Neurotoxicology and Teratology 22 (2000) 541–545

0892-0362/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved.PII: S 0 8 9 2 - 0 3 6 2 ( 0 0 ) 0 0 0 7 2 - 6

Retinoic acid exposure on gestational days 11 to 13 impairsswallowing in rat offspring

R. Robert Holson

a,

*, Jane Adams

b

, Sherry A. Ferguson

c

, Frank M. Scalzo

d

a

Department of Psychology, New Mexico Tech, Socorro, NM 87801, USA

b

Department of Psychology, University of Massachusetts, Boston Harbor Campus, Boston, MA 02125, USA

c

Division of Neurotoxicology, National Center for Toxicological Research, Jefferson, AR 72079, USA

d

Department of Pediatrics, Arkansas Children’s Hospital Research Institute, University of Arkansas for Medical Sciences, Little Rock, AR 72202, USA

Received 22 October 1999; accepted 14 January 2000

Abstract

We have previously reported that exposure to 10 mg/kg of all-

trans

-retinoic acid (RA) daily on the 11th, 12th, and 13th days of rat ges-tation is lethal to all fetuses so exposed, due to an inability to suckle [R.R. Holson et al., Neurotoxicol Teratol 19 (1997) 347–353]. Becausethis lethal RA effect could be due to any of a variety of causes, from olfactory problems in locating the nipple to a motor problem in suckingor swallowing, we performed the following experiment. Albino dams were exposed to 10-mg/kg RA or vehicle daily over gestational days(GDs) 11 to 13. On the afternoon of GD 21 all pups were delivered by c-section. Tongue cannulae were inserted into the oral cavity of theseoffspring, and used to infuse a solution of condensed milk directly into the mouth. During and after each of four infusions, the behavioralresponse to the infusion (typically rolling and curling) was recorded. Controls responded well to this procedure, typically swallowing allmilk so infused. In contrast, almost no RA-exposed neonates were able to swallow milk infused into the oral cavity. In such cases the milksimply dribbled out of the mouth, while the stomach was found to be empty at autopsy. However, the RA-treated animals did seem awarethat milk was entering their mouths, because they showed a normal behavioral response to milk infusion. We conclude that GD 11–13 re-tinoid lethality is due to motor not sensory problems in the control of swallowing. © 2000 Elsevier Science Inc. All rights reserved.

Keywords:

Retinoids; All-

trans

-retinoic acid; Swallowing; Cranial nerves

1. Introduction

It is clear that gestational retinoid exposure can causehigh neonatal lethality in rats [4,14,24,27]. The nature ofthis lethality has been little investigated outside of a seriesof studies that attributed deaths from late gestational expo-sure to problems with lung development and breathing [24–27]. Recently we investigated this phenomenon further. Wefound that the gestational day (GD) 11 to 13 period was es-pecially sensitive to such lethality, with a 10-mg/kg dailyretinoic acid (RA) dose inevitably killing the entire litterwithin 48 h following normal vaginal delivery. Such deathstook two forms: some neonates were found intact but deadin the cage shortly after delivery, while others lived for upto 48 h, but never had milk in their stomachs [14].

To better understand this lethality, we conducted furtherexperiments in which pups were delivered by late-gestationcesarean section, then immediately cross-fostered to treat-ment-naive dams. Interactions between control or treatedpups and foster dams were observed for 6 h, and the next

day nipple attachment was also assessed [14]. Late c-sectionrevealed that all fetuses were alive in utero just prior to nor-mal vaginal delivery. However, RA-exposed pups had diffi-culty in initiating breathing, and had to be repeatedly jarredand shaken before sustained breathing commenced [14].This difficulty in initiation of breathing probably accountsfor the neonates found dead shortly after normal vaginal de-livery. Moreover, despite repeated attempts by the fosterdam, no RA-exposed neonate was ever able to nurse, or toattach to the anesthetized foster dam’s nipple. Within 48 hall RA-exposed offspring died without any indication ofsuccessful nursing [14].

