- Email: [email protected]
Factors influencing lumbar spine bone mineral density assessment by dual-energy X-ray absorptiometry: Comparison with lumbar spinal radiogram
J Orthop Sci (2000) 5:390–396
Development of the bones and synovial joints in the rat model of the VATER association Ghassan Abu-Hijleh, Bao-Quan Qi, Andrew K. Williams, and Spencer W. Beasley Department of Paediatric Surgery, Christchurch Hospital, Private Bag 4710, Christchurch, New Zealand
Abstract The adriamycin-induced rat model of the Vertebral, Anorectal, Tracheo-Esophageal, Radial and Renal (VATER) association produces a variety of vertebral, rib, and limb abnormalities. This study was designed to document accurately the nature of these abnormalities and to determine whether synovial joints are affected. Fetuses from pregnant Sprague Dawley rats that had received intraperitoneal injections of 1.75 mg/kg of adriamycin on days 6–9 or 10–13 of gestation were harvested. Double-stained skeletal preparations and histological sections were examined for vertebral, rib, and limb anomalies. The incidence of anomalies was high in the group treated on gestational days (GD) 6–9, while it was low in the GD 10–13 group. The length and thickness of the long bones were reduced, with bowing and reduction in their endochondral ossification. Sirenomelia occurred in the group treated on GD 6–9, and was often associated with a short tail and anal atresia. The joint cavities, and intra-articular structures such as menisci and the cruciate ligaments developed normally from the mesenchymal interzone. These data indicate that adriamycin inhibits skeletal growth and differentiation without any interference in the differentiation of the mesenchymal interzone, thus producing normal synovial joints.
esophageal atresia (EA), vertebral defects, congenital heart disease, radial aplasia, anal atresia, and urinary tract anomalies.15,20,21,30 The range of anomalies is similar to that seen in the human infant with the Vertebral, Anorectal, Tracheo-Esophageal, Radial and Renal (VATER) association,10,22,28 for which reason the adriamycin-exposed prenatal rat is considered a suitable model to study the VATER association.13,15,19–20 The type and severity of abnormality depends on the dose and gestational age at the time of adriamycin injection.19,29 Although a number of bony defects, including radial, vertebral, and rib anomalies, have been described,13,15 the nature and pattern of involvement of joints in relation to adriamycin-induced limb malformations are not known. The present study was designed to document accurately the range of bony defects and to examine the structure of the joints of rat fetuses exposed to adriamycin.
Materials and methods Key words Adriamycin · Rat · Embryo · VATER association · Synovial joint · Bones · Limbs · Vertebra · Sirenomelia
Introduction Adriamycin (doxorubicin) has significant articular activity against a variety of malignancies including sarcomas, lymphomas, leukemia, neuroblastoma, and breast cancer.26 It is also known to induce a variety of congenital malformations in rats,13,15,19,20,29 including
Offprint requests to: S.W. Beasley Received: November 11, 1999 / Accepted: February 16, 2000
Female Sprague-Dawley rats of about 300 g body weight were mated overnight. The presence of a vaginal plug, identified on the following morning, was designated day 0 of gestation. The rats were housed in a temperature-controlled (21 6 1°C) room with a 12 : 12h light : dark cycle. Laboratory chow (Animal Laboratory Christchurch School of Medicine Christchurch, New Zealand) and tap water were provided ad libitum. A daily dose of 1.75 mg/kg of adriamycin in sterile saline (0.5 mg/ml) was injected intraperitoneally into six rats on each of days 6–9 of gestation. A second group of five rats were injected intraperitoneally with the same dose per kg on days 10–13 of gestation. Control rats were injected with comparable volumes of saline only, on days 6–9 and 10–13. The rats were killed on day 20 of gestation. The numbers of resorbed and live fetuses
G. A.-Hijleh et al.: Bones and joints in the VATER Association
391
were recorded. A total of 59 experimental (11 litters) and 31 control (3 litters) fetuses were collected, blotted, weighed individually, and then fixed in either 10% formalin for histological study, or buffered formalin for a double stain with alcian blue and alizarin red-S. All fetuses were examined under a dissecting microscope for external malformations. Half of each litter was used for a double stain, and the remainder for histological study. The skeletons of buffered formalin-fixed fetuses were stained with alcian blue and alizarin red-S, cleared of soft tissues in 0.5% potassium hydroxide, dehydrated in graded glycerol, and stored in 100% glycerol.12 Both fore and hindlimbs were amputated, at the shoulder and hip joints, respectively, embedded in paraffin wax, and serially sectioned in the sagittal plane at 5-µm intervals. The sections were stained with hematoxylin and eosin and examined by brightfield microscopy. All studies had approval from the Christchurch School of Medicine Ethical Committee.
