TERATOLOGY 45:65-74 (1992)
Teratogenicity of All-trans Retinoic Acid During Early Embryonic Development in the Cynomolgus Monkey (Macaca fascicularis) ANDREW G. HENDRICKX AND HANS HUMMLER California Regional Primate Research Center, University of California, Davis, California 95616 (A.G.H.);F . Hoffinann-La Roche Ltd., Basel, Switzerland (H.H.)
ABSTRACT The embryotoxic and teratogenic potential of all-trans retinoic acid was assessed following exposure prior to and during early organogenesis in the cynomolgus monkey (Macaca fascicularis). Sixteen pregnant females were orally administered all-trans retinoic acid (Tretinoin, Hoffmann-La Roche) once daily from GD 10-20 and twice daily from GD 21-24 at three different dosages, 5 (n=9), 10 (n=6) and 20 mg/kg ( n = l ) . Adverse clinical signs resembling hypervitaminosis A were observed in one animal a t 5 mg/kg, in three animals at 10 mg/kg, and in the animal treated with 20 mglkg all-trans retinoic acid. Maternal weight loss was observed in the 10- and 20-mgkg groups. A dosedependent increase in embryolethality was observed, with 22% (2/9), 50% (3/6), and 100% (1/1)occurring at 5, 10, and 20 mg/kg, respectively. The majority of embryonic deaths occurred between GD 16 and 20; the incidence of these early losses was higher than in historical and concurrent controls. No malformations, but a single growth-retarded fetus, was observed in the 5-mgikg group. Craniofacial malformations, consisting of external ear defects, mandibular hypoplasia, cleft palate, and temporal bone abnormalities, were seen in three viable fetuses in the 10-mg/kggroup. Skeletal variations were common to the majority (70%, 7/10) of viable fetuses in both dose groups and were increased relative to historical controls (32%, 25/77). Unlike previous studies with 13-cis-retinoic acid during the pre- and early organogenic stages of development (Hummler et al., Teratology 42~263-272, 1990), no thymic hypo- or aplasia or heart anomalies were observed, which may be attributable to the slightly longer 13-cis retinoic acid treatment period, i.e., GD 10-27. However, external ear and temporal bone defects were common to both all-trans and 13-cis retinoic acid. The similarity observed in the malformation syndrome induced by both all-trans and 13-cis retinoic acid in the cynomolgus monkey and 134s retinoic acid embryopathy in humans supports this macaque species as a model for further developmental toxicity studies of vitamin A-related compounds. al., '90; Eichele, '89). The teratogenicity of Vitamin A and its derivatives (retinoids) certain vitamin A-related compounds has have been the subject of intense investiga- been well documented in laboratory anition over the past 30 years because of their mals (archival data from Hoffmann-La beneficial effects in a variety of dermatolog- Roche, '77-730; Howard and Willhite, '86). ical disorders as well as in the treatment of To date, the developmental toxicity of isoneoplastic diseases. The interest in these compounds has been recently enhanced by Received May 28,1991;accepted August 27,1991. the growing evidence that retinoids are also Address reprint requests to Dr. Andrew G. Hendricku, Caliin the Of devel- fornia Regional Primate Research Center, University of Califoropmental processes (Brockes, '89; Doll6 et nia, Davis, CA 95616-8542. INTRODUCTION
0 1992 WILEY-LISS,
INC.
66
A.G. HENDRICKX AND H. HUMMLER
tretinoin (13-cis-retinoic acid) and etreti- cages and maintained in accordance with nate has also been confirmed in humans standards established by the United Stages (Lammer et al., '85; Rosa et al., '86; Ellis Federal Animal Welfare Act and Institute and Voorhees, 1987). Retinoic acid embry- for Laboratory Animal Resources (ILAR). opathy consists of craniofacial (especially Monkeys were fed Purina Monkey Chow the external ear), heart and great vessel, (25% protein) twice daily with fruit supplethymic, and central nervous system (CNS) mentation during the treatment period, abnormalities. The potential teratogenicity provided with water ad libitum from autoof megadoses of vitamin A (retinol) has also matic lixit devices, and maintained on a 12-hr lighting schedule in animal rooms been of recent concern (Miller e t al., '87). Tretinoin (all-trans retinoic acid) is a nat- having a year-round temperature of approxural metabolite of retinol that has been imately 22°C and 60% relative humidity. important in the topical treatment of acne All animals in the CRPRC research colony vulgaris and, more recently, has shown a are hand-caught and handled routinely beneficial effect in photodamaged skin. without sedation or anesthesia, at least Owing to severe systemic toxicity of oral twice monthly, for weighing, cage cleaning, preparations, it is used clinically in topical testing for tuberculosis, collection of blood, form only and does not appear to be associ- ultrasound examination, and drug treatated with abnormal development (Boyd, ment. The first day of menstrual bleeding was '89). However, questions remain concerning long-term exposure to this drug, and designated as day 1 of the cycle. Animals well-controlled epidemiological studies are were mated for a 2-hr period with a fertile needed to clarify this issue (SCRIP, 1989). male on two alternate days during midcycle Although the pattern of malformations is (between cycle days 10 and 14).The second similar, the increased teratogenic potency of (last) day of positive mating, indicated by all-trans retinoic acid relative to 13-cis the presence of sperm in a vaginal smear, retinoic acid has been demonstrated for sev- was considered GD 0. Pregnancy was eral species, including the mouse, rat, and confirmed by ultrasound on GD 16-19 hamster (Howard and Willhite, '86; archival (Tarantal and Hendrickx, '88a). All pregnancies were monitored sonographically in data from Hoffmann-La Roche, '77-'80). Studies in the mouse indicate that this phe- order to verify embryoniclfetal growth and nomenon may be attributable to differential to confirm viability as evidenced by the transplacental pharmacokinetics of these presence of cardiac activity (Tarantal and two compounds (Nau et al., '87; Creech Hendrickx, '88b). Examinations on handKraft et al., '87). held animals were performed approxiWe recently investigated the developmen- mately every 3-5 days between GD 16 and tal toxicity of 13-cis retinoic acid orally ad- 70, then every 10 days until GD 100 k 2. ministered to the cynomolgus monkey Abortions and resorptions were verified by (Macaca fascicularis) prior to and during ultrasonography and/or by recovery of the early organogenesis (gestational days [GDI products of conception. 10-27). Several features of the human synAll-trans retinoic acid (Tretinoin, Hoffdrome, i.e., ear, thymic, and cardiovascular mann-La Roche, LTD., No. RO 01-54881 defects, were observed at a similar dose (2.5 000,99.8% purity), was stored a t 4°C in ammg/kg) to that used clinically (Hummler et ber vials under nitrogen or argon gas. Drug al., '90). The present study was carried out purity was certified by chemical analysis to compare the developmental toxicity of prior to experimental application. The drug all-trans retinoic acid following a similar was prepared a s a suspension in rape-seed treatment regimen in the cynomolgus mon- oil that had been previously gassed with N, and allowed to sit for approximately 12 hr. key. Suspensions of dosing solutions were prepared fresh daily under dimmed lighting usMATERIALS AND METHODS ing a magnetic stirrer. Dosing concentraFemale cynomolgus monkeys (M. fascicu- tions of 2,4, and 8 mglml were administered laris) with a proven reproductive history in a constant dosing volume of 2.5 ml/kg. and weighing 2.7-4.6 kg were individually Appropriate volumes of each suspension housed a t the California Regional Primate were based on the most recent body weight Research Center (CRPRC) in aluminum and administered to unfasted animals by
67
ALL-TRANS RETINOIC ACID TERATOGENICITY IN MACAQUES
Dose (mg/kg) 5
TABLE 1 . Developmental toxicity of all-trans retinoic acid in M. fascicularis Treatment Outcome No. of Embryo(%) Fetuses Malformations lethality Days pregnancies 10-20 9 2l 22 7 0
2x5 10 2x10 20 2x20 2 x 10
21-24 10-20 21-24 10-20 21 22-24
(%I 0
6
3*
50
3
34
100
1
13
100
0
-
-
'GD 16-18 and GD 18-20. 'GD 18-19, GD 26-30, and GD 48. 3Before GD 18. 4Craniofacial (ear defects, n = 2; mandibular hypoplasia, n = 2; cleft palate, n = 1).
