Misexpression of BRE gene in the developing chick neural tube affects neurulation and somitogenesis.

The brain and reproductive expression (BRE) gene is expressed in numerous adult tissues and especially in the nervous and reproductive systems. However, little is known about BRE expression in the developing embryo or about its role in embryonic development. In this study, we used in situ hybridization to reveal the spatiotemporal expression pattern for BRE in chick embryo during development. To determine the importance of BRE in neurogenesis, we overexpressed BRE and also silenced BRE expression specifically in the neural tube. We established that overexpressing BRE in the neural tube indirectly accelerated Pax7(+) somite development and directly increased HNK-1(+) neural crest cell (NCC) migration and TuJ-1(+) neurite outgrowth. These altered morphogenetic processes were associated with changes in the cell cycle of NCCs and neural tube cells. The inverse effect was obtained when BRE expression was silenced in the neural tube. We also determined that BMP4 and Shh expression in the neural tube was affected by misexpression of BRE. This provides a possible mechanism for how altering BRE expression was able to affect somitogenesis, neurogenesis, and NCC migration. In summary, our results demonstrate that BRE plays an important role in regulating neurogenesis and indirectly somite differentiation during early chick embryo development.

INTRODUCTION

The brain and reproductive expression () gene is expressed in a variety of tissues, including the brain, ovary, testis, heart, kidney, and adrenal glands (Li ; Miao ). The gene is most highly expressed in the nervous and reproductive systems, hence its name. is now considered to be an adaptor protein involved in stress response and DNA- damage-repair response by some yet-unknown mechanisms. It is also believed to be a homeostatic or housekeeping protein (Ching ), since the gene is capable of modulating the action of hormones and cytokines in stress response, cell survival, and various pathological conditions, such as inflammation, infection, and cancers (Tang ). Moreover, has also been called “TNFRSF1A modulator” because it can directly bind to tumor necrosis factor receptor 1 (TNFR-1) and modulate TNF signaling (Gu ). Recently we reported that plays an important role in regulating stem cell differentiation by helping to maintain stemness (Chen ). However, the function of during embryo development has not yet been investigated. Hence, in this study, we examined how overexpressing and silencing expression in the chick embryo neural tube affected development.

The neural tube develops from a bilateral pair of neural plate elevations at the early embryonic stage and subsequently fuses to form a tubular structure extending cranial-caudally. This is a complex morphogenetic process involving cell induction, proliferation, and apoptosis. The neural tube will eventually form the spinal cord and brain at late embryogenesis. During neural tube formation, neural crest cells (NCCs) are formed at the crests of neural plate elevations. These NCCs undergo an epithelial-to-mesenchymal transition (EMT) as the neural tube closes and migrate throughout the embryo, where they differentiate into a host of different cell types (Huang and Saint-Jeannet, 2004; Bronner, 2012; Bronner and LeDouarin, 2013). These dynamic processes associated with NCC emergence, migration, and development could be regarded as a continual process of cellular differentiation, that is, the transition of premigratory NCCs in the dorsal neural tube into migratory NCCs and then their migration beneath the neuroepithelium or between the somites and neural tube. Like NCCs, somites are also transient embryonic structures. They are derived from the paraxial mesoderm and take the form of a paired spherical body localized on each side of neural tube (Christ ; Noden ). The somites appear cranially and extend caudally to the tail end of the chick embryo. They differentiate to form the dermatome, myotome, and sclerotome, which in turn become the dermis, skeletal muscles, and cartilage and connective tissues (Christ and Ordahl, 1995; Kageyama ; Eckalbar ).