Results The incidence of resorption was 75.5% in the group treated on gestational day (GD) 6 to 9, while it was 30.5% in the experimental group injected on GD 10 to 13, and only 2% in the control animals. Embryos treated on GD 10–13 appeared normal externally, while the group treated on GD 6–9 had external malformations, including severe edema, short limbs, short or rudimentary tail (pigtailed) (Fig. 1A,B), anal atresia, hemorrhagic spots along the vertebral column covered by the skin only (Fig. 1C) especially in the lumbosacral region, and sirenomelia (fusion of the hindlimbs) (Fig. 1A,B). The control skeleton Axial and appendicular skeletal development was normal. The vertebral column was differentiated into 7 cervical, 13 thoracic, 6 lumbar, 4 sacral, and 30 caudal vertebrae. There were 13 pairs of ribs, the last 3 being short and floating. The humerus had a great deltoid tuberosity. There was a third trochanter on the postaxial side of the femur, in addition to the greater and lesser trochanters. The histology of serial sections of both fore and hindlimbs revealed the development of distinct joint cavities in all synovial joints (Fig. 2A,E). The articular surfaces of the long bones were covered by multiple layers of flattened cells arranged tangential to the surface. The joint cavities had a synovial lining, and in some areas, synovial folds projected into the cavity. A vascular network lay immediately peripheral to the synovial cells. In the knee joint, a few capillaries
Fig. 1A–D. Experimental fetuses (treated on gestational day [GD] 6–9), collected on GD 20, showing severe edema, short limbs, and rudimentary tail (A and B), hemorrhagic spots under the skin (arrows in C), and sirenomelia with malrotation of both feet (B). The double-stained fetus seen in D showed that the fusion between the hindlimbs was by soft tissue only
penetrated the peripheral regions of the menisci and the cruciate ligaments. The menisci had chondrogenic cells in their central parts and fine collagenous fibers in their superficial layers. The intra-articular ligaments were dense fibrous structures, while the femoropatellar cavity and supra-patellar recess were well proportioned and continuous. The patella was penetrated by blood vessels from its anterior and superior surfaces. The shafts of the humerus, ulna, femur, and tibia had radiating bony trabeculae extending approximately to the level of the epiphysis. Ossification in the radius, and fibula, and in the metacarpal and metatarsal bones was confined to the mid-shaft. The experimental skeleton The double-stained embryos revealed multiple vertebral anomalies (Fig. 3B–D). The incidence of
392
G. A.-Hijleh et al.: Bones and joints in the VATER Association
Fig. 2A–F. Sagittal sections of some synovial joints in both fore and hindlimbs of GD 20 fetuses. Control fetuses (A, E), as well as the experimental fetuses (B, C, D, and F) showed normal development of the synovial joints. The knee joint is shown in A–C, the ankle joint in D, the elbow joint in E, and the wrist joint in F. The joint cavities are well formed in all the joints (arrows in A–F). F, Femur; T, tibia; f, fibula; C, cruciate ligament; m, meniscus; P, patella; U, ulna; R, radius; t, talus; CAL, calcaneus; Ca, carpal; MCa, metacarpal bones. A and C–F 340; B 3100
vertebral anomalies was 58% in the GD 6–9 group and 23% in the group treated on GD 10–13 (Table 1). The vertebrae showed anomalies that included absence of the vertebral bodies (Fig. 3D), especially the thoracic vertebrae, and hemi- and butterfly (Fig. 3B) vertebrae. Others had clefting or absence of the vertebral centra (Fig. 3B,D). Open vertebral arches were seen in the lumbosacral region (Fig. 4B,C). Embryos with open vertebral arches had hemorrhagic spots (spina bifida occulta) under the overlying skin (Fig. 1C). Embryos treated on GD 6–9 had a 30% incidence of rib anomalies, compared with a 10% incidence in the GD 10–13 fetuses (Table 1). The ribs were irregular or angulated in shape (Fig. 3C), or there were accessory unilateral lumbar ribs attached to the first lumbar vertebra (Fig. 3D). Tail malformations were present in 85% of embryos treated on GD 6–9. The short or rudimentary tail (Fig. 1A,B) was usually associated with anal agenesis and sirenomelia. Fetuses treated on
Table 1. Incidence of the major skeletal abnormalities in adriamycin-exposed fetuses Period of exposure to adriamycin Gestational days 6–9 Gestational days 10–13
Vertebra (%)
Rib (%)
Limb (%)
Tail (%)
58
30
36
85
23
10
10
-
GD 10–13 had a short tail, but no obvious tail malformations. Examination of cleared limb skeletons under a dissection microscope revealed shortening, bowing, and deformity of the long bones (humerus, radius, ulna, femur, tibia, and fibula) (Fig. 5B,D). The length of ossified segments was markedly reduced, but the cartilaginous ends of the long bones were normal. The
G. A.-Hijleh et al.: Bones and joints in the VATER Association
393
incidence of limb anomalies was 36% in the GD 6–9 group and 10% in the GD 10–13 group (Table 1). Sirenomelia, the most dramatic limb abnormality in the GD 6–9 group, ranged from soft-tissue fusion between the proximal segments to fusion of the entire length of the hindlimbs, including the feet, which were malrotated. The length and thickness of the long bones were reduced (Fig. 6B–D). Endochondral ossification was markedly reduced, and, thus, most of the shaft was cartilaginous (Fig. 6E). The articular surfaces of the synovial joints were normal, and the intra-articular structures (menisci and cruciate ligaments) appeared normal, but were poorly vascularized, as was the synovial membrane (Fig. 2B–D,F). No apoptosis was seen. Some fetuses had syndactyly (Fig. 6F).
