J. Embryol. exp. Morph. Vol. 32, 2, pp. 505-514, 1974
Printed in Great Britain
505
The effect of excess vitamin A on the development
of rat embryos in culture
By G I L L I A N M. M O R R I S S 1 AND C. E. S T E E L E 2
Department of Anatomy and Department of Physiology, Cambridge
SUMMARY
Primitive streak stage rat embryos were cultured in serum containing added vitamin A
alcohol at concentrations ranging from 0-5 to 20/tg/ml. At the higher concentrations there
was an overall growth retardation, and differentiation was inhibited. This was thought to be
due to inhibition of DNA synthesis. At 0-5 /*g/ml the embryos developed abnormally in a
manner similar to that previously seen in vivo. This result, which was supported by an ultrastructural study, provides evidence that the teratogenic effects of vitamin A are due to a
direct action on the embryos.
Comparison of embryonic development in vitro at different concentrations of vitamin A
with the results obtained in vivo suggests that in the latter, only a very small proportion of the
amount injected was reaching the embryos in an active form. The maternal factors involved
are discussed, and excess vitamin A is defined as an excess of free and non-specifically bound
retinol. The levels of vitamin A which affected differentiating embryos in vitro are lower than
the minimum levels previously shown to affect differentiated tissues. This suggests that
relatively minor increases in serum levels during pregnancy may cause malformations without
affecting the maternal organism. Since raised serum vitamin A levels have been shown to be
associated with impaired liver function in man, these factors may be significant in human
teratogenesis.
INTRODUCTION
T h e a d m i n i s t r a t i o n of large a m o u n t s of v i t a m i n A t o m a m m a l s d u r i n g early
pregnancy causes malformations in the offspring (Cohlan, 1953; Giroud &
Martinet, 1959). Although this is firmly established, it is not clear whether the
teratogenic effect is brought about by a direct action on the embryos, or via an
effect on the maternal organism.
In support of the latter, Woollam & Milien (1960) found that the effects of
vitamin A were enhanced by cortisone and reduced by insulin and thyroxine,
and postulated a mechanism involving interference with maternal carbohydrate
metabolism. However, Cohlan & Stone (1961) failed to confirm these results;
furthermore they found that parathyroidectomy, thymectomy, or bilateral
adrenalectomy themselves had no influence on gross foetal development, nor
did they affect the incidence of vitamin A-induced malformations.
In a study of the ultrastructural changes occurring in embryos following the
1
Author's Address: Department of Anatomy, Downing Street, Cambridge CB2 3DY, U.K.
Author's Address: Department of Physiology, Downing Street, Cambridge CB2 3DY,
U.K.
2
506
G. M. MORRISS AND C. E. STEELE
maternal vitamin A injection, Morriss (1973 à) interpreted the results as supportive evidence for a direct action on the embryo, since the cytological effects
were similar to those seen in in vitro studies of mammalian cells and tissues in
the presence of excess vitamin A (Glauert, Daniel, Lucy & Dingle, 1963;
Daniel, Dingle, Glauert & Lucy, 1966).
This study of the effects of vitamin A on the development of embryos in vitro
was undertaken in order to test the hypothesis that vitamin A exerts its teratogenic effects by direct action on the embryos.
A further question concerns the large amounts of vitamin A which are needed
to produce malformations in vivo, since these are quite unrelated to normal
physiological requirements and may affect the maternal organism as well as the
embryos: Langman & Welch (1966) gave three injections of 30000 i.u. ; Kochhar
(1968) gave three injections of 60000 i.u. ; Morriss (1972) gave a single injection
of 100000 i.u. If vitamin A affects embryos directly, the amount actually reaching
them may be only a small fraction of the amount given to the dam ; determination of the effective concentrations in vitro would give an indication of the
serum levels of active vitamin A in the in vivo situation.
MATERIALS AND METHODS
Rat ' egg-cylinders ' were explanted during the afternoon of day 8 of pregnancy
(day 0 being the day of the positive vaginal smear) and cultured in circulating
or static medium by the technique of New (1971). Explantation was carried out
in Tyrode saline and Reichert's membrane opened. Those embryos to be cultured
in circulating medium were anchored to a piece of fabric by means of the
ectoplacental cone and Reichert's membrane before being placed in the circulator.