This study further demonstrated that the GD 11–13 neo-nates had a fundamental problem in nursing, one not attrib-utable to maternal behavior. It did not reveal the nature ofthis lethal impairment. Neuroanatomical investigation of thebrainstem of these neonates revealed abnormalities in thepontine nucleus, inferior olive, and area postrema, but not inthe hypoglossal nucleus [14]. Combined with the reduction incerebellar weight caused by nonlethal RA doses on GD 11–13 [13,15,16], these findings provide further evidence for anRA-induced impairment of the wider cerebellar system.They do not throw much light on the lethal inability to nurse.

* Corresponding author. Tel.: 505-835-5862; fax: 505-835-5826.

E-mail address

: rholson@nmt.edu (R.R. Holson).

542

R.R. Holson et al. / Neurotoxicology and Teratology 22 (2000) 541–545

One problem in resolving this issue is that a variety ofdeficits can disrupt suckling. For example, an intact olfac-tory system is necessary for nipple location and attachment[7,30,33], and there are suspicions that RA exposure candisrupt olfactory bulb development [3,22,31,36]. Sensorystimuli from the oral cavity are also important in elicitingsucking and swallowing, whereas the motor act of suckingand swallowing depends upon coordinated input from pace-makers in the nucleus of the solitary tract and motor neu-rons in the nucleus ambiguus, the hypoglossal nuclei, andthe trigeminal motor nucleus [17].

As a first step in identifying the nature of this problem, wehave undertaken the experiments reported here. To assessswallowing ability independent of maternal input, olfactorycues, and the sucking reflex, milk was infused directly intothe posterior oral cavity of c-section-derived neonates. Re-sponses to the sensory cues provided by this infusion and theact of swallowing were then evaluated by direct observation.

2. Methods

2.1. Experimental design

Prospective dams were exposed to either 10 mg/kg of RAor vehicle daily over GDs 11–13. Thus there were only twoconditions, litters exposed to in utero retinoids and controls.Consequently parametric treatment effects were evaluatedwith a simple between-groups

t

-test, whereas categorical vari-ables were assessed with the

x

2

test. Because litter was the unitof exposure, one male was evaluated per litter. A total of eightcontrol litters and ten RA-exposed litters were so evaluated.

2.2. Animals and animal husbandry

Rats derived from the albino Crl:COBS CD (SD) BRstrain were housed in 45

3

22

3

20 cm acrylic cages withwood chip bedding. NIH-31 rat feed and filtered tap waterwere available

ad libitum

throughout the study. Cages wereplaced four to a shelf on five-shelf wheeled stainless steelracks. The temperature- and humidity-controlled vivariumin which these animals were housed (23

8

C

6

2

8

C, humidity50%

6

10%) was on a 12-h light-dark cycle (lights off at1800 h). Potential dams (nulliparous females at least 70days old) were date mated by housing overnight with an ex-perienced male breeder in a cage with a wire-mesh bottom.If a copulatory plug was discovered below the cage the nextmorning, females were assumed to be in GD 0 of preg-nancy. Such potentially pregnant females were then as-signed randomly to treatment conditions and housed indi-vidually until c-section on the afternoon of GD 21.Treatments were counterbalanced for rack position, withtwo treated and two control dams housed individually oneach shelf of the rack.

2.3. Retinoid exposure

Potential dams were gavaged once daily, between 0800 and1000 h for 3 consecutive days on GDs 11 through 13. Both

treated and control groups were given 6 ml of sesame oil perkilogram of body weight. For retinoid exposure, 100 mg ofall-

trans

-retinoic acid (Sigma chemical) were suspended in60 ml of sesame oil, just prior to daily gavage. This RA con-centration provided a 10-mg RA dose per kilogram of bodyweight. The suspension was prepared and administered underlow illumination to prevent photodegradation.

2.4. Delivery by cesarean section

Dams were rendered unconscious by brief exposure toCO

2

, then quickly killed by cervical dislocation. Teams ofresearchers rapidly performed a laparotomy and removedthe uterine horns. Each researcher then proceeded to removepups from placental membrane in one of the two horns. Notmore than 6 min elapsed from initial exposure to CO

2

anddelivery of the last pup by this technique. Pups were placedin shallow petrie dishes, lined with paper soaked in physio-logical saline, and placed in an incubator at 30

8

C and 80%humidity. The petrie dishes were repeatedly shaken andtapped until all pups in the litter were breathing spontane-ously. One hour later a single male pup was chosen at ran-dom for cannulation.