Discussion
Fig. 3A–D. Double-stained skeletons of GD 20 fetuses. A Control and B–D experimental fetuses. Note missing or clefting of the vertebral centra (B, D), butterfly vertebrae (arrows in B), or absence of vertebral bodies (D). Rib anomalies are shown (C, D) as wavy and angulated, as in (C), or an accessory unilateral lumbar rib (arrow in D)
This study confirms that adriamycin injected into Sprague-Dawley rats on GD 6–9 causes skeletal malformations in a high percentage of embryos (Table 1), consistent with findings in previous studies.13 Skeletal anomalies occurred less frequently when the fetuses were exposed to adriamycin during GD 10–13. Sirenomelia has not been documented previously in studies of skeletal abnormalities in this animal model, except by Qi et al.,19 who observed it in two fetuses. The present study reveals that sirenomelia is caused by softtissue connection between the two hindlimbs, and not by skeletal union. The skeletal abnormalities are strikingly similar to those seen in infants with the VATER association.5,21,27 Whereas retinoic acid administered on GD 12 in the mouse consistently causes limb malformations, including abnormalities of the joints, in a high percentage of
Fig. 4A–C. Double-stained skeletons of GD 20 fetuses. A Control and B, C experimental fetuses. Note open vertebral arches in the sacral region (arrows in B) or along the lumbosacral region (arrows in C)
394
G. A.-Hijleh et al.: Bones and joints in the VATER Association
Fig. 5A–D. Double-stained skeletons of forelimbs (A, B) and hindlimbs (C, D). Note that control limbs are shown in A and C, while the limbs of experimental fetuses are shown in B and D. The experimental fetuses showed shortness and bowing of the humerus, radius, and ulna (B) and femur, tibia, and fibula (D)
embryos,1 the skeletal effect of adriamycin is restricted to the vertebra, ribs, and long bones; the synovial joints, including the joint cavities, menisci, and cruciate ligaments, develop normally. The axial development of the embryos is prone to complex malformation,4 including neural tube defects, oral and facial clefts at the rostral end, and sirenomelia and anorectal abnormalities at the caudal end. The caudal eminence is an area of active cell proliferation and differentiation, and of morphogenetic tissue interaction.22,23 It provides mesenchyme for the notochord, somites (including caudal vertebrae), hindgut, neural tube, hindlimbs, and blood vessels.18 The cluster of developmental abnormalities that arises from faulty differentiation or regression of this region constitutes caudal regression syndrome (CRS). Our Sprague-Dawley rat model includes many of the features of CRS.9,16 Alles and Sulik4 have highlighted the association of Potter sequence, VATER or Vertebral, Anorectal, Cardiac, Tracheo-Esophageal, Radial and Renal, Limb (VACTERL) association and Omphalocele, Exstrophy, Imperforate anus, Spinal
defects (OEIS) complex with CRS. Past theories of the pathogenesis of these abnormalities include: (1) lateral compression by amniotic folds (mechanical);18,24 (2) defective and/or deficient caudal mesoderm;18 (3) defective tissue interaction;7 (4) vascular steal;14,24 (5) dilatation of the neural tube;11 and (6) a combination of vascular disruption, mesodermal injury, and defective microperfusion.14 Alternatively, aberrations of the timing or location of apoptosis (programmed cell death) may contribute to dysmorphogenesis of the embryonic structures.3,29 Abnormal embryos induced by retinoic acid have accentuated apoptosis in the limb buds.25 Histological sections of our adriamycin-treated embryos did not show excessive apoptosis in the mesenchymal interzone areas, possibly because of the late stage (gestational day 20) at which the rats were examined. Myelocystocele, spina bifida aperta, sympodia, and skeletal anomalies in hamsters, and axial skeletal defects associated with vascular lesions in the mouse and chick are reported to be induced by exposure to retinoic acid.28 Vascular disruption and consequent hypoxia, edema, and nutritional deficiency is now fairly
G. A.-Hijleh et al.: Bones and joints in the VATER Association
395
Fig. 6A–F. Serial sections of fore and hindlimb bones of GD 20 fetuses. Note the well developed ossification in control tibia (A), and delay of ossification and bowing of the shaft of the experimental tibia (B), fibula (C), and radius (D). Absence of ossification in the shaft of the metatarsal bone (E) was noted. Delay in clefting (syndactyly) of the interdigital space (arrows in F) was also noted. T, Tibia; f, fibula; R, radius; Mta, metatarsal A–D and F 340; E 3100
well established as a teratogenic mechanism.6,8 The observations of short tail, hemorrhagic spots over the lumbosacral region, and sirenomelia in a group treated on GD 6–9 in our study may be the result of a direct effect of adriamycin or may have occurred via events involving a combination of vascular disruption and mesodermal injury. The avascular mesenchymal interzone differentiates into a three-layered structure in which the outer layers become the articular surfaces and the intermediate layer gives rise to the intra-articular structures and joint cavity.2,10 The vascular disruption with mesodermal injury could be a plausible hypothesis to explain the malformations seen in the present study. This may explain why the joint cavity, intra-articular structures, and articular surfaces of the long bones escaped any abnormalities — because they arise from avascular mesenchymal tissue. The normal development of the synovial joints in the treated embryos suggests that there is no abnormal differentiation of the interzone mesenchyme. Well-developed joint cavities
and intra-articular structures imply complete differentiation of the avascular mesenchymal interzone, a concept that is in agreement with the observations in large numbers of VATER association patients who have no documented deformity or limitation of movements in the synovial joints. Acknowledgments. The authors sincerely thank the Robert McClelland Trust for supporting this work.