No more than five embryos were cultured in each circulator or watchglass and
embryos from different litters were divided between different treatments in
order to minimize initial differences.
The nutrient medium was immediately centrifuged (i.e.) serum containing
50 /*g/ml streptomycin, which has been shown to support good development of
egg-cylinders in vitro (Steele, 1972; Steele & New, 1974). This was prepared from
blood drawn from the dorsal aorta of an anaesthetized rat. The blood was
immediately centrifuged and the clot (free of erythrocytes) removed. After
recentrifugation the serum was decanted. All centrifugations were carried out
at 2500 rev/min for 5 min. For most cultures the medium was prepared on the
same day as the explantation.
In circulator cultures 6-8 ml of medium was circulated at 5 ml/min and
equilibrated with a gas mixture containing 5 % C0 2 in air. In watchglass cultures
1 ml medium was equilibrated with 40% 0 2 . Sterile procedure was observed
throughout the culture period.
Control embryos were grown in i.e. serum to which 0-01 ml absolute alcohol/
ml was added. Experimental embryos were grown in i.e. serum containing
Excess vitamin A and rat embryos
507
Table 1. Embryonic development after 48 h in culture
Average yolk-sac
No. with
No. of embryos
diameter
No. with heartbeat protruding axis
K
K
Vitamin A ,
*
,
,
» ,
, ,
*
*
(/ig/ml) A-treated Controls* A-treated Controls A-treated Controls A-treated Controls
20
10
5
3
1
0-5
4
4
3
16
3
32
3
3
3
20
9
28
1-2
1-5
1-6
1-6
1-65
1-7
1-7
1-7
1-7
1-7
1-8
1-8
0/4
0/4
0/3
4/16
1/3
18/32
3/3
3/3
1/3
14/20
9/9
26/28
2/4t
4/4
3/3
10/16
3/3
9/32
0/4
0/4
0/3
1/20
1/9
5/28
* Where more than one experimental culture shared the same control culture, the controls
appear more than once in this column. Total no. control embryos = 49.
f Other two 'embryos.' indistinct.
retinol (vitamin A alcohol) dissolved in absolute alcohol, in final concentrations
ranging from 0-5-20-0 /*g/ml; the volume added was 001 ml/ml of medium in
all cultures.
The embryos were examined with a dissecting microscope after 24 and 48 h
and the yolk-sac diameter, activity of the heart and general appearance recorded.
Paraffin sections of some embryos were prepared for more detailed morphological investigation.
Some cultured embryos were prepared for electron microscopy. At 1 /*g
vitamin A/ml, embryos were examined after 1 h (8), 2 h (2) and 6 h (2) (numbers
in brackets = embryos examined in each case). At 0-5/tg/ml embryos were
examined after 10 min (2), 20 min (7), 30 min (2), 1 h (8), 2 h (7), 4 h (2) and
6 h (3). (These were all watchglass cultures, so that some embryos in the same
dish could be left to continue developing to 48 h.) The embryos were rinsed in
Tyrode saline, then fixed in 4 % cacodylate-buffered glutaraldehyde, post-fixed
in osmium tetroxide, and embedded in Araldite. Thin sections were double
stained and viewed in a Philips EM 300.
RESULTS
Development achieved by embryos after culture for 48 hours (see Table 1)
Embryos cultured in medium containing 5, 10 or 20/^g/ml vitamin A developed very poorly. After 48 h none had a beating heart, and the embryo
proper was represented only by a thickened area on the yolk-sac wall in which
little or no differentiation was apparent when viewed with a dissecting microscope. There was a definite longitudinal axis, the posterior part of which formed
a straight and rigid protrusion from the yolk sac. As in previous studies on
culture techniques (Le Goascogne & Brun, 1969; Steele, 1972), this protrusion
was occasionally observed in control embryos (7 out of 49 of all controls) but
in these it was always smaller, less rigid, and therefore curved.
