Pediatr Surg Int (2004) 20: 529533

DOI 10.1007/s00383-004-1212-9

O R I GI N A L A R T IC L E

Bao Quan Qi Spencer W. Beasley Dejan Arsic

Abnormalities of the vertebral column and ribs associated

with anorectal malformations

Accepted: 2 February 2004 / Published online: 16 June 2004

Springer-Verlag 2004

Abstract Lumbosacral vertebral abnormalities are a

common association of anorectal malformations
(ARMs) and are one of the determinants of the eventual level of fecal continence that can be achieved. This
study used a fetal rat model to investigate the spectrum
of axial skeletal maldevelopment that may occur with
ARMs. Time-mated pregnant rats received 125 mg/kg
of 1% ethylenethiourea (ETU) (experimental group) or
vehicle only (control). Their fetuses were examined for
external malformations and prepared for staining of
their skeletons using Alcian blue and Alizarin red S.
ARMs developed in 67/68 (98%) of ETU-exposed fetuses, of which 28 (42%) also developed rachischisis,
mainly involving the lumbosacral vertebrae. No skeletal abnormality was found in control fetuses. ETUexposed fetuses with ARMs and rachischisis had
abnormal ossication of the vertebral centrum, abnormal fusion between the neural arches of vertebrae,
localized narrow or interrupted thoracic vertebral canal, a widely open vertebral canal in the lumbosacral
area (rachischisis), and absence of the lower two sacral
and coccygeal vertebrae. Rib abnormalities included
absence of two to three oating ribs, abnormal fusion
of adjacent proximal segments, and abnormal ramication, irregularity, and angulation of their distal segments. The vertebral and rib abnormalities found in
ETU-exposed fetuses with ARMs but no rachischisis
were much less severe. In addition to the lumbosacral
anomalies that are common with ARMs, severe
abnormalities of the thoracic vertebrae and their corresponding ribs may occur also. Fetuses with both
ARM and rachischisis tend to have more extensive and
severe vertebral and rib anomalies. These observations
imply a possible common aetiology for ARMs and

vertebral anomalies and are consistent with our

understanding of the perceived role of the notochord in
axial development.
Keywords Anorectum Imperforate anus
Vertebral Rachischisis Ethylenethiourea

Introduction
Vertebral defects, especially those of the lumbosacral
region, are common with many congenital malformations, including anorectal malformations (ARMs) [1],
esophageal atresia [2, 3], urinary tract anomalies [4], and
as part of several recognized associations, such as the
VATER (or VACTERL) association [5, 6] and Currarino triad [7, 8, 9, 10]. The presence of concomitant sacral abnormalities in infants with ARMs is highly
relevant because it is believed to be one of the main
determinants of the eventual level of fecal continence
that can be achieved. However, relatively little is known
about involvement of the rest of the vertebral column
and adjacent axial skeleton. The present study used the
fetal rat model of ARMs induced chemically with
ethylenethiourea (ETU) to stain the intact vertebral
column and adjoining axial skeletons without altering
their anatomical position to identify the range of
abnormalities produced.

Materials and methods

Teratogen preparation

B. Q. Qi S. W. Beasley (&) D. Arsic

Department of Paediatric Surgery,
Christchurch Hospital, Private Bag 4710,
Christchurch, New Zealand
E-mail: spencer.beasley@cdhb.govt.nz
Tel.: +64-3-3640486
Fax: +64-3-3641584

Ethylenethiourea,
ETU
(2-imidazolidinethione,
C3H6N2S) powder produced by Aldrich Chemical
Company in Milwaukee, WI, USA, was dissolved in
distilled water to a concentration of 1%. The solution
was saved at 4C in lightproof conditions.

