American Journal of Medical Genetics 87:115–127 (1999) Caudal Dysgenesis in Staged Human Embryos: Carnegie Stages 16–23 Rengasamy Padmanabhan,1* Ichiro Naruse,2 and Kohei Shiota2 1 Department of Anatomy, Faculty of Medicine, UAE University, Al Ain, United Arab Emirates Congenital Anomaly Research Center and Department of Anatomy and Developmental Biology, Graduate School of Medicine, Kyoto University, Kyoto, Japan 2 The severity of expression of malformations of the median axis in the caudal region of human embryos is highly variable and ranges from caudal dysgenesis and sirenomelia to simple sacral hypoplasia. Several forms of sacral dysgenesis may be discovered later in life. This shows that caudal malformations of relatively lesser severity should occur at a greater frequency than actually reported. In the present study we looked at the morphology and histology of some human embryos with caudal dysgenesis. Several developmental alterations of the median axis were observed. These included significant reduction in the craniofacial mesenchyme characterized by hypoplasia of the pharyngeal arches, palatal shelves, and agenesis or hypoplasia of the auricular hillocks at the rostral end, absence of the caudal trunk from midsacral to all coccygeal segments, vertebral fusion or agenesis, defective development of the primary and secondary neural tubes, rectal and urinary tract dysgenesis, and deficiency, malrotation, and deficiency of the limbs at the caudal end. Hindlimb malformations included bilateral agenesis (one case), meromelia, and various forms of abnormal rotation, but no instances of sirenomelia were present. Radial dysgenesis has been reported to be associated with caudal dyplasia in the literature, however, we observed agenesis of the ulna in one and of the fibula in another embryo. There was an impressive association between limb malfor- Contract grant sponsor: Japanese Ministry of Education, Science, Sports and Culture; Contract grant sponsor: Ministry of Health and Welfare; Contract grant sponsor: Heiwa Nakajima Foundation (Tokyo). *Correspondence to: Dr. R. Padmanabhan, Department of Anatomy, Faculty of Medicine and Health Sciences, United Arab Emirates University, PO Box 17666, Al Ain, United Arab Emirates. E-mail: [email protected] Received 6 April 1999; Accepted 14 July 1999 © 1999 Wiley-Liss, Inc. mations and body wall defects. The histological studies demonstrated caudal vascular deficiency and hemorrhagic lesions in the limbs of the dysplastic embryos. The data suggest that these polytopic field defects arise very early in development possibly as result of disturbances to fundamental developmental events that share common molecular and cellular mechanisms. Am. J. Med. Genet. 87:115–127, 1999. © 1999 Wiley-Liss, Inc. KEY WORDS: human embryos; caudal dysgenesis; limb-body wall defects; morphological and histological study INTRODUCTION The “syndrome” of caudal dysgenesis comprises a complex group of malformations of the caudal embryonic axis including abnormalities of the lumbosacral spine, imperforate anus, rectovesical and rectourethral fistulae, agenesis of the kidneys, internal genitalia with the exception of the gonads, and flexion, inversion, and external rotation of the lower limbs [Duhamel, 1961]. Although Duhamel’s name is most commonly cited in the literature on caudal dysgenesis, the condition was first described by Rocheus in 1542 [cited by Kampmeier, 1927]. This anomaly has been variously and often synonymously (although incorrectly) referred to as caudal dysgenesis, caudal dysplasia, sacral agenesis, sacral dysplasia, caudal spinal aplasia/ agenesis, lumbosacral agenesis, etc. [Pang and Hoffman, 1980; Welch and Aterman, 1984; Pappas et al., 1989; Alles and Sulik, 1993]. The severity of expression of the anomaly is variable [Pappas et al., 1989]. In extreme forms the lower extremities are represented by a median limb with toes turned backward (symmelia/ sirenomelia) giving the appearance of the mythological mermaid, hence the fanciful term mermaid anomaly. The severe visceral malformations associated with caudal dysgenesis are incompatible with extrauterine life, whereas symmelia is the least grave and most variable abnormality. It has an earlier onset than lumbosacral 116 Padmanabhan et al. agenesis. The incidence is reported to be 1 in 60,000 births. The frequency is 8 to 15% higher in twinning and approximately 100 to 150 times higher among monozygotic twins than among dyzygotic twins or singletons. Concordance for sex is almost a rule although concordance for sirenomelia is unusual [Young et al., 1986]. Familial occurrence has been reported, but no Mendelian pattern of inheritance has been established. A recent study provides evidence for the homeobox gene HLXB9 involvement in dominantly inherited sacral agenesis, but the possible role of this gene in caudal dysgenesis remains to be established [Ross et al., 1998]. The fact that sacral dysgenesis may be discovered later in life indicates that caudal malformations of relatively lesser severity, which have no apparent clinical manifestations, occur at a greater frequency than actually reported. Clinical data indicate that there may be many causes for the defective differentiation of the caudal end of the embryonic axis during gestation days 13 to 22 leading to caudal deficiencies [Hoyme, 1988; Young et al., 1986]. Cadmium, lead, trypan blue, retinoic acid, and ochratoxin A have been shown to induce caudal dysgenesis in experimental animals [Rajala and Kaplan, 1980; Padmanabhan and Hameed, 1990; Alles and Sulik, 1993; Wei and Sulik, 1996; Padmanabhan, 1998]. The facts that insulin may cause rumplessness in chick embryos, that sacral dysgenesis occurs when pregnant rats are experimentally made diabetic, and that there exist occasional clinical observations of sacral hypoplasia in infants of diabetic mothers, suggest that maternal diabetes may contribute to sacral hypoplasia in the offspring [Landauer, 1945; Duraiswamy, 1950; Wilson and Vallance-Owen, 1966; Pedersen et al., 1971; Deuchar, 1977; Young et al., 1986; Welch and Aterman, 1984; Perrot et al., 1987; Padmanabhan and Al Zuhair, 1989]. However, the incidence in the babies of established diabetics is surprisingly as low as 1% or < 1% [Sarnath et al., 1976; Pappas et al., 1989; Mills, 1982; Kalter, 1993]. In addition to diabetes, embryonal trauma, maternal fever, nutritional deficiency, toxic substances, and genetic factors have been considered