Neural Crest Specification Regulated by the Helix-Loop-Helix Repressor Id2

See allHide authors and affiliations

Science  14 Aug 1998:
Vol. 281, Issue 5379, pp. 988-991
DOI: 10.1126/science.281.5379.988


Vertebrate neural crest cells, derived from the neural folds, generate a variety of tissues, such as cartilage, ganglia, and cranial (intramembranous) bone. The chick homolog of the helix-loop-helix transcriptional regulator Id2 is expressed in cranial but not trunk neural folds and subsequently in some migrating cranial neural crest cells. Ectopic expression of Id2 with recombinant retroviruses converted ectodermal cells to a neural crest fate, demonstrating that proper regulation of Id2 is important for sustaining epidermal traits. In addition, overexpression of Id2 resulted in overgrowth and premature neurogenesis of the dorsal neural tube. These results suggest that Id2 may allocate ectodermal precursors into neural rather than epidermal lineages.

Neural crest cells follow migratory pathways and generate final cell types in accordance with their original rostrocaudal position along the axis (1,2). Although both cranial and trunk neural crest cells can form melanocytes, glia, sensory neurons, and autonomic neurons (3, 4), only cranial neural crest cells have the ability to form cartilage and bone (2, 5). In screening for transcripts that are selectively expressed in cranial but not trunk neural folds (6), we isolated the chick homolog of Id2 (7) (Fig. 1, A and B). Id proteins (inhibitors of DNA binding) are negative regulators of basic helix-loop-helix (bHLH) transcription factors (7–10) that are involved in developmental processes such as myogenesis (9,11–13), neurogenesis (14,15), bone morphogenesis (16), lymphopoiesis (17), hematopoiesis (18), myeloid differentiation (19), and cell growth (20–22). We show that Id2 directs ectodermal precursor toward neural crest and neurogenic fates.

Figure 1

Amino acid sequence of chick Id2. (A) Comparison of the deduced amino acid sequences of chicken Id2 (GenBank accession number AF068831), human Id2 (35), rat Id2 (31), and mouse Id2 (7). Only residues differing from the chicken Id2 sequence are shown. Chick Id2 is 94% identical to human Id2, 91% identical to rat Id2, and 88% identical to mouse Id2. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr. The complete amino acid sequence comparison is available (B) A phylogenetic tree illustrating the relationship between chick Id2 and other Id proteins. The tree was prepared with the Clustal method with a PAM250 residue weight table.

In situ hybridization revealed Id2 transcripts in the most dorsal cranial neural fold region of the presumptive brain (Fig. 2, A and B). The caudal border of expression approximately corresponds to the midvagal region (Fig. 2, A and C) at the rhombomere 7/8 boundary and is maintained beyond stage 17. As neural crest cells migrated out of the neural tube,Id2 was expressed by a subpopulation of migrating cranial neural crest cells at the midbrain level (Fig. 2D). Transverse sections show Id2 mRNA in recently emigrated neural crest cells as they move toward the branchial arches (Fig. 2E). At later stages, Id2 was expressed in the region of forming cranial ganglia, such as the ciliary and trigeminal ganglia. Within the neural tube, expression became restricted to the alar plate in stage 28 embryos. In addition to expression in cranial neural fold derivatives, Id2 was expressed in the somites and subsequently in vertebral cartilage.

Figure 2

Expression pattern of chick Id2, which is selectively expressed in the cranial neural folds and neural crest. (A) Whole-mount in situ hybridization of a 5 somite stage embryo reveals that Id2 is expressed in the cranial neural folds, with the caudal limit of expression in the caudal hindbrain. (B) In a transverse section of the embryo shown in (A), at the level of the dashed line, Id2-expressing cells (arrow) are within the cranial neural folds. NT, neural tube. (C) By the 11 somite stage, Id2 is expressed in the neural tube and in subpopulations of migrating neural crest cells. The caudal limit of expression remains in the caudal hindbrain. The dark spot of asymmetric expression on the left side of the embryo is in the splanchnopleure, in the region of foregut closure just caudal to the heart. (D) Higher magnification of (C) shows subpopulations of Id2-expressing neural crest cells (arrows) migrating from the neural tube. (E) In a transverse section of the embryo shown in (C) and (D) at the level of the dashed line, Id2-expressing cells appear to migrate away (arrowhead) from the neural tube and into the branchial arches. (F) After retrovirally mediated overexpression of chick Id2 in early embryos, staining with an antibody to the viral protein p27 reveals near uniform viral expression in a transverse section through an embryo injected at the 1 somite stage and fixed 1 day later. Note the ectopic ganglion (EG)–like aggregate of cells within the neural tube. (G andH) A 3-day embryo stained with antibody to myc after retroviral infection of myc-tagged Id2 reveals nuclear Id2 expression in cells within the lumen of the neural tube at low (G) and high (H) magnification; arrows in (H) indicate myc-tagged nuclei. Scale bar, 250 μm in (A), 150 μm in (B), 650 μm in (C), 130 μm in (D), 100 μm in (E) and (F), 120 μm in (G), and 20 μm in (H).

