Determination of the Optimal Time andDosage of All-Trans Retinoic Acid forInduction of Murine ExencephalyNAOHIKO KUNO,1,2 KENJI KADOMATSU,1 AND TAKASHI MURAMATSU1*1Department of Biochemistry, Nagoya University School of Medicine, Showa-ku, Nagoya 466, Japan2Department of Obstetrics and Gynecology, Nagoya University School of Medicine, Showa-ku, Nagoya 466, Japan

ABSTRACT Murine exencephaly correspondsto human anencephaly, and provides a model forstudying the mechanism of development of the centralnervous system. A system which induces exencephalyat an extremely high rate is required in order to examineembryos, before the stage of neural tube closure, assamples of future exencephaly. Herein, we report on asystem which is close enough to the best conditions forinduction of this malformation, involving ICR mice andall-trans retinoic acid. The intraperitoneal administrationof 30 mg/kg of all-trans retinoic acid at 03:00 hr on day8 (copulatory plug, day 0) induced exencephaly in81.6% of live embryos, as evaluated on day 10, with a41.7% embryonic death rate. Earlier administration(more than 3 hr before) greatly increased the rate ofembryonic death, whereas later administration resultedin a reduction in the rate of exencephaly. These findingssuggest that a specific time during early developmentis crucial for neural tube closure, and provide a systemwhich may facilitate the study of neural developmentand the pathophysiology of human anencephaly. Teratol-ogy 60:63–67, 1999. r 1999 Wiley-Liss, Inc.

Anencephaly is one of the most common humanmalformations of the central nervous system, and therelationship between its occurrence and maladministra-tion of vitamin A derivatives has been proposed (Geelenet al., ’80; Rosa et al., ’94). Murine exencephaly, which isinducible by all-trans retinoic acid (RA), provides anideal model for investigating the mechanism of develop-ment of the nervous system, and the pathophysiology ofhuman anencephaly. In most studies, RA was adminis-tered between days 7–9, but murine exencephaly isnot induced at a high frequency in live embryos, i.e.,14% in SWV/Bc mice and 22% in ICR/Bc (Tom et al.,’91), 6.6% in CD-1 (Rasco and Hood, ’95a), 5–28% inCD-1 (Sulik et al., ’95), 13–26% in SWV (Kapron-Brasand Trasler, ’84), and 18.6% in NMRI (Elmazar et al.,’96). In the case of SELH/Bc mice, which is a straingenetically liable to exencephaly, the rate is 53% (Tomet al., ’91).

To examine future exencephalic embryos before thestage of anterior neural tube closure, it is essential todetermine the fate of each RA-treated embryo: itsanterior neuropore will either close or remain open.

However, this seems to be practically difficult to achieve,because the fates of neural tissues are morphologicallyindistinguishable at such an early stage. Alternatively,one could induce exencephaly at an extremely highrate. If such a system were established, we couldconsider more embryos, treated with RA at stagesearlier than neural tube closure, as reliable samples foranalyzing the mechanism of development of murineexencephaly. This necessity prompted us to determinethe optimal conditions for induction of this malforma-tion.

MATERIALS AND METHODS

ICR mice, purchased from SLC (Shizuoka, Japan),were kept in an air-conditioned animal room at atemperature of 23–24°C with light control 12 hr dark(21:00–09:00 hr) and 12 hr light (09:00–21:00 hr). Theywere fed tap water and laboratory mouse chow (CE-2,CLEA Japan Co.) ad libitum. Adult females were cagedwith males in pairs overnight, and the animals with avaginal plug the next morning were considered to be atday 0 of pregnancy.

All-trans retinoic acid (RA) and sunflower seed oilwere purchased from Sigma (St. Louis, MO). RA wassuspended in 600 µl of sunflower seed oil and adminis-tered to the pregnant mice intraperitoneally. The dos-age (0–70 mg/kg) of RA was determined according to theweight on the day of administration. The suspensionwas given only once to each animal between days 5–9 ofpregnancy at a selected time. To the control mice, 600 µlof sunflower seed oil alone were given.

The treated females were sacrificed on either day 10or 16 by cervical dislocation, and embryos were ob-served under a dissecting microscope. The numbers ofimplantations, dead embryos, and live implants withand without exencephaly were recorded.