causes of caudal malformations [Pang and Hoffman, 1980]. Kampmeier  thought that an abnormal umbilical artery of vitelline origin could deprive the caudal region of the embryo of essential nutrients resulting in caudal dysgenesis. This abnormal artery was found in a number of sirenomelic fetuses [Kapur et al., 1991; Stocker and Heifetz, 1987]. These reports were based on observations on late-stage fetuses and newborn infants, and there was no evidence that the vascular abnormality preceded the observed dysgenesis sequence. O’Rahilly and Müller  examined over 100 normal embryos of Carnegie stage 8–18 and four synophthalmic embryos of stages 16–20 and commented on the pathogenesis of several median anomalies including sirenomelia. Studies on younger human embryos with caudal dysgenesis are of great importance in delineating the pathogenetic mechanisms, but published data are lacking. Recent studies have shown that vascular disruption precede caudal dysgenesis in the mouse [Seller and Wallace, 1993; Padmanabhan, 1998]. The objectives of the present study were to ex- amine human embryos with caudal dysgenesis and evaluate the morphological and histological alterations that might subsequently contribute to caudal dysgenesis. Here we report on the gross and histological abnormalities of Carnegie stages 16–23 embryos with caudal dysgenesis and discuss their possible pathogenetic mechanisms. MATERIALS AND METHODS The embryos used in this study were from a large collection of conceptuses obtained mostly from apparently normal pregnancies terminated by dilation and curettage for economic reasons and preserved at the Congenital Anomaly Research Centre of the Kyoto University Faculty of Medicine. A few cases were also obtained from threatened abortions. The collection was begun in 1961 by Prof. Hideo Nishimura and assisted by many of his experienced colleagues and obstetricians. Details of these embryos have been described earlier [Nishimura, 1975]. They were staged and preserved in formalin, Bouins or Lillies fluid. They were first examined thoroughly under a stereomicroscope and external anomalies were photographed meticulously from several aspects. Both age-matched normal and abnormal embryos were sectioned and stained with hematoxylin and eosin (H&E) and Mallory azan trichrome. Included in the present report are 10 embryos with caudal dysgenesis ranging in age from Carnegie stages 16 to 23. Two of them (one of stage 21 and one of stage 22) had been sectioned serially for histological examination. Alternate sections were stained with H&E and trichrome. They were examined with an Olympus BH-2 microscope. Both gross and histological specimens were compared and contrasted with agematched normal embryos [Nishimura, 1983; O’Rahilly and Muller, 1987]. RESULTS Macroscopic Observations All embryos shown in Table I lacked or had an extremely small trunk distal to the postaxial border of the hindlimbs. The absent segments included midsacral to all coccygeal vertebrae. Virtually all of them had one or more craniofacial malformations such as holoprosencephaly (1 case), hypoplasia of pharyngeal arches, nasal processes and premaxilla, and reduction in number, malpositioning, and hypoplasia of the auricular hillocks (Fig. 1 B–D,F). One of the embryos had severe mandibular retrognathism and unfused lateral nasal and maxillary processes (Fig. 2D). Meromelia of the upper and lower limbs, unilateral cleft hand and cleft foot (case no 50641 stage 18), and retardation of the development of digits were also observed. The lower limbs were commonly oriented horizontally with the postaxial borders facing medially (inversion) (Fig. 2B). The foot plates adjoining each other in one case (case 215052, stage 19) gave the appearance as if the perineum was covered by the hindlimbs. One of the embryos of stage 16 (case 37430) with a remarkable deficiency in the caudal segment of the trunk, had failed to develop lower limb buds; it had a fish-like Caudal Dysgenesis in Human Embryos 117 TABLE I. Anomalies Associated With Caudal Dysgenesis in 10 Human Embryos of Carnegie Stage 16–23 Embryo number Photographic examination Gross examination Histological examination 14535 + + – 37430 + – – 17 38638 + + – 18 382666 + + – 18 50641 + + – 19 21502 + – – 19 28410 + – – 21 3191 + – + 22 21743 + – + 23 33511 + – – Stage 16 appearance (Fig. 1B). The genital tubercle was located on the summit of a median genital swelling. The anococcygeal distance appeared to be reduced in conformity to the caudal narrowing of the trunk. The external genitalia of most embryos were hypoplastic and situated very close to the umbilicus possibly because of underdevelopment of the infraumbilical portion of the body wall (Fig. 2 E,F). The ventral abdominal wall and the lower chest wall appeared thin and translucent through which the liver, heart, and pericardium could be easily seen. Microscopic Observations Embryo of Carnegie Stage 21 (Case 3191). The cerebral ventricles of this embryo were rather hypoplastic and slightly collapsed. Structurally the neuro- Abnormalities Holoprosencephaly, two auricular hillocks in pharyngeal arch 1 and one in arch 2 - all hypoplastic; fore- and hindlimb meromelia; narrow hand plate and absence of digital rays in hand plate and foot plate. Rudimentary first and second pharyngeal arches with no auricular hillocks; hand plate as broad as the arm, which is also hypoplastic; absence of both hindlimbs giving rise to fish-like appearance; the trunk ends at the genital tubercle; thin transparent ventral abdominal wall, through which the liver could be seen. Hypoplastic pharyngeal arch 1 and 2; arch 1 has no auricular hillocks and arch 2 has two rudimentary hillocks; medial and lateral nasal processes hypoplastic; ventral body wall comprises a transparent membrane; tail consists of two somites whereas the normal embryos of this stage contains about 10 somites; the chorionic villi appeared withered. Four auricular hillocks surround the first pharyngeal cleft; forelimb buds are hypoplastic; left hand plate is pointed like a tongue, with 1/4 of the postaxial portion missing; right hand plate is like a narrow spade and corresponds in width to two digital rays; leg portion of the right hind limb is absent and foot plate is represented by a nipple-like stump at apex of the thigh; the left hindlimb lacks leg segment and has a second limb-like appendage on the preaxial aspect. Left side split hand and split foot plate; the