nasogastric intubation under dimmed lighting conditions. An aliquot of each dosing suspension was frozen (-20°C) for retrospective confirmation of retinoic acid concentration. All-trans retinoic acid was administered a t 5, 10, and 20 mglkg to a total of 16 pregnant females that were randomly assigned to treatment groups (Table 1). Because of a high prenatal loss rate a t dosages 210 mgl kg, only one animal was assigned to the 20mgikg treatment group. The animals were dosed prior to and during early organogenesis; i.e., once daily from GD 10-20 and twice daily a t approximately 8-hr intervals from GD 21-24. This dosing regimen was chosen to resemble that used previously to study the teratogenicity of 13-cis retinoic acid in the same species, M. fascicularis (Hummler et al., '90). The change in drug dosing from once to twice daily was done coincident with the onset of organogenesis (i.e., the highly teratogenic susceptible period) to increase the chance for teratogenicity yet limiting embryolethality during the preorganogenic period. Blood samples (2 ml) were obtained from all animals prior to treatment on GD 5 and during treatment a t 2 , 3 , 8 , and 24 hr on GD 15 and 17 and at 2,3, and 8 hr on GD 22 and 24. Samples were collected in tubes containing ammonium and potassium oxalate as anticoagulants, centrifuged at 10,000 rpm, and plasma samples were stored frozen (-20°C) until assayed. All blood sampling and storage procedures were carried out under dimmed lighting conditions. Evaluation of plasma retinoic acid levels will be the subject of a separate publication. Maternal body weights were obtained
once pretreatment, weekly during the treatment period and monthly post-treatment until pregnancy termination. All animals were examined twice daily for adverse physical and behavioral signs, food intake, and vaginal bleeding. Hysterotomies were performed on GD 100 ? 2. Fetuses were weighed, sexed, measured, photographed and examined for external malformations prior to evisceration. A standard teratological examination included measurements of body weight, crown-rump and crown-hip lengths, femur length, foot length, biparietal and occipitofrontal diameters, head and chest circumference, and anogenital distance; evaluation of the gross appearance and weight of the brain, thymus, spleen, liver, adrenals, and kidneys; a placental examination; and a skeletal evaluation of Alizarin Red Sstained carcasses. The hearts were examined as previously described by Hendrickx et al., ('85). A total of 77 vehicle control pregnancies on various teratology treatment regimens was used to evaluate the maternal (body weights and clinical signs) and fetal (weights and measurements) data. Of these, 10 pregnant animals were concurrently handled and treated (e.g., nasogastric intubation) in a manner similar to that of the treated animals in the current study, except that the vehicle was methylcellulose. Weights and measurements of individual fetuses below the 95% confidence interval of control values (X 2 SD range) were considered significant. The probability that the incidence of skeletal variations occurred in the absence of a treatment effect was determined using the chi-square test (Daniel, '83).
*
68
A.G. HENDRICKX AND H. HUMMLER RESULTS
Maternal toxicity One animal in the 5-mg/kg dose group exhibited chapped mouth on 4 consecutive days at the end of the treatment period. No other significant signs of maternal toxicity were noted in the low dose group. Chapped mouth was also observed for 4-6 days in two animals treated with 10 mgkg all-trans retinoic acid. An increased incidence of other adverse clinical signs resembling hypervitaminosis A, i.e., swollen eyelids, poor appetite, and mild diarrhea, were also observed in this dose group. Swollen eyelids were reported for 1-10 days during treatment in three animals. All six treated animals experienced poor appetite for 2-12 days at the end of and after the period of drug administration. Diarrhea (1-2 days duration) occurred in four animals in this group. The one animal treated with 20 mg/ kg all-trans retinoic acid experienced swollen eyelids (1day), chapped mouth (2 days), poor appetite (8 days), and diarrhea (3 days) during the period of drug treatment. Normal weight gain was observed in the 5-mg/kg dose group during and after treatment relative to historical controls. By contrast, most of the animals treated with 10 mg/kg retinoic acid, as well as the animal in the 20-mg/kg dose group, experienced weight loss during and/or after the period of treatment. However, normal weight gain was restored by hysterotomy in two of three animals (insufficient weight data on one animal) with viable fetuses at GD 100 ? 2. Embryolethality An increased incidence of prenatal mortality was observed with increasing dose of all-trans retinoic acid: 2/9 (22%)at 5 mg/kg, 3/6 (50%)at 10 mg/kg, and 1/1(100%)at 20 mg/kg (Table 1). This dose-related embryolethality prevented further animal assignments to the high (20-mgikg) dose group. The majority of losses (two at 5 mg/kg and one each at 10 and 20 mg/kg)) were detected by ultrasonography as early resorptions between GD 16 and 20, and no tissue was available for examination. Two later losses (GD 26-30 resorption and GD 48 abortion with tissue fragments only) occurred in the 10-mg/kg dose group. Teratology No malformations were observed in the seven fetuses exposed to 5 mg/kg all-trans
retinoic acid; however, one fetus exhibited intrauterine growth retardation (IUGR). All three viable fetuses treated with 10 mg/ kg all-trans retinoic acid exhibited craniofacial defects. Bilateral external ear defects and mandibular hypoplasia were observed in one fetus (Fig. 1).The ear defects consisted of absent helix and crura of antihelix (bilateral), absent tragus (right side only), and ectopic tissue tags and pits (bilateral). Skeletal staining of this fetus revealed an asymmetric mandible with a shorter mandibular body on the left, bilateral reduction of coronoid and condylar processes, and a more obtuse-than-normal angle between the right side of the mandibular body and ramus. Temporal bone defects included reduced ossification of the squamae, abnormally shaped zygomatic processes (i.e., abnormal thickening), as well as irregular ossification in the subarcuate fossa region (i.e., absence of arcuate eminence) (Fig. 2). Less severe unilateral (right side) auricular defects, consisting of a poorly defined tragus in addition to a deep pit behind the antitragus and an ectopic tissue tag on the inner surface of the lobule, were observed in a second fetus. A third fetus displayed cleft palate and mandibular hypoplasia (Fig. 3). All weights and measurements of the three fetuses exposed to 10 mg/kg all-trans retinoic acid were within the normal control range. No malformations were observed in the control fetuses (n = 77). Most viable fetuses in this study exhibited one or more skeletal variations. The 71% (5/7) incidence in the 5 mg/kg dose group consisted of extra ribs on the 7th cervical vertebrae (n = 3) and variations in normal vertebrae count (n = 3) or morphology (i.e., asymmetry, n = 1). One fetus exposed to 10 mg/kg all-trans retinoic acid exhibited a hypoplastic 12th rib pair. The other fetus in this dose group had an absent rib pair (11 pairs), hypoplastic ribs (1st and l l t h , unilateral) and delayed ossification of the medial one-third of the scapulae. One or more variations in ribhertebral number or morphology were observed in 25 of the 77 controls fetuses (32%). DISCUSSION
Maternal toxicity The signs of maternal toxicity in the present study (swollen eyelids, chapped mouth, and gastrointestinal disorders) encompass part of a wider range of side effects
ALL-TRANS RETINOIC ACID TERATOGENICITY IN MACAQUES
69
A
Fig. 1. A,B. Diagram and photograph, respectively, of the right auricle from a control GD 100 cynomolgus macaque fetus. In the diagram, first pharyngeal arch derivatives are designated by diagonal shading; the remaining auricle is derived from the second pharyngeal arch. C,D. Abnormal right and left auricles, respectively, from a fetus administered all-trans retinoic acid. Bilateral defects include absence of the helix and crura of the antihelix, and ectopic tissue tags (arrowheads). The auricle in (C) also has additional tissue tags behind the scapha region (obscured in photograph) and a n absent tragus.
associated with oral administration of alltrans retinoic acid in humans. Since therapeutic levels of this drug are associated with symptoms of hypervitaminosis A (including inflammatioddrying of the skin and mucous membranes, hair loss, headache, and gastrointestinal disorders), all-trans retinoic acid is almost exclusively administered topically rather than systemically for the treatment of various dermatological disorders (Lucek and Colburn, '85; Biesalski, '89).
A high incidence of hyperemic eyelids and facial erythema was observed in rhesus monkeys treated with higher doses of alltrans retinoic acid (20 or 40 mg/kg/day for 4-8 days during organogenesis) than used
in the current study (Hendrickx et al., '80). Maternal toxicity was not reported as an experimental endpoint in two additional nonhuman primate teratology studies with 7.5-80 mg/kg all-trans retinoic acid administered during organogenesis (Wilson, '74; Fantel et al., '77). It should be noted that drug purity and protection against biodegradation (light, air exposure) were not indicated in these earlier studies. Dose-related maternal toxicity observed in a previous teratology study with 13-cis retinoic acid in cynomolgus monkeys (Hummler et al., '90) included weight reduction, diarrhea, encrusted nose and/or eyelids, increased lacrimation, rhinorrhea, and swollen eyelids and forehead in animals treated with 10 and 25
70
A.G. HENDRICKX AND H. HUMMLER
Fig. 2. A. Control GD 100 cynomolgus macaque fetus stained with Alizarin Red-S. Labeled features include the temporal squama (S), ramus (R) and body (B) of the mandible, and zygomatic process of the temporal bone (arrowhead). B. In this fetus administered alltrans retinoic acid, note the reduction in the temporal squama and mandibular ramus and body. Also note the abnormal thickening at the root of the zygomatic process (between arrowheads). C. Inside the cranial fossa of
mgkg for 11 days. Less severe side effects (reduction in food consumption and diarrhea) were associated with lower doses (2-5 mg/kg) and/or a shorter treatment period during early pregnancy.
the control fetus in A, the periosteal membrane has been partially removed from the temporal bone to reveal the region of the arcuate eminence (arrow) and medial aspect of the subarcuate fossa (asterisk). This arched region transmits the osseous labyrinth of the anterior semicircular canal. D. Compare the absence of the arcuate eminence and abnormal configuration of the subarcuate fossa in this treated fetus from B.
