Antisense Oligonucleotide Down-Regulation of E-Cadherin in the Yolk ...

BIOLOGY OF REPRODUCTION 53, 1229-1238 (1995)

Antisense Oligonucleotide Down-Regulation of E-Cadherin in the Yolk Sac and Cranial Neural Tube Malformations' Beiyun Chen 3 and Barbara F. Hales 2 Department of Pharmacology& Therapeutics, McGill University, Montreal, Qu6bec, Canada ABSTRACT The cadherins are a family of calcium-dependent cell adhesion molecules that are regulated both spatially and temporally during development. Epithelial cadherin (E-cadherin) is present in epithelial cells in both the embryo and yolk sac during organogenesis. The consequences of disrupting the expression of E-cadherin at this stage of development are poorly understood. We report here our studies on the effects of antisense oligonucleotides on E-cadherin in the rat whole embryo culture system. Four 18-base single strand phosphorothioate oligodeoxynucleotides (AS-oligos), complementary to various regions of the mouse E-cadherin cDNA sequence, were dissolved insaline and injected into the amniotic cavities of 5-7 somite rat embryos; asense (S-oligo) to oligo-1, an 18-base random sequence oligo (C-oligo), and PBS were used as controls. Embryos were cultured for up to 45 h; embryo morphology and the relative concentrations of E-cadherin protein were examined. All six oligonucleotides (AS-oligos and control oligos) induced malformations when amounts ranging from 25 to 50 pmol of oligonucleotide were injected per embryo. The malformations induced by all the oligos included craniofacial hypoplasia, an enlarged pericardium, twisted spinal cord, swelling of the rhombencephalon, and underdeveloped forelimb. Injection of AS-oligo-1, asequence starting at the tenth base downstream from the translation initiation codon (ATG), resulted in malformed embryos with a high incidence of cranial neural tube malformations. The effects of AS-oligo-1 on the relative abundance of E-and neural (N)cadherin proteins were examined by Western blot analysis. In the AS-oligo-1-exposed malformed embryos, the relative abundance of Eand N-cadherin proteins was not altered up to 24 h after injection; E-and N-cadherin concentrations inthe embryo were decreased at 45 h postinjection. In contrast, the relative abundance of the E-cadherin protein in the yolk sac was reduced at 1-2 h after injection of AS oligo-1 and returned to control levels by 4 h. S-oligo-1 did not induce any change in the relative abundance of E-or N-cadherins. Thus, there was a tissue-specific and temporary "knockdown" of E-cadherin expression in the yolk sac of embryos exposed to antisense (ASoligo-1); the down-regulation of yolk sac E-cadherin appears to lead to the induction of neural tube defects in the embryo. The exposure of whole embryos in culture to antisense oligonucleotides provides a model system in which the roles of developmentally important molecules and their spatial and temporal contributions to embryogenesis can be elucidated.

INTRODUCTION The cadherins are a family of calcium-dependent cell adhesion molecules mediating cell-cell interactions [1, 2]. Two members of this family are epithelial cadherin (E-cadherin) and neural cadherin (N-cadherin). During development, the expression of each cadherin is spatio-temporally regulated and associated with a variety of morphogenetic events such as neural tube formation [1, 31. In the process of neural tube formation, the expression of E- and N-cadherins changes dynamically. The undifferentiated ectoderm expresses only E-cadherin. When the neural plate invaginates, it gradually loses E-cadherin and acquires N-cadherin during separation from the overlying ectoderm; the ectoderm continues to express E-cadherin [4]. E-cadherin plays a pivotal role during early embryonic development. Embryos that are homozygously negative for E-cadherin, or have a targeted mutation in the E-cadherin gene, show severe abnormalities before implantation [5, 6]; Accepted July 6, 1995. Received March 10, 1995. 'This work was supported by a grant from the Medical Research Council of Canada. B. Chen is the recipient of a fellowship from the Medical Research Council of Canada. 2 Correspondence: Dr. Barbara F. Hales, Department of Pharmacology and Therapeutics, McGill University, 3655 Drummond Street, Montreal, Quebec, Canada H3G Y6. FAX: (514) 398-7120; e-mail: [email protected] 3Current address: Department of Pathology, New York Medical College, Valhalla, NY 10595.