preaxial digit precociously marked out in the left hand; postaxial borders of hindlimbs are medial and preaxial borders are lateral in orientation; the trunk ends at the level of the root of hindlimbs; lower lumbar and all distal segments are absent; genital tubercle and umbilicus adjoin each other; all umbilical vessels are present although small in caliber. Perineum is covered by hypoplastic hindlimbs postaxial borders directed medially, separated only by the genital tubercle, which lies at the center of the perineum; sacral and all coccygeal segments of the trunk are absent; there is a median dorsal elevation in the lower trunk. Medial and lateral nasal processes are hypoplastic and widely separated (bilateral cleft upper lip?); preaxial polydactyly of right lower limb (extra digital ray; deep interdigital notches are seen; mid-sacral to all coccygeal segments of the trunk are missing. There are four nipple primordia (polythelia). Hypoplastic medial and lateral nasal processes; cleft left hand; ventral abdominal wall deficient; more details in the text. Lower sacral and all coccygeal segments of the trunk absent; lower limb meromelia; more details in the text. Narrow perineum; lower limb meromelia; midsacral to all distal segments of the trunk absent. epithelium of the brain was just as mature as that of the normal embryos of stage 21. The medial and lateral nasal processes were severely hypoplastic indicative of the possibility of giving rise to facial clefts later in development. The vomeronasal organs were present. The palatal shelves were on either side of the tongue pointing to the floor of the oral cavity. The primordia of the pituitary, thyroid, parathyroid, heart, digestive organs, kidneys, suprarenals, and gonads (testes) appeared to be normal and appropriate for the stage of development. The clavicles showed evidence of ossification in the lateral and midshaft regions. The cartilaginous models of the humerus and scapula were normally developed. However, the radii and ulnae were remarkably short. There were six carpal primordia and five metacarpals. The 4th and 5th metacarpals distally 118 Padmanabhan et al. Fig. 1. Human embryos of Carnegie stage 16 (A and B), 17(C), and 18(D–F). A: A normal embryo. Observe in the caudal dysgenesis embryos marked hypoplasia of the pharyngeal arches and absence of auricular hillocks (B–F), retarded development of forelimbs (C,F), absence of hind limbs (B), absence of caudal trunk (B,C,D,F), and split hand (D). E: Limbs of embryo F. It has an accessory preaxial digit (arrow in lower picture) and no digital rays in the left foot plate; there is a single median digit in the right hand plate (upper picture). reached the ring finger. Metacarpals 1 and 2 entered the middle finger. The thumb consisted of only a skin appendage. The normal right hand possessed no first metacarpal and all fingers including the thumb had one phalanx each. The sternal bars were absent except most proximally. Starting from a little lateral to the midclavicular plane, the ventral abdominal wall was deficient in musculature. The liver and the gut with the mesentery protruded through a median defect of considerable size. The right edge of this defect in its caudal aspect showed a clear skin covering, although the left edge appeared to have been damaged during curettage. The viscera were found to project outside the lower abdomen. The umbilical cord contained two arteries and a vein. The mesonephric ducts and ureters entered the vesicourethral portion of the urogenital sinus (the bladder). The rectum ended abruptly at the level between the first and second sacral segments. Pubic bones were absent. The genital tubercle was small; the glandular urethra consisted of an epithelial plate. In the lower limbs, the primordia of the femur and tibia were present but not that of the fibula (Fig. 3 A,B). Hemorrhage was extensive in the thigh and leg but most prominent at the knee (Fig. 3C). The spinal cord (primary neural tube) had no skeletal protection below the midsacral level and became continuous with the short, kinky secondary neural tube at an angle. Irregular vascular proliferation and hemorrhage were promi- Caudal Dysgenesis in Human Embryos 119 Fig. 2. Embryos of Carnegie stage 19 (A–C), 21 (D), and 22 (E: lateral view; F: ventral view) showing caudal dysgenesis. The hypoplastic hindlimbs of the embryo A are attached to the caudal tip of the trunk with their roots closely positioned in the narrow perineum (B) and postaxial borders facing each other. The genital tubercle is either absent (B) or adjoins the caudal tip (arrow in C). Varying degrees of malrotation and hypoplasia of hindlimbs, deficiency of lower abdominal wall (E,F), oblique facial cleft (nonfusion of maxillary and lateral nasal processes), and extreme mandibular hypoplasia (D) are also obvious. (B) Perineum of embryo A. (F) Ventral view of embryo E. nent in the area of union between the primary and secondary neural tubes (Fig. 3 D,E). There was no evidence of a tail. Embryo of Carnegie Stage 22 (Case 21743). The histological structure of the head and neck of this embryo appeared normal except for the fact that the palatal shelves were asymmetric. The sternal bars were hypoplastic, widely separated, and gradually disappeared more distally (Fig. 4 C,D), whereas the normal embryos of this stage had well-formed primordia of costal cartilages, ribs, and sternal bars, and well-differ- entiated abdominal musculature (Figs. 4 and 5). The ventral abdominal wall was open at the midlumbar region where the parietal peritoneal lining was found to be continuous with the epidermis. As a result, the abdominal contents were outside the abdomen (Fig. 5 A,B). The abdominal wall ventral to the midaxillary plane lacked in differentiation and organization of the musculature, and as a result, the lower pericardium, heart, and liver remained unprotected below the costal margin (Fig. 5 C,D). The gonads (testes), kidneys, spleen, and digestive 120 Padmanabhan et al. Fig. 3. (A–C) Transverse sections of an embryo of Carnegie stage 21 with caudal dysgenesis. Primordia of femur (F) and tibia (T) are present but not that of fibula. Note the presence of prominent hematoma (arrow in A and B) at knees. (C) A high power view of hemorrhage (arrows) seen in A. The caudal end of the same embryo (D) shows kinky union between the primary (arrow) and secondary (arrowheads) neural tubes. The neuroepithelial infoldings of secondary neural