monkeys during this period of development. Furthermore, the very low incidence of spontaneous abortions before GD 100 (3.5%, 4/113) in the vehicle-treated controls compared with 17.2% (34/198) in our cynomolgus breeding colony in which no conditionEmbryolethality ing is done, verifies that experimental The overall incidence of dose-related em- handling alone has no adverse effects on bryolethality in the current study was pregnancy outcome. 37.5% (6/16). Most losses occurred prior to All-trans retinoic acid has been associated the onset of organogenesis (i.e., GD 16-20). with increased embryolethality in previous This 25% (4/16) incidence of early loss con- nonhuman primate studies. Wilson ('74) retrasts with the 4.2% (5/119) rate of resorp- ported a 44% (7/16) rate of intrauterine tion during this developmental period in un- death in rhesus pregnancies exposed to 10treated cynomolgus monkeys as detected 80 mg/kg on various treatment regimens beby ultrasound a t the CRPRC. Two losses tween GD 20 and 46. By contrast, no signif(12.5%)occurred during organogenesis (GD icant increase in prenatal mortality (3/15, 20-50) in the 10 mgkg dose group. This 20%) was observed following treatment of rate also exceeds the CRPRC 7-year (1983- rhesus monkeys with 20-40 mg/kg all-trans 1990) prenatal mortality rate (2.7%, 3/113) retinoic acid for 4- to 8-day periods during observed in vehicle-treated cynomolgus organogenesis (Hendrickx et al., '80). Treat-
ALL-TRANS RETINOIC ACID TERATOGENICITY IN MACAQUES
71
Fig. 3. A. Profile of a control GD 100 cynomolgus macaque fetus. B. Mandibular hypoplasia in a fetus administered all-trans retinoic acid. C. Cleft palate was also evident in this treated fetus (arrowhead).
ment of pigtail monkeys with 7.5-10 mgkg of all-trans retinoic acid during the entire period of organogenesis resulted in a high level of prenatal loss -63% (12/19).Shorter treatment periods (1-16 days) with higher doses (10-45 mg/kg) were less embryolethal -26% loss, 5/19 (Fantel et al., '77). Teratology Craniofacial malformations consisting of abnormal ears, hypoplastic mandible and temporal bone, as well as cleft palate, characterized the primary effects of all-trans retinoic acid administered a t 10 mgikg prior to and during early organogenesis in the cynomolgus monkey. The threshold dose of all-trans retinoic acid in this study is considered to be 5 mg/kg, based on the increased frequency of skeletal variations and the case of IUGR in the absence of frank malformations observed in this dose group. It is of note that no signs of hypervitaminosis A were associated with the 5 mgikg dose. It is uncertain whether the production of both maternal hypervitaminosis A and fetal malformations at 10 mg/kg is coincidental or implies a cause-and-effect relationship. The high incidence (70%, 7/10) of skeletal variations observed in viable fetuses in the 5- and lO-mg/kg dose groups significantly exceeds (Pc0.02) the 32% (25/77) frequency of rib and vertebral variations in vehicletreated control cynomolgus monkeys in CRPRC developmental toxicity studies. This increased level of variations is consis-
Fig. 3B,C.
tent with the observed teratogenic potential of all-trans retinoic acid in the current study. It has been suggested that certain skeletal variations (e.g., extra 14th ribs) in rodents may be regarded as indicators of teratogenic potency of a drug a t some higher dosage (Kimmel and Wilson, '73). Increased skeletal variations have also been associated with the treatment with retinoic acid, acitretin, etretinate and other retinoids in rodent studies carried out a t Hoffmann-La Roche Laboratories (archival data from Hoffmann-La Roche, '77-730). Thus, retinoids may provide a good example of a class of compounds that produce skeletal variations at subteratogenic doses. A high incidence (71%) of ear anomalies has also been associated with early oral exposure t o 13-cis retinoic acid (2.5 mgkg, GD 10-25 and 2 x 2.5 mg/kg, GD 26 and 27) in the cynomolgus monkey. The most severely affected structures were the superior aspect of the helix, crura of the antihelix, antitra-
72
A.G. HENDRICKX AND H. HUMMLER
gus, and lobule, all derivatives of the second pharyngeal arch (Hummler et al., '90). In addition to these second arch structures, first arch derivatives (i.e., crus, ascending portion of the helix, and tragus), were also adversely affected in the current study. Hypoplasia of the temporal squamae, abnormally shaped zygomatic arches, and irregular ossification in the subarcuate fossa region were also evident in one all-trans fetus. Similar or slightly later susceptibility to mandibular hypoplasia, temporozygomatic hypo- and hyperplasia, and ear dysplasias was observed following l-day thalidomide exposures in bonnet and african green monkeys during early gestation (i.e., GD 24-28) (Newman and Hendrickx, '81, '85; Hendrickx and Sawyer, '78). Thymic hypo- and aplasia and heart anomalies (VSD and transposition of the great vessels) seen in the fetuses treated with 13-cis retinoic acid were not evident in those exposed to all-trans retinoic acid. Only a low incidence (2/52, 4%) of thymic abnormalities (accessory thymus, athymia) was observed in earlier studies in primates with all-trans retinoic acid when treatment encompassed GD 17-45 (Fantel et al., '77; Hendrickx et al., '80; Wilson, '74). The mandible and palate defects observed in the current study were not associated with exposure to 13-cis retinoic acid, as reported earlier by Hummler et al. ('90). These differences in malformation syndromes may be related to the slightly longer treatment regimen (18 vs. 15 days, i.e., GD 10-27 vs. GD 10-24) with 13-cis retinoic acid or inherent differences in potency of the two isomers. The effective teratogenic dose of 13-cis retinoic acid, 2.5 mglkg, is lower than that of all-trans retinoic acid (10 mg/ kg). Pharmacokinetic and placental transfer data in the cynomolgus monkey are required to ascertain whether these dose differences are related to differences in embryonic exposure to each of these compounds. Preliminary data in nonpregnant cynomolgus monkeys indicate differences in kinetic profiles for 13-cis and all-trans retinoic acid. Creech Kraft et al. ('89) reported a more rapid plasma elimination of the alltrans isomer compared with the 13-cis isomer following 10 days of dosing with 10 mg/ kg of each compound. Although the malformation syndromes for all-trans and 13-cis retinoic acid are similar in mice, rats, and hamsters, there are
marked differences in the teratogenic potencies of the two isomers. The lowest teratogenic doses reported for all-trans retinoic acid in these species are 4, 0.4-2, and 12.5 mg/kg/day, respectively, compared with 100, 150, and 25 mg/kg/day, respectively, for 13-cis retinoic acid (Miller et al., '87). These observed differences in potency may be related to differential absorption, distribution, metabolic fate, placental permeability, or to interaction with putative embryonic intracellular receptors (Howard et al., '89). Craniofacial malformations have also been described following administration of all-trans retinoic acid in other nonhuman primate studies. Hendrickx et al. ('80) observed midfacial and mandibular asymmetry and hypoplasia, arched palate, abnormal cranial dimensions, and malformed ears in rhesus monkeys treated with 20-40 mg/kg (p.0.) for 4-8 days between GD 17 and 45. The ear anomalies, which occurred at a 29% (4114) incidence, consisted of cupping of the helix and lobule, bilateral microtia with atresia of the external meatus and low-set ears. Brain and appendicular and axial skeletal defects were also observed in this study. Of four malformed rhesus monkey fetuses exposed for different periods of time to 10-40 mg/kg (p.0.) all-trans retinoic acid during organogenesis, one exhibited an abnormal ear and two had cleft palate (Wilson, '74). A retinoic acid syndrome, characterized by craniofacial and skeletal defects, was described by Fantel and colleagues in 58% (11/19) of pig-tail monkey fetuses exposed to oral doses of 7.5 and 10 mgkg from GD 18-44 (Fantel et al., '77; Newell-Morris et al., '80). The ear anomalies observed in seven fetuses consisted of unfused ear tags and displacement of the external auditory meatus. In addition to malformed ears and eyes, cleft palate, and abnormalities of the mid-face and mandible, axial and appendicular skeletal defects and muscular-joint contractures were also evident. Similar to the current study, visceral defects were uncommon in these nonhuman primate studies with all-trans retinoic acid. The increased variety of malformations in these earlier studies may be attributable to higher doses and/or longer treatment periods during organogenesis. Table 2 summarizes the developmental toxicity associated with all-trans retinoic acid reported in macaques to date.
ALL-TRANS RETINOIC ACID TERATOGENICITY IN MACAQUES
73
TABLE 2 . Primate developmental toxicity studies with all-trans retinoic acid' Current study Hendrickx et al. ('80) Fantel et al. ('77) Wilson ('74) (Macaca mulatta) (Macaca fascicularis) (Macaca mulatta) (Macaca nemestrina) Dose (mglkgiday, p.0.) 10-80 7.5-10 20-40 5-20 Treatment davs 20-46 (variable) 18-44 17-45 (4-8 davs) 10-24 Maternal toxi"city Weight loss X Gastrointestinal X X X Facial Embryo/fetal lethality X X X Malformations Craniofacial X X X X Ears X X X Mandible X X X X Palate X X Skeletal X X Thymus X X2 Heart X Brain X2 'x, present; -, not reported
%=I.