these embryos decompact after forming a morula and do not form a trophectodermal epithelium or blastocyst cavity. A number of studies have demonstrated that normal cadherin expression and function are also important during organogenesis. Shimamura and Takeichi [7] demonstrated that an anti-E-cadherin antibody altered the overall morphology of the developing mouse brain in vitro. Other studies [8, 9] have shown that antibodies to cadherins can disturb the morphology of cultured embryonic tissues. A recent study by Levine et al. [10] showed that the expression of a mutant E-cadherin that lacked a cytoplasmic tail caused lesions in the ectoderm during gastrulation in the early Xenopus embryo. The local overexpression or ectopic expression of Ncadherin in Xenopus embryos has been shown to affect tissue morphology [11, 12]. These data demonstrate that the amount and the temporal-spatial distribution of cadherins in tissues are crucial for normal embryonic development. Antisense oligonucleotides are useful tools in achieving specific inhibition of targeted gene expression [13]. Oligonucleotides may inhibit gene expression through several mechanisms; these include the prevention of new protein synthesis by translational arrest, the promotion of mRNA degradation by an RNase H-dependent mechanism, the inhibition of mRNA maturation by the masking of sequences required for formation of the spliceosome, the inhibition of mRNA transportation out of the nucleus, and the inhibition of gene tran1229

1230

CHEN AND HALES

69 537

5'

2166

1

2

3'

MI

EC

P

2267

3

4

oligo-1: 5'-GGAAAAGCTGCGGCACCG-3' (bp 78-95) oligo-2: 5'-ATGACCCAGTCTCGMC-3' (bp 530-547) oligo-3: 5'-CTTGACCCTGATACGTGCTC-3' (bp 1205-1224) oligo-4: 5'-ATIT1GTATTCGCCAATC-3' (bp 2051-2068) FIG. 1. Schematic diagram of the mouse E-cadherin cDNA. The box represents the open reading frame. P:prepeptide; EC: extracellular domain of the mature E-cadherin protein; TM: transmembrane domain of E-cadherin; IC: intracellular domain of E-cadherin. Numbers 1,2,3, and 4 represent the sequences targeted by AS oligos-1, -2,-3,and -4,respectively. The sequences shown below are the antisense sequences complementary to the selected target sites.

scription by formation of a triple helix structure [14, 151. The mechanism involved may depend on the cell or tissue type targeted, the particular mRNA targeted, the target site on the mRNA, and the chemical nature of the oligonucleotide. A previous study from our laboratory showed that during rat organogenesis, E-cadherin is expressed in both embryo and yolk sac, with particularly high levels found in the yolk sac of cultured embryos [16]. To determine the consequences of specifically disrupting E-cadherin at a particular stage of development, antisense oligonucleotides to the Ecadherin cDNA sequence were injected into the amniotic sacs of rat embryos, which were then cultured for up to 45 h. This antisense approach in the whole embryo culture system has been very useful in investigating the role of certain molecules during organogenesis and the mechanisms of action of teratogens [17, 18]. In the present paper, we describe the effects of antisense phosphorothioate oligonucleotides targeted to different sequences of the E-cadherin mRNA on rat embryonic development during organogenesis in vitro.

stant, PQ). Rats were housed in the McIntyre Animal Centre (McGill University, Montreal, PQ, Canada) and given Purina rat chow and water ad libitum. The day on which spermatozoa were found in the vaginal smear was defined as Day 0 of pregnancy. The embryo culture procedure used in this study was based on the system described by New [19]. The uteri of etherized pregnant rats were removed on the morning of Day 10 of gestation; the embryos were dissected free of maternal tissue and of Reichert's membrane, leaving the ectoplacental cone and yolk sac intact. The dissection was done in Hanks' Balanced Salt Solution (Gibco Laboratories, Burlington, ON, Canada) under aseptic conditions. Single-strand phosphorothioate oligodeoxynucleotides were purchased from Oligo Etc Inc. (Wilsonville, OR). A series of four 18-base single-strand phosphorothioate oligodeoxynucleotides (AS-oligos), complementary to four different regions of the mouse E-cadherin cDNA sequence [201 as illustrated in Figure 1, were dissolved in PBS and injected into the amniotic cavities of 5-7 somite rat embryos. The volume of injection was kept constant at 100 nl/embryo. The sense (S-oligo-1) sequence to AS-oligo-1 (Fig. 1), an 18base random sequence oligo (5'-ATGACCCCGGCTCGTGTC-3'; C-oligo), and PBS were used as controls. The sequences of all chosen oligonucleotides were compared to the rodent sequence databases of GenBank and EMBL bank to ensure a low (below 70%) homology between the chosen sequences and other rodent sequences. Three embryos were placed in each sterile 60-ml culture bottle containing 4.8 ml of medium, which was composed of 90% heat-inactivated filtered rat serum and 10% Tyrode's saline (Gibco Laboratories). The contents of the bottles were gassed with a mixture of 20% 02:5% CO2:75% N2. The bottles were placed on a rotator (New Brunswick Scientific Company, Edison, NJ) at 30 rpm, and the embryos were cultured for up to 45 h at 37°C. After the first 24 h, the embryos were regassed with 95% 02:5% CO 2. Embryo growth and morphology were evaluated after 24 or 45 h of culture. Western Blot Analysis