tube seen in higher magnification (E) are possibly the result of kinking. Also note the patchy hemorrhages (hm in D and E) around the neural tube. The neural tube lacks skeletal protection. Bar ⳱ 450 m (A,B); 90 m (C,E); 225 m (D). organs appeared to have been normally formed. The separation of the bladder and rectum was complete although the rectum ended blindly. No ureters could be traced to the bladder. The Wolffian ducts appeared to be rather dilated. The phallus was rudimentary. There was a single umbilical artery directly continuous with the dorsal aorta on the left of the median plane (Fig. 6A). It was almost as large in size as the dorsal aorta itself. Distal to this site of continuity the dorsal aorta was absent. The internal iliac vessels were also lacking. The third, fourth, and fifth sacral vertebral centra were fused with reduced disc space, which was confined to the periphery. There were no coccygeal vertebrae. In the normal embryos of this stage, there were five sacral vertebrae and as many as seven coccygeal vertebrae (Fig. 6B); situated on the pelvic floor were abundant autonomic elements extending from the sympathetic chains towards the cloaca. The affected embryo presented a caudal tissue mass comprising cells similar to those of the ganglion impar but with no definitive organization (Fig. 6A). The caudal portion of the primary neural tube was kinky and dilated with vacuolated neuroepithelium, and the secondary neural tube was rudimentary at the beginning and partially duplicated distally. Here both the primary and secondary neural tubes lacked skeletal protection (Fig. 6 D,E). The clavicles showed signs of ossification in the mid and lateral segments of the shaft. The primordia of the scapula, humerus, and radius were well defined, but the ulnae were absent. The humeri were relatively short and curved. The distal end of the humerus and proximal end of the radius were broad. The lower limbs were short (meromelic) with inadequately developed muscles. The femora were short and fused with the hip bone without a proper cavity of the hip joint (Fig. 6F). The pubic bones were absent. Caudal Dysgenesis in Human Embryos 121 Fig. 4. Transverse sections of a normal embryo (A,C) and an embryo with caudal dysgenesis (B,D) at stage 22. The sternal bars (open arrows in C,D) are well formed in the normal embryo in contrast to the hypoplastic and widely separated sternal bars of the caudal dysgenesis embryo (B,D). O, oesophagus; T, trachea; AO, aorta; CV, cardinal vein; H, heart. Bar ⳱ 450 m (A,B); 225 m (C); 360 m (D). 122 Padmanabhan et al. Fig. 5. Transverse sections of normal (A,C) and caudal dysplastic (B,D) embryos of stage 22. The ventral abdominal wall of caudal dysgenesis embryo is deficient allowing viscera to protrude externally (B). The musculature of the abdominal wall of the normal embryo has differentiated and organized well into the external oblique (EO), internal oblique (IO), and transversus (TA) abdominis. In sharp contrast, the caudal dysgenesis embryo shows a gross deficiency and lack of differentiation of this musculature (open arrow in D). Bar ⳱ 450 m (A,B); 90 m (C); 225 m (D). Caudal Dysgenesis in Human Embryos 123 Fig. 6. Sections of embryos of Carnegie stage 22. The embryo with caudal dysgenesis shows a gross reduction in number and fusion of caudal vertebral primordia (S in A), disorganized caudal tissue mass (CTM in A), a single umbilical artery (UA in A), kinky and dilated central canal of the caudal portion of the neural tube (CC in D), duplication of the secondary neural tube (SNT in D), and vacuolated ventricular zone of the neuroepithelium (NE in E). The femur (F) remains fused with the ilium (IL) in the caudal dysgenesis embryo (F). (B,C) Normal embryos of stage 22 showing prominent caudal vertebrae (B) and hip joint (C). B, bladder; R, rectum; U, urethra. Bar ⳱ 225 m (A, C, D–F); 360 m (B). DISCUSSION Several theories have been postulated to explain the pathogenesis of caudal dysgenesis. The “lateral compression” theory originally proposed by Dareste [1891, cited by Stevenson et al., 1986], holds that abnormal amniotic folds compress the caudal end of the embryo and suppress the development of the pelvic structures and cause abnormal rotation of the limbs. This theory is supported by observations on the mutant sirenomelic mice [Hornbeek, 1970; Orr et al., 1982]. Matsunaga and Shiota  reported a high incidence of unilateral limb defects in tubal pregnancies and caudal dys- genesis in pregnancies with uterine myomata possibly as result of spatial restriction and associated vascular compression. According to Kallen and Winberg  deficiency of the caudal mesoderm and/or irregularities of the notochord and somites might lead to duplication of the notochord and neural tube, chordoma of the sacrococcygeal region, anorectal dysgenesis (atresia, stenosis, and ectopia), and urogenital and vertebral malformations. Narrow defects of the caudal mesoderm would result in median anomalies such as sacral agenesis and more extensive defects lead to lateralized malformations such as limb defects. Extreme median tissue deficiency will result in failure of fission of the limb 124 Padmanabhan et al. fields resulting in symmelia [O’Rahilly and Müller, 1989]. Failure of the caudal somites to develop and the lack of midline mesoderm promote fusion of the hindlimb buds. Whether it is merging, fusion of hindlimb buds, or failure of fission of limb fields that leads to sympodia is controversial [Hoyme, 1988; Barr, 1988]. The hindlimb buds of human embryos do not appear until stage 13 (postconception day 28), whereas the critical period for caudal dysgenesis is between day 13 and 22 [O’Rahilly and Müller, 1988]. Therefore, the probability for fusion of formed limb buds is rather low. Our recent study in which retinoic acid (RA)-treated mouse embryos were examined at early stages did not show any evidence for fusion of limb buds at all, although symmelia was observed in a number of embryos at term [Padmanabhan, 1998]. The vascular steal theory originally proposed by Weigert  was supported by many subsequent investigators [Kampmeier, 1927; Stevenson et al., 1986; Kapur et al., 1991; Murphy et al., 1992]. These authors observed in babies with caudal dysgenesis a single umbilical artery that arose from the abdominal aorta slightly proximal to its bifurcation. Distal to this origin, the aorta was extremely hypoplastic. The caudal anomalies were therefore interpreted as