We explored the functional importance of Id2 in vivo by overexpressing it in the chick embryo using retrovirally mediated gene transfer with RCASBP (B) virus (replication-competent avian leukemia virus LTR splice acceptor Bryan polymerase, subgroup B envelope genes) (23–26). Id2-expressing retrovirus (either untagged or myc-tagged) was applied to the surface ectoderm (including both ectoderm and open neural plate) of embryos ranging in age from Hamburger-Hamilton (HH) (27) stages 5 to 6 (head process stage) to HH stage 11 (12 somite stage); some embryos also received a subectodermal injection of virus into the underlying mesenchyme. Near ubiquitous expression of the retroviral gene product was observed in embryonic cells examined 1 to 3 days after infection, as assayed by staining either with antibodies to the viral protein p27 (Fig. 2F) or with the myc tag (Fig. 2, G and H).

Profound abnormalities were observed in the neural tube at the levels of the forebrain, midbrain, and hindbrain in embryos analyzed 1 to 9 days after infection. For viral infections performed at the youngest stages (HH stages 5 to 8), the ectoderm directly overlying the neural tube was absent, apparently converted to a neural crest phenotype. The normally cuboidal epithelium of the dorsal ectoderm was replaced by mesenchymal cells that express the neural crest marker, HNK-1 (Fig. 3, B and C). In contrast, embryos infected with a control virus (encoding Id2 in the antisense orientation) displayed the normal pattern of HNK-1 (Fig. 3A) and Slug (Fig. 3I) immunoreactivity, with neural crest cells clearing the dorsal surface of the neural tube by 1 day after infection. To confirm that the ectopic HNK-1–reactive cells were derived from the ectoderm, we labeled the surface ectoderm with the lipophilic dye 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (DiI) after neural tube closure (28). Retrovirus was applied at the 1 to 4 somite stage followed by DiI labeling of the ectoderm after neural tube closure (∼11 somite stage). In control embryos, this DiI application resulted in labeling of the surface ectoderm and a small population of invaginating placodal cells (Fig. 3D) but not of neural crest cells. In embryos infected with Id2-expressing virus, ectodermal cells close to the neural tube already appeared multilayered by the time of DiI application (11 somite stage); some DiI-labeled ectodermal cells expressed the HNK-1 epitope as early as the 14 somite stage (Fig. 3, E and F). By 1 day after infection, there was a marked overgrowth of the DiI-labeled ectoderm with numerous DiI-labeled cells in the lateral ectoderm undergoing extensive overgrowth, becoming multilayered and HNK-1 immunoreactive (Fig. 3, G and H). These results demonstrate that ectopic Id2 expression converts the DiI-labeled surface ectoderm to a neural crest phenotype. The ectoderm-to-neural crest conversion was observed in embryos infected with retrovirus before HH stage 8 (5 somite stage) and markedly declined thereafter (Fig. 3K). As it takes about 15 to 18 hours after infection to achieve high levels of ectopic protein expression (26), the time of competence to convert ectoderm to neural crest appears to decline by HH stage 11 (13 to 14 somite stage).

Figure 3

Retrovirally mediated overexpression of chick Id2 in early embryos. (A) An embryo infected with control virus encoding antisense Id2 displays the normal pattern of HNK-1 immunoreactivity, with neural crest cells (arrow) migrating away from and clearing the dorsal surface of the neural tube. (B) A transverse section through an embryo injected with Id2 retrovirus at the 1 somite stage, fixed after 1 day, and stained with HNK-1 antibody (green) and neurofilament antibody (red). The ectoderm overlying the neural tube was absent and was replaced by large masses of HNK-1–positive cells (curved arrow). A large aggregate of neurofilament-positive cells appeared to form an ectopic ganglion within the neural tube. (C) Another section through the embryo in (B), showing that the cuboidal epithelium of the ectoderm ends and meets (arrow) mesenchymal HNK-1–immunoreactive cells. (D) The surface ectoderm of the embryo in (A) was labeled with the lipophilic dye DiI after neural tube closure, resulting in labeled cells (arrow) in the surface ectoderm and a small population of invaginating placodal cells. (E and F) DiI labeling (E) and HNK-1 immunoreactivity (F) of an embryo infected with Id2-expressing virus showing a marked overgrowth of the ectoderm as early as 5 hours after DiI labeling (14 somite stage); some of the DiI-labeled cells [arrow in (E)] are also HNK-1–positive [arrow in (F)]. (G and H) DiI labeling (G) and HNK-1 immunoreactivity (H) of an embryo 1 day after being infected with Id2-expressing virus; there was a marked over growth of the ectoderm (Ect.), some of which is HNK-1–positive. (I) Slug antibody staining of an embryo 1 day after infection of control virus shows expression in the dorsal neural tube and migrating neural crest cells. (J) Slug antibody staining of an embryo 1 day after infection of Id2 retrovirus shows that a mass of Slug-expressing cells has invaginated into the neural tube (insets). After 1 day, abnormal ectoderm neural tube (or both) phenotypes were observed in 36 of 39 embryos infected with myc-Id2 virus and 9 of 12 embryos with untagged Id2, confirming that the myc tag did not interfere with function of the Id2 construct; 2 of 16 control embryos had abnormalities in the neural tube (typical of the background level of defects observed in noninfected embryos). (K) Conversion of the embryonic ectoderm into neural crest and formation of ectopic ganglia after infection with Id2 retrovirus. Yellow graph shows the percentage of embryos in which the surface ectoderm over the neural tube has been converted into HNK-1–positive cells as a function of stage of infection. Embryos were analyzed 1 day after infection. Between head process (H.P.) and 5 somite stage, 89% of embryos exhibited an ectoderm-to-neural conversion. Between the 6 and 8 somite stages, this phenotype rapidly declined. Given that protein expression takes 15 to 18 hours (24, 26), this decline suggests that the ectoderm loses the ability to be converted to neural crest cells at the 14 somite stage. Maroon graph shows the percentage of embryos with aggregates of ectopic neural crest–like ganglia within or above the dorsal neural tube as a function of stage of infection. Similar phenotypes were observed for infections performed between head fold and the 12 somite stage, the last stage examined. The total number of embryos is shown within each bar. (L) An embryo injected with Id2 retrovirus at the 4 somite stage and allowed to develop for 9 days. The ectoderm over the dorsal neural tube (arrow) was absent, and the forebrain (FB) was not contained within the cranium, perhaps because of overgrowth of the neural tissue. (M) A control embryo injected with antisense Id2 retrovirus at the 4 somite stage and fixed after 9 days. The ectoderm over the neural tube is continuous. In embryos fixed between 2 and 9 days after infection, abnormalities were observed in 8 of 8 Id2-infected embryos and in 0 of 6 control embryos. Scale bars, 90 μm in (A), (D), (G), and (H); 80 μm in (B); 40 μm in (C); 70 μm in (E) and (F); 50 μm in (I) and (J); and 2.5 mm in (K) and (L).