References 1. Abu-Hijleh G, Padmanabhan R. Retinoic acid induced abnormal development of hindlimb joints in the mouse. Eur J Morphol 1997;35:327–36. 2. Abu-Hijleh G, Reid O, Scothorne RJ. Cell death in the developing chick knee joint: 1. Spatial and temporal patterns. Clin Anat 1997;10:183–200. 3. Alles AJ, Sulik KK. Retinoic acid induced spina bifida: evidence for a pathogenetic mechanism. Development 1990;408:73–81.
396 4. Alles AJ, Sulik KK. A review of caudal dysgenesis and its pathogenesis as illustrated in an animal model. Birth Defects 1993;29:83–102. 5. Barnes JC, Smith WL. The VATER association. Radiology 1978;126:445–9. 6. Bouwes JNB, Weaver DD. Subclavian artery supply disruption sequence: hypothesis of a vascular etiology for Poland, KlippelFeil, and Mobius anomalies. Am J Med Genet 1986;23:903–18. 7. Chandebois R, Brunet C. Origin of abnormality in a human symmelian foetus as elucidated by our knowledge of vertebrate development. Teratology 1987;36:11–22. 8. Danielson MK, Danielsson BRG, Marchner H, et al. Histopathological and hemodynamic studies supporting hypoxia and vascular disruption as explanation to phenytoin teratogenicity. Teratology 1992;46:485–97. 9. Duhamel B. From mermaid to anal imperforation: the syndrome of caudal regression. Arch Dis Child 1961;36:152–5. 10. Fernbach SK, Glass RBJ. The expanded spectrum of limb anomalies in the VATER association. Pediatr Radiol 1988;18: 215–20. 11. Gardner WJ. Hypothesis: overdistension of the neural tube may cause anomalies of nonneural organs. Teratology 1980;22:229–38. 12. Inouye M. Differential staining of cartilage and bone in fetal mouse skeleton by alcian blue and alizarin red-S. Cong Anom 1976;16:171–3. 13. Kotsios C, Merei J, Hutson JM, et al. Skeletal anomalies in the adriamycin-exposed prenatal rat: a model for VATER association. J Orthop Res 1998;16:50–3. 14. McCoy MC, Chescheir NC, Kuller JA, et al. A fetus with sirenomelia, omphalocele, and meningomyelocele, but with normal kidneys. Teratology 1994;50:168–71. 15. Merei J, Hasthorpe S, Farmer P, et al. Relationship between esophageal atresia with tracheoesophageal fistula and vertebral anomalies in mammalian embryos. J Pediatr Surg 1998;33:58–63. 16. Murphy JJ, Fraser GC, Blair GK. Sirenomelia: case of the surviving mermaid. J Pediatr Surg 1992;27:1265–8. 17. O’Rahilly R, Muller F. Interpretation of some median anomalies as illustrated by cyclopia and symmelia. Teratology 1989;40:409– 21.
G. A.-Hijleh et al.: Bones and joints in the VATER Association 18. Orr BY, Long SY, Steffek AI. Craniofacial, caudal and visceral anomalies associated with mutant sirenomelic mice. Teratology 1982;26:311–7. 19. Qi B, Diez-Pardo JA, Navarro C, et al. Narrowing the embryologic window of the adriamycin-induced fetal rat model of esophageal atresia and tracheoesophageal fistula. Pediatr Surg Int 1996;11:444–7. 20. Qi B, Merei J, Farmer P, et al. Cardiovascular malformations in rat fetuses with esophageal atresia and tracheoesophageal fistula induced by adriamycin. Pediatr Surg Int 1997;12:556–64. 21. Quan L, Smith DW. The VATER association: vertebral defects, anal atresia, T-E fistula with esophageal atresia, radial and renal dysplasia: a spectrum of associated defects. J Pediatr 1973;82:104– 7. 22. Schoenwolf GC, Garcia-Martinez V, Dias MS. Mesoderm movement and fate during avian gastrulation and neurulation. Dev Dyn 1992;193:235–48. 23. Smith JL, Gestland KM, Schoenwolf GC. Prospective fate map of the mouse primitive streak at 7.5 days of gestation. Dev Dyn 1994;201:279–89. 24. Stevenson RE, Jones KL, Phelan MC, et al. Vascular steal: the pathogenetic mechanism producing sirenomelia and associated defects of the viscera and soft tissues. Pediatrics 1986;78:451– 7. 25. Sulik K, Dehart DB. Retinoic acid-induced limb malformations resulting from apical ectodermal ridge cell death. Teratology 1988;37:527–37. 26. Tan C, Etcubanas E, Wollner N, et al. Adriamycin — an antitumor antibiotic in the treatment of neoplastic diseases. Cancer 1973;32:9–17. 27. Temtamy SA, Miller JD. Extending the scope of the VATER association: definition of the VATER syndrome. J Pediatr 1974; 85:345–9. 28. Tibbles L, Wiley MJ. A comparative study of the effects of retinoic acid given during the critical period for induction of spina bifida in mice and hamsters. Teratology 1988;37:113–25. 29. Zhou B, Hutson J, Farmer P, et al. Apoptosis in tracheoesophageal embryogenesis in rat embryos with or without adriamycin treatment. J Pediatr Surg 1999;24:872–6.