530

Animal preparation
Young virgin female Sprague-Dawley rats with body
weight 200250 g (Animal Laboratory, Christchurch
School of Medicine) were mated with a male rat overnight. The next morning, on which sperm was found on
a vaginal smear, was designated as day 0 of gestation.
The pregnant rats were divided randomly into control
(four rats) and experimental (eight rats) groups. Experimental rats received 1% ETU (125 mg/kg) by gavage
on gestational day 10 under a light inhalational anesthetic. Control rats received the vehicle only, at the same
volume. The rats were kept individually in an air-conditioned, 12-h light/dark cycle animal laboratory
(Christchurch School of Medicine) and fed normal rat
chow and tap water ad libitum. Dams were sacriced
with an overdose inhalation of halothane on gestational
day 21, and their fetuses were recovered by Caesarean
section. The abdomen was opened to eviscerate the
internal organs without damaging the ribs. The carcasses
were xed in 95% ethanol.
Bone staining
On the 2nd day of xation, the skin and adipose tissue
were removed and the thoracic organs eviscerated. The
specimens were xed for a further 45 days before
treatment with acetone for 1 day or longer to remove fat.
The specimens were stained for 23 days in a preprepared solution (70% ethanol containing 0.015% ltered alcian blue, 0.005% ltered alizarin red S, and
5% acetic acid [11, 12]) at room temperature. After
being washed in water, the carcasses were cleared in a
1% aqueous solution of KOH for 48 h or longer until
the skeletons were clearly visible through the surrounding tissue. The specimens were then placed in an
aqueous solution of 20% glycerin containing 1% KOH
for 35 days, during which time the fat tissue was further removed by microscopic dissection to facilitate the
future observation of the skeletons. The specimens went
through 20% aqueous glycerin containing 1% KOH,
50% glycerin, and 80% glycerin, and stored in
100% glycerin for examination. Cartilage was stained
blue and ossied skeleton was stained red.
The skeletons were examined under the microsurgical
microscope and abnormalities recorded by photography. Specic attention was paid to the vertebral column
and adjacent axial skeletons.

Results
Control fetuses
On gestational day 21, all segments of the vertebral
column were easy to recognize and looked normal.
There were seven cervical vertebrae, 13 thoracic, six
lumbar, four sacral, and 45 caudal. From the cervical

to sacral segments, the vertebrae were mainly composed

of three ossication centers: the centrum (for the vertebral body) and two lateral ossifying centers (for the
neural arches that will form the vertebral canal around
the spinal cord). The future spinal canal was formed
evenly. There were 13 pairs of ribs joining the corresponding 13 thoracic vertebrae. Their breadth and
length were even, and ossication was uniform
(Fig. 1A). The sacrum was not fused yet, unlike in the
adult rat. The distal two to three caudal vertebrae presented just one point of ossication (Fig. 2A). The bones
of the forelimb were normally dierentiated (upper image in Fig. 3). In the hind limb, the ilium, ischium, and
pubis were ossied but not yet fused. The appearance of
the normal limb is shown in Fig. 2A.
There were 68 fetuses harvested from eight ETUtreated rats, of which 67 (98.5%) had ARMs. The
remaining fetus had an anteriorly placed anus. Rachischisis occurred in 28 (42%) and mainly involved the
lumbosacral area. The main features of the vertebral and
axial skeletal anomalies of ETU-exposed fetuses are
summarized in Table 1.
ETU fetuses with ARMs and rachischisis
The centrum (vertebral body) of the 28 ETU-exposed
fetuses with ARMs and rachischisis had two ossication
centers instead of one, extending from the lower thoracic
region to the sacral segment. Some fetuses had no ossication center for the centrum. The size and shape of the
two lateral ossication centers (neural arches) varied,
and abnormal fusion between them made it dicult to
count their number precisely. The seven cervical vertebrae were less severely aected. Five fetuses had localized thoracic vertebral canal stenosis at the 5th8th
thoracic vertebrae level. In three fetuses, the thoracic
vertebral canal was interrupted with abnormal bone
pieces bridging the ribs on both sides. The vertebral
canal in the lower lumbosacral area was widely open
posteriorly, corresponding to the location of the rachischisis. The lower two sacral and caudal vertebrae were
always absent (Fig. 1B, C; Fig. 2B). All fetuses had rib
abnormalities. The average number of ribs was 11 on the
left and 10 on the right. The lowest two to three oating
ribs were frequently missing. The rst three or four pairs
of ribs were usually normal, whereas ribs from the 4th
10th thoracic vertebrae exhibited a variety of abnormalities, including abnormal fusion and branching,
irregular, or angulated ribs and free rib segments.
(Fig. 1B, C). Four of 28 fetuses had radial hypoplasia
on the right forelimb with abnormal elbow angulation
(Fig. 3). In two fetuses, the femur and bula were absent
on the right hindlimb (Fig. 2C).
ETU-exposed fetuses with ARMs only
These fetuses (n=39) had less severe skeletal abnormalities. The number and shape of the cervical, thoracic,