the result of arrested development because of hypoperfusion. It is important to point out here that none of these studies provide evidence to show that the vascular abnormality preceded caudal regression. This assumption cannot also explain adequately the more cranial anomalies (e.g., anencephaly, limb-body wall defects, VATER, or VACTERL abnormalities) observed in many of the sirenomelic fetuses [Rodriguez et al., 1991; Tang et al., 1991; Murphy et al., 1992; McCoy et al., 1994]. The possible role of vascular steal in the pathogenetic mechanism has earned attention possibly because of the frequently reported association of a single umbilical artery of vitelline origin with sirenomelia [Kampmeier, 1927; Stocker and Heifetz, 1987; Kapur et al., 1991; McCoy et al., 1994]. This has also initiated lively discussions on the role of vascular disruption and hypoperfusion as a teratogenic mechanism [Van Allen, 1981; Hoyme et al., 1981; Hoyme, 1988; Barr, 1988; Chandebois and Brunet, 1987; 1988; Alles and Sulik, 1993; McCoy et al., 1994; Padmanabhan, 1998]. Chandebois [Chandebois and Brunet, 1988] did not think that hypoperfusion secondary to vascular disruption contributed to caudal dysplasia possibly because the fetus whose pelvic histology he examined was already at an advanced stage of development. Although trypan blue-induced hematomas in chick embryos disappear quickly, rumplessness (the chicken equivalent of caudal dysgenesis) is generally observed at hatching. Therefore, Barr  argued that hematoma in caudal median axis was an appropriate conceptual model to explain the diversity of malformations observed in sirenomelic fetuses. The two early embryos we studied histologically provided evidence that some vascular anomalies were associated and that hemorrhage occurred in the malformed (or deformed) tissues including the limbs. Two more recent studies in mouse embryos [Seller and Wallace, 1993; Padmanabhan, 1998] also provide particular evidence that hemorrhage does occur prior to regression of the median axis. However, it is important to understand that vascular disruption and subsequent hemorrhage are possibly secondary pathogenetic mechanisms and that what initiates the cascade is not clear from these studies. While the debate continues, more work needs to be done to identify the factors that contribute to the vulnerability of the blood vessels of the median axis. The hypothesis of overdistention of neural tube advocated by Gardner  holds that the reopening of a closed neural tube from which neural tube fluid escapes can impair a number of developmentally immature structures such as the primordia of the vertebral column, neural tube, gut, limb buds, and metanephros. Thus, he explained cranioschisis, anenencephalus, and Klippel-Feil syndrome at the cranial end, VATER or VACTERL anomalies in the mid-body, and caudal dysgenesis including sirenomelia at the caudal end as consequences of reopening of the closed neural tube. According to Gardner , overdistention of the caudal neural tube would move the hindlimb buds dorsally and allow their fusion leading to sirenomelia. Our previous studies in rat embryos [Padmanabhan, 1984, 1988, 1991] provided experimental evidence for reopening of cranial neural tube and for several non-neural malformations, but there was no limb bud fusion. Neither Shenefelt’s  hamster embryos nor our mouse embryos [Padmanabhan, 1998] that were exposed to RA after caudal neural tube closure had any manifestation of sirenomelia, suggesting that Gardner’s theory could not be proved experimentally. McCoy et al.  proposed that a combination of defective cell proliferation, cell migration or differentiation, and excessive cell death could be triggered by a transient hypoperfusion; the consequent mesodermal tissue deficiency and a hypoplastic vasculature might additionally contribute to the malformation. A close look at the anatomy of the structures affected in our embryos shows that dysgenesis is not confined to the caudal end of the embryonic axis but includes several median and paramedian structures such as the body wall and limbs, observations confirmed by O’Rahilly and Muller  in human embryos and our studies on RA-treated mouse embryos [Padmanabhan, 1998]. It is also apparent that no single theory could accommodate the pathogenesis of all these malformations. The midline of the embryo is considered to be an important part of the primary field characterized by active cell proliferation, migration, and differentiation and tissue organization and induction [Opitz and Gilbert, 1982a; Opitz, 1993]. These developmental processes at the rostral end contribute precursor cell populations involved in the formation of the pharyngeal arches and craniofacial development. At the caudal end is another area of similarly intense morphogenetic activity called the caudal eminence, which is first identified at stage 9 [O’Rahilly and Müller, 1989; Opitz, 1993]. It is different from the end bud of avian embryos. It extends from the neurenteric canal to the cloacal membrane and provides precursor mesenchyme for the formation of the notochord, somites, caudal vertebrae, blood vessels, hindgut, neural tube, and hindlimbs. Caudal Dysgenesis in Human Embryos Blood vessels enter the caudal eminence at stages 14– 16, and the tissues formed in excess begin to regress at stage 17. The portion of the trunk derived from the caudal eminence and that derived from the primitive streak meet at the upper sacral level at the site of closure of the caudal neuropore [Müller and O’Rahilly, 1987; O’Rahilly and Müller, 1989]. These developmental activities occurring along the median plane are spatially and temporally coordinated, delicately balanced, and possibly regulated by several genes [Kessel and Gruss, 1991; Kessel, 1992; Gruss and Walther, 1992; Holland and Hogan, 1988], which may be interfered with by physical or chemical agents or by vascular accidents. Clinical data show that the primary field is uniquely susceptible to complex malformations [Opitz and Gilbert, 1982a, 1982b; Lubinsky and Moeschler, 1987]. The results of this study have clearly established the early onset of several developmental alterations associated with caudal dysgenesis in the human embryos. These include deficiency of the craniofacial mesenchyme characterized by hypoplasia of the pharyngeal arches, medial and lateral nasal processes, palatal shelves, agenesis or hypoplasia of the auricular hillocks at the rostral