Grant sponsor: Ministry of Education, Science, Sports, and Culture ofJapan.

*Correspondence to: Takashi Muramatsu, Department of Biochemis-try, Nagoya University School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466, Japan. E-mail: tmurama@tsuru.med.nagoya-u.ac.jp

Received 26 February 1998; Accepted 18 February 1999

TERATOLOGY 60:63–67 (1999)

r 1999 WILEY-LISS, INC.

RESULTS

Pregnant mice were treated with different combina-tions of timing and doses of RA, and the outcome wasevaluated on day 10 (Fig. 1). The results are listed inTable 1 (total number of mice was 240). The rate of

exencephalic embryos among live embryos was low onday 5, and then increased until 00:00 hr or 03:00 hr onday 8, when most of the live embryos were exence-phalic. After this period, the rate of exencephaly de-creased. In contrast, concerning the embryonic death

Fig. 1. Mice were injected with RA (30 mg/kg) or vehicle only(sunflower seed oil) at 03:00 hr on day 8 postcoitum, and sacrificed at12:00 hr on day 10 postcoitum. Embryos were fixed with 10%formaldehyde, stained with 0.5% Nileblue sulfate in water for 3 sec,and then washed with phosphate-buffered saline. A: Lateral view of a

control embryo. B: Frontal view of A. C: Lateral view of an RA-treatedexencephalic embryo. Arrow indicates edge of the extroverted neuraltube. D: Frontal view of C. In B and D, the dorsal side of the embryofaces upward (scale bars, 200 µm).

64 N. KUNO ET AL.

rate, the dose inducing total embryonic loss on day 5was 70 mg/kg, and the dose giving the same outcomehad decreased to 40 mg/kg at 12:00 hr on day 6, 40mg/kg at 12:00 hr on day 7, 30 mg/kg at 18:00 hr on day7, and 20 mg/kg at 21:00 hr on day 7. Finally, mostembryos died with 20 mg/kg of RA given at 00:00 hr onday 8, and thereafter, the rate of abortion rapidlydecreased. As a result, the best time of RA administra-tion to acquire exencephalic embryos most efficientlywas regarded as 03:00 hr on day 8: RA administrationat a dose of 30 mg/kg at 03:00 hr on day 8 inducedexencephaly in 81.6% of the live embryos, with anembryonic death rate of 41.7% (Fig. 1). As to the dosageof RA, 30, 40, and 50 mg/kg seemed preferable.

Although development of many dead embryos ceasedtoo early to proceed through neural closure, we alsoobserved exencephalic dead embryos macerated in theuterus on day 10, e.g., 2 of 35 dead embryos and 2 of 24dead embryos in the cases treated with 20 and 30 mg/kgof RA, respectively, at 03:00 hr on day 8. They werealmost at the same developmental stage as were theirlive littermates. This indicates that these dead embryosdeveloped almost until day 10, but neural tube closurewas not completed.

To identify the developmental stage at which theinfluence of the administered RA was most effective,untreated embryos were collected at 03:00 hr, 06:00 hr,and 09:00 hr on day 8. We examined three litters at

TABLE 1. Results of retinoic acid treatment in ICR mice, as analyzed on day 10of gestation*

DayTime(hr)

Dose(mg/kg)

Number of Rate of (%)

Li Im L.E. D.E. Ex E.D. Ex/L.E.

5 12 0 1 7 5 2 0 28.6 0.030 7 64 55 9 4 14.1 7.350 6 84 64 20 9 23.8 14.160 5 67 37 30 13 44.8 35.170 3 54 0 54 0 100.0

6 12 0 1 17 17 0 0 0.0 0.030 7 101 75 26 10 25.7 13.340 5 63 0 63 0 100.050 4 56 0 56 0 100.070 4 53 0 53 0 100.0

7 12 0 1 14 12 2 0 14.3 0.010 9 101 34 67 13 66.3 38.220 15 190 26 164 1 86.3 3.830 15 200 12 188 9 94.0 75.040 9 123 0 123 0 100.050 8 106 0 106 0 100.0

18 20 3 47 7 40 6 85.1 85.730 3 45 0 45 0 100.040 2 18 0 18 0 100.0

21 20 2 25 0 25 0 100.030 3 38 0 38 0 100.0

8 0 20 4 56 1 55 1 98.2 100.030 5 75 13 62 12 82.7 92.340 6 71 0 71 0 100.050 3 39 0 39 0 100.0