MATERIALS AND METHODS

Embryo Culture and Microinjection Timed-gestation pregnant Sprague-Dawley rats (200-225 g) were purchased from Charles River Canada, Inc. (St. Con-

After culture for 0.5, 1, 2, 4, 8, 24, or 45 h, embryos and yolk sacs were collected separately and were directly dissolved in sample loading buffer (62.5 mM Tris-HCl [pH 6.81, 2% SDS, 10% glycerol, 5% -mercaptoethanol) and boiled for 10-20 min. Samples (20 gg protein/lane) were fractionated

TABLE 1. Malformations induced by the injection of rat embryos with E-cadherin antisense oligonucleotides. Amount (pmol)

C-oligo

25 30 40 50

17 (12) 45 (11) 50 (10) 100 (7)

Percentage of malformed/cultured embryosb (number per group) AS-oligo-2 AS-oligo-3 AS-oligo-1 56 (9) 72 (50) 92 (12) 100 (7)

33 (9) 17 (12) 62 (13) 100 (6)

aAmount of oligonucleotide injected into the amniotic cavity of each 5-7 somite rat embryo. bEmbryo malformations were evaluated 45 h after injection of each oligonucleotide.

25 (8) 64 (11) 27 (11) 100 (6)

AS-oligo-4 20 (10) 27 (11) 22 (9) 67 (6)

E-CADHERIN ANTISENSE AND NEURAL TUBE MALFORMATIONS

1231

FIG. 2. Embryos treated with PBS (A), AS-oligo-1 (B), or AS-oligos -2, -3, or -4 (C-F, respectively) 45 h after exposure. The types of malformations observed included an open anterior neuropore and hypoplasia of the brain (B), an enlarged pericardium (C and E), a swollen hindbrain (D), an underdeveloped forelimb (E), and an indented or abnormally curved spinal cord (F). The malformations induced bythe different oligos were similar, except that AS-oligo-1 induced a high incidence of open anterior neuropore (B). The bar represents 0.5 mm.

by SDS-PAGE in 7.5% acrylamide gels [21]. The fractionated proteins were transferred to a nitrocellulose membrane [22]. The membrane was cut in half; the top half was used for immunoblotting of E- or N-cadherin, and the bottom half was used for immunoblotting of actin as described below. The blot was blocked with 5% milk in TBS-T (137 mM NaCl, 20 mM Tris [pH 7.4], 0.1% Tween 20) at room temperature for 2 h. Rabbit anti-mouse uvomorulin (E-cadherin) and anti-chicken N-cadherin sera were kindly provided by Dr. R. Kemler (Max Planck Institut fir Immunobiologie, Freiburg, Germany) [23] and Dr. G.B. Grunwald (Thomas Jefferson University, Philadelphia, PA) [24], respectively. The mouse monoclonal antibody against actin (Amersham Canada Ltd., Oakville, ON) was a gift from Dr. G. Almazan (McGill University). The blots were incubated with anti-E-cadherin (1:100), anti-N-cadherin

(1:500), or anti-actin (1:1000) in 5% milk/TBS overnight at 40 C. Each blot was then washed three times in TBS-T and incubated with biotinylated secondary antibody at room temperature for 1 h. The antibody binding was visualized with streptavidin alkaline phosphatase using 5-bromo-4-chloro3-indolyl phosphate and nitro blue tetrazolium as substrates (Amersham Canada Ltd.). Immunoblotting was done at least three times with independent batches of samples. Quantification of Datafrom Western Blots Black and white prints of the immunoblots were scanned with an LKB laser densitometer (Pharmacia, Montreal, PQ, Canada). Values are expressed as a percentage of the control values obtained from the PBS-treated samples at each time point and represent means SD. The data were ex-

1232

CHEN AND HALES A

A

100

g

80

*

u,

*

0

.o

E 40 0

a

20

0

50 3 10 30 C Amount of oligo injected per embryo (pmoles)

B o100 FIG. 3. Percentage of embryos with open anterior neuropore 45 h after injection of antisense oligos-1, -2,-3, or -4.Amount 0 represents PBS injection. The number of embryos per group is given in parentheses inTable 1. There were no embryo deaths.