end, reduction or absence of the caudal trunk from midsacral to all coccygeal segments, vertebral agenesis and/or fusion, defective development of the primary and secondary neural tubes, rectal and urinary tract dysgenesis, and deficiency, malrotation, and malformations of the limbs at the caudal end of the embryonic axis. Total absence (one case), meromelia, and various forms of abnormal rotation of hindlimbs were also observed, and no instance of sirenomelia was present. Whereas radial dysgenesis has been reported to be associated with caudal dysgenesis [Quan and Smith, 1973; Young et al., 1986; Alles and Sulik, 1993], we observed agenesis of the ulna in one of our embryos. The fibula was absent in the other embryo. The most common associated defect below the knee is reported to involve an absence of the fibula and comparable abnormalities of the upper limb are relatively uncommon [Passarge and Lenz, 1966; Welch and Aterman, 1984], and according to the filed theory, this homology (radius and tibia or ulna and fibula) is one of the attributes of developmental field defects [MartínezFrías et al., 1998]. The occurrence of limb malformations in combination with body wall defects in our study is impressive and consistent. The histological studies have also demonstrated caudal vascular deficiency and hemorrhagic lesions in the limbs of the dysplastic embryos. Basically the embryonic origin of precursors of these seemingly diverse structures are to be found in the median and paramedian planes during blastogenesis [Opitz and Gilbert, 1982a; O’Rahilly and Müller, 1989; Opitz, 1993; Martínez-Frías et al., 1998]. These include the neural plate, neural tube, notochord, gut, prechordal plate, the primitive streak, caudal eminence, connecting stalk, etc., in the median plane and the bilaterally paired structures such as the somites, limb fields, heart tubes, etc. At this stage, the whole embryo forms a primary field and responds to intrinsic or extrinsic noxious stimuli largely as a pleuripotent single 125 unit. Pattern formation establishes upstream domains of specific transcription, signaling, and growth factor gene expression resulting in gradual subdivision of the primary field into several progenitor fields, the primordia of all final structures [Opitz, 1993; Martínez-Frías et al., 1998]. The HOX, PAX, T-Box, and sonic hedgehog (SHH) genes and the RA molecule, which emanates from the organizer, play major roles in regionalization, positional information, and anteropoterior and dorsoventral axis formation [Kessel and Grus, 1991; Kessel, 1992; Gruss and Walther, 1992; Conlan, 1995; Smith, 1999; Mansouri et al., 1999]. After the initiation of gastrulation, HOX genes and other homeobox-containing genes begin to be expressed in specific anteriorposterior domains [Conlan, 1995]. In response to RA signals generated from midline structures, cells ingressing from the primitive streak, sequentially activate Hox genes leading to overlapping nonidentical expression domains of Hox genes along the anteroposterior axis. The combination of functionally active Hox genes, known as the Hox code, specifies the identity of body region [Kessel, 1992]. Exogenous RA alters Hox codes resulting in homeotic transformations [Kessel and Grus, 1991]. Misexpression of the Hox genes or lack of morphogens such as RA during this critical period of development may result in segmentation defects such as fusion or absence of vertebrae. There is evidence for involvement of HLXB9 gene in hereditary sacral agenesis [Ross et al., 1998] The expression of Shh is both temporally and spatially restricted and in developing vertebrate body axis is tightly correlated with expression of HNF3. Targeted mutation of HNF3 in the mouse results in defective development of the notochord and loss of Shh expression. Exposure of chicken wing bud to RA leads to activation of Shh indicating that both HNF3 and RA receptors regulate Shh gene [Chang et al., 1997]. Shh emanates from the axial midline structures such as the notochord and floor plate, and influences dorsal-ventral specification of somite derivatives [Johnson et al., 1994]. Targeted gene disruption in the mouse has demonstrated that the Shh plays a crucial role in patterning of embryonic structures such as the brain, spinal cord, spinal column, and the limbs [Chiang et al., 1996]. Abnormalities are found to occur first in the establishment and maintenance of midline structures such as the notochord and floor plate, and later, defects of the neural tube and limbs and cyclopia develop. Mutations in the C-terminal domain of SHH cause holoprosencephaly [Roessler et al., 1997]. These findings suggest that the polytopic abnormalities observed in our embryos arose very early in development and represent defective signaling and pattern formation. The literature on caudal regression indicates the frequent association of VATER, VACTERL (vertebral malformations, anal atresia, tracheoesophageal fistula, cardiac defects, and renal/radial agenesis), OEIS complex (omphalocele, extrophy of the bladder, imperforate anus, and spinal defects), and Potter sequence with caudal dysgenesis [Gardner, 1980; Young et al., 1986]. It is important to point out here that such combinations of multisystemic, often severe and complex anomalies that affect predominantly the midline, arise 126 Padmanabhan et al. as a response to a dysmorphogenetic stimulus during blastogenesis or at a time when progenitor fields are being established and share common molecular and cellular mechanisms [Duncan and Shapiro, 1993]. Therefore, they should logically be called primary polytopic developmental field defects [Martínez-Frías et al., 1998] rather than multiple idiopathic anomalies of blastogenesis [Opitz, 1993]. This concept receives credence from epidemiological [Martínez-Frías, 1994] and experimental [Alles and Sulik, 1993; Rutledge et al., 1994; Padmanabhan, 1998] studies. In our mouse model of caudal dysgenesis [Padmanabhan, 1998], administration of single doses of RA during blastogenesis resulted not only in sirenomelia but also in defects of a host of structures that could be grouped together as primary polytopic field defects. The severity differed according to the dose and the developmental stage at the time of RA-administration. When embryos of early somite stage were exposed to a low dose of RA, hematomas developed in the tail bud within 24 hr of treatment. This was accompanied by pronounced cell death, edema, and