Development of the bones and synovial joints in the rat model of the VATER association Ghassan Abu-Hijleh, Bao-Quan Qi, Andrew K. Williams, and Spencer W. Beasley Department of Paediatric Surgery, Christchurch Hospital, Private Bag 4710, Christchurch, New Zealand
Abstract The adriamycin-induced rat model of the Vertebral, Anorectal, Tracheo-Esophageal, Radial and Renal (VATER) association produces a variety of vertebral, rib, and limb abnormalities. This study was designed to document accurately the nature of these abnormalities and to determine whether synovial joints are affected. Fetuses from pregnant Sprague Dawley rats that had received intraperitoneal injections of 1.75 mg/kg of adriamycin on days 6–9 or 10–13 of gestation were harvested. Double-stained skeletal preparations and histological sections were examined for vertebral, rib, and limb anomalies. The incidence of anomalies was high in the group treated on gestational days (GD) 6–9, while it was low in the GD 10–13 group. The length and thickness of the long bones were reduced, with bowing and reduction in their endochondral ossification. Sirenomelia occurred in the group treated on GD 6–9, and was often associated with a short tail and anal atresia. The joint cavities, and intra-articular structures such as menisci and the cruciate ligaments developed normally from the mesenchymal interzone. These data indicate that adriamycin inhibits skeletal growth and differentiation without any interference in the differentiation of the mesenchymal interzone, thus producing normal synovial joints.
esophageal atresia (EA), vertebral defects, congenital heart disease, radial aplasia, anal atresia, and urinary tract anomalies.15,20,21,30 The range of anomalies is similar to that seen in the human infant with the Vertebral, Anorectal, Tracheo-Esophageal, Radial and Renal (VATER) association,10,22,28 for which reason the adriamycin-exposed prenatal rat is considered a suitable model to study the VATER association.13,15,19–20 The type and severity of abnormality depends on the dose and gestational age at the time of adriamycin injection.19,29 Although a number of bony defects, including radial, vertebral, and rib anomalies, have been described,13,15 the nature and pattern of involvement of joints in relation to adriamycin-induced limb malformations are not known. The present study was designed to document accurately the range of bony defects and to examine the structure of the joints of rat fetuses exposed to adriamycin.
Materials and methods Key words Adriamycin · Rat · Embryo · VATER association · Synovial joint · Bones · Limbs · Vertebra · Sirenomelia
Introduction Adriamycin (doxorubicin) has significant articular activity against a variety of malignancies including sarcomas, lymphomas, leukemia, neuroblastoma, and breast cancer.26 It is also known to induce a variety of congenital malformations in rats,13,15,19,20,29 including
Offprint requests to: S.W. Beasley Received: November 11, 1999 / Accepted: February 16, 2000
Female Sprague-Dawley rats of about 300 g body weight were mated overnight. The presence of a vaginal plug, identified on the following morning, was designated day 0 of gestation. The rats were housed in a temperature-controlled (21 6 1°C) room with a 12 : 12h light : dark cycle. Laboratory chow (Animal Laboratory Christchurch School of Medicine Christchurch, New Zealand) and tap water were provided ad libitum. A daily dose of 1.75 mg/kg of adriamycin in sterile saline (0.5 mg/ml) was injected intraperitoneally into six rats on each of days 6–9 of gestation. A second group of five rats were injected intraperitoneally with the same dose per kg on days 10–13 of gestation. Control rats were injected with comparable volumes of saline only, on days 6–9 and 10–13. The rats were killed on day 20 of gestation. The numbers of resorbed and live fetuses
G. A.-Hijleh et al.: Bones and joints in the VATER Association
391
were recorded. A total of 59 experimental (11 litters) and 31 control (3 litters) fetuses were collected, blotted, weighed individually, and then fixed in either 10% formalin for histological study, or buffered formalin for a double stain with alcian blue and alizarin red-S. All fetuses were examined under a dissecting microscope for external malformations. Half of each litter was used for a double stain, and the remainder for histological study. The skeletons of buffered formalin-fixed fetuses were stained with alcian blue and alizarin red-S, cleared of soft tissues in 0.5% potassium hydroxide, dehydrated in graded glycerol, and stored in 100% glycerol.12 Both fore and hindlimbs were amputated, at the shoulder and hip joints, respectively, embedded in paraffin wax, and serially sectioned in the sagittal plane at 5-µm intervals. The sections were stained with hematoxylin and eosin and examined by brightfield microscopy. All studies had approval from the Christchurch School of Medicine Ethical Committee.