It is now well established that the neural tube, somites, and notochord closely interact with each other to regulate normal development. During somitogenesis, inhibitory and stimulatory signals generated from the surrounding tissues (such as the notochord, floor plate, neural tube, dorsal ectoderm, and lateral mesoderm) regulate somite morphogenesis and differentiation (Lee ; Francetic and Li, 2011). Signals from the dorsal neural tube affect the development of the somite-derived dermomyotome and myotome (Munsterberg ; Marcelle ; Sela-Donenfeld and Kalcheim, 2002; Serralbo and Marcelle, 2014), and it has been reported that migrating NCCs affect somite differentiation (Rios ; Serralbo and Marcelle, 2014). Reciprocally, the medial lip of the dermomyotome inhibits the transcription of noggin in the neural tube, which relieves repression from bone morphogenic protein (BMP) signaling and stimulates the emigration of NCCs. The migration of NCCs to the ventral and dorsal sides of the somites is involved in regulating somitic myogenesis (Kalcheim, 2011). The notochord and floor plate generate sonic hedgehog (Shh) protein, an important morphogen, which controls many important morphogenetic events in the embryo, including somite development. Shh can transiently control somite formation (Resende ) and promote somitic chondrogenesis (Murtaugh ) and myogenesis by inducing Myf5 expression directly and MyoD indirectly (Chiang ). Inversely, factors produced in a rostrocaudal pattern by the somites could confine the movement of spinal motor axons and NCCs to the rostral half of the somitic sclerotome (Koblar ).

In this study, we first defined the spatiotemporal expression pattern of during early chick embryo development. We then examined the effects of overexpressing and silencing in the neural tube to elucidate the importance of this gene in NCC migration, somite development, neurite outgrowth, and cell cycle.

RESULTS

is expressed in the developing neural tube and somites

In situ hybridization was performed on chick embryos to establish where was spatiotemporally expressed at the primitive streak (HH4) stage to the heart formation (HH12) stage (Figure 1). In stage HH4 chick embryos, the primitive streak has elongated to attain its overall length. The neural plate, which will give rise to the neural tube, has also fully formed. is found mainly expressed in the neural plate at this stage (Figure 1A). In transverse sections, could be seen strongly expressed in the epiblast cells around the primitive streak (neural plate; Figure 1A, 1–4). was also expressed in the mesoderm layer directly beneath the epiblast, although weaker when compared with the neural plate. In stage HH8–10 embryos, was expressed mainly in closing (Figures 1B, 1 and 2) and closed (Figure 1C1) rostral neural tube (derived from the neural plate), which indicates continuous expression for in the neural tissues. Cranial NCCs also expressed (Figure 1C2). This suggests that might be involved in the migration of NCCs. was also expressed in developing somites, although relatively weakly (Figure 1C3). In stage HH12 chick embryos, was expressed in the neural tube (Figure 1D) but not as intensely as in earlier-stage embryos (Figure 1D1). In contrast, expression dramatically increased in head mesenchymal tissues and trunk somites (Figure 1D, 1–3), implying that might also be involved in somite development and differentiation.

FIGURE 1:

Expression pattern of BRE in the early chick embryo. Whole-mount in situ hybridization for expression was performed on stage HH4, 8, 10, and 12 embryos. (A) In HH4 chick embryos, was expressed mainly in the neural plate. The embryo was transversely sectioned along the rostral–caudal axis (A, A1–A4) as indicated by the dotted white lines. (A1–A4) was mainly expressed in the ectoderm (EC) of the neural plate region and weakly in the lateral mesoderm (M) along the embryonic rostral–caudal axis. (B) In HH8 chick embryos, was mainly expressed in the head folds, closing neural folds, and regressing primitive streak. (B1–B3) Transverse sections of B, showing expressed in the closing neural folds and the neuroepithelial site adjacent to the underlying mesoderm layer. (C) was mainly expressed in the forebrain, midbrain, hindbrain, and trunk neural tube of HH10 embryos. (C1–C3) Transverse sections of C showing that was expressed in the cranial neural tube and in NCCs located at the heart tube and trunk neural tube levels. (D) In HH12 chick embryo, was dispersedly expressed in the neural tube and more intensely expressed in mesodermal structures, such as somites. (D1–D3) Transverse sections of D showing that was mainly expressed in somites, brain, and trunk neural tube (although the last two tissues more weakly compared with younger embryos). (E) Schematic drawing depicting how our gene constructs were transfected into the neural tube of chick embryos in ovo. (F) In situ hybridization showing that the endogenous BRE was silenced after BRE-siRNA transfection (left). (G, G1) In situ hybridization showing that BRE expression (white arrows) was increased after unilateral cotransfection of the neural tube with BRE-wt and GFP (left). (H) Semiquantitative RT-PCR analysis confirmed that BRE expression was increased after transfection with BRE-wt construct and decreased after transfection with BRE-siRNA (N = 3). Scale bars, 500 μm (A–D), 20 μm (A1–A4), 50 μm (B1–B3, C1, C2, D2, D3, F, G1), 25 μm (C3), 100 μm (D1). EC, ectoderm; EN, endoderm; M, endoderm; NT, neural tube; PS, primitive streak; So, somite.