tissue disruption resulting in total regression of the tail, complete in the following 3 to 4 days of gestation. Most of the embryos of the moderately highdose group failed to develop a tail bud, whereas others developed hemorrhagic or avascular tail buds that subsequently degenerated. In the high-dose group, extensive cell death was observed in the caudal median axis as early as 6 hr post-treatment. Subsequently, cell death also affected the hindgut, neural tube, and sclerotome-myotome junction area, but not the notochord. The profound tissue loss in the median area appears to have led to agenesis of the tail and caudal vertebrae. The dilated vertebral canals of the spina bifida fetuses was also associated with vertebral arch fusion, hemivertebrae, reduction in number of ribs and or fusion of lower ribs indicating defective segmentation, and/or loss of precursor tissues because of cell death. One of the important new findings of this study was the role of vascular disruption in caudal regression resembling the vascular disruption sequence described in the tail short variable [Seller and Wallace, 1993]. Teratogenic exposure-related tissue damage and the consequent repair and regenerative events might also contribute to abnormalities [Snow, 1984; Van Allen, 1981]. The concept of a combination of developmental vulnerability inherent in the primary field and teratogen interaction with consequent alterations in molecular and cellular mechanisms of development may account for the variable expression of the malformations that occur at the rostral and caudal ends of the embryo. This view will also provide an explanation for the diversity and combinations of anomalies with caudal dysgenesis observed in our series of early embryos and reported by earlier investigators in late fetuses and newborn infants. gratefully appreciated. Dr. Padmanabhan’s visit to the Center was supported by a fellowship from the Heiwa Nakajima Foundation (Tokyo). This work was supported by grants from the Japanese Ministry of Education, Science, Sports, and Culture and from the Ministry of Health and Welfare. ACKNOWLEDGMENTS Kessel M. 1992. Respecification of vertebral identities by retinoic acid. Development 115:487–501. We thank the collaborating obstetricians and the previous and present members of staff of the Congenital Anomaly Research Center of Kyoto University. The excellent technical support of Ms. Chigako Uwabe is Kessel M, Gruss P. 1991. Homeotic transformations of murine vertebrae and concomitant alterations of Hox codes induced by retinoic acid. Cell 67:89–104. REFERENCES Alles AJ, Sulik KK. 1993. A review of caudal dysgenesis and its pathogenesis as illustrated in an animal model. Birth Defects Orig Art Ser 29:83–102. Barr M Jr. 1988. Comments on origin of abnormality in a human symmelian foetus as elucidated by our knowledge of vertebrate development. Teratology 38:487–488. Chandebois R, Brunet C. 1987. Origin of abnormality in a human symmelian foetus as elucidated by our knowledge of vertebrate development. Teratology 36:11–22. Chandebois R, Brunet C. 1988. Response to comments on Origin of abnormality in a human symmelian foetus as elucidated by our knowledge of vertebrate development. Teratology 38:489–491. Chang BE, Blader P, Fischer N, Inghan PW, Strahle U. 1997. Axial (HNF3) and retinoic acid receptors are regulators of the zebrafish sonic hedgehog promoter. The EMBO J 16:3955–3964. Chiang C, Litingtung Y, Lee E, Young K, Corden JL, Westphal H, Beachy PA. 1996. Cyclopia and defective axial patterning in mice lacking Sonic hedgehog gene function. Nature 383:407–413. Conlan RA. 1995. Retinoic acid and pattern formation in vertebrates. Trends Genet 11:314–319. Dareste C. 1891. La production artificielle des monstruosites. Paris: cited by Stevenson et al., 1986 vide infra. p 420. Deuchar EM. 1977. Embryonic malformations in rats, resulting from maternal diabetes: preliminary observations. J Emb Exp Morphol 41:93– 99. Duhamel B. 1961. From mermaid to anal imperforation: the syndrome of caudal regression. Arch Dis Child 36:152–155. Duncan PA, Shapiro LR. 1993. Interrelationships of the hemifacial microsomia-VATER and sirenomelia phenotypes. Am J Med Genet 47:75–84. Duraiswamy PK. 1950. Insulin induced skeletal abnormalities in developing chickens. Br Med J 2:384–390. Gardner WJ. 1980. Hypothesis: overdistension of the neural tube may cause anomalies of nonneural organs. Teratology 22:229–238. Gruss P, Walther C. 1992. Pax in development. Cell 69:719–722. Holland PWH, Hogan BLM. 1988. Expression of homeobox genes during mouse development: a review. Genes Dev 2:773–782 Hoornbeek EK. 1970. A gene producing symmelia in the mouse. Teratology 3:1–10. Hoyme HE. 1988. The pathogenesis of sirenomelia: an editorial comment. Teratology 38:485–486. Hoyme HE, Higginbottom MC, Jones KL. 1981 The vascular pathogenesis of gastroschisis: intrauterine interruption of the omphalomesenteric artery. J Pediatr 98:228–231. Johnson RL, Laufer W, Riddle RD, Tabin C. 1994. Ectopic expression of sonic hedgehog alters dorsal-ventral patterning of somites. Cell 79: 1165–1173. Kallen B, Winberg J. 1974. Caudal mesoderm pattern of anomalies: from renal agenesis to sirenomelia. Teratology 9:99–112. Kalter H. 1993. Case reports of malformations associated with maternal diabetes: history and critique. Clin Genet 43:174–179. Kampmeier OF. 1927. On sireniform monsters, with a consideration of the causation and the predominance of the male sex among them. Anat Rec 34:365–389. Kapur RP, Mahony BS, Nyberg DA, Resta RG Shepard TH. 1991. Sirenomelia with a “vanishing twin.” Teratology 43:103–108. Landauer W. 1945. Rumplessness of chicken embryos produced by injection of insulin and other chemicals. J Exp Zool 98:65–77. Caudal Dysgenesis in Human Embryos Lubinsky M, Moeschler J. 1987. Different clusters within the VATER association distinguished through cardiac defects: Possible effects of teratogenic timing. Dysmorphol Clin Genet 1:80–82. Mansouri A, Goudreau G, Gruss P. 1999. Pax genes and their role in organogenesis. Cancer Res 59:1707s–1710s. Martínez-Frías ML. 1994. Developmental field defects and associations: epidemiologic evidence of their relationship. Am J Med Genet 56:374– 381. Martínez-Frías ML, Frias JL, Opitz JM. 1998. Errors of morphogenesis and developmental field theory. Am J Med Genet 76:291–296. Matsunaga E, Shiota K. 1980. Ectopic pregnancy and myomata uteri: teratogenic effects and maternal characteristics. Teratology 