531

Fig. 1AC Posterior view of the thoracic and lumbar vertebrae. A

Control fetus: Showing uniform ossication centers of the thoracic
and lumbar vertebrae and 13 pairs of ribs. B ETU-exposed fetus:
Showing a gap (arrow) in the thoracic column. The lower thoracic
and lumbar vertebrae are widely open posteriorly and their
ossication irregular. One rib on the left and four on the right
are absent. Note the abnormal angulation and fusion of the ribs. C
Additional anomalies in another ETU-exposed fetus include
abnormal bony bridges between the ribs (white arrows) and free
rib segment (black arrow) (TV thoracic vertebrae, LV lumbar
vertebrae, SV sacral vertebrae)

Fig. 2AC Posterior view of the lumbosacral vertebrae and lower

limbs. A Control fetus: Showing normal lumbosacral ossication.
The three caudal vertebrae are starting to ossify. The ilium,
ischium, pubis, and long bones of the lower limbs are well formed.
B ETU-exposed fetus: Abnormal ossication of the lumbar and
sacral vertebrae and lumbosacral rachischisis. The lower sacral and
caudal vertebrae are absent. C Additional anomalies in another
ETU-exposed fetus include hypoplasia of the right femur (arrow)
and agenesis of right bula (LV lumbar vertebrae, SV sacral
vertebrae, CV caudal vertebrae, Pub pubis, Isch ischium, Fem
femur, Tib tibula, Fib bula, MT metatarsal)

and lumbar vertebrae were relatively normal in most of

them. The upper two sacral vertebrae were smaller than
normal. The lower two sacral and caudal vertebrae were
absent in 37 fetuses. In 35 of 39 (90%) fetuses without
rachischisis, only minor rib abnormalities were recorded,
mostly at the 6th10th thoracic levels. All fetuses had
one to three missing oating ribs. No obvious abnormalities were observed in the long bones of the limbs.

Discussion
Vertebral anomalies, especially of the lumbosacral
area, are known to be associated with ARMs. In the
human, the incidence of coexisting lumbosacral defects
ranges from 17% in simple ARMs to 46% in complex
ARMs, such as cloacal malformations [13]. Our study
shows that the ETU animal model of ARMs consistently produces sacral deciency but also causes other

532

Fig. 3 Posterior view of the right upper limbs of control (upper)

and ETU-exposed (lower) fetuses. The ETU-exposed fetus had
radial agenesis and abnormal angulation of the elbow joint (arrow)
(Scap scapula, Hum humerus, Rad radius, Uln ulna, MC
metacarpal)

vertebral anomalies and spinal defects. Fetuses with

both ARMs and rachischisis have more severe and
extensive vertebral anomalies as well as other axial
skeletal abnormalities. When ARMs occur without
rachischisis, the associated skeletal anomalies include
rib abnormalities and sacral agenesis or absent sacral
segments. Sacral defects may aect normal physiological function (i.e., level of continence) but are not lifethreatening.
However, the coexistence of ARMs and rachischisis
has more severe implications. In ETU-exposed fetal rats,
these vertebral and neural tube defects almost always
occurred in the lumbosacral region, suggesting that the
caudal segment of the embryos is more vulnerable, although some severe vertebral anomalies also occurred in
the thoracic area. The same developmental phenomenon