2.5. Tongue cannula implantation

Implantation was conducted according to standard tech-niques [10,20]. Briefly, polyethylene tubing (PE-10) wasflared at one end by heating. Curved 0.255 mm wire was in-serted into the base of the jaw, through the digastric muscleand tongue, exiting the top of the tongue at a midline poste-rior location at the margin of the intermolar eminence. Thewire was then pulled out of the mouth, and the nonflaredend of the polyethylene tube was fitted snugly over the wire.Wire and tubing were rubbed with vegetable oil, and thewire was used to pull the tubing back through the tongueand out of the skin of the lower jaw. The flared end of thepolyethylene tubing was pulled flat against the tongue. Fol-lowing cannulation, pups were returned to the incubator.Three hours later, the cannulae were used for infusion of di-luted condensed milk into the oral cavity.

2.6. Milk infusion and behavioral observations

For behavioral assessment of the response to milk infu-sion into the posterior oral cavity, pups were placed on aheating pad (30

8

C), and the entire session was videotaped.A syringe loaded with a 50% aqueous dilution (v/v) of Bor-den’s

®

condensed milk was attached to the cannula, andused for four consecutive 5-

m

l infusions of the diluted milk.Each infusion lasted 5 s, and was followed by 45 s of behav-ioral observations. Interinfusion interval was 2.5 min. Usinga standard behavioral check list [10,20], the observer re-corded which of a range of possible behaviors occurred inthat 45-s period. Behaviors so evaluated included (1) failureto swallow with milk running out of mouth, (2) stretching,(3) rolling and curling, (4) mouth movements, (5) pawtreading, (6) myoclonic jerks, (7) pivoting, and 8) forward

R.R. Holson et al. / Neurotoxicology and Teratology 22 (2000) 541–545

543

movement. Following the final set of observations pupswere killed by decapitation, and the stomach was openedunder a dissecting microscope to determine whether any ofthe infused milk had been swallowed.

3. Results

As in previous experiments, all treated pups were alive atc-section delivery, but required longer to begin spontaneousbreathing. Subjects in both groups responded well to cannula-tion, and the control neonates typically swallowed the entiretyof each infusion. As Table 1 shows, the most common behav-ioral responses to milk infusions (roll, mouth, paw tread)averaged across all four trials did not differ significantly be-tween controls and RA-treated neonates. In contrast, thenumber of trials on which treated pups did not entirely swallowthe 0.2-ml milk infusion was significantly lower, with onlytwo of ten RA-exposed pups ever swallowing. In contrast,only one of ten control pups failed to swallow without drib-bling on at least two of the four trials (

x

2

5

8.10,

p

,

0.005). Not surprisingly, only three of ten retinoid-exposedneonates had any milk in the stomach at autopsy, whileseven of eight controls did so (

x

2

5

5.95,

p

,

0.02; Fig. 1).Further, when the behavioral responses to milk infusionwere pooled (total trials on which pups rolled, mouthed, ortread), there was again no difference between groups [Fig.1;

t

(16)

5

0.46,

p

5

0.65]. Thus the retinoid group appearedto react normally to infusion of milk into the mouth, butwere severely impaired in their ability to swallow.

4. Discussion

These results demonstrate a devastating impairment of theability to swallow following gestational exposure to 10 mg/kg of all-

trans

-retinoic acid daily on GDs 11 through 13. In-terestingly, this deficit did not appear to be sensory in nature,because the RA-exposed animals responded to milk infusionwith normal levels of rolling, paw treading, and mouthing.This impairment in swallowing could account for the fact thatso many neonates exposed to this dosing regime die in thecourse of normal vaginal delivery, and that even after cesar-ean delivery pups require exceptional mechanical stimulationbefore initiating unsupported breathing. Presumably, the ani-mals that die suffocate due to an inability to clear the airways.