13-cis Retinoic acid is a known human teratogen that produces a characteristic pattern of malformations involving craniofacial, cardiac, thymic, and CNS structures (Lammer et al., '85;Rosa et al., '86). Most of the features of the human syndrome have been duplicated in the cynomolgus monkey with early exposure to all-trans retinoic acid in the current study or 13-cis retinoic acid (Hummler et al., '90). The most common craniofacial defect in human infants involves the external ear (microtia or anotia); other facial anomalies include micrognathia, cleft palate, flattened nasal bridge, and hypertelorism (Lammer et al., '85;Rosa et al., '86). While both isomers induce a high level of ear defects in the cynomolgus monkey, all-trans retinoic acid also produces micrognathia and cleft palate. Most of the cardiac malformations in humans are categorized as conotruncal defects, including transposition of the great vessels, tetralogy of Fallot, double-outlet ventricle, truncus arteriosus, and VSD. Transposition of the great vessels and VSD have been observed in cynomolgus monkey fetuses exposed to 13-cis retinoic acid. Human infants with cardiac defects often exhibit abnormalities of the thymus, including ectopia, hypoplasia, or aplasia. 13-cis Retinoic acid induces a high level (57%) of thymic hypoplasia or aplasia in the cynomolgus monkey. CNS defects associated with prenatal exposure to retinoic acid in humans include hydrocephalus, microcephaly, and cerebellar abnormalities, especially agenesis of the vermis. Further examination of cynomolgus mon-
key fetuses treated with 13-cis retinoic acid (Hummler et al., '90) revealed two cases of vermis hypoplasia (unpublished results). On the basis of limited epidemiologic data, there does not appear to be any evidence of teratogenicity associated with topical all-trans retinoic acid (tretinoin). However, further studies involving blood levels in pregnant women are warranted, since questions remain about long-term exposure to this drug (SCRIP, '89). Little, if any, topically applied all-trans retinoic acid is absorbed systemically (Lucek and Colburn, '85).Based on the extensive animal literature, one could speculate that the differences in the teratogenic potencies of 13-cis and all-trans retinoic acid in humans may be largely related to their different routes of exposure (oral vs. dermal) and pharmacokinetic profiles. The clinical importance of vitamin A and its derivatives in anticancer and dermatologic therapy is well recognized. Currently, additional synthetic retinoids are being developed in an effort to enhance efficacy and diminish drug-related toxicity, particularly in women of child-bearing age (Kerker and Farmer, '90; Boyd, '89). The present investigation verifies the similarity of the malformation syndromes induced by both alltrans and 13-cis retinoic acid in the cynomolgus monkey and 13-cis retinoic acid embryopathy in humans. The reproductive, endocrinologic, developmental, and metabolic similarities between the nonhuman primate and humans further support their use as a model for developmental toxicity
74
A.G. HENDRICKX AND H. HUMMLER
studies of vitamin A-related compounds (Hendrickx and Binkerd, '90; Hummler et al., '90). ACKNOWLEDGMENTS
We wish to acknowledge the excellent assistance of Pam Peterson, Jeff Rowland, and Dr. Alice Tarantal in conducting the study and in preparing the manuscript. LITERATURE CITED Archival Data from Hoffmann-La Roche (1977 and 1980). Biesalski. H.K. (1989) Comuarative assessment of the toxicolo'gy of vitamin A anh retinoids in man. Toxicology, 57t117-161. Boyd, A.S. (1989) An overview of the retinoids. Am. J . Med., 86.568474. Brockes, J.P. (1989) Retinoids, homeobox genes, and limb morphogenesis. Neuron, 2:1285-1294. Creech Kraft, J., D.M. Kochhar, W.J. Scott, and H. Nau (1987) Low teratogenicity of 13-cis retinoic acid (isotretinoin) in the mouse corresponds to low embryo concentrations during organogenesis: Comparison to the all-trans isomer. Toxicol. Appl. Pharmacol., 87: 474-482. Creech Kraft, J.,L. Roberts, J.R. Baily, H. Nau, and W. Slikker, Jr. (1989) Pharmacokinetics of 13-cis- and all-trans-retinoicacid in the cynomolgus monkey. Teratology, 39:447A. Daniel, W. W. (1983) Biostatistics: A Foundation for Analysis in the Health Sciences, 3rd Ed. John Wiley & Sons, Inc., New York. Dolle, P., E. Ruberte, P. Leroy, G. Morris-Kay, and P. Chambon (1990) Retinoic acid receptors and cellular retinoid binding proteins. I. A systematic study of their differentia1 pattern of transcription during mouse organogenesis. Development, 110:1133-1151. Eichele, G. (1989) Retinoids and vertebrate limb pattern formation. Trends Genet., 5:246-251. Ellis, C.N., and J.J. Voorhees (1987) Continuing medical education (Therapy): Etretinate therapy. J. Am. Acad. Dermatol., 16:267-291. Fantel, A.G., T.H. Shepard, L.L. Newell-Morris, and B.C. Moffett (1977) Teratogenic effects of retinoic acid in pigtail monkeys (Macaca nemestrinu). I. General features. Teratology, 15:65-72. Hendrickx, A.G., and P.E. Binkerd (1990) Nonhuman primates and teratological research. J . Med. Primatol., 19:81-108. Hendrickx, A.G., M. Cukierski, S. Prahalada, G. Janos, and J . Rowland (1985) Evaluation of bendectin embryotoxicity in nonhuman primates. I. Ventricular septa1 defects in prenatal macaques and baboon. Teratology, 32r179-189. Hendrickx, A.G., and R.H. Sawyer (1978) Developmental staging and thalidomide teratogenicity in the green monkey (Cercopithecus aethiopsi. Teratology, 18:393-404. Hendrickx, A.G., S. Silverman, M. Pellegrini, and A.J. Steffek (1980) Teratological and radiocephalometric analysis of craniofacial malformations induced with retinoic acid in rhesus monkeys (Mucaca mulatta). Teratology, 22:13-22.
Howard, W.B., and C.C. Willhite (1986) Toxicity of retinoids in humans and animals. J . Toxicol. Toxin Rev., 5.55-94. Howard. W.B.. C.C. Willhite, S.T. Omaye, and R.P. S h a k a (1989) Comparative distribution, pharmacokinetics and placental permeabilities of all-transretinoic acid, 13-cis-retinoic acid, all-trans-4-0x0retinoic acid, retinyl acetate and 9-cis-retinal in hamsters. Arch. Toxicol., 63:112-120. Hummler, H., R. Korte, and A.G. Hendrickx (1990) Induction of malformations in the cynomolgus monkey with 13-cis-retinoic acid. Teratology, 42:263-272. Kerker, B.J., and E.R. Farmer (1990) Update on retinoids. Saudi Med. J., 11:165-170. Kimmel, C.A., and J.G. Wilson (1973) Skeletal deviations in rats: Malformations or variations? Teratology, 8:309-316. Lammer, E.J., D.T. Chen, R.M. Hoar, N.D. Agnish, P.J. Benke, J.T. Braun, C.J. Curry, P.M. Fernhoff, A.W. Grix, I.T. Lott, J.M. Richard, and S.C. Sun (1985) Retinoic acid embryopathy. N. Engl. J . Med., 313: 837-841. Lucek, R.W., and W.A. Colburn (1985) Clinical pharmacokinetics of the retinoids. Clin. Pharmacokinet., 10: 38-62. Miller, R.K., K. Brown, J . Cordero, D. Dayton, B. Hardin, M. Greene, C. Grabowski, A. Hendrickx, E. Hook, R. Jensh, G. Kimmel, J. Manson, K.O. Shea, J. Schardein, and R. Staples (1987) Teratology Society Position Paoer: Recommendations for Vitamin A Use During Pregnancy. Teratology, 35:269-275. Nau, H., D.M. Kochhar, J.M. Creech Kraft, B. Loefberg, J . Reiners. and T. Sparenberg (1987) Teratogenesis of retinoids: Aspects o i species aifferences andtransplacental pharmacokinetics. In: Approaches to Elucidate Mechanisms in Teratogenesis. F. Welsch, ed. Hemisphere Publishing Corporation, Washington, D.C., pp. 1-15. Newell-Morris, L., J.E. Sirianni, T.H. Shepard, A.G. Fantel, and B.C. Moffett (1980) Teratogenic effects of retinoic acid in pigtail monkeys (Macaca nemestrina). XI. Craniofacial features. Teratology, 22:87-101. Newman, L.M., and A.G. Hendrickx (1981) Fetal ear malformations induced by maternal ingestion of thalidomide in the bonnet monkey (Mucaca radiata). Teratology 23:351-364. Newman, L.M., and A.G. Hendrickx (1985) Temporomandibular malformations in the bonnet monkey (Macacu radiata) fetus following maternal ingestion of thalidomide. J . Craniofac. Genet. Dev. Biol., 5~147157. Rosa, F.W., A.L. Wilk, and F.O. Kelsey (1986) Teratogen update: Vitamin A congeners. Teratology, 33: 355-364. SCRIP, No. 1374/5:17 (1989) FDA and teratogenic potential of Retin-A. Tarantal, A.F., and A.G. Hendrickx (1988a) Use of ultrasound for 'early pregnancy detection in the rhesus and cynomolgus macaque (Macacu mulatta and Macaca fascicularis). J . Med. Primatol., I7:105-112. Tarantal, A.F., and A.G. Hendrickx (1988b) Characterization of prenatal growth and development in the crab-eating macaque (Macaca fasciculuris) by ultrasound. Anat. Rec., 222:177-184. Wilson, J.G. (1974)Teratologic causation in man and its evaluation in non-human primates. In: Birth Defects. A.G. Motulsky, ed. American Elsevier Publishing Co., New York, p i . 191-203. ~~
Teratogenicity of All-trans Retinoic Acid During Early Embryonic Development in the Cynomolgus Monkey (Macaca fascicularis) ANDREW G. HENDRICKX AND HANS HUMMLER California Regional Primate Research Center, University of California, Davis, California 95616 (A.G.H.);F . Hoffinann-La Roche Ltd., Basel, Switzerland (H.H.)