may be the basis of spinal bida occulta, buttery vertebrae, or abnormal angulation of the vertebral column
seen clinically. This model highlights the possibility of
concurrence of signicant thoracic anomalies with
ARMs in addition to the common lumbosacral anomalies.
The concurrence of ARMs and congenital vertebral
anomalies is not surprising, given their embryological
origin. In the 3rd week of embryogenesis, a middle layer
of mesoderm forms between the dorsal ectoderm and
ventral endoderm. However, there is no mesoderm in the
areas of the pharynx, cloacal membranes, and notochord lying in the midline at this stage [14]. The mesoderm immediately lateral to the notochord develops into
the paraxial mesoderm and condenses into paired segments, the somites that will dierentiate into three distinct areas: the dermatome, myotome, and sclerotome.
The sclerotome grows medially, ventrally, and dorsally
to form the vertebral mesenchymal anlages [15]. The
mesenchymal cells from the medial sclerotome around
the notochord form the centrum and most of the vertebral body. The dorsal mesenchymal cells form the
neural arches. The ventrolateral mesenchyme develops
into the ribs in the thorax, transverse processes of the
lumbar spine, and alae of the sacrum [14]. These different parts of the mesenchyme eventually fuse and ossify to form the bones of the vertebrae and ribs. Any
teratogen or other factors that disturb the mesenchymal
formation of the vertebral elements and the processes of
their fusion and ossication produce congenital anomalies of the vertebral column and adjacent axial skeletons, such as the ribs.
There is accumulating evidence that the notochord
serves as a key central organizer during early organogenesis. Vertebral abnormalities and some congenital
gastrointestinal malformations follow abnormal development of the notochord [16, 17, 18]. In the Adriamycin-induced rat model of esophageal atresia, the
notochord is abnormally located (e.g., anterior deviation) or branched and remains in close contact with the
foregut. It is located away from the medial mesenchyme for the centrum [19, 20]. A variety of foregut

Table 1 Main features of skeletal anomalies of ETU-exposed fetuses

Thoracolumbar vertebrae
Canal stenosis
Canal interruption
Ossication center for centrum
Lateral ossifying centres
Ribs
Pairs of ribs
Floating ribs
General features
Sacral vertebrae
Free caudal vertebrae
Upper limb
Lower limb

Control
(N=8)

ETU with ARMs only

(N=39)

ETU ARMs and rachschisis

(N=28)

0
0
1
2

0
0
1
2

5
3
2 or absent
Variable/abnormal fusion

13
4
Uniform
4
34
Normal
Normal

1112
13
Branching, angulation
4
Absent
Normal
Normal

811
0
Branching, fusion, angulation
Lower 2 missing, widely open posteriorly
Absent
Radial agenesis (4/28)
Femur/bular agenesis (2/28)

533

malformations (including esophageal atresia) and cervicothoracic vertebral anomalies ensue [2, 21].
The situation appears to be similar in the hindgut. In
the rat model of ARMs induced by ethylenethiourea,
lumbosacral defects in conjunction with rachischisis and
tail abnormalities have been observed in more than half
the fetuses. The notochord is abnormal in the lumbosacral area (e.g., ventral deviation, ventrodorsal branching,
and ectopic notochordal tissue) as in the foregut [19, 20].
The notochord is in abnormally close contact with the
posterior wall of the cloaca or neural tube [22]. Observations from both Adriamycin-induced and ETU-induced rat models suggest that an abnormal notochord
may lead to the development of foregut malformations
and cervicothoracic vertebral defects at its rostral end, or
hindgut malformations and lumbosacral vertebral
anomalies at its caudal end. Persistence and overgrowth
of the notochord may prevent anterior fusion of chondrication centers of the medial sclerotome, producing
buttery vertebrae, hemivertebral fusion anomalies,
block vertebrae and congenital absence of vertebrae
[16]. It is now known that the notochord, as an organizer
of axial development, releases Sonic hedgehog protein,
an important signal that controls the development,
growth, and dierentiation of the foregut, hindgut, and
the vertebra by activating its downstream genes, such as
Gli2, Gli3, Hoxa-13, Hoxd-13, and Bmp4 [23, 24, 25, 26].
Orford et al. have demonstrated that abnormal localization of the notochord in Adriamycin-exposed rat embryos is also accompanied by downregulation of the Shh
activity at the site of tracheo-esophageal separation [20].
It is possible that ethylenethiourea may also induce
malformations of the notochord, hindgut, vertebrae, and
caudal neural tube by interfering with the Shh-Gli or
Shh-Hox gene pathways.