21:61–69. McCoy MC, Chescheir NC, Kuller JA, Altman GA, Flannagan LM. 1994. A fetus with sirenomelia, omphalocele, and meningomyelocele, but with normal kidneys. Teratology 50:168–171. Mills JL. 1982. Malformations in infants of diabetic mothers. Teratology 25:385–394. Muller F, O’Rahilly R. 1987. Development of the human brain, the closure of the caudal neuropore, and the beginning of the secondary neurulation at stage 12. Anat Embryol 176:413–430. Murphy JJ, Fraser GC, Blair GK. 1992. Sirenomelia: case of the surviving mermaid. J Pediatr Surg 27:1265–1268. Nishimura H. 1975. Prenatal versus postnatal malformations based on the Japanese experience on induced abortions in the human being. In: Blandau RJ, editor. Ageing Gamete. Basel: Karger. p 349–368. Nishimura H. 1983. An atlas of human prenatal histology. Tokyo: IgakuShon. Opitz JM. 1993. Blastogenesis and the “primary filed” in human development. In: Opitz JM, Paul NW, editors. Blastogenesis-normal and abnormal. Birth Defect Orig Art Ser 29:3–37. Opitz JM, Gilbert EF. 1982a. CNS anomalies and the midline as a developmental field. Am J Med Genet 12:443–455. Opitz JM, Gilbert EF. 1982b. Pathogenetic analysis of congenital anomalies in humans. Pathol Ann 12:301–349. O’Rahilly R, Muller F. 1987. Developmental stages in human embryos. Publication 637. Washington, D.C.: Carnegie Institution of Washington. p 141–158. 127 of the caudal regression syndrome (lumbosacral agenesis) with magnetic resonance imaging. Neurosurgery 25:462–465. Passarge E, Lenz W. 1966. Syndrome of caudal regression in infants of diabetic mothers: observations of further cases. Pediatrics 37:672–674. Pedersen ML, Tygstrup I, Villumsen AL, Pedersen J. 1971. Congenital malformations in the offspring of diabetic women. Diabetologia 7:404A. Perrot LJ, Williamson S, Jimenez JF. 1987. The caudal regression syndrome in infants of diabetic mothers. Ann Clin Lab Sci 17:211–220. Quan L, Smith DW. 1973. The VATER association, vertebral defects, anal atresia, tracheoesophageal fistula with esophageal atresia, radial and renal dysplasia: a spectrum of associated defects. J Pediatr 82:104. Rajala GM. Kaplan S. 1980. The formation of caudal haematoma in trypan blue-treated chick embryos as a function of morphological stage at treatment. Teratology 21:265–269. Rodriguez JL, Palacios J, Razquin S. 1991. Sirenomelia and anencephaly. Am J Med Genet 39:35–37. Roessler E, Belloni E, Gaudenz K, Vargas F, Scherer SW, Tsui LC, Muenke M 1997. Mutations in the C-terminal domain of Sonic hedgehog cause holoprosencephaly. Hum Mol Genet 6:1847–1853. Ross AJ, Ruiz-Perez V, Wang Y, Hagan DM, Scherer S, Lynch SA, Lindsay S, Gustard E, Belloni E, Wilson DI, Wadey R, Goodman F, Orstavik KH, Monclair T, Robson S, Reardon W, Burn J, Scambler P, Strachan T. 1998. A homeobox gene, HLXB9, is the major locus for dominantly inherited sacral agenesis. Nat Genet 20:358–361. Rutledge JC, Shourbaji AG, Hughes LA, Polifka JE, Cruz YP, Bishop JB, Generoso WM. 1994. Limb and lower-body duplications induced by retinoic acid in mice. Proc Natl Acad Sci USA 91:5436–5440. Sarnath HB, Case ME, Graviss R. 1976. Sacral agenesis: neurologic and neuropathologic features. J Neurol 26:1124–1129. Seller MJ, Wallace ME. 1993. Tail short variable: characterization of a new mouse mutant and its possible analogy to certain human vascular disruption defects. Teratology 48:383–391. Shenefelt RD. 1972. Morphogenesis of malformations in hamsters caused by retinoic acid: relation to dose and stage of treatment. Teratology 5:103–118. Smith J. 1999. T-box genes: what they do and how they do it. Trends Genet 15:154–158. O’Rahilly R, Muller F. 1989. Interpretation of some median anomalies as illustrated by cyclopia and symmelia. Teratology 40:409–421. Snow MHL. 1984. Restorative growth in mammalian embryos. In: Kalter H, editor. Issues and reviews. New York: Plenum Press. p 251–273. Orr BY, Long SY, Steffek AI. 1982. Craniofacial, caudal and visceral anomalies associated with mutant sirenomelic mice. Teratology 26: 311–317. Padmanabhan R. 1984. Experimental induction of cranioschisis aperta and exencephaly after neural tube closure - a rat model. J Neurol Sci 66: 235–243. Padmanabhan R. 1988. Light microscopic studies on the pathogenesis of exencephaly and cranioschisis induced in the rat after neural tube closure. Teratology 37:29–36. Padmanabhan R. 1991. Is exencephaly the forerunner of anencephaly: an experimental study on the effect of prolongation of gestation on the exencephaly induced after neural tube closure in the rat. Acta Anat 141:182–192. Padmanabhan R. 1998. Retinoic acid-induced caudal regression syndrome in the mouse fetus. Reprod Toxicol 12:139–151. Padmanabhan R, Al Zuhair AGH. 1989. Effect of streptozotocin induced maternal diabetes on fetal skeletal development in the rat. Cong Anom 29:151–163 Padmanabhan R, Hameed MS. 1990. Characteristics of the limb malformations induced by maternal exposure to cadmium in the mouse. Reprod Toxicol 4:291–304. Pang D, Hoffman HJ. 1980. Sacral agenesis with progressive neurological deficit. Neurosurgery 7:118–126. Pappas CTE, Seaver L, Carrion C, Rekate H. 1989. Anatomical evaluation Stevenson RE, Jones KL, Phelan MC, Jones MC, Barr M Jr, Clericuzio C, Harley RA, Benirschke K. 1986. Vascular steal: the pathogenetic mechanism producing sirenomelia and associated defects of the viscera and soft tissues. Pediatrics 78:451–457. Stocker JT, Heifetz SA. 1987. Sirenomelia: a morphological study of 33 cases and review of the literature. Perspect Pediatr Pathol 10:7–50. Tang TT, Oechler HW, Hinke, DH, Segura AD, Franciosi RA. 1991. Limbbody wall complex in association with sirenomelia sequence. Am J Med Genet 41:21–25. Van Allen MI. 1981. Fetal vascular disruptions: mechanisms and some resulting birth defects. Pediatr Ann 10:219–233. Wei X, Sulik KK. 1996. Pathogenesis of caudal dysgenesis induced by ochratoxin A in chick embryos. Teratology 53:378–391. Weigert C. 1886. Zwei Falle von Missbildung eines Ureter und einer Samenblase mit Bemerkungen uber einfache Nabelarterien. Virchows Arch Pathol Anat 104:10–20. Welch IP, Aterman K. 1984. The syndrome of caudal dysplasia: a review including etiologic considerations and evidence of heterogeneity. Pediatr Pathol 2:313–327. Wilson JSP, Vallance-Owen J. 1966. Congenital deformities and insulin antagonism. Lancet 2:940–942. Young ID, O’Reilly KM, Kendall CH. 1986. Etiological heterogeneity in sirenomelia. Pediatr Pathol 5:31–43.