ABSTRACT The embryotoxic and teratogenic potential of all-trans retinoic acid was assessed following exposure prior to and during early organogenesis in the cynomolgus monkey (Macaca fascicularis). Sixteen pregnant females were orally administered all-trans retinoic acid (Tretinoin, Hoffmann-La Roche) once daily from GD 10-20 and twice daily from GD 21-24 at three different dosages, 5 (n=9), 10 (n=6) and 20 mg/kg ( n = l ) . Adverse clinical signs resembling hypervitaminosis A were observed in one animal a t 5 mg/kg, in three animals at 10 mg/kg, and in the animal treated with 20 mglkg all-trans retinoic acid. Maternal weight loss was observed in the 10- and 20-mgkg groups. A dosedependent increase in embryolethality was observed, with 22% (2/9), 50% (3/6), and 100% (1/1)occurring at 5, 10, and 20 mg/kg, respectively. The majority of embryonic deaths occurred between GD 16 and 20; the incidence of these early losses was higher than in historical and concurrent controls. No malformations, but a single growth-retarded fetus, was observed in the 5-mgikg group. Craniofacial malformations, consisting of external ear defects, mandibular hypoplasia, cleft palate, and temporal bone abnormalities, were seen in three viable fetuses in the 10-mg/kggroup. Skeletal variations were common to the majority (70%, 7/10) of viable fetuses in both dose groups and were increased relative to historical controls (32%, 25/77). Unlike previous studies with 13-cis-retinoic acid during the pre- and early organogenic stages of development (Hummler et al., Teratology 42~263-272, 1990), no thymic hypo- or aplasia or heart anomalies were observed, which may be attributable to the slightly longer 13-cis retinoic acid treatment period, i.e., GD 10-27. However, external ear and temporal bone defects were common to both all-trans and 13-cis retinoic acid. The similarity observed in the malformation syndrome induced by both all-trans and 13-cis retinoic acid in the cynomolgus monkey and 134s retinoic acid embryopathy in humans supports this macaque species as a model for further developmental toxicity studies of vitamin A-related compounds. al., '90; Eichele, '89). The teratogenicity of Vitamin A and its derivatives (retinoids) certain vitamin A-related compounds has have been the subject of intense investiga- been well documented in laboratory anition over the past 30 years because of their mals (archival data from Hoffmann-La beneficial effects in a variety of dermatolog- Roche, '77-730; Howard and Willhite, '86). ical disorders as well as in the treatment of To date, the developmental toxicity of isoneoplastic diseases. The interest in these compounds has been recently enhanced by Received May 28,1991;accepted August 27,1991. the growing evidence that retinoids are also Address reprint requests to Dr. Andrew G. Hendricku, Caliin the Of devel- fornia Regional Primate Research Center, University of Califoropmental processes (Brockes, '89; Doll6 et nia, Davis, CA 95616-8542. INTRODUCTION
0 1992 WILEY-LISS,
INC.
66
A.G. HENDRICKX AND H. HUMMLER
tretinoin (13-cis-retinoic acid) and etreti- cages and maintained in accordance with nate has also been confirmed in humans standards established by the United Stages (Lammer et al., '85; Rosa et al., '86; Ellis Federal Animal Welfare Act and Institute and Voorhees, 1987). Retinoic acid embry- for Laboratory Animal Resources (ILAR). opathy consists of craniofacial (especially Monkeys were fed Purina Monkey Chow the external ear), heart and great vessel, (25% protein) twice daily with fruit supplethymic, and central nervous system (CNS) mentation during the treatment period, abnormalities. The potential teratogenicity provided with water ad libitum from autoof megadoses of vitamin A (retinol) has also matic lixit devices, and maintained on a 12-hr lighting schedule in animal rooms been of recent concern (Miller e t al., '87). Tretinoin (all-trans retinoic acid) is a nat- having a year-round temperature of approxural metabolite of retinol that has been imately 22°C and 60% relative humidity. important in the topical treatment of acne All animals in the CRPRC research colony vulgaris and, more recently, has shown a are hand-caught and handled routinely beneficial effect in photodamaged skin. without sedation or anesthesia, at least Owing to severe systemic toxicity of oral twice monthly, for weighing, cage cleaning, preparations, it is used clinically in topical testing for tuberculosis, collection of blood, form only and does not appear to be associ- ultrasound examination, and drug treatated with abnormal development (Boyd, ment. The first day of menstrual bleeding was '89). However, questions remain concerning long-term exposure to this drug, and designated as day 1 of the cycle. Animals well-controlled epidemiological studies are were mated for a 2-hr period with a fertile needed to clarify this issue (SCRIP, 1989). male on two alternate days during midcycle Although the pattern of malformations is (between cycle days 10 and 14).The second similar, the increased teratogenic potency of (last) day of positive mating, indicated by all-trans retinoic acid relative to 13-cis the presence of sperm in a vaginal smear, retinoic acid has been demonstrated for sev- was considered GD 0. Pregnancy was eral species, including the mouse, rat, and confirmed by ultrasound on GD 16-19 hamster (Howard and Willhite, '86; archival (Tarantal and Hendrickx, '88a). All pregnancies were monitored sonographically in data from Hoffmann-La Roche, '77-'80). Studies in the mouse indicate that this phe- order to verify embryoniclfetal growth and nomenon may be attributable to differential to confirm viability as evidenced by the transplacental pharmacokinetics of these presence of cardiac activity (Tarantal and two compounds (Nau et al., '87; Creech Hendrickx, '88b). Examinations on handKraft et al., '87). held animals were performed approxiWe recently investigated the developmen- mately every 3-5 days between GD 16 and tal toxicity of 13-cis retinoic acid orally ad- 70, then every 10 days until GD 100 k 2. ministered to the cynomolgus monkey Abortions and resorptions were verified by (Macaca fascicularis) prior to and during ultrasonography and/or by recovery of the early organogenesis (gestational days [GDI products of conception. 10-27). Several features of the human synAll-trans retinoic acid (Tretinoin, Hoffdrome, i.e., ear, thymic, and cardiovascular mann-La Roche, LTD., No. RO 01-54881 defects, were observed at a similar dose (2.5 000,99.8% purity), was stored a t 4°C in ammg/kg) to that used clinically (Hummler et ber vials under nitrogen or argon gas. Drug al., '90). The present study was carried out purity was certified by chemical analysis to compare the developmental toxicity of prior to experimental application. The drug all-trans retinoic acid following a similar was prepared a s a suspension in rape-seed treatment regimen in the cynomolgus mon- oil that had been previously gassed with N, and allowed to sit for approximately 12 hr. key. Suspensions of dosing solutions were prepared fresh daily under dimmed lighting usMATERIALS AND METHODS ing a magnetic stirrer. Dosing concentraFemale cynomolgus monkeys (M. fascicu- tions of 2,4, and 8 mglml were administered laris) with a proven reproductive history in a constant dosing volume of 2.5 ml/kg. and weighing 2.7-4.6 kg were individually Appropriate volumes of each suspension housed a t the California Regional Primate were based on the most recent body weight Research Center (CRPRC) in aluminum and administered to unfasted animals by
67
ALL-TRANS RETINOIC ACID TERATOGENICITY IN MACAQUES
Dose (mg/kg) 5
TABLE 1 . Developmental toxicity of all-trans retinoic acid in M. fascicularis Treatment Outcome No. of Embryo(%) Fetuses Malformations lethality Days pregnancies 10-20 9 2l 22 7 0
2x5 10 2x10 20 2x20 2 x 10
21-24 10-20 21-24 10-20 21 22-24
(%I 0
6
3*
50
3
34
100
1
13
100
0
-
-
'GD 16-18 and GD 18-20. 'GD 18-19, GD 26-30, and GD 48. 3Before GD 18. 4Craniofacial (ear defects, n = 2; mandibular hypoplasia, n = 2; cleft palate, n = 1).