Results The incidence of resorption was 75.5% in the group treated on gestational day (GD) 6 to 9, while it was 30.5% in the experimental group injected on GD 10 to 13, and only 2% in the control animals. Embryos treated on GD 10–13 appeared normal externally, while the group treated on GD 6–9 had external malformations, including severe edema, short limbs, short or rudimentary tail (pigtailed) (Fig. 1A,B), anal atresia, hemorrhagic spots along the vertebral column covered by the skin only (Fig. 1C) especially in the lumbosacral region, and sirenomelia (fusion of the hindlimbs) (Fig. 1A,B). The control skeleton Axial and appendicular skeletal development was normal. The vertebral column was differentiated into 7 cervical, 13 thoracic, 6 lumbar, 4 sacral, and 30 caudal vertebrae. There were 13 pairs of ribs, the last 3 being short and floating. The humerus had a great deltoid tuberosity. There was a third trochanter on the postaxial side of the femur, in addition to the greater and lesser trochanters. The histology of serial sections of both fore and hindlimbs revealed the development of distinct joint cavities in all synovial joints (Fig. 2A,E). The articular surfaces of the long bones were covered by multiple layers of flattened cells arranged tangential to the surface. The joint cavities had a synovial lining, and in some areas, synovial folds projected into the cavity. A vascular network lay immediately peripheral to the synovial cells. In the knee joint, a few capillaries
Fig. 1A–D. Experimental fetuses (treated on gestational day [GD] 6–9), collected on GD 20, showing severe edema, short limbs, and rudimentary tail (A and B), hemorrhagic spots under the skin (arrows in C), and sirenomelia with malrotation of both feet (B). The double-stained fetus seen in D showed that the fusion between the hindlimbs was by soft tissue only
penetrated the peripheral regions of the menisci and the cruciate ligaments. The menisci had chondrogenic cells in their central parts and fine collagenous fibers in their superficial layers. The intra-articular ligaments were dense fibrous structures, while the femoropatellar cavity and supra-patellar recess were well proportioned and continuous. The patella was penetrated by blood vessels from its anterior and superior surfaces. The shafts of the humerus, ulna, femur, and tibia had radiating bony trabeculae extending approximately to the level of the epiphysis. Ossification in the radius, and fibula, and in the metacarpal and metatarsal bones was confined to the mid-shaft. The experimental skeleton The double-stained embryos revealed multiple vertebral anomalies (Fig. 3B–D). The incidence of
392
G. A.-Hijleh et al.: Bones and joints in the VATER Association
Fig. 2A–F. Sagittal sections of some synovial joints in both fore and hindlimbs of GD 20 fetuses. Control fetuses (A, E), as well as the experimental fetuses (B, C, D, and F) showed normal development of the synovial joints. The knee joint is shown in A–C, the ankle joint in D, the elbow joint in E, and the wrist joint in F. The joint cavities are well formed in all the joints (arrows in A–F). F, Femur; T, tibia; f, fibula; C, cruciate ligament; m, meniscus; P, patella; U, ulna; R, radius; t, talus; CAL, calcaneus; Ca, carpal; MCa, metacarpal bones. A and C–F 340; B 3100
vertebral anomalies was 58% in the GD 6–9 group and 23% in the group treated on GD 10–13 (Table 1). The vertebrae showed anomalies that included absence of the vertebral bodies (Fig. 3D), especially the thoracic vertebrae, and hemi- and butterfly (Fig. 3B) vertebrae. Others had clefting or absence of the vertebral centra (Fig. 3B,D). Open vertebral arches were seen in the lumbosacral region (Fig. 4B,C). Embryos with open vertebral arches had hemorrhagic spots (spina bifida occulta) under the overlying skin (Fig. 1C). Embryos treated on GD 6–9 had a 30% incidence of rib anomalies, compared with a 10% incidence in the GD 10–13 fetuses (Table 1). The ribs were irregular or angulated in shape (Fig. 3C), or there were accessory unilateral lumbar ribs attached to the first lumbar vertebra (Fig. 3D). Tail malformations were present in 85% of embryos treated on GD 6–9. The short or rudimentary tail (Fig. 1A,B) was usually associated with anal agenesis and sirenomelia. Fetuses treated on
Table 1. Incidence of the major skeletal abnormalities in adriamycin-exposed fetuses Period of exposure to adriamycin Gestational days 6–9 Gestational days 10–13
Vertebra (%)
Rib (%)
Limb (%)
Tail (%)
58
30
36
85
23
10
10
-
GD 10–13 had a short tail, but no obvious tail malformations. Examination of cleared limb skeletons under a dissection microscope revealed shortening, bowing, and deformity of the long bones (humerus, radius, ulna, femur, tibia, and fibula) (Fig. 5B,D). The length of ossified segments was markedly reduced, but the cartilaginous ends of the long bones were normal. The
G. A.-Hijleh et al.: Bones and joints in the VATER Association
393
incidence of limb anomalies was 36% in the GD 6–9 group and 10% in the GD 10–13 group (Table 1). Sirenomelia, the most dramatic limb abnormality in the GD 6–9 group, ranged from soft-tissue fusion between the proximal segments to fusion of the entire length of the hindlimbs, including the feet, which were malrotated. The length and thickness of the long bones were reduced (Fig. 6B–D). Endochondral ossification was markedly reduced, and, thus, most of the shaft was cartilaginous (Fig. 6E). The articular surfaces of the synovial joints were normal, and the intra-articular structures (menisci and cruciate ligaments) appeared normal, but were poorly vascularized, as was the synovial membrane (Fig. 2B–D,F). No apoptosis was seen. Some fetuses had syndactyly (Fig. 6F).