3 20 5 51 27 24 19 47.1 70.430 6 84 49 35 40 41.7 81.640 5 72 20 52 17 72.2 85.050 3 42 18 24 17 57.1 94.4

6 20 5 68 35 33 9 48.5 25.730 5 47 32 15 13 31.9 40.640 5 54 28 26 16 48.1 57.150 1 12 0 12 0 100.0

9 20 2 19 16 3 2 15.8 12.530 4 56 48 8 11 14.3 22.940 9 115 46 69 10 60.0 21.750 8 78 60 18 26 23.1 43.3

12 0 1 14 14 0 0 0.0 0.010 5 75 69 6 4 8.0 5.820 4 41 29 12 0 29.3 0.030 8 104 93 11 10 10.6 10.840 8 115 92 23 15 20.0 16.350 9 110 84 26 14 23.6 16.760 12 149 100 49 5 32.9 5.0

9 12 50 4 48 44 4 0 8.3 0.0

*D.E., dead embryos; E.D., embryonic death; Ex, exencephaly; Im, implantations; L.E., liveembryos; Li, litters.

INDUCTION OF MURINE EXENCEPHALY BY RA 65

each time and obtained 30–40 embryos in each group.Although inter- and intralitter variations of developmen-tal stages were observed, the majority of embryos(60–70%) in each group exhibited an external appear-ance as follows. The embryos were just at the beginningof the cranial neural fold elevation at 03:00 hr, and nosomites were observed (data not shown). The caudalneural fold was not clear. At 06:00 hr, the neural foldwas more elevated, and 2 or 3 somites were recognized(data not shown). At 09:00 hr, the cranial neural foldtowered, and the caudal neural fold was distinctlyidentifiable (data not shown). The number of somiteswas 5 or 6 at this point. However, we also observedsome range of variation of the developmental stages ineach group. For example, a few embryos showed 6somites at 03:00 hr, whereas an embryo with no somiteswas also detected at 09:00 hr.

We further studied the remote effect of retinoic acidadministration. Thirty mg/kg of RA were given to 5pregnant mice at 03:00 hr on day 8, and the embryoswere examined on day 16 (Table 2). The embryonicdeath rate against total implantation was 52.1%, andthe rate of exencephalic embryos against live embryoswas 68.6%. As the latter value was slightly lower thanthat on day 10 (Table 1), some exencephalic embryosmay have expired between days 10–16. However, mostaffected embryos may remain alive at this mid- andlate-gestational period.

DISCUSSION

In the present study, we demonstrated that exen-cephaly can be induced at a rate of more than 80% oflive embryos on day 10 with a relatively low rate ofembryonic death. Taking 40% as the embryonic deathrate into account, about 50% of embryos were expectedto be ‘‘exencephalic embryos in the future’’ on day 8 or8.5. In addition, we observed exencephalic dead em-bryos on day 10, indicating that the rate of ‘‘exence-phalic embryos in the future’’ could be greater. Bycomparing with results previously reported for theinduction of exencephaly (Kapron-Bras and Trasler, ’84;Tom et al., ’91; Rasco and Hood, ’95a; Sulik et al., ’95;Elmazar et al., ’96), the present rate of exencephaly vs.live embryos is extremely high, and that of exencephalyvs. total embryos is similar to that in SELH/Bc mice,which are genetically liable to exencephaly (Tom et al.,’91). The present system would facilitate study on themechanism of neural tube closure, because at least halfof the embryos before day 9, when neural tube closure

normally occurs, can be analyzed as future exence-phalic embryos.