Another phenotype observed with infections performed between HH stages 5 and 11 was overgrowth of the dorsal neural tube and the presence of ectopic masses of cells dorsal to or within the cranial neural tube (Figs. 2, F and G, and 3, B and J). Large ganglionlike aggregates, which contained numerous neurofilament-immunoreactive cells (Fig. 3B), were observed within the lumen of the neural tube in embryos infected with either myc-Id2 or untagged Id2 virus. In embryos fixed between 2 and 9 days after infection, ectopic masses were found within the cerebellum and brain vesicles, which protruded from the cranium lacking dorsal ectoderm (Fig. 3L). In contrast, the majority of embryos infected with control virus displayed no obvious abnormalities (Fig. 3M).

When retroviral constructs were injected into the cranial mesenchyme instead of the neural tube, all tissues except the neural tube expressed the viral protein p27 or the myc tag (or both), suggesting that the virus does not cross the basal lamina of the neural tube. No defects were observed in these embryos. Thus, ectopic expression of Id2 appears only to affect ectodermally derived cells at or near the ectoderm-neuroectoderm border, where Id2 is normally expressed. No morphological abnormalities were observed in the trunk of infected embryos. Because Id2 heterodimerizes with other HLH transcription factors, the ability to detect a phenotype may be directly related to the availability of the appropriate partner.

Here, we have shown that the HLH transcriptional regulator Id2 directs ectodermal precursors into neural instead of epidermal lineages. In the chick, Id2 is differentially expressed in the cranial, but not trunk, neural folds. Consistent with the expression of Id2 in the cranial neural folds, Id proteins are often associated with sites of active epithelial-mesenchymal transitions (29). Overexpression of Id2 results in an epidermal-to-neural crest conversion, suggesting that the bHLH partner to Id2 may promote epidermalization. Because individual neural fold cells can become epidermis, neural crest, or neural tube cells (30), the expression of Id2 in the neural folds may maintain a balance between these fates such that the appropriate proportion of cells enters the neural crest and neural tube lineages. Cranial ganglia are derived from both the neural crest and the ectodermal placodes; furthermore, both placode- and neural crest–derived cells express the HNK-1 epitope. Therefore, we cannot rule out the possibility that overexpression of Id2 drives some ectodermal cells into a placodal fate; however, Id2 is not normally expressed in the placodes. The ectopic, ganglionlike aggregates observed after Id2 overexpression suggest an effect on promotion of neurogenesis, consistent with studies demonstrating Id expression in dividing neuroblasts (15, 31), presumptive interneurons, and motor neurons in the developing spinal cord (15). In Drosophila melanogaster, the Id homolog extramacrochaete regulates sensory organ patterning in the adult fly epidermis by negatively regulating the bHLH proteins achaete and scute (32, 33) in a dose-dependent manner (33, 34). In vertebrates, our results demonstrate that Id2 plays stage-dependent roles in ectodermal development, likely depending on the presence of other interacting bHLH molecules. Thus, Id2 heterodimers may maintain the balance between differentiation and proliferation in neural crest and ectodermal precursors.

  • * To whom correspondence should be addressed. E-mail: mbronner{at}


View Abstract

Navigate This Article