80 '2,

a 60

50 pmoles 30 pmoles

0

amined by one-way ANOVA using the Complete Statistical System software program (Statsoft Inc., Tulsa, OK). The level of significance was p - 0.05.

t) ,0

I a

RESULTS Effects of E-Cadherin Antisense Oligonucleotides on Embryonic Development C-oligo, the random sequence phosphorothioate control oligo, caused embryo death when injected in amounts greater than 63 pmol/embryo. The C-oligo at less than 10 pmol/embryo did not increase the percentage of malformed embryos above that seen with saline (PBS) injections (10%). However, in the range of 25-50 pmol/embryo, C-oligo did increase the incidence of embryo malformations (Table 1). In addition, all four AS-oligos were capable of inducing embryo malformations when doses from 25 to 50 pmol were injected per embryo. AS-oligo-1 induced a higher percentage of embryo malformations than did AS oligos-2, -3, and -4 or C-oligo (Table 1). Pictures of embryos treated with the AS-oligos are shown in Figure 2. Embryo A is a PBS-treated control embryo. Embryo B was treated with AS-oligo-1 and had an open neural tube defect as a result of the failure of neural tube closure 2 [25, 26]; this resulted in a flaring of the neural folds over the prosencephalic-mesencephalic border region. Treatment with AS-oligo-1 at 30-50 pmol/embryo resulted in this type of open neural tube defect in approximately 50% of the embryos after 45 h of culture (Fig. 3); C-oligo and the other antisense oligos in a similar dosage range produced

40

0

20 1~ ~"'~"""""~ ~

10 pmoles

0

24

45

Time after injection (hour) FIG. 4. A) Rates of overall embryo malformations 45 h after the injection of Srepresents PBS injection. B)Percentage of embryos oligo-1 or AS-oligo-1. Amount 0O with an open anterior neuropore 24 or 45 h after the injection of AS-oligo-1. Solid circles: 10 pmol injected per embryo; solid squares: 30 pmol injected per embryo; solid triangles: 50 pmol injected per embryo. The percentage of embryos with an open anterior neuropore among the untreated and sense-treated embryos was always less than 10%, at either 24 or 45 h.

this particular malformation in less than 10% of the exposed embryos. Other malformations were also observed in Coligo- and AS-oligo-treated embryos. Embryo C (Fig. 2C) had an enlarged pericardium. Embryo D (Fig. 2D) had hypoplastic facial prominences as compared to control embryos and a pronounced edema over the rhombencephaIon, with failure of the closure 4 epithelium to fuse. The tail of this embryo failed to assume a proper rotation beyond the head of the embryo. Embryo E (Fig. 2E) had an enlarged pericardium and an underdeveloped forelimb. Embryo F also had abnormal craniofacial development, with a gross underdevelopment of the craniofacial structures and an abnormal curvature of the posterior neural tube (closure 1). The percentages of embryos with open anterior neuropore 45 h after injection of antisense oligos-1, -2, -3, or -4


1233

FIG. 5. Embryos after the injection of 30 pmol/embryo of S-oligo-1 orAS-oligo-1: 24 h after the injection of S-oligo-1 (A)orAS-oligo-1 (Band C)and 45 h after the injection of S-oligo-1 (D)or AS-oligo-1 (Eand F). Arrows indicate an open anterior neuropore. Arrowhead indicates hypoplasia of the forebrain. The bar represents 0.5 mm.

are presented in Figure 3. Only the injection of oligo-1 resulted in a consistently elevated incidence of open anterior neuropore defects; AS-oligo-l-treated embryos (e.g., Fig. 2B) were selected for further study. Effects of AS-Oligo-1 on Neural Tube Closure The sense sequence of AS-oligo-1, S-oligo-1, was used as a control for further investigations of the induction of neural tube defects by AS-oligo-1. Figure 4A shows the overall rates of embryo malformations observed after 45 h of culture when different amounts of AS-oligo-1 or S-oligo-1 were injected into 5-7 somite rat embryos. The injection of PBS or 3 pmol of either the antisense or sense oligonucleotide caused approximately 10% of the embryos to be malformed. When 10 to 50 pmol of these oligos were injected, AS-oligo-1 caused a significantly higher percentage of embryo malformations than