Further, lesser impairments may be seen in animals that sur-vive lower retinoid doses at this exposure period. Several lab-oratories have reported abnormal ultrasonic vocalizations insuch animals [1,2,18,19], while children surviving gestationalexposure to Accutane

®

often have problems with gastric re-flux (J. Adams, unpublished data).

The neurological locus of this loss of motor control of theswallowing reflex remains to be determined. In a previousstudy, volume of the hypoglossal nucleus was normal in suchanimals [14], but this finding does not preclude abnormalitiesin cranial nerve XII, or even subtle abnormalities in functionnot accompanied by neuronal cell loss. On the other hand,these treated brains did have abnormalities in the principal nu-cleus of the inferior olive and in the pontine nucleus [14], twomajor cerebellar relay nuclei. Even at lower, nonlethal RAdoses on GDs 11–13, it is also the case that the cerebellumis reliably stunted [13,15,16]. Thus there is strong evidencefor a primary impact of GD 11–13 exposure on the developingcerebellar system. It is possible that impaired cerebellarfunction could contribute to the inability to swallow. A recentelectrophysiological study has implicated the inferior olivein the motor control of tongue movements [35], whereaschemical ablation of the developing inferior olive evidentlyalso produces an inability to eat [37].

Other possibilities must also be considered. In additionto the hypoglossal nucleus, motor output neurons in andnear the nucleus ambiguus and in the trigeminal motor nu-cleus are important in the motor control of swallowing[5,6,17], and these latter regions have yet to be histologi-cally assessed following GD 11–13 exposure to lethal RAdoses. Moreover, neurons in the nucleus of the solitary tract

Table 1Behavioral response to milk infusion

Behavior Control RA

Ingest

a

2.88

6

0.52 0.35

6

0.26Roll 0.94

6

0.27 1.35

6

0.37Mouth 3.63

6

0.26 2.80

6

0.45Paw tread 1.75

6

0.16 1.75

6

0.38

Shown are the number of trials (out of a maximum of four) on which aparticular response occurred,

6

SEM, for eight control and ten pups ex-posed to RA on GDs 11–13.

a

Ingest: dribble

5

0, swallow

5

1.

Fig. 1. Retinoid exposure on GD 11–13 impaired swallowing without alter-ing the behavioral response to oral milk infusion. (Left) Percentage of con-trol and GD 11–13 retinoid-exposed litters found to have no milk in stom-achs after four consecutive infusions. (Right) Mean number of behavioralresponses (roll, mouth, or tread) summed over the four trials 6 SEM forthe eight control (open bar) and ten RA-exposed (closed bar) litters.

544

R.R. Holson et al. / Neurotoxicology and Teratology 22 (2000) 541–545

are important pacemakers for motor neurons actually con-trolling swallowing in nucleus ambiguus [8]. Presumablydamage here could also result in impaired swallowing, butthis possibility is more remote in that this nucleus is also themajor recipient of taste input from the tongue, a sensorysystem that appears to be largely intact in our animals. Inany case, a variety of abnormalities in brainstem are possi-ble, given recent findings concerning the role played by en-dogenous retinoids in determining the appropriate develop-ment of brainstem rhombomeres in regions including thoseimplicated in the motor control of swallowing [9,11,12,21,23,29,34].

In conclusion, we have demonstrated that rats exposed to10 mg/kg of RA on GDs 11–13 suffer a lethal insult to as-yetunidentified motor systems important for the control ofswallowing. This deficit is not accompanied by any majorsensory deficiency, because the animals clearly respond to thepresence of milk in the oral cavity. Further examination of thisproblem is best conducted by injection of retrograde tracerinto the muscles controlling tongue, jaw movements, palateand esophagus [28], but this approach is handicapped by thevery rapid death of animals so treated. Hence studies are nowunderway to determine whether survival can be extended byartificial feeding sufficiently to provide time for retrogradetracers to be transported to the relevant brainstem nuclei.

References

[1] J. Adams, Ultrasonic vocalizations as diagnostic tools in studies ofdevelopmental toxicity: An investigation of the effects of hypervita-minosis A, Neurobehav Toxicol Teratol 4 (1982) 299–304.

[2] J. Adams, Structure-activity and dose-response relationships in theneural and behavioral teratogenesis of retinoids, Neurotoxicol Teratol15 (1993) 193–202.