nasogastric intubation under dimmed lighting conditions. An aliquot of each dosing suspension was frozen (-20°C) for retrospective confirmation of retinoic acid concentration. All-trans retinoic acid was administered a t 5, 10, and 20 mglkg to a total of 16 pregnant females that were randomly assigned to treatment groups (Table 1). Because of a high prenatal loss rate a t dosages 210 mgl kg, only one animal was assigned to the 20mgikg treatment group. The animals were dosed prior to and during early organogenesis; i.e., once daily from GD 10-20 and twice daily a t approximately 8-hr intervals from GD 21-24. This dosing regimen was chosen to resemble that used previously to study the teratogenicity of 13-cis retinoic acid in the same species, M. fascicularis (Hummler et al., '90). The change in drug dosing from once to twice daily was done coincident with the onset of organogenesis (i.e., the highly teratogenic susceptible period) to increase the chance for teratogenicity yet limiting embryolethality during the preorganogenic period. Blood samples (2 ml) were obtained from all animals prior to treatment on GD 5 and during treatment a t 2 , 3 , 8 , and 24 hr on GD 15 and 17 and at 2,3, and 8 hr on GD 22 and 24. Samples were collected in tubes containing ammonium and potassium oxalate as anticoagulants, centrifuged at 10,000 rpm, and plasma samples were stored frozen (-20°C) until assayed. All blood sampling and storage procedures were carried out under dimmed lighting conditions. Evaluation of plasma retinoic acid levels will be the subject of a separate publication. Maternal body weights were obtained
once pretreatment, weekly during the treatment period and monthly post-treatment until pregnancy termination. All animals were examined twice daily for adverse physical and behavioral signs, food intake, and vaginal bleeding. Hysterotomies were performed on GD 100 ? 2. Fetuses were weighed, sexed, measured, photographed and examined for external malformations prior to evisceration. A standard teratological examination included measurements of body weight, crown-rump and crown-hip lengths, femur length, foot length, biparietal and occipitofrontal diameters, head and chest circumference, and anogenital distance; evaluation of the gross appearance and weight of the brain, thymus, spleen, liver, adrenals, and kidneys; a placental examination; and a skeletal evaluation of Alizarin Red Sstained carcasses. The hearts were examined as previously described by Hendrickx et al., ('85). A total of 77 vehicle control pregnancies on various teratology treatment regimens was used to evaluate the maternal (body weights and clinical signs) and fetal (weights and measurements) data. Of these, 10 pregnant animals were concurrently handled and treated (e.g., nasogastric intubation) in a manner similar to that of the treated animals in the current study, except that the vehicle was methylcellulose. Weights and measurements of individual fetuses below the 95% confidence interval of control values (X 2 SD range) were considered significant. The probability that the incidence of skeletal variations occurred in the absence of a treatment effect was determined using the chi-square test (Daniel, '83).
*
68
A.G. HENDRICKX AND H. HUMMLER RESULTS
Maternal toxicity One animal in the 5-mg/kg dose group exhibited chapped mouth on 4 consecutive days at the end of the treatment period. No other significant signs of maternal toxicity were noted in the low dose group. Chapped mouth was also observed for 4-6 days in two animals treated with 10 mgkg all-trans retinoic acid. An increased incidence of other adverse clinical signs resembling hypervitaminosis A, i.e., swollen eyelids, poor appetite, and mild diarrhea, were also observed in this dose group. Swollen eyelids were reported for 1-10 days during treatment in three animals. All six treated animals experienced poor appetite for 2-12 days at the end of and after the period of drug administration. Diarrhea (1-2 days duration) occurred in four animals in this group. The one animal treated with 20 mg/ kg all-trans retinoic acid experienced swollen eyelids (1day), chapped mouth (2 days), poor appetite (8 days), and diarrhea (3 days) during the period of drug treatment. Normal weight gain was observed in the 5-mg/kg dose group during and after treatment relative to historical controls. By contrast, most of the animals treated with 10 mg/kg retinoic acid, as well as the animal in the 20-mg/kg dose group, experienced weight loss during and/or after the period of treatment. However, normal weight gain was restored by hysterotomy in two of three animals (insufficient weight data on one animal) with viable fetuses at GD 100 ? 2. Embryolethality An increased incidence of prenatal mortality was observed with increasing dose of all-trans retinoic acid: 2/9 (22%)at 5 mg/kg, 3/6 (50%)at 10 mg/kg, and 1/1(100%)at 20 mg/kg (Table 1). This dose-related embryolethality prevented further animal assignments to the high (20-mgikg) dose group. The majority of losses (two at 5 mg/kg and one each at 10 and 20 mg/kg)) were detected by ultrasonography as early resorptions between GD 16 and 20, and no tissue was available for examination. Two later losses (GD 26-30 resorption and GD 48 abortion with tissue fragments only) occurred in the 10-mg/kg dose group. Teratology No malformations were observed in the seven fetuses exposed to 5 mg/kg all-trans
retinoic acid; however, one fetus exhibited intrauterine growth retardation (IUGR). All three viable fetuses treated with 10 mg/ kg all-trans retinoic acid exhibited craniofacial defects. Bilateral external ear defects and mandibular hypoplasia were observed in one fetus (Fig. 1).The ear defects consisted of absent helix and crura of antihelix (bilateral), absent tragus (right side only), and ectopic tissue tags and pits (bilateral). Skeletal staining of this fetus revealed an asymmetric mandible with a shorter mandibular body on the left, bilateral reduction of coronoid and condylar processes, and a more obtuse-than-normal angle between the right side of the mandibular body and ramus. Temporal bone defects included reduced ossification of the squamae, abnormally shaped zygomatic processes (i.e., abnormal thickening), as well as irregular ossification in the subarcuate fossa region (i.e., absence of arcuate eminence) (Fig. 2). Less severe unilateral (right side) auricular defects, consisting of a poorly defined tragus in addition to a deep pit behind the antitragus and an ectopic tissue tag on the inner surface of the lobule, were observed in a second fetus. A third fetus displayed cleft palate and mandibular hypoplasia (Fig. 3). All weights and measurements of the three fetuses exposed to 10 mg/kg all-trans retinoic acid were within the normal control range. No malformations were observed in the control fetuses (n = 77). Most viable fetuses in this study exhibited one or more skeletal variations. The 71% (5/7) incidence in the 5 mg/kg dose group consisted of extra ribs on the 7th cervical vertebrae (n = 3) and variations in normal vertebrae count (n = 3) or morphology (i.e., asymmetry, n = 1). One fetus exposed to 10 mg/kg all-trans retinoic acid exhibited a hypoplastic 12th rib pair. The other fetus in this dose group had an absent rib pair (11 pairs), hypoplastic ribs (1st and l l t h , unilateral) and delayed ossification of the medial one-third of the scapulae. One or more variations in ribhertebral number or morphology were observed in 25 of the 77 controls fetuses (32%). DISCUSSION
Maternal toxicity The signs of maternal toxicity in the present study (swollen eyelids, chapped mouth, and gastrointestinal disorders) encompass part of a wider range of side effects
ALL-TRANS RETINOIC ACID TERATOGENICITY IN MACAQUES
69
A
Fig. 1. A,B. Diagram and photograph, respectively, of the right auricle from a control GD 100 cynomolgus macaque fetus. In the diagram, first pharyngeal arch derivatives are designated by diagonal shading; the remaining auricle is derived from the second pharyngeal arch. C,D. Abnormal right and left auricles, respectively, from a fetus administered all-trans retinoic acid. Bilateral defects include absence of the helix and crura of the antihelix, and ectopic tissue tags (arrowheads). The auricle in (C) also has additional tissue tags behind the scapha region (obscured in photograph) and a n absent tragus.
associated with oral administration of alltrans retinoic acid in humans. Since therapeutic levels of this drug are associated with symptoms of hypervitaminosis A (including inflammatioddrying of the skin and mucous membranes, hair loss, headache, and gastrointestinal disorders), all-trans retinoic acid is almost exclusively administered topically rather than systemically for the treatment of various dermatological disorders (Lucek and Colburn, '85; Biesalski, '89).
A high incidence of hyperemic eyelids and facial erythema was observed in rhesus monkeys treated with higher doses of alltrans retinoic acid (20 or 40 mg/kg/day for 4-8 days during organogenesis) than used
in the current study (Hendrickx et al., '80). Maternal toxicity was not reported as an experimental endpoint in two additional nonhuman primate teratology studies with 7.5-80 mg/kg all-trans retinoic acid administered during organogenesis (Wilson, '74; Fantel et al., '77). It should be noted that drug purity and protection against biodegradation (light, air exposure) were not indicated in these earlier studies. Dose-related maternal toxicity observed in a previous teratology study with 13-cis retinoic acid in cynomolgus monkeys (Hummler et al., '90) included weight reduction, diarrhea, encrusted nose and/or eyelids, increased lacrimation, rhinorrhea, and swollen eyelids and forehead in animals treated with 10 and 25
70
A.G. HENDRICKX AND H. HUMMLER
Fig. 2. A. Control GD 100 cynomolgus macaque fetus stained with Alizarin Red-S. Labeled features include the temporal squama (S), ramus (R) and body (B) of the mandible, and zygomatic process of the temporal bone (arrowhead). B. In this fetus administered alltrans retinoic acid, note the reduction in the temporal squama and mandibular ramus and body. Also note the abnormal thickening at the root of the zygomatic process (between arrowheads). C. Inside the cranial fossa of
mgkg for 11 days. Less severe side effects (reduction in food consumption and diarrhea) were associated with lower doses (2-5 mg/kg) and/or a shorter treatment period during early pregnancy.
the control fetus in A, the periosteal membrane has been partially removed from the temporal bone to reveal the region of the arcuate eminence (arrow) and medial aspect of the subarcuate fossa (asterisk). This arched region transmits the osseous labyrinth of the anterior semicircular canal. D. Compare the absence of the arcuate eminence and abnormal configuration of the subarcuate fossa in this treated fetus from B.