Vitamin A is a lipophilic reagent, and RA, a deriva-tive of vitamin A, might be stored in fat tissue and actfor a long time through slow release. However, our dataruled out this idea, and suggested the brief action of RAwhen it is administered intraperitoneally, being consis-tent with the previous report (Satre and Kochhar, ’89).We injected RA at intervals of 3 hr, and, according to thetime of administration, the reaction dramaticallychanged. If it acts for a long period, a change in reactionshould be obscure. In addition, repetitive administra-tion of 13-cis RA, even though the total dosage was thesame, was reported to increase the induction rate ofmalformation (Jiang et al., ’94). It was also reportedthat the maximum RA concentration in embryonictissue is attained 3 hr after administration, and thattissue RA level returns to pretreatment level within 8hr (Satre and Kochhar, ’89). These data indicate thatthe effect of each RA administration lasts for a shortperiod. Taking the rapid growth of mice and shortduration of the effect of RA into consideration, we gaveRA at intervals of 3 hr throughout the sensitive period,and determined the optimal conditions. The presentoptimal conditions turned out to be much better thanthose previously reported, which were determinedthrough experiments involving the administration ofRA at 6-hr intervals (Webster et al., ’86; Rasco andHood, ’95b). Because of the rapid embryonic develop-ment and short duration of RA action, the presentfindings probably indicate that finer experiments interms of time are necessary to determine the optimalconditions for the establishment of animal models inearly stages of development with RA.

We should also note that the present study demon-strated the importance of fine setting of time for RAadministration, but did not necessarily provide the bestconditions applicable to strains other than ICR. Be-cause of strain differences, those who use other mousestrains should find the best conditions by fine time andRA dose settings.

We serially observed the external appearance ofnormal embryos at 03:00, 06:00, and 09:00 hr on day 8and identified the developmental stages of embryoswhen they were most susceptible to RA and displayedexencephaly. During this period, the elevation of theneural folds and the appearance of somites were ob-served. Taking into account the previous report (Satreand Kochhar, ’89) showing the maximum tissue RA

TABLE 2. Results of retinoic acid treatment in ICR mice, as analyzed on day 16of gestation*

DayTime(hr)

Dose(mg/kg)

Number of Rate of (%)

Li Im L.E. D.E. Ex E.D. Ex/L.E.

8 3 30 5 73 35 38 24 52.1 68.6

*D.E., dead embryos; E.D., embryonic death; Ex, exencephaly; Im, implantations; L.E., liveembryos; Li, litters.

66 N. KUNO ET AL.

level to be at 3 hr after administration, the tissue RAconcentration may reach its peak around 06:00 hr if RAis administered under the best conditions for inducingexencephaly, namely at 03:00 hr on day 8. Since themajority of embryos were at the 2–3-somite stage at06:00 hr on day 8, we consider this developmental stageto be the appropriate representation for 06:00 hr on day8 and the most susceptible target for RA in the induc-tion of exencephaly. RA has been reported to exert itsaction on neurulation through its morphogenic abilityto change the shape of neuroepithelial cells or its abilityto induce excessive cell death (Geelen et al., ’80; Sulik etal., ’88; Ahuja et al., ’97). RA also induces posterioriza-tion, as revealed by expression changes of Hox genes(Simeone et al., ’95; Papalopulu and Kintner, ’96). Inaddition, it should be noted that the key moleculesrelevant to RA-induced exencephaly would be retinoicacid receptor (RAR)g, sonic hedgehog, and AP2. RARgknockout mice are resistant to RA-induced embryole-thality, craniofacial malformations, and neural tubedefects if RA is administered at day 7.3, indicating thatat least some of the signals related to RA-inducedmalformation are mediated by this receptor (Iulianellaand Lohnes, ’97). Sonic hedgehog is downregulated byteratogenic doses of RA for craniofacial malformations(Helms et al., ’97). It is of particular note that mice nullfor the RA-inducible transcription factor AP2 geneexhibit exencephaly, craniofacial defects, and thoraco-abdominoschisis (Zhang et al., ’96). The exencephaly inAP2 knockout mice is manifestated at day 9.5, beingreminiscent of the present study. RA administeredaround day 8 may bind to RARg and induce disorga-nized expressions of sonic hedgehog and AP2, leading tomalformations, including exencephaly.

We also examined embryos on day 16 after injectionof 30 mg/kg of RA at 03:00 hr on day 8, and found thatnot many exencephalic embryos died between days10–16. Being exencephalic throughout the developmen-tal period may have an adverse effect on other organs.The importance of the pituitary endocrine system forthe development of gonadal organs (Rabinovici andJaffe, ’90), and the role of the brain above the medullaoblongata in fetal breathing movements, have beenreported (Kurauchi et al., ’95). Using the present sys-tem, it would be interesting to clarify the importance ofnormal development of the brain as compared to that ofother organs.