[3] R.M. Anchan, D.P. Drake, C.F. Haines, E.A. Gerwe, A.S LaMantia,Disruption of local retinoid-mediated gene expression accompaniesabnormal development in the mammalian olfactory pathway, J CompNeurol 379 (1997) 171–184.

[4] V.T. Armenti, E.M. Johnson, Effects of maternal hypervitaminosis Aon perinatal rat lung histology, Biol Neonate 36 (1979) 305–310.

[5] X. Bao, E.B. Wiedner, S.M. Altschuler, Transsynaptic localization ofpharyngeal premotor neurons in rat, Brain Res 696 (1995) 246–249.

[6] D. Bieger, D.A. Hopkins, Viscerotopic representation of the upper al-imentary tract in the medulla oblongata in the rat: The nucleus ambig-uus, J Comp Neurol 262 (1987) 546–562.

[7] E.M. Blass, M.H. Teicher, C.P. Cramer, J.P. Bruno, W.G. Hall, Ol-factory, thermal and tactile controls of suckling in preauditory andprevisual rats, J Comp Physiol Psychol 91 (1977) 1248–1260.

[8] D.L. Broussard, R.B. Lynn, E.B. Wiedner, S.M. Altschuler, Solitarialpremotor neuron projections to the rat esophagus and pharynx: Impli-cations for control of swallowing, Gastroenterology 114 (1998)1268–1275.

[9] A.J. Durston, J.P.M. Timmermans, W.J. Hage, H.F.J. Hendriks, N.J.de Vries, M. Heideveld, P.D. Nieuwkoop, Retinoic acid causes an an-teroposterior transformation in the developing central nervous sys-tem, Nature 340 (1989) 140–144.

[10] W.G. Hall, J.S. Rosenblatt, Suckling behavior and intake control inthe developing rat pup, J Comp Physiol Psychol 91 (1977) 1232–1247.

[11] J. Hill, J.D.W. Clarke, N. Vargesson, T. Jowett, N. Holder, Exoge-nous retinoic acid causes specific alterations in the development of

the midbrain and hindbrain of the zebrafish embryo including re-specification of the Mauthner neuron, Mech Dev 50 (1995) 3–16.

[12] N. Holder, J. Hill, Retinoic acid modifies development of the mid-brain-hindbrain border and affects cranial ganglion formation in ze-brafish embryos, Development 113 (1991) 1159–1170.

[13] R.R. Holson, R.A. Gazzara, S.A. Ferguson, J. Adams, Behavioral ef-fects of low-dose gestational day 11–13 retinoic acid exposure, Neu-rotoxicol Teratol 19 (1997) 335–346.

[14] R.R. Holson, R.A. Gazzara, S.A. Ferguson, J. Adams, A behavioraland neuroanatomical investigation of the lethality caused by gesta-tional day 11–13 retinoic acid exposure, Neurotoxicol Teratol 19(1997) 347–353.

[15] R.R. Holson, R.A. Gazzara, S.A. Ferguson, S.F. Ali, J.B. LaBorde, J.Adams, Gestational retinoic acid exposure: A sensitive period for ef-fects on neonatal mortality and cerebellar development, NeurotoxicolTeratol 19 (1997) 355–362.

[16] R.R. Holson, J. Adams, S.A. Ferguson, Gestational stage-specific ef-fects of retinoic acid exposure in the rat, Neurotoxicol Teratol 21(1999) 393–402.

[17] A. Jean, Brainstem organization of the swallowing network, BrainBehav Evol 25 (1984) 109–116.

[18] R.P. Jensh, D.M. Kocchar, M.K. Till, M.B. Eskesen, Postnatal behav-ioral sequlae of prenatal exposure to 13-

cis

retinoic acid, Teratology41 (1990) 621–622.

[19] R.P. Jensh, D.M. Kocchar, M.K. Till, M.B. Eskesen, Effects of prena-tal exposure to isotretinoin (13-

cis

retinoic acid) on neonatal vocal-izations, Teratology 43 (1991) 497.