S-oligo-1. The injection of 50 pmol of S-oligo-1 resulted in malformed embryos with various phenotypes, such as swelling of the hindbrain, blebs in the heart region, an underdeveloped limb bud, and a shortened tail; these malformations resembled those caused by AS-oligos 2, 3, 4, and C-oligo (Fig. 2). In contrast, the injection of 30 pmol of AS-oligo-1 resulted in open cranial neural tube malformations in 50% of the embryos after 45 h in culture (Fig. 4B). Interestingly, the incidence of open neural tube defects in AS-oligo-l-treated embryos was lower after 45 h of culture than after 24 h of culture (Fig. 4B). The incidences of the other malformations (e.g., swelling of the hindbrain, blebs in the heart region, underdeveloped limb buds) observed were not affected or were increased by the length of time in culture (data not shown). It is possible that the hindbrain swelling, pericardial blebs, limb malformations, and tail shortening produced by all the

1234

CHEN AND HALES

FIG. 6. A)Western blot analysis of E-cadherin protein in the embryos after injection of PBS (lanes 1, 4, 7, 10, 13, 16, and 19), 30 pmol of S-oligo-1 (lanes 2, 5, 8, 11, 14, 17, and 20), or 30 pmol of AS-oligo-1 (lanes 3, 6, 9, 12, 15, 18, and 21). Samples were collected at the indicated time points after the injections, and 20 g of protein was loaded on each lane. The 120 x 103 M, protein band is E-cadherin. The other protein bands in the Western blot were nonspecific; these bands were also found when equivalent concentrations of control rabbit serum were substituted for the primary antibody. The bottom half of the same blot was used for immunoblot analysis of actin to serve as an internal control for protein loading. B)Densitometric analysis of the 120 x 103 M, E-cadherin protein band. Values obtained from PBS-treated samples were designated as 100%. Data are expressed as means ± SD, n = 4. *Significantly different from PBS control at the same time point (p < 0.05).

oligonucleotides represent generalized defects induced by many phosphororthioate oligonucleotides, rather than effects on the cadherins per se. Embryos treated with 30 pmol of S-oligo-1 or AS-oligo-1 are shown in Figure 5. Embryos A, B, and C were cultured for 24 h, while embryos D, E, and F were cultured for 45 h. Embryos treated with S-oligo-1 (embryos A and D) developed normally. AS-oligo-1-treated embryos showed open

neural tube malformations and abnormal craniofacial development. Embryo B (Fig. 5B) failed to achieve closure 2 over the rostral portion of the prosencephalon and had a significant underdevelopment of the facial prominence. Embryo C (Fig. 5C) failed to achieve closures 3. 2, or 4 and had only rudimentary craniofacial development. Embryo E (Fig. 5E) had abnormal closure 2. Embryo F (Fig. 5F), with grossly underdeveloped anterior neural development, was


1235

FIG. 7. A) Western blot analysis of E-cadherin protein inthe yolk sacs after injection of PBS (lanes 1, 4, 7, 10, 13, 16, and 19), 30 pmol of S-oligo-1 (lanes 2, 5, 8, 11, 14, 17, and 20), or 30 pmol of AS-oligo-1 (lanes 3, 6, 9, 12, 15, 18, and 21). Samples were collected at the indicated time points after the injections, and 20 Ig of protein was loaded on each lane. The bottom half of the same blot was used for immunoblot analysis of actin to serve as an internal control for protein loading. B) Densitometric analysis of 3 the 120 x 10 M, E-cadherin protein band. Values obtained from PBS-treated samples were designated as 100%. Data are expressed as means ± SD, n = 4. *Significantly different from PBS control at the same time point (p < 0.05).

similar to embryo B in Figure 2B. This embryo exhibited failure of neural tube closure 2 and no epithelial covering suggestive of closure 4; the neural folds were widely separated over the rhombencephalon. Thus, the injection of AS-oligo-1 into the amniotic cavity of 5-7 somite rat embryos resulted in a delay in neural tube closure, resulting in a spectrum of neural tube defects. The next question was whether the injection of AS-oligo-1 into the amniotic cavity

of 5-7 somite rat embryos also altered the regulation of Ecadherin protein in these embryos. Relative Concentrationsof E-Cadherin Proteinin the Embryo after AS-Oligo-1 Exposure The steady state concentrations of E-cadherin protein in the embryos treated with the S-oligo-1 oligonucleotide were

1236

CHEN AND HALES

than a primary effect of the injected E-cadherin antisense oligonucleotide. Relative Concentrationsof E-CadherinProtein in the Yolk Sac afterAS-Oligo-1 Exposure The steady state concentrations of E-cadherin protein in the yolk sac throughout the 45-h culture period after injection of S-oligo-1 were similar to those after the injection of PBS (Fig. 7, A and B). However, the injection of 30 pmol of AS-oligo-1 dramatically down-regulated E-cadherin protein in the yolk sac (Fig. 7, A and B). Injection of AS-oligo-1 resulted in a decrease in the steady state concentration of E-cadherin protein to 25% of the control value after 1 h; this decrease was still present but was attenuated after 2 h (37% of control) and disappeared by 4 h postinjection. Despite the down-regulation of E-cadherin protein in the yolk sac, the concentrations of actin remained relatively unchanged in this tissue after AS-oligo-1 injection (Fig. 7A). Thus, Ecadherin protein concentrations in the yolk sac are downregulated after exposure in culture to E-cadherin AS-oligo-1.