Discussion
Fig. 3A–D. Double-stained skeletons of GD 20 fetuses. A Control and B–D experimental fetuses. Note missing or clefting of the vertebral centra (B, D), butterfly vertebrae (arrows in B), or absence of vertebral bodies (D). Rib anomalies are shown (C, D) as wavy and angulated, as in (C), or an accessory unilateral lumbar rib (arrow in D)
This study confirms that adriamycin injected into Sprague-Dawley rats on GD 6–9 causes skeletal malformations in a high percentage of embryos (Table 1), consistent with findings in previous studies.13 Skeletal anomalies occurred less frequently when the fetuses were exposed to adriamycin during GD 10–13. Sirenomelia has not been documented previously in studies of skeletal abnormalities in this animal model, except by Qi et al.,19 who observed it in two fetuses. The present study reveals that sirenomelia is caused by softtissue connection between the two hindlimbs, and not by skeletal union. The skeletal abnormalities are strikingly similar to those seen in infants with the VATER association.5,21,27 Whereas retinoic acid administered on GD 12 in the mouse consistently causes limb malformations, including abnormalities of the joints, in a high percentage of
Fig. 4A–C. Double-stained skeletons of GD 20 fetuses. A Control and B, C experimental fetuses. Note open vertebral arches in the sacral region (arrows in B) or along the lumbosacral region (arrows in C)
394
G. A.-Hijleh et al.: Bones and joints in the VATER Association
Fig. 5A–D. Double-stained skeletons of forelimbs (A, B) and hindlimbs (C, D). Note that control limbs are shown in A and C, while the limbs of experimental fetuses are shown in B and D. The experimental fetuses showed shortness and bowing of the humerus, radius, and ulna (B) and femur, tibia, and fibula (D)
embryos,1 the skeletal effect of adriamycin is restricted to the vertebra, ribs, and long bones; the synovial joints, including the joint cavities, menisci, and cruciate ligaments, develop normally. The axial development of the embryos is prone to complex malformation,4 including neural tube defects, oral and facial clefts at the rostral end, and sirenomelia and anorectal abnormalities at the caudal end. The caudal eminence is an area of active cell proliferation and differentiation, and of morphogenetic tissue interaction.22,23 It provides mesenchyme for the notochord, somites (including caudal vertebrae), hindgut, neural tube, hindlimbs, and blood vessels.18 The cluster of developmental abnormalities that arises from faulty differentiation or regression of this region constitutes caudal regression syndrome (CRS). Our Sprague-Dawley rat model includes many of the features of CRS.9,16 Alles and Sulik4 have highlighted the association of Potter sequence, VATER or Vertebral, Anorectal, Cardiac, Tracheo-Esophageal, Radial and Renal, Limb (VACTERL) association and Omphalocele, Exstrophy, Imperforate anus, Spinal
defects (OEIS) complex with CRS. Past theories of the pathogenesis of these abnormalities include: (1) lateral compression by amniotic folds (mechanical);18,24 (2) defective and/or deficient caudal mesoderm;18 (3) defective tissue interaction;7 (4) vascular steal;14,24 (5) dilatation of the neural tube;11 and (6) a combination of vascular disruption, mesodermal injury, and defective microperfusion.14 Alternatively, aberrations of the timing or location of apoptosis (programmed cell death) may contribute to dysmorphogenesis of the embryonic structures.3,29 Abnormal embryos induced by retinoic acid have accentuated apoptosis in the limb buds.25 Histological sections of our adriamycin-treated embryos did not show excessive apoptosis in the mesenchymal interzone areas, possibly because of the late stage (gestational day 20) at which the rats were examined. Myelocystocele, spina bifida aperta, sympodia, and skeletal anomalies in hamsters, and axial skeletal defects associated with vascular lesions in the mouse and chick are reported to be induced by exposure to retinoic acid.28 Vascular disruption and consequent hypoxia, edema, and nutritional deficiency is now fairly
G. A.-Hijleh et al.: Bones and joints in the VATER Association
395
Fig. 6A–F. Serial sections of fore and hindlimb bones of GD 20 fetuses. Note the well developed ossification in control tibia (A), and delay of ossification and bowing of the shaft of the experimental tibia (B), fibula (C), and radius (D). Absence of ossification in the shaft of the metatarsal bone (E) was noted. Delay in clefting (syndactyly) of the interdigital space (arrows in F) was also noted. T, Tibia; f, fibula; R, radius; Mta, metatarsal A–D and F 340; E 3100
well established as a teratogenic mechanism.6,8 The observations of short tail, hemorrhagic spots over the lumbosacral region, and sirenomelia in a group treated on GD 6–9 in our study may be the result of a direct effect of adriamycin or may have occurred via events involving a combination of vascular disruption and mesodermal injury. The avascular mesenchymal interzone differentiates into a three-layered structure in which the outer layers become the articular surfaces and the intermediate layer gives rise to the intra-articular structures and joint cavity.2,10 The vascular disruption with mesodermal injury could be a plausible hypothesis to explain the malformations seen in the present study. This may explain why the joint cavity, intra-articular structures, and articular surfaces of the long bones escaped any abnormalities — because they arise from avascular mesenchymal tissue. The normal development of the synovial joints in the treated embryos suggests that there is no abnormal differentiation of the interzone mesenchyme. Well-developed joint cavities
and intra-articular structures imply complete differentiation of the avascular mesenchymal interzone, a concept that is in agreement with the observations in large numbers of VATER association patients who have no documented deformity or limitation of movements in the synovial joints. Acknowledgments. The authors sincerely thank the Robert McClelland Trust for supporting this work.