monkeys during this period of development. Furthermore, the very low incidence of spontaneous abortions before GD 100 (3.5%, 4/113) in the vehicle-treated controls compared with 17.2% (34/198) in our cynomolgus breeding colony in which no conditionEmbryolethality ing is done, verifies that experimental The overall incidence of dose-related em- handling alone has no adverse effects on bryolethality in the current study was pregnancy outcome. 37.5% (6/16). Most losses occurred prior to All-trans retinoic acid has been associated the onset of organogenesis (i.e., GD 16-20). with increased embryolethality in previous This 25% (4/16) incidence of early loss con- nonhuman primate studies. Wilson ('74) retrasts with the 4.2% (5/119) rate of resorp- ported a 44% (7/16) rate of intrauterine tion during this developmental period in un- death in rhesus pregnancies exposed to 10treated cynomolgus monkeys as detected 80 mg/kg on various treatment regimens beby ultrasound a t the CRPRC. Two losses tween GD 20 and 46. By contrast, no signif(12.5%)occurred during organogenesis (GD icant increase in prenatal mortality (3/15, 20-50) in the 10 mgkg dose group. This 20%) was observed following treatment of rate also exceeds the CRPRC 7-year (1983- rhesus monkeys with 20-40 mg/kg all-trans 1990) prenatal mortality rate (2.7%, 3/113) retinoic acid for 4- to 8-day periods during observed in vehicle-treated cynomolgus organogenesis (Hendrickx et al., '80). Treat-
ALL-TRANS RETINOIC ACID TERATOGENICITY IN MACAQUES
71
Fig. 3. A. Profile of a control GD 100 cynomolgus macaque fetus. B. Mandibular hypoplasia in a fetus administered all-trans retinoic acid. C. Cleft palate was also evident in this treated fetus (arrowhead).
ment of pigtail monkeys with 7.5-10 mgkg of all-trans retinoic acid during the entire period of organogenesis resulted in a high level of prenatal loss -63% (12/19).Shorter treatment periods (1-16 days) with higher doses (10-45 mg/kg) were less embryolethal -26% loss, 5/19 (Fantel et al., '77). Teratology Craniofacial malformations consisting of abnormal ears, hypoplastic mandible and temporal bone, as well as cleft palate, characterized the primary effects of all-trans retinoic acid administered a t 10 mgikg prior to and during early organogenesis in the cynomolgus monkey. The threshold dose of all-trans retinoic acid in this study is considered to be 5 mg/kg, based on the increased frequency of skeletal variations and the case of IUGR in the absence of frank malformations observed in this dose group. It is of note that no signs of hypervitaminosis A were associated with the 5 mgikg dose. It is uncertain whether the production of both maternal hypervitaminosis A and fetal malformations at 10 mg/kg is coincidental or implies a cause-and-effect relationship. The high incidence (70%, 7/10) of skeletal variations observed in viable fetuses in the 5- and lO-mg/kg dose groups significantly exceeds (Pc0.02) the 32% (25/77) frequency of rib and vertebral variations in vehicletreated control cynomolgus monkeys in CRPRC developmental toxicity studies. This increased level of variations is consis-
Fig. 3B,C.
tent with the observed teratogenic potential of all-trans retinoic acid in the current study. It has been suggested that certain skeletal variations (e.g., extra 14th ribs) in rodents may be regarded as indicators of teratogenic potency of a drug a t some higher dosage (Kimmel and Wilson, '73). Increased skeletal variations have also been associated with the treatment with retinoic acid, acitretin, etretinate and other retinoids in rodent studies carried out a t Hoffmann-La Roche Laboratories (archival data from Hoffmann-La Roche, '77-730). Thus, retinoids may provide a good example of a class of compounds that produce skeletal variations at subteratogenic doses. A high incidence (71%) of ear anomalies has also been associated with early oral exposure t o 13-cis retinoic acid (2.5 mgkg, GD 10-25 and 2 x 2.5 mg/kg, GD 26 and 27) in the cynomolgus monkey. The most severely affected structures were the superior aspect of the helix, crura of the antihelix, antitra-
72
A.G. HENDRICKX AND H. HUMMLER
gus, and lobule, all derivatives of the second pharyngeal arch (Hummler et al., '90). In addition to these second arch structures, first arch derivatives (i.e., crus, ascending portion of the helix, and tragus), were also adversely affected in the current study. Hypoplasia of the temporal squamae, abnormally shaped zygomatic arches, and irregular ossification in the subarcuate fossa region were also evident in one all-trans fetus. Similar or slightly later susceptibility to mandibular hypoplasia, temporozygomatic hypo- and hyperplasia, and ear dysplasias was observed following l-day thalidomide exposures in bonnet and african green monkeys during early gestation (i.e., GD 24-28) (Newman and Hendrickx, '81, '85; Hendrickx and Sawyer, '78). Thymic hypo- and aplasia and heart anomalies (VSD and transposition of the great vessels) seen in the fetuses treated with 13-cis retinoic acid were not evident in those exposed to all-trans retinoic acid. Only a low incidence (2/52, 4%) of thymic abnormalities (accessory thymus, athymia) was observed in earlier studies in primates with all-trans retinoic acid when treatment encompassed GD 17-45 (Fantel et al., '77; Hendrickx et al., '80; Wilson, '74). The mandible and palate defects observed in the current study were not associated with exposure to 13-cis retinoic acid, as reported earlier by Hummler et al. ('90). These differences in malformation syndromes may be related to the slightly longer treatment regimen (18 vs. 15 days, i.e., GD 10-27 vs. GD 10-24) with 13-cis retinoic acid or inherent differences in potency of the two isomers. The effective teratogenic dose of 13-cis retinoic acid, 2.5 mglkg, is lower than that of all-trans retinoic acid (10 mg/ kg). Pharmacokinetic and placental transfer data in the cynomolgus monkey are required to ascertain whether these dose differences are related to differences in embryonic exposure to each of these compounds. Preliminary data in nonpregnant cynomolgus monkeys indicate differences in kinetic profiles for 13-cis and all-trans retinoic acid. Creech Kraft et al. ('89) reported a more rapid plasma elimination of the alltrans isomer compared with the 13-cis isomer following 10 days of dosing with 10 mg/ kg of each compound. Although the malformation syndromes for all-trans and 13-cis retinoic acid are similar in mice, rats, and hamsters, there are
marked differences in the teratogenic potencies of the two isomers. The lowest teratogenic doses reported for all-trans retinoic acid in these species are 4, 0.4-2, and 12.5 mg/kg/day, respectively, compared with 100, 150, and 25 mg/kg/day, respectively, for 13-cis retinoic acid (Miller et al., '87). These observed differences in potency may be related to differential absorption, distribution, metabolic fate, placental permeability, or to interaction with putative embryonic intracellular receptors (Howard et al., '89). Craniofacial malformations have also been described following administration of all-trans retinoic acid in other nonhuman primate studies. Hendrickx et al. ('80) observed midfacial and mandibular asymmetry and hypoplasia, arched palate, abnormal cranial dimensions, and malformed ears in rhesus monkeys treated with 20-40 mg/kg (p.0.) for 4-8 days between GD 17 and 45. The ear anomalies, which occurred at a 29% (4114) incidence, consisted of cupping of the helix and lobule, bilateral microtia with atresia of the external meatus and low-set ears. Brain and appendicular and axial skeletal defects were also observed in this study. Of four malformed rhesus monkey fetuses exposed for different periods of time to 10-40 mg/kg (p.0.) all-trans retinoic acid during organogenesis, one exhibited an abnormal ear and two had cleft palate (Wilson, '74). A retinoic acid syndrome, characterized by craniofacial and skeletal defects, was described by Fantel and colleagues in 58% (11/19) of pig-tail monkey fetuses exposed to oral doses of 7.5 and 10 mgkg from GD 18-44 (Fantel et al., '77; Newell-Morris et al., '80). The ear anomalies observed in seven fetuses consisted of unfused ear tags and displacement of the external auditory meatus. In addition to malformed ears and eyes, cleft palate, and abnormalities of the mid-face and mandible, axial and appendicular skeletal defects and muscular-joint contractures were also evident. Similar to the current study, visceral defects were uncommon in these nonhuman primate studies with all-trans retinoic acid. The increased variety of malformations in these earlier studies may be attributable to higher doses and/or longer treatment periods during organogenesis. Table 2 summarizes the developmental toxicity associated with all-trans retinoic acid reported in macaques to date.
ALL-TRANS RETINOIC ACID TERATOGENICITY IN MACAQUES
73
TABLE 2 . Primate developmental toxicity studies with all-trans retinoic acid' Current study Hendrickx et al. ('80) Fantel et al. ('77) Wilson ('74) (Macaca mulatta) (Macaca fascicularis) (Macaca mulatta) (Macaca nemestrina) Dose (mglkgiday, p.0.) 10-80 7.5-10 20-40 5-20 Treatment davs 20-46 (variable) 18-44 17-45 (4-8 davs) 10-24 Maternal toxi"city Weight loss X Gastrointestinal X X X Facial Embryo/fetal lethality X X X Malformations Craniofacial X X X X Ears X X X Mandible X X X X Palate X X Skeletal X X Thymus X X2 Heart X Brain X2 'x, present; -, not reported
%=I.