508
G. M. MORRISS AND C. E. STEELE
Figs. 1-4. Embryos cultured for 48 h, to show height of cephalic neural folds, amount
and density of cephalic mesenchyme, and extent of foregut development.
Fig. .1. L.s. control embryo, 12 sections lateral to the mid-sagittal plane.
Fig. 2. L.S. embryo cultured in serum containing 0-5 fig vitamin A/ml, 12 sections
lateral to the mid-sagittal plane. Flatter neural folds and less cephalic mesoderm than
control ; posterior part of axis protrudes from yolk sac.
Fig. 3. Oblique coronal plane, control embryo, position of greatest height of
neural fold.
Fig. 4. Coronal plane, embryo cultured in serum containing 0-5 fig vitamin A/ml,
showing flattened neural folds and great extension of foregut into head.
Embryos cultured in medium containing 3 /^g/ml vitamin A developed more
normally. Four out of 16 developed a heartbeat, although at 48 h the rate was
only 35 beats/min compared with 60-80 beats/min in the controls.
Development was most similar to that of control embryos at a vitamin A
concentration of 0-5 /tg/ml; the protrusion of the posterior part of the axis was
greatly reduced in extent, and occurred in only 9 out of 32 embryos. The
development of the head differed from that of the controls in that the neural
folds wereflatand widely open. This was confirmed in sections (Figs. 1-4), which
also showed (not illustrated) that the paraxial mesoderm was diffuse and had
Excess vitamin A and rat embryos
509
FIGURES 5-7
Fig. 5. 0-5 fig vitamin A/ml, 10 min culture. Extra-embryonic endoderm: distorted
nuclei with swollen nuclear membranes.
Fig. 6. 0-5 fig vitamin A/ml, 30 min culture. Embryonic ectoderm (with basement
membrane) and mesoderm: 'buds' (b), vacuoles (arrowed) and condensed mitochondria.
Fig. 7. 0-5 fig vitamin A/ml, 30min culture. Extra-embryonic endoderm: large
lipid droplet (/), increased intercellular spacing.
510
G. M. M O R R I S S A N D C. E. S T E E L E
not formed discrete somites. The closed area of the neural tube was never
straight, but had multiple irregular lateral flexures. The foregut was deeper in
experimental embryos at this concentration than in the controls.
Electron microscopy (Figs. 5-7)
In all embryos cultured for 30 min or more at vitamin A concentrations of
1 /Ag/ml and 0-5 /*g/ml, all the cytological abnormalities previously observed in
embryos treated in vivo with vitamin A were seen. These were: increase in the
surface area of the plasma membrane and intracellular membranes; autophagic
and/or heterophagic vacuoles; small vacuoles close to the cell surface; cell
death; 'budding' of cells, particularly the basal surface of ectoderm cells;
distorted ectodermal nuclei with enlarged intermembranous spaces ; condensed
mitochondria; wider spacing of cells with an increased volume of extracellular
fluid ; increase in the number of large lipid droplets of medium electron density.
Except where stated, these effects occurred in the cells of all three germ layers.
The embryos viewed after only 10 min of culture at 0-5 jag vitamin A/ml were
similar except that the effects were apparent only in the endoderm, and the
increase in spacing of these cells was slight; advanced cell degeneration and
death was only apparent in the parietal endodermal (Reichert's membrane) cells,
which cannot be considered a functional part of the extra-embryonic membrane
system in the culture situation.
In the controls, two of the above effects were constantly observed: from
30 min of culture onwards, all embryos contained some cells with autophagic
or heterophagic vacuoles, and the mitochondria appeared to be condensed
compared with those of embryos of the same age fixed immediately after
explantation.
DISCUSSION
Allowing for differences directly attributable to development in vitro, the
morphology of embryos cultured for 48 h in serum containing 0-5 jug vitamin
A/ml is strikingly similar to that of embryos explanted 48 h after the maternal
administration of 100000 i.u. (60000 jug) vitamin A (Morriss, 1972). This
suggests that vitamin A affects embryonic development in vivo by direct action
on the embryos. In both of these experimental situations the development of the
cephalic neural folds was retarded; this was associated with reduction of
cephalic mesenchyme, exaggeration of the cephalic flexure and greater extension
of the foregut into the head than in the controls.