References 1. Abu-Hijleh G, Padmanabhan R. Retinoic acid induced abnormal development of hindlimb joints in the mouse. Eur J Morphol 1997;35:327–36. 2. Abu-Hijleh G, Reid O, Scothorne RJ. Cell death in the developing chick knee joint: 1. Spatial and temporal patterns. Clin Anat 1997;10:183–200. 3. Alles AJ, Sulik KK. Retinoic acid induced spina bifida: evidence for a pathogenetic mechanism. Development 1990;408:73–81.
396 4. Alles AJ, Sulik KK. A review of caudal dysgenesis and its pathogenesis as illustrated in an animal model. Birth Defects 1993;29:83–102. 5. Barnes JC, Smith WL. The VATER association. Radiology 1978;126:445–9. 6. Bouwes JNB, Weaver DD. Subclavian artery supply disruption sequence: hypothesis of a vascular etiology for Poland, KlippelFeil, and Mobius anomalies. Am J Med Genet 1986;23:903–18. 7. Chandebois R, Brunet C. Origin of abnormality in a human symmelian foetus as elucidated by our knowledge of vertebrate development. Teratology 1987;36:11–22. 8. Danielson MK, Danielsson BRG, Marchner H, et al. Histopathological and hemodynamic studies supporting hypoxia and vascular disruption as explanation to phenytoin teratogenicity. Teratology 1992;46:485–97. 9. Duhamel B. From mermaid to anal imperforation: the syndrome of caudal regression. Arch Dis Child 1961;36:152–5. 10. Fernbach SK, Glass RBJ. The expanded spectrum of limb anomalies in the VATER association. Pediatr Radiol 1988;18: 215–20. 11. Gardner WJ. Hypothesis: overdistension of the neural tube may cause anomalies of nonneural organs. Teratology 1980;22:229–38. 12. Inouye M. Differential staining of cartilage and bone in fetal mouse skeleton by alcian blue and alizarin red-S. Cong Anom 1976;16:171–3. 13. Kotsios C, Merei J, Hutson JM, et al. Skeletal anomalies in the adriamycin-exposed prenatal rat: a model for VATER association. J Orthop Res 1998;16:50–3. 14. McCoy MC, Chescheir NC, Kuller JA, et al. A fetus with sirenomelia, omphalocele, and meningomyelocele, but with normal kidneys. Teratology 1994;50:168–71. 15. Merei J, Hasthorpe S, Farmer P, et al. Relationship between esophageal atresia with tracheoesophageal fistula and vertebral anomalies in mammalian embryos. J Pediatr Surg 1998;33:58–63. 16. Murphy JJ, Fraser GC, Blair GK. Sirenomelia: case of the surviving mermaid. J Pediatr Surg 1992;27:1265–8. 17. O’Rahilly R, Muller F. Interpretation of some median anomalies as illustrated by cyclopia and symmelia. Teratology 1989;40:409– 21.
G. A.-Hijleh et al.: Bones and joints in the VATER Association 18. Orr BY, Long SY, Steffek AI. Craniofacial, caudal and visceral anomalies associated with mutant sirenomelic mice. Teratology 1982;26:311–7. 19. Qi B, Diez-Pardo JA, Navarro C, et al. Narrowing the embryologic window of the adriamycin-induced fetal rat model of esophageal atresia and tracheoesophageal fistula. Pediatr Surg Int 1996;11:444–7. 20. Qi B, Merei J, Farmer P, et al. Cardiovascular malformations in rat fetuses with esophageal atresia and tracheoesophageal fistula induced by adriamycin. Pediatr Surg Int 1997;12:556–64. 21. Quan L, Smith DW. The VATER association: vertebral defects, anal atresia, T-E fistula with esophageal atresia, radial and renal dysplasia: a spectrum of associated defects. J Pediatr 1973;82:104– 7. 22. Schoenwolf GC, Garcia-Martinez V, Dias MS. Mesoderm movement and fate during avian gastrulation and neurulation. Dev Dyn 1992;193:235–48. 23. Smith JL, Gestland KM, Schoenwolf GC. Prospective fate map of the mouse primitive streak at 7.5 days of gestation. Dev Dyn 1994;201:279–89. 24. Stevenson RE, Jones KL, Phelan MC, et al. Vascular steal: the pathogenetic mechanism producing sirenomelia and associated defects of the viscera and soft tissues. Pediatrics 1986;78:451– 7. 25. Sulik K, Dehart DB. Retinoic acid-induced limb malformations resulting from apical ectodermal ridge cell death. Teratology 1988;37:527–37. 26. Tan C, Etcubanas E, Wollner N, et al. Adriamycin — an antitumor antibiotic in the treatment of neoplastic diseases. Cancer 1973;32:9–17. 27. Temtamy SA, Miller JD. Extending the scope of the VATER association: definition of the VATER syndrome. J Pediatr 1974; 85:345–9. 28. Tibbles L, Wiley MJ. A comparative study of the effects of retinoic acid given during the critical period for induction of spina bifida in mice and hamsters. Teratology 1988;37:113–25. 29. Zhou B, Hutson J, Farmer P, et al. Apoptosis in tracheoesophageal embryogenesis in rat embryos with or without adriamycin treatment. J Pediatr Surg 1999;24:872–6.