FIG. 8. A) Immunoblot analysis of the N-cadherin protein in embryos after injection of PBS (lanes 1 and 4), 30 pmol of S-oligo-1 (lanes 2 and 5), or 30 pmol of ASoligo-1 (lanes 3 and 6). Samples were collected at the indicated time points after the injections, and 20 g of protein was loaded on each lane. The bottom half of the same blot was used for immunoblot analysis of actin to serve as an internal control for protein loading. B)Densitometric analysis of the 130 x 103 Mr N-cadherin protein band. Values obtained from PBS-treated samples were designated as 100%. Data are expressed as means SD, n = 3. *Significantly different from PBS control at the same time point (p < 0.05).

similar to those in their PBS-treated counterparts (Fig. 6, A and B). In the embryos that had been given an injection of ASoligo-1, there were no differences in the concentrations of Ecadherin protein up to 24 h postinjection as compared to those in the PBS- or sense-treated embryos (Fig. 6, A and B). However, at 45 h postinjection, the concentration of Ecadherin protein was decreased to 22% of the control level in embryos with an open neural tube defect. In the antisense-treated embryos in which the cranial neural tube was closed, the steady state concentrations of E-cadherin were not decreased (data not shown). Actin concentrations remained relatively unchanged in these embryos at all times assessed (Fig. 6A). Thus, E-cadherin protein concentrations were decreased in the malformed embryos with an open neural tube defect, but not in the AS-oligo-l-treated embryos with a closed cranial neural tube. Moreover, this decrease was a relatively late effect (45 h). On the basis of these observations, we hypothesize that the decrease in embryo E-cadherin is likely to be a secondary result of the embryo malformation rather

Relative Concentrationsof N-Cadherin in the Embryo after AS-Oligo-1 Injection The relative concentrations of N-cadherin were studied to determine whether another cadherin was affected by the E-cadherin antisense oligonucleotides. Forty-five hours after the injection of 30 pmol of AS-oligo-1, a dramatic decrease in the relative abundance of N-cadherin protein was detected in the embryos with open neural tube malformations (Fig. 8A). The relative abundance of N-cadherin protein in these embryos was decreased to 34% of the control N-cadherin level (Fig. 8B). N-cadherin protein was not detected in yolk sac samples using the available antibody. The decrease in N-cadherin protein in the embryo coincides with the decrease observed in embryonic E-cadherin rather than in yolk sac E-cadherin; this decrease may also be a consequence rather than a cause of the embryo malformations. DISCUSSION In the present study, the injection of E-cadherin antisense oligonucleotides topically in the vicinity of the embryonic neural folds in rat embryos led to malformations of the cranial neural tube and facial prominences as well as of rotation and heart and limb development. Interestingly, the injection of different E-cadherin antisense oligonucleotides appeared to result in the production of embryos with different craniofacial abnormalities. The most dramatic malformation, a flaring of the neural folds over the prosencephalic-mesencephalic border regions, was seen primarily with one antisense oligonucleotide, AS-oligo-1; these anterior neural tube malformations resulted from a failure of closure 2 [25]. Injection of the E-cadherin antisense oligonucleotides had less effect on the