References
1. Heij HA, Nievelstein RAJ, Zwart I, Verbeeten BWJM, Valk J,
Vos A (1996) Abnormal anatomy of the lumbosacral region
imaged by magnetic resonance in children with anorectal malformations. Arch Dis Child 74:441444
2. Xia HM, Migliazza L, Montedonico S, Rodriguez JI, Diezpardo JA, Tovar JA (1999) Skeletal malformations associated
with esophageal atresia: clinical and experimental studies. J
Pediatr Surg 34:13851392
3. Merei J, Hasthorpe S, Farmer P, Hutson J (1998) Relationship
between oesophageal atresia with tracheoesophageal stula and
vertebral anomalies in mammalian embryos. J Pediatr Surg
33:5863
4. Loder RT, Dayioglu MM (1990) Association of congenital
vertebral malformations with bladder and cloacal exstrophy. J
Pediatr Orthop 10:389393
5. Quan L, Smith A (1972) The VATER association: vertebral
defects, anal atresia, tracheoesophageal stula with esophageal
atresia, radial dysplasia. Birth Defects 8:7578

6. Botto LD, Khoury MJ, Mastroiacovo P, Castilla EE, Moore

CA, Skjaerven R (1997) The spectrum of congenital anomalies
of the VATER association: an international study. Am J Med
Genet 71:815
7. Currarino G, Coln D, Votteler T (1981) Triad of anorectal,
sacral and presacral anomalies. AJR Am J Roentgenol
137:395398
8. Tsuchida Y, Watanasupt W, Nakajo N (1989) Anorectal malformations associated with a presacral tumor and sacral defects. Pediatr Surg Int 4:398402
9. Lee SC, Chun YS, Jung SE, Park KW, Kim WK (1997) Currarino triad: anorectal malformations, sacral bony abnormalities, and presacral massa review of 11 cases. J Pediatr Surg
32:5861
10. Dias MS, Azizkhan RG (1998) A novel embryogenetic mechanism for Currarino triad: inadequate dorsoventral separation
of the caudal eminence from hindgut endoderm. Pediatr Neurosurg 28:223229
11. Inouye M (1976) Dierential staining of cartilage and bone in
fetal mouse skeleton by alcian blue and alizarin res S. Cong
Anom 16:171173
12. Wassersug RJ (1976) A procedure for dierential staining of
cartilage and bone in whole formaline-xed vertebrates. Stain
Technology 51:131134
13. Kiely EM, Pena A (1998) Anorectal malformations. In: ONeill
JA, Rowe MI, Grosfeld JL, Fonkalsrud EW, Coran AG (eds)
Pediatric surgery. Mosby, St. Louis, 5th edn, vol II, pp 1425
1460
14. Lonstein JE (1983) Spinal embryology. In: Winter RB (ed)
Congenital deformities of the spine. Thieme Stratton, New
York, pp 110
15. Tsou PM (1977) Embryology of congenital kyphosis. Clin
Orthop Rel Res 128:1825
16. Elliott GB, Tredwell SJ, Elliott KA (1970) The notochord as an
abnormal organizer in production of congenital intestinal defect. Am J Roentgenol Radium Ther Nucl Med 110:628634
17. Faris JC, Crowe JE (1975) The split notochord syndrome. J
Pediatr Surg 10:467472
18. Bremer JL (1952) Dorsal intestinal stula: accessory neuroenteric canal; diastematomyelia. Arch Pathol 54:132
19. Qi BQ, Beasley SW (1999) Relationship of the notochord to
foregut development in the fetal rat model of esophageal
atresia. J Pediatr Surg 34:15931598
20. Orford J, Manglick P, Cass DT, Tam PP (2001) Mechanisms
for the development of esophageal atresia. J Pediatr Surg
36:985994
21. Abu-Hijleh G, Qi BQ, Williams A, Beasley SW (2000) Development of the bones and synovial joints in the rat model of the
VATER association. J Orthop Sci 5:391397
22. Qi BQ, Beasley SW: Evidence that the notochord may be pivotal in the development of sacral and anorectal malformations.
J Pediatr Surg [submitted]
23. Litingtung Y, Lei L, Westphal H, Chiang C (1998) Sonic
hedgehog is essential to foregut development. Nature Gen
20:5861
24. Riddle R, Johnson RL, Laufer E, Tabin C (1993) Sonic
hedgehog mediates the polarising activity of the ZPA. Cell
75:14011416
25. Kimmel SG, Mo R, Hui CC, Kim PCW (2000) New mouse
models of congenital anorectal malformations. J Pediatr Surg
35:227231
26. Kim PCW, Mo R, Hui CC (2001) Murine models of VACTERL syndrome: role of Sonic hedgehog signalling pathway. J
Pediatr Surg 36:381384