Further evidence for direct action was provided by the electron microscopic
observations. At the lowest concentrations (1-0 and 0-5 /*g/ml) all the cytological
effects previously seen in vivo (Morriss, 1973 a) were present in the cultured
embryos within 30 min of the start of culture. Two phenomena (autophagic or
heterophagic vacuoles and condensed mitochondria) were present in both
Excess vitamin A and rat embryos
511
experimental and control cultured embryos, and therefore cannot be ascribed
to the action of vitamin A. In the previous in vivo study they were seen in embryos
from vitamin A-treated but not control dams ; since they had not been described
in other electron microscopic studies on excess vitamin A, they were thought
to be maternal-mediated effects (Morriss, 1973 A, b).
At higher concentrations (5-20 /ig vitamin A/ml serum) embryonic development was retarded far more than at 0-5 /*g/ml, but in addition there was an
almost complete failure of differentiation. Both of these effects may be attributable to a decrease in DNA synthesis, since this has been demonstrated in the
presence of excess vitamin A in embryonic chick-limb cartilage (Dingle, Lucy &
Fell, 1961) and in various tissues of 13-day rat embryos (Kochhar, 1968).
The results also indicate that young differentiating embryos are more sensitive
to vitamin A excess than are differentiated tissues. Cultured pig articular
cartilage is sensitive to minimum levels of 5 i.u. (3 fig vitamin A/ml medium)
(Barratt, 1973) and embryonic chick epidermis to 2-5 i.u. (1-5/tg/ml) (Fell &
Rinaldini, 1965). Embryos cultured at lower concentrations and for shorter
exposure times would yield further information on this point.
Since such low concentrations affected in vitro development, it seems that only
very small blood concentrations of active vitamin A were reaching the embryos
affected in vivo. The great discrepancy between effective amounts in vivo and
in vitro may be related to factors involved in maternal vitamin A storage,
transport, and metabolism.
In contrast to vitamin A alcohol (retinol), retinyl esters have little or no
action on tissues in vitro (Fell, Dingle & Webb, 1962; Roels, 1969). Preliminary
work (Morriss & Steele, unpublished) indicated that this holds true for whole
embryos under culture conditions similar to those of the present study. It thus
appears likely that vitamin A affects embryos in vivo in the alcohol form. Vitamin
A is stored in the liver as the ester (Moore, 1957). It is hydrolysed to retinol in
the liver cells (Mahadevan, Ayyoub & Roels, 1966) and transported in the plasma
attached to a lipoprotein, retinol-binding protein (RBP) (Raz, Kanai & Goodman, 1968). RBP is synthesized in the liver, and its secretion into the plasma is
related to the availability of vitamin A (Goodman, Muto, Smith & Milch, 1972).
Retinol-RBP may also be bound to a prealbumin ; this complex releases retinol
less easily than retinol-RBP alone, so may serve to stabilize the retinol in the
bloodstream (Goodman, 1969). Dingle, Fell & Goodman (1972) found that
when bound to RBP, retinol has no effect on cultured embryonic chick limbbones, although it does have an effect when bound to non-specific serum proteins. Thus excess serum vitamin A should be defined in terms of the concentration of free and non-specifically bound retinol. The free fraction is not indicated
in reported serum assays, although it is clearly critical for the adverse systemic
as well as developmental effects.
The fact that very low levels of retinol affect embryonic development may be
a significant factor in human teratogenesis. Normal human serum vitamin A
33
E M B 32
512
G. M. MORRISS AND C. E. STEELE
levels are 30-50 /tg/100 ml (Russell, Bagheri & Boyer, 1973). High serum vitamin
A levels due to decreased storage capacity have been associated with liver
injuries such as cirrhosis and carbon tetrachloride poisoning (Baker & Frank,
1968). Chronic vitamin A intoxication is followed by high serum levels only
when the liver capacity has been exceeded, but once this stage is reached serum
levels may remain high for months or years after vitamin A ingestion has
ceased (Muenter, Perry & Ludwig, 1971). Large doses of vitamin A may result
in extensive fibrosis, portal hypertension, and ascites (Russell et al. 1973), so
that relatively small amounts of vitamin A bring about rapid increases in
serum levels.