closure of the neural tube in the caudal region. The mechanism of closure of the posterior portion of the neural tube is likely to be different from that of the anterior portion. In addition, there may be a specificity of the effect of the Ecadherin antisense oligonucleotides for neural fold closure at one level of the body axis. Further studies were done to begin to investigate the mechanism by which one antisense oligonucleotide, ASoligo-1, interferes with closure of the neural folds in the prosencephalic-mesencephalic regions. It is interesting that there appeared to be a decrease in the percentage of AS-oligo-1exposed embryos with open neural tubes after 45 h of culture as compared to 24 h of culture. Such a decrease may be attributable to a repair or recovery in neural tube closure. In the AS-oligo-l-exposed malformed embryos with an open cranial neural tube, the relative concentrations of Eand N-cadherin proteins were decreased 45 h after antisense injection. This is likely to be a secondary result of the embryo malformations rather than a direct effect of the injected Ecadherin antisense oligonucleotide on the synthesis of these cadherins. Augustine et al. [18] showed that a 20-base phosphorothioate oligonucleotide injected into the amniotic cavity of 5-6 somite mouse embryos was detected in the embryonic tissue 30 min after injection and that it reached its maximal concentration in the embryo by 3 h; much lower amounts of intact oligonucleotide were detected in the embryos after 24 h. The oligonucleotide was distributed throughout the embryo, in virtually all cells, although some concentration was observed in the neural folds of the spinal cord. Thus, the AS-oligo-1 E-cadherin oligonucleotide used in the present study would be expected to be concentrated and to exert its maximal effect in the embryonic tissues within a few hours after the injection. Although we do not have information on the regulation of the message, newly synthesized E-cadherin (135-kDa polypeptide) has a t-1/2 of -45 min in cultured Madin-Darby canine kidney epithelial cells; this precursor protein is rapidly and efficiently processed to the 120-kDa mature polypeptide [27]. After a 15-min pulse label with [3 5 S]methionine, newly synthesized E-cadherin was maximally detected at the cell surface following a chase period of 45-60 min; approximately 50% reached the cell surface within 30 min. E-cadherin present at the cell surface also had quite a rapid turnover (t-1/2 -5 h). The lack of a change in the steady state concentration of E-cadherin protein in the embryo within the first few hours after antisense oligo-1 injection implies either that this oligonucleotide is unable to down-regulate E-cadherin in the embryo or that any change in E-cadherin regulation in the embryo is either relatively small or localized in area. Attempts to examine this question further by means of whole-mount immunohistochemistry were not successful because of the high background staining obtained with use of our anti-E-cadherin antibody. In contrast, we did detect a dramatic decrease in the abun-

1237

dance of E-cadherin protein in the yolk sac 1-2 h after the injection of AS-oligo-1. The yolk sac is composed of only two cell layers and expresses high levels of E-cadherin during this stage of development [161. Oligonucleotides injected into the amniotic cavity could easily reach the yolk sac through the amnion and inhibit E-cadherin synthesis in the yolk sac. The finding that an antisense E-cadherin oligonucleotide is able to down-regulate E-cadherin in the yolk sac, and not in the embryo, suggests that the expression of this gene is subject to tissue-specific regulation. Antisense down-regulation may involve different mechanisms in these two tissues. The steady state concentration of E-cadherin protein in the yolk sac returned to control levels 4 h after the antisense injection. A return to normal yolk sac E-cadherin concentrations may contribute to the partial recovery from cranial neural tube defects that was observed in some of the embryos. The decreased concentration of yolk sac E-cadherin 1-2 h after AS-oligo-1 injection may alter yolk sac function in such a manner as to have a profound effect on neural tube closure in the embryo. Yolk sac function is critical for normal development at this stage of organogenesis, but for the most part, dysfunction has been associated with growth retardation or reduction malformations [28]. That the down-regulation of a cell adhesion molecule, E-cadherin, in the yolk sac may lead specifically to cranial neural tube defects is a novel observation. Further studies are needed to elucidate the mechanisms involved. This study demonstrated, for the first time, that temporary "knockdown" of the expression of E-cadherin during a critical period of development led to specific neural tube defects. The exposure of whole embryos in culture to antisense oligonucleotides provides a model system in which the roles of developmentally important molecules and their spatial and temporal contributions to embryogenesis can be studied. ACKNOWLEDGMENTS We thank Dr. R. Kemler for the E-cadherin antibody, Dr. G.B. Grunwald for the Ncadherin antibody, and Dr. R. Finnell for his assistance in the assessment of the cranial neural tube malformations.

REFERENCES 1. Takeichi M. The cadherins: cell-cell adhesion molecules controlling animal morphogenesis. Development 1988; 102:639-655. 2. Anderson H. Adhesion molecules and animal development. Experientia 1990; 46:2-13. 3. Takeichi M.Cadherins: a molecular family important in selective cell-cell adhesion. Annu Rev Biochem 1990; 59:237-252. 4. Takeichi M, Inuzuka H, Shimamura K, Fujimori T, Nagafuchi A. Cadherin subclasses: differential expression and their roles in neural morphogenesis. Cold Spring Harbor Symp Quant Biol 1990; LV:319-325. 5. Larue L, Ohsugi M,HirchenhainJ, Kemler R. E-cadherin null mutant embryos fail to form a trophectoderm epithelium. Proc Natl Acad Sci USA 1994; 91:8263-8267. 6. Riethmacher D, Brinkmann V,Birchmeier C. A targeted mutation in the mouse E-cadherin gene results in defective preimplantation development. Proc Natl Acad Sci USA 1995; 92:855-859. 7. Shimamura K, Takeichi M. Local and transient expression of E-cadherin involved in mouse embryonic brain morphogenesis. Development 1992; 116:1011-1019.

