, R. M. (1995). The nuclear receptor superfamily:2562.The second decade. Cell 83, 835–839.

Streit, A., Sockanathan, S., Perez, L., Rex, M., Scotting, P. J., Sharpe,Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982). ‘‘MolecularP. T., Lovell-Badge, R., and Stern, C. (1997). Preventing the lossCloning: A Laboratory Manual.’’ Cold Spring Harbor Laboratoryof competence for neural induction: HGF/SF,L5 and Sox2. Devel-Press, Cold Spring Harbor, NY.opment 124, 1191–1202.

Mitrani, E., and Eyal-Giladi, H. (1982). Cells from early chick em-Tini, M., Tsui, L. C., and Giguere, V. (1994). Heterodimeric interac-

bryos in culture. Differentiation 21, 56–61. tion of the retinoic acid and thyroid hormone receptors in tran-Mitrani, E., and Shimoni, Y. (1989). Retinoic acid inhibits growth scriptional regulation on the gamma F-crystallin everted retinoic

in agarose of early chick embryonic cells and may be involved acid response element. Mol Endocrinol 8, 1494–1506.in regulation of axis formation. Development 107, 275–280. Tsou, H. C., Si, S. P., Lee, X., Gonzalez-Serva, A., and Peacocke,

Modak, S. P., Ghatpande, S. K., Rane, R. K., and Mulherkar, L. M. (1994). A b2-RARE-lacZ transgene identifies retinoic acid-(1993). Caudalization by retinoic acid is correlated with inhibi- mediated transcriptional activation in distinct cutaneous sites.tion of cell population growth and expansion of chick blasto- Exp. Cell. Res. 214, 27–34.derms cultured in vitro. Int. J. Dev. Biol. 37, 601–607. Twal, W., Roze, L., and Zile, M. H. (1995). Anti-retinoic acid mono-

Nunez, J. (1984). Effects of thyroid hormones during brain differen- clonal antibody localizes all trans-retinoic acid in target cells andtiation. Mol. Cell. Endocrinol. 37, 125–132. blocks normal development in early quail embryo. Dev. Biol.

Old, R., Jones, E. A., Sweeney, G., and Smith, D. P. (1992). Preco- 168, 225–234.cious synthesis of a thyroid hormone receptor in Xenopus em- Umesono, K., Giguere, V., Glass, C. K., Rosenfeld, M. G., and Ev-bryos causes hormone-dependent developmental abnormalities. ans, R. M. (1988). Retinoic acid and thyroid hormone induce geneDev. Biol. 201, 312–321. expression through a common responsive element. Nature 336,

Over, M., Medium, A., and Scrambled, E. (1980). Injection of cell 262–265.supernatants in the subgerminal cavity of chicken embryos can Vieira, A. V., Sanders, E. J., and Schneider, W. J. (1995). Transport ofprevent axis formation. Br. J. Nutr. 3, 154–155. serum transthyrenin into chicken oocytes. A receptor-mediated

Pain, B., Clark, M. E., Shen, M., Nakazama, H., Sakurai, M., Sa- mechanism. J. Biol.Chem. 270, 2952–2956.Waddington, C. H. (1936). Organizers in mammalian development.marut, J., and Etches, R. (1996). Long-term culture and characteri-

Nature 138, 125.sation of avian embryonic stem cells with multiple morphogene-Witzgall, R., O’Leary, E., and Bonventre, J. V. (1994). A mammaliantic potentialities. Development 122, 2339–2348.

expression vector for the expression of GAL4 fusion proteins withPapalopulu, N., Clarke, J. D. W., Bradley, L., Wilkinson, D., Krum-an epitope tag and histidine tail. Anal. Biochem. 223, 291–298.lauf, R., and Holder, N. (1991). Retinoic acid causes abnormal

Zavacki, A. M., Harney, J. W., Brent, G. A., and Larsen, P. R. (1993).development and segmental patterning of the anterior hindbrainDominant negative inhibition by mutant thyroid hormone recep-in Xenopus embryos. Development 113, 1145–1158.tors is thyroid hormone response element and receptor isoformPfahl, M. (1993). Nuclear receptor/AP-1 interaction. Endocrin. Rev.specific. Mol. Endocrinol. 7, 1319–1330.14, 651–658.

Zimmer, A., and Zimmer, A. (1992). Induction of RARb-lacZPrati, M., Calvo, R., and Morreale de Escobar, G. (1991). L-Thyrox-transgene by retinoic acid reflects the neuromeric organizationine and 3,5,3*-triiodothyronine concentrations in the chicken eggof the central nervous system. Development 116, 977–983.and in the embryo before and after the onset of thyroid function.

Endocrinology 130, 2651–2659. Received for publication September 8, 1997Riddle, R. D., Johnson, R. L., Laufer, E., and Tabin, C. (1994). Sonic Accepted February 2, 1998


AID DB 8872 / 6x3c$$$141 04-03-98 08:01:56 dbal

Documents

Involvement of Thyroid Hormone and Its & Receptor in Avian Neurulation · 2017-02-25 · neurulation a marked increase was observed in neurectoderm. A reporter cell line was then