LITERATURE CITEDAhuja HS, James W, Zakeri Z. 1997. Rescue of the limb deformity in

hammertoe mutant mice by retinoic acid-induced cell death. DevDynam 208:466–481.

Elmazar MMA, Reichert U, Shroot B, Nau H. 1996. Pattern ofretinoid-induced teratogenic effects: possible relationship with rela-tive selectivity for nuclear retinoid receptors RARa, RARb, RARg.Teratology 53:158–167.

Geelen JAG, Langman J, Lowdon JD. 1980. The influence of excessvitamin A on neural tube closure in the mouse embryo. AnatEmbryol (Berl) 159:223–234.

Helms JA, Kim CH, Hu D, Minkoff R, Thaller C, Eichele G. 1997. Sonichedgehog participates in craniofacial morphogenesis and is down-regulated by teratogenic doses of retinoic acid. Dev Biol 187:25–35.

Iulianella A, Lohnes D. 1997. Contribution of retinoic acid receptor g toretinoid-induced craniofacial and axial defects. Dev Dynam 209:92–104.

Jiang H, Gyda M III, Harnish DC, Chandraratna RA, Soprano KJ,Kochhar DM, Soprano DR. 1994. Teratogenesis by retinoic acidanalogs positively correlates with elevation of retinoic acid receptor-beta 2 mRNA levels in treated embryos. Teratology 50:38–43.

Kapron-Bras CM, Trasler DG. 1984. Gene-teratogen interaction andits morphological basis in retinoic acid-induced mouse spina bifida.Teratology 30:143–150.

Kurauchi O, Ohno Y, Mizutani S, Tomoda Y. 1995. Longitudinalmonitoring of fetal behavior in twins when one is anencephalic.Obstet Gynecol 86:672–674.

Papalopulu N, Kintner C. 1996. A posteriorising factor, retinoic acid,reveals that anteroposterior patterning controls the timing ofneuronal differentiation in Xenopus neuroectoderm. Development122:3409–3418.

Rabinovici J, Jaffe RB. 1990. Development and regulation of growthand differentiated function in human and subhuman primate fetalgonads. Endocr Rev 11:532–557.

Rasco JF, Hood RD. 1995a. Maternal restraint stress-enhanced terato-genicity of all-trans retinoic acid in CD-1 mice. Teratology 51:57–62.

Rasco JF, Hood RD. 1995b. Enhancement of the teratogenicity ofall-trans-retinoic acid by maternal restraint stress in mice as afunction of treatment timing. Teratology 51:63–70.

Rosa F, Piazza-Hepp T, Goetsch R. 1994. Holoprosencephaly with firsttrimester topical tretinoin. Teratology 49:418–419.

Satre MA, Kochhar DM. 1989. Elevations in the endogenous levels ofthe putative morphogen retinoic acid in embryonic mouse limb-budsassociated withlimb dysmorphogenesis. Dev Biol 133:529–536.

Simeone A, Avantaggiato V, Moroni MC, Mavilio F, Arra C, Cotelli F,Nigro V, Acampora D. 1995. Retinoic acid induces stage-specificantero-posterior transformation of rostral central nervous system.Mech Dev 51:83–98.

Sulik KK, Cook CS, Webster WS. 1988. Teratogens and craniofacialmalformations: relationships to cell death. Development 103:213–232.

Sulik KK, Dehart DB, Rogers JM, Chernoff N. 1995. Teratogenicity oflow doses of all-trans retinoic acid in presomite mouse embryos.Teratology 51:398–403.

Tom C, Juriloff DM, Harris MJ. 1991. Studies of the effect of retinoicacid on anterior neural tube closure in mice genetically liable toexencephaly. Teratology 43:24–40.

Webster WS, Johnston MC, Lammer EJ, Sulik KK. 1986. Isotretinoinembryopathy and the cranial neural crest: an in vivo and in vitrostudy. J Craniofac Genet Dev Biol 6:211–222.

Zhang J, Hagopian-Donaldson S, Serbedzija G, Elsemore J, Plehn-Dujowich D, McMahon AP, Flavell RA, Williams T. 1996. Neuraltube, skeletal and body wall defects in mice lacking transcriptionfactor AP-2. Nature 381:238–241.

INDUCTION OF MURINE EXENCEPHALY BY RA 67