In human pregnancy, the maternal serum vitamin A determinations reported
in the literature do not cover the period during which the developing embryo is
most vulnerable to teratogens (e.g. Wild, Schorah & Smithells, 1974; Gal,
Sharman & Pryse-Davies, 1972). Appropriate timing is difficult to achieve
since the embryo is at the primitve streak stage at the time of the first missed
period, and neural tube closure is complete 11 days later (Langman, 1969).
However, congenitally malformed infants have been reported following actual
ingestion of large amounts of vitamin A by women in the first month of pregnancy; in one case the urogenital system was affected (Pilotti & Scorta, 1965),
in another, the heart and palate and in a third the palate alone (D. E. Poswillo,
personal communication).
Clearly, even in the absence of exogenous sources of the vitamin, liver damage
leading to raised serum vitamin A during early human pregnancy might raise
the free retinol levels enough to cause congenital malformation. Whether this
does in fact occur has yet to be established.
The authors wish to acknowledge the support of the Wellcome Trust and Medical Research
Council. We are grateful to Professors R. J. Harrison and C. R. Austin for the provision of
laboratory facilities, to Mr G. Owen for photography, to Dr A. D. Thomson for providing
the clinical references, and to Dame Honor Fell, Professor C. R. Austin, Dr J. Herbert and
Professor D. E. Poswillo for valuable criticism of the manuscript. The vitamin A was
supplied by Dr N. T. Pollitt of Roche Products Ltd.
REFERENCES
BAKER, H. & FRANK, O. (1968). Clinical Vitaminology. New York: Interscience.
BARRATT, M. E. J. (1973). The role of soft connective tissue in the response of pig
articular
cartilage in organ culture to excess of retinol. J. Cell Sei. 13, 205-219.
COHLAN, S. Q. (1953). Excessive intake of vitamin A as a cause of congenital anomalies in
the rat. Science, N.Y. 117, 533-536.
COHLAN, S. Q. & STONE, S. M. (1961). Observations on the effect of experimental endocrine
procedures on the teratogenic action of hypervitaminosis-A in the rat. Biologia Neonat. 3,
330-342.
DANIEL, M. R., DINGLE, J. T., GLAUERT, A. M. & LUCY, J. A. (1966). The action of excess
vitamin A alcohol on the fine structure of rat dermal fibroblasts. /. Cell Biol. 30, 465-475.
DINGLE, J. T., LUCY, J. A. & FELL, H. B. (1961). Studies on the mode of action of excess of
vitamin A. 1. Effect of excess of vitamin A on the metabolism and composition of embryonic
chick-limb cartilage grown in organ culture. Biochem. J. 79, 497.
Excess vitamin A and rat embryos
513
DINGLE, J. T., FELL, H. B. & GOODMAN, D. S. (1972). The effect of retinol and of retinol-
binding protein on embryonic skeletal tissue in organ culture. J. Cell Sei. 11, 393-402.
FELL, H. B., DINGLE, J. T. & WEBB, M. (1962). Studies on the mode of action of excess of
vitamin A. 4. The specificity of the effect on embryonic chick-limb cartilage in culture and
on isolated rat liver lysosomes. Biochem. J. 83, 63-69.
FELL, H. B. & RINALDINI, L. M (1965). The effects of vitamins A and C on cells and tissues
in culture. In Cells and Tissues in Culture, vol. i (ed. E. N. Wilmer), pp. 659-699. New
York : Academic Press.
GAL, I., SHARMAN, I. M. & PRYSE-DAVIES, J. (1972). Vitamin A in relation to human congenital malformations. Adv. Terat. 5, 143-159.
GIROUD, A. & MARTINET, M. (1959). Extension a plusieurs espèces de mammifères des
malformations embryonnaires par hypervitaminose A. C. r. Séanc. Soc. Biol. 153, 201.
GLAUERT, A. M., DANIEL, M. R., LUCY, J. A. & DINGLE, J. T. (1963). Studies on the mode of
action of excess of vitamin A. 7. Changes in the fine structure of erythrocytes during
haemolysis. / . Cell Biol. 17, 111-121.