[20] P. Kehoe, E.M. Blass, Gustatory determinants of suckling in albinorats 5–20 days of age, Dev Psychobiol 18 (1985) 67–82.

[21] R. Krumlauf, N. Papalopulu, J.D.W. Clarke, N. Holder, Retinoic acidand the xenopus hindbrain, Semin Dev Biol 2 (1991) 181–188.

[22] A.S. LaMantia, M.C. Colbert, E. Linney, Retinoic acid induction andregional differentiation prefigure olfactory pathway formation in themammalian forebrain, Neuron 10 (1993) 1035–1048.

[23] H. Marshall, S. Nonchev, M.H. Sham, I. Muchamore, A. Lumsden,R. Krumlauf, Retinoic acid alters hindbrain Hox code and inducestransformation of rhombomeres 2/3 into a 4/5 identity, Nature 360(1992) 737–741.

[24] I.M. Matthews Jr., E.M. Johnson, L.M. Newman, Introduction of lategestational teratogenesis in rat lung by hypervitaminosis A, Teratol-ogy 23 (1981) 253–258.

[25] L.M. Newman, E.M. Johnson, A.S.A. Cadogan, Intralitter variationof neonatal rat lung morphology following prenatal maternal hypervi-taminosis A, Anat Rec 204 (1982) 153–160.

[26] L.M. Newman, E.M. Johnson, Abnormal lung function induced byprenatal insult, Exp Pharmacol 65 (1983) 237–261.

[27] L.M. Newman, E.M. Johnson, A.S.A. Cadogan, Oxygen consump-tion and survival prediction in neonatal rats exposed to prenatal hy-pervitaminosis A, Teratology 28 (1983) 219–227.

[28] P.A. Núñez-Abades, R. Pásaro, A.L. Bianchi, Study of the topograph-ical distribution of different populations of motoneurons within rat’snucleus ambiguus, by means of four different fluorochromes, Neuro-sci Lett 135 (1992) 103–107.

[29] N. Papalopulu, J.D.W. Clarke, L. Bradley, D. Wilkinson, R. Krum-lauf, N. Holder, Retinoic acid causes abnormal development and seg-mental patterning of the anterior hindbrain in xenopus embryos, De-velopment 113 (1991) 1145–1158.

[30] P.E. Pedersen, E.M. Blass, Olfactory control over suckling in albinorats, in: R.N. Aslin, J.R. Alberts, M.R. Petersons (Eds.), Developmentof Perception: Psychobiological Perspectives, Academic Press, NewYork, 1981, pp. 359–381.

[31] R.E. Shenefelt, Morphogenesis of malformations in hamsters causedby retinoic acid: Relation to dose and stage at treatment, Teratology 5(1972) 103–118.

[32] P.J. Singh, E. Tobach, Olfactory bulbectomy and nursing behavior inrat pups, Dev Psychobiol 8 (1975) 151–164.

[33] M.H. Teicher, E.M. Blass, First suckling response of the new-born al-

R.R. Holson et al. / Neurotoxicology and Teratology 22 (2000) 541–545

545

bino rat: The roles of olfaction and amniotic fluid, Science 198(1977) 635–636.

[34] J. van der Wees, J.G. Schilthuis, C.H. Koster, H. Diesveld-Schipper,G.E. Folkers, P.T. van der Saag, M.I. Dawson, K. Shudo, B. van derBurg, A.J. Durston, Inhibition of retinoic acid receptor-mediated sig-nalling alters positional identity in the developing hindbrain, Devel-opment 125 (1998) 545–556.

[35] J.P. Welsh, E.J. Lang, I. Sugihara, R. Llinas, Dynamic organization

of motor control within the olivocerebellar system, Nature 374 (1995)453–457.

[36] J. Whitesides, M. Hall, R. Anchan, A.S. LaMantia, Retinoid signalingdistinguishes a subpopulation of olfactory receptor neurons in the de-veloping and adult mouse, J Comp Neurol 394 (1998) 445–461.

[37] P. Woodhams, R. Rodd, R. Balazs, Age-dependent susceptibility ofinferior olive neurones to 3-acetylpyridine in the rat. Brain Res. 153(1978) 194–198.