1238

CHEN AND HALES

8. Matsunaga M, Hatta K, Takeichi M. Role of N-cadherin cell adhesion molecules in the histogenesis of neural retina. Neuron 1988; 1:289-295. 9. Hirai Y, Nose A, Kobayashi S, Takeichi M. Expression and role of E- and P-cadherin adhesion molecules in embryonic histogenesis. I. Lung epithelial morphogenesis. Development 1989; 105:263-270. 10. Levine E, Lee CH, Kintner C, Gumbiner BM. Selective disruption of E-cadherin function in early Xenopusembryos by a dominant negative mutant. Development 1994; 120:901909. 11. Detrick RJ,Dickey D, Kintner CR. The effects of N-cadherin misexpression on morphogenesis in Xenopus embryos. Neuron 1990; 4:493-506. 12. Fujimori T, Miyatani S, Takeichi M.Ectopic expression of N-cadherin perturbs histogenesis in Xenopusembryos. Development 1990: 110:97-104. 13. Stein CA, Tonkinson JL, Yakubov L. Phosphorothioate oligodeoxynucleotides-antisense inhibitors of gene expression. Pharmacol &Ther 1991; 52:365-384. 14. Chiang M-Y, Chan H, Zounes MA, Freier SM, Lima WF, Bennett CE Antisense oligonucleotides inhibit intercellular adhesion molecule 1 expression by two distinct mechanisms. J Biol Chem 1991; 266:18162-18171. 15. Stein CA, Cheng Y-C. Antisense oligonucleotides as therapeutic agents-is the bullet really magical? Science 1993; 261:1004-1012. 16. Chen B, Blaschuk OW, Hales BE Cadherin mRNAs during rat embryo development in vivo and in vitro. Teratology 1991; 44:581-590. 17. Shepard TH, Fantel A, Mirkes PE. Somite-stage mammalian embryo culture-use in study of normal physiology and mechanisms of teratogenesis. In: Banbury Report 26: Developmental Toxicology: Mechanisms and Risks. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 1987: 29-44. 18. Augustine K,Liu ET, Sadler TW. Antisense attenuation of Wnt-1 and Wnt-3a expression in whole embryo culture reveals roles for these genes in craniofacial, spinal cord, and cardiac morphogenesis. Dev Genet 1993; 14:500-520.

19. New DAT. Whole embryo culture and the study of mammalian development during organogenesis. Biol Rev 1978; 53:81-122. 20. Nagafuchi A, Shirayoshi Y, Okazaki K. Yasuda K, Takeichi M. Transformation of cell adhesion properties by exogenously introduced E-cadherin cDNA. Nature 1987; 329:341-343. 21. Laemmli U. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970; 227:680-685. 22. Towbin H, Staehlin T,GordonJ. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci USA 1979; 76:4350-4354. 23. Vestweber D, Kemler R. Rabbit antiserum against a purified surface glycoprotein decompacts mouse preimplantation embryos and reacts with specific adult tissues. Exp Cell Res 1984; 152:169-178. 24. Lagunowich LA, Donoso LA, Grunwald GB. Identification of mammalian and inverte brate analogue of the avian calcium-dependent cell adhesion protein N-cadherin with synthetic peptide-directed antibodies against a conserved cytoplasmic domain. Biochem Biophys Res Commun 1990; 172:313-320. 25. Golden JA, Chernoff GF. Intermittent pattern of neural tube closure in two strains of mice. Teratology 1993; 47:73-80. 26. Van Allen MI. Kalousek DK, Chernoff GF, Juriloff D, Harris M, McGillivray BC, Yong SL, Langlois S, MacLeod PM, Chitayat D, Friedman JM, Wilson RD, McFadden D, PontzarJ, Ritchie S, Hall JG. Evidence for multi-site closure of the neural tube in humans. Am J Med Genet 1993; 47:723-743. 27. Shore EM, Nelson WJ. Biosynthesis of the cell adhesion molecule uvomorulin (E-cadherin) in Madin-Darby canine kidney epithelial cells. J Biol Chem 1991: 266:1967219680. 28. Freeman SJ, Brown N. An in vitro study of teratogenicity in the rat due to antibodyinduced yolk sac dysfunction. Roux's Arch Dev Biol 1986; 195:236-242.