GOODMAN, D E W. S. (1969). Retinol transport in human plasma. Am. J. clin. Nutr. 22,
911-912.
GOODMAN, D E W. S., MUTO, Y., SMITH, J. E. & MILCH, P. (1972). Retinol transport in rat
plasma and its nutritional regulation. Biochem. J. 128, 15-16P.
KOCHHAR, D. M. (1968). Studies of vitamin A-induced teratogenesis : effects on embryonic
mesenchyme and epithelium, and on incorporation of HMhymidine. Teratology 1,299-310.
LANGMAN, J. (1969). Medical Embryology, pp. 46-290. Baltimore: Williams and Wilkins.
LANGMAN, J. & WELCH, G. W. (1966). Effect of vitamin A on development of the central
nervous system. / . comp. Neurol. 128, 1-16.
LE GOASCOGNE, C. & BRUN, J. L. (1969). Développement in vitro de l'embryon de rat.
C. r. hebd. Séanc. Acad. Sei., Paris 268, 3195-3200.
MAHADEVAN, S., AYYOUB, N. I. & ROELS, O. A. (1966). Hydrolysis of retinol palmitate by
rat liver. / . biol. Chem. 241, 57-64.
MOORE, T. (1957). Vitamin A. Amsterdam: Elsevier.
MORRISS, G. M. (1972). Morphogenesis of the malformations induced in rat embryos by
maternal hypervitaminosis A. J. Anat. 113, 241-250.
MORRISS, G. M. (1973a). The ultrastructural effects of excess maternal vitamin A on the
primitive streak stage rat embryo. / . Embryol. exp. Morph. 30, 219-242.
MORRISS, G. M. (19736). An electron microscopic study of the effects of hypervitaminosis A
on the maternal-embryonic relationship of the rat at 8 days of pregnancy. / . Reprod. Fert.
33, 451-456.
MUENTER, M. D., PERRY, H. O. & LUDWIG, J. (1971). Chronic vitamin A intoxication in
adults (hepatic, neurologic and dermatologie complications). Am. J. Med. 50, 129-136.
NEW, D. A. T. (1971). Methods for the culture of post-implantation embryos of rodents.
In Methods in Mammalian Embryology (ed. J. C. Daniel), pp. 305-319. San Francisco:
Freeman.
PILOTTI, G. & SCORTA, A. (1965). Minerva Ginecologica 17, 1103. (Quoted in Wild et al. 1974.)
RAZ, A., KANAI, A. & GOODMAN, D E W. S. (1968). Retinol binding protein (RBP): a newly
recognized plasma lipoprotein. Fedn Proc. Fedn Am. Socs exp. Biol. 27, 594. (Abstract.)
ROELS, O. A. (1969). The fifth decade of vitamin A research. Am. J. din. Nutr. 22, 903907.
RUSSELL, R. M., BAGHERI, S. A. & BOYER, J. C. (1973). The hepatic lesion of hypervitaminosis
A - a unique pattern of hepatic fibrosis with portal hypertension and ascites. Gastroenterology 65, 544/568.
STEELE, C. E. (1972). Improved development of'rat egg-cylinders' in vitro as a result of fusion
of the heart primordia. Nature New Biology, Lond. 237, 150-151.
STEELE, C. E. & NEW, D. A. T. (1974). Serum variants causing the formation of double
hearts and other abnormalities in explanted rat embryos. / . Embryol. exp. Morph. 31,
707-719.
WILD, J., SCHORAH, C. J. & SMITHELLS, R. W. (1974). Vitamin A, pregnancy, and oral
contraceptives. Br. med. J. i, 57-59.
33-2
514
G. M. MORRISS AND C. E. STEELE
D. H. M. & MILLEN, J. W. (1960). The modification of the activity of certain
agents exerting a deleterious effect on the development of the mammalian embryo. In
CibaFdn. Symp. Congenital Malformations (ed. G. E. W. Wolstenholme and M. O'Connor),
pp. 158-172. London: J. and A. Churchill.
WOOLLAM,
{Received 9 March 1974)