13-cis Retinoic acid is a known human teratogen that produces a characteristic pattern of malformations involving craniofacial, cardiac, thymic, and CNS structures (Lammer et al., '85;Rosa et al., '86). Most of the features of the human syndrome have been duplicated in the cynomolgus monkey with early exposure to all-trans retinoic acid in the current study or 13-cis retinoic acid (Hummler et al., '90). The most common craniofacial defect in human infants involves the external ear (microtia or anotia); other facial anomalies include micrognathia, cleft palate, flattened nasal bridge, and hypertelorism (Lammer et al., '85;Rosa et al., '86). While both isomers induce a high level of ear defects in the cynomolgus monkey, all-trans retinoic acid also produces micrognathia and cleft palate. Most of the cardiac malformations in humans are categorized as conotruncal defects, including transposition of the great vessels, tetralogy of Fallot, double-outlet ventricle, truncus arteriosus, and VSD. Transposition of the great vessels and VSD have been observed in cynomolgus monkey fetuses exposed to 13-cis retinoic acid. Human infants with cardiac defects often exhibit abnormalities of the thymus, including ectopia, hypoplasia, or aplasia. 13-cis Retinoic acid induces a high level (57%) of thymic hypoplasia or aplasia in the cynomolgus monkey. CNS defects associated with prenatal exposure to retinoic acid in humans include hydrocephalus, microcephaly, and cerebellar abnormalities, especially agenesis of the vermis. Further examination of cynomolgus mon-
key fetuses treated with 13-cis retinoic acid (Hummler et al., '90) revealed two cases of vermis hypoplasia (unpublished results). On the basis of limited epidemiologic data, there does not appear to be any evidence of teratogenicity associated with topical all-trans retinoic acid (tretinoin). However, further studies involving blood levels in pregnant women are warranted, since questions remain about long-term exposure to this drug (SCRIP, '89). Little, if any, topically applied all-trans retinoic acid is absorbed systemically (Lucek and Colburn, '85).Based on the extensive animal literature, one could speculate that the differences in the teratogenic potencies of 13-cis and all-trans retinoic acid in humans may be largely related to their different routes of exposure (oral vs. dermal) and pharmacokinetic profiles. The clinical importance of vitamin A and its derivatives in anticancer and dermatologic therapy is well recognized. Currently, additional synthetic retinoids are being developed in an effort to enhance efficacy and diminish drug-related toxicity, particularly in women of child-bearing age (Kerker and Farmer, '90; Boyd, '89). The present investigation verifies the similarity of the malformation syndromes induced by both alltrans and 13-cis retinoic acid in the cynomolgus monkey and 13-cis retinoic acid embryopathy in humans. The reproductive, endocrinologic, developmental, and metabolic similarities between the nonhuman primate and humans further support their use as a model for developmental toxicity
74
A.G. HENDRICKX AND H. HUMMLER
studies of vitamin A-related compounds (Hendrickx and Binkerd, '90; Hummler et al., '90). ACKNOWLEDGMENTS
We wish to acknowledge the excellent assistance of Pam Peterson, Jeff Rowland, and Dr. Alice Tarantal in conducting the study and in preparing the manuscript. LITERATURE CITED Archival Data from Hoffmann-La Roche (1977 and 1980). Biesalski. H.K. (1989) Comuarative assessment of the toxicolo'gy of vitamin A anh retinoids in man. Toxicology, 57t117-161. Boyd, A.S. (1989) An overview of the retinoids. Am. J . Med., 86.568474. Brockes, J.P. (1989) Retinoids, homeobox genes, and limb morphogenesis. Neuron, 2:1285-1294. Creech Kraft, J., D.M. Kochhar, W.J. Scott, and H. Nau (1987) Low teratogenicity of 13-cis retinoic acid (isotretinoin) in the mouse corresponds to low embryo concentrations during organogenesis: Comparison to the all-trans isomer. Toxicol. Appl. Pharmacol., 87: 474-482. Creech Kraft, J.,L. Roberts, J.R. Baily, H. Nau, and W. Slikker, Jr. (1989) Pharmacokinetics of 13-cis- and all-trans-retinoicacid in the cynomolgus monkey. Teratology, 39:447A. Daniel, W. W. (1983) Biostatistics: A Foundation for Analysis in the Health Sciences, 3rd Ed. John Wiley & Sons, Inc., New York. Dolle, P., E. Ruberte, P. Leroy, G. Morris-Kay, and P. Chambon (1990) Retinoic acid receptors and cellular retinoid binding proteins. I. A systematic study of their differentia1 pattern of transcription during mouse organogenesis. Development, 110:1133-1151. Eichele, G. (1989) Retinoids and vertebrate limb pattern formation. Trends Genet., 5:246-251. Ellis, C.N., and J.J. Voorhees (1987) Continuing medical education (Therapy): Etretinate therapy. J. Am. Acad. Dermatol., 16:267-291. Fantel, A.G., T.H. Shepard, L.L. Newell-Morris, and B.C. Moffett (1977) Teratogenic effects of retinoic acid in pigtail monkeys (Macaca nemestrinu). I. General features. Teratology, 15:65-72. Hendrickx, A.G., and P.E. Binkerd (1990) Nonhuman primates and teratological research. J . Med. Primatol., 19:81-108. Hendrickx, A.G., M. Cukierski, S. Prahalada, G. Janos, and J . Rowland (1985) Evaluation of bendectin embryotoxicity in nonhuman primates. I. Ventricular septa1 defects in prenatal macaques and baboon. Teratology, 32r179-189. Hendrickx, A.G., and R.H. Sawyer (1978) Developmental staging and thalidomide teratogenicity in the green monkey (Cercopithecus aethiopsi. Teratology, 18:393-404. Hendrickx, A.G., S. Silverman, M. Pellegrini, and A.J. Steffek (1980) Teratological and radiocephalometric analysis of craniofacial malformations induced with retinoic acid in rhesus monkeys (Mucaca mulatta). Teratology, 22:13-22.
Howard, W.B., and C.C. Willhite (1986) Toxicity of retinoids in humans and animals. J . Toxicol. Toxin Rev., 5.55-94. Howard. W.B.. C.C. Willhite, S.T. Omaye, and R.P. S h a k a (1989) Comparative distribution, pharmacokinetics and placental permeabilities of all-transretinoic acid, 13-cis-retinoic acid, all-trans-4-0x0retinoic acid, retinyl acetate and 9-cis-retinal in hamsters. Arch. Toxicol., 63:112-120. Hummler, H., R. Korte, and A.G. Hendrickx (1990) Induction of malformations in the cynomolgus monkey with 13-cis-retinoic acid. Teratology, 42:263-272. Kerker, B.J., and E.R. Farmer (1990) Update on retinoids. Saudi Med. J., 11:165-170. Kimmel, C.A., and J.G. Wilson (1973) Skeletal deviations in rats: Malformations or variations? Teratology, 8:309-316. Lammer, E.J., D.T. Chen, R.M. Hoar, N.D. Agnish, P.J. Benke, J.T. Braun, C.J. Curry, P.M. Fernhoff, A.W. Grix, I.T. Lott, J.M. Richard, and S.C. Sun (1985) Retinoic acid embryopathy. N. Engl. J . Med., 313: 837-841. Lucek, R.W., and W.A. Colburn (1985) Clinical pharmacokinetics of the retinoids. Clin. Pharmacokinet., 10: 38-62. Miller, R.K., K. Brown, J . Cordero, D. Dayton, B. Hardin, M. Greene, C. Grabowski, A. Hendrickx, E. Hook, R. Jensh, G. Kimmel, J. Manson, K.O. Shea, J. Schardein, and R. Staples (1987) Teratology Society Position Paoer: Recommendations for Vitamin A Use During Pregnancy. Teratology, 35:269-275. Nau, H., D.M. Kochhar, J.M. Creech Kraft, B. Loefberg, J . Reiners. and T. Sparenberg (1987) Teratogenesis of retinoids: Aspects o i species aifferences andtransplacental pharmacokinetics. In: Approaches to Elucidate Mechanisms in Teratogenesis. F. Welsch, ed. Hemisphere Publishing Corporation, Washington, D.C., pp. 1-15. Newell-Morris, L., J.E. Sirianni, T.H. Shepard, A.G. Fantel, and B.C. Moffett (1980) Teratogenic effects of retinoic acid in pigtail monkeys (Macaca nemestrina). XI. Craniofacial features. Teratology, 22:87-101. Newman, L.M., and A.G. Hendrickx (1981) Fetal ear malformations induced by maternal ingestion of thalidomide in the bonnet monkey (Mucaca radiata). Teratology 23:351-364. Newman, L.M., and A.G. Hendrickx (1985) Temporomandibular malformations in the bonnet monkey (Macacu radiata) fetus following maternal ingestion of thalidomide. J . Craniofac. Genet. Dev. Biol., 5~147157. Rosa, F.W., A.L. Wilk, and F.O. Kelsey (1986) Teratogen update: Vitamin A congeners. Teratology, 33: 355-364. SCRIP, No. 1374/5:17 (1989) FDA and teratogenic potential of Retin-A. Tarantal, A.F., and A.G. Hendrickx (1988a) Use of ultrasound for 'early pregnancy detection in the rhesus and cynomolgus macaque (Macacu mulatta and Macaca fascicularis). J . Med. Primatol., I7:105-112. Tarantal, A.F., and A.G. Hendrickx (1988b) Characterization of prenatal growth and development in the crab-eating macaque (Macaca fasciculuris) by ultrasound. Anat. Rec., 222:177-184. Wilson, J.G. (1974)Teratologic causation in man and its evaluation in non-human primates. In: Birth Defects. A.G. Motulsky, ed. American Elsevier Publishing Co., New York, p i . 191-203. ~~