Overexpressing in the neural tube accelerates somitogenesis

We have validated that is expressed in the neural tube, using in situ hybridization. Hence we asked whether neural tube development would be affected when we misexpressed . We injected and electroporated constructs containing the full-length (-wt) or small interfering RNA (-siRNA) into the neural tube to overexpress or silence expression, respectively (Figure 1E). Electroporation was performed so that one side of the neural tube was transfected with the construct and the contralateral half was untransfected and served as the control. In situ hybridization was performed to confirm that expression has been ipsilaterally silenced using our -siRNA (Figure 1F) or ipsilaterally overexpressed after transfection with our -wt construct (Figure 1, G and G1) in the neural tube. Using semiquantitative real-time (RT)-PCR and quantitative PCR, we confirmed that expression could be significantly up-regulated after -wt transfection (Figure 1H and Supplemental Figure S1A; control–green fluorescent protein [GFP], 0.98 ± 0.04, and -wt, 1.28 ± 0.09; p < 0.05, N = 8) or down-regulated after -siRNA transfection (Figure 1H, N = 3).

Then immunofluorescence staining was performed for Pax7, which labels the dorsal side of neural tube, somites, and dermomyotome derivatives (Figure 2, A–G; Otto ; Galli ). We determined that the neural tube unilaterally cotransfected with control-siRNA plus GFP constructs has no effect on the Pax7+ somites and neural tube (Figure 2, A–B2, I, and J). Surprisingly, we found Pax7+ somites were larger on the side of the neural tube that overexpressed than the control side (N = 8/10; Figure 2, C and J). In corresponding transverse sections, the affected somites have differentiated to form a wide strip of Pax7+ dermomyotome, whereas on the control side, the dermomyotome was smaller and less differentiated (Figure 2, D1, D2, and I). At more caudal levels, where the somites are less mature, the larger somites on the -overexpressed side of the neural tube were still clearly evident (Figure 2, E1, E2, and I; transfection-side somite area vs. control-side somite area: control-GFP, 0.98 ± 0.09, and -wt, 1.57 ± 0.15; p < 0.001). Next we silenced BRE expression in half of the neural tube; this inhibited somite development (N = 7/10; Figure 2, F and J). In corresponding transverse sections, we found that the length of the Pax7+ dermomyotome on the BRE-silenced side of the neural tube (left) was smaller than on the contralateral control side (Figure 2, G1 and G2), suggesting that reduced BRE expression in the neural tube indirectly interfered with somite development and differentiation (Figure 2I; transfection-side somite area vs. control-side somite area: BRE-siRNA, 0.81 ± 0.04; p < 0.001). Furthermore, the level of Pax7 expression was confirmed by semiquantitative RT-PCR and quantitative PCR analysis after electroporation with BRE-wt and BRE-siRNA (Figure 2H; control-GFP, 0.98 ± 0.04; BRE-wt, 1.39 ± 0.16; p < 0.05, N = 7; Supplemental Figure S1B). The quantitative PCR results showed that Pax3, another gene involved in somite differentiation, was also increased after BRE overexpression in the neural tube compared with control samples (control-GFP, 0.91 ± 0.14, BRE-wt, 2.63 ± 0.63; p < 0.05, N = 7; Supplemental Figure S1C).

FIGURE 2:

Somite differentiation is affected by misexpressing BRE in neural tube. (A) Representative whole-mount embryo showing the neural tube unilaterally cotransfected with control-siRNA plus GFP constructs, cultured, and then immunofluorescently stained for Pax7 expression. The somites of the transfection side (white arrows) are similar to the somites on the contralateral side of the untransfected neural tube. N = 10/10 embryos. (B1, B2) Transverse section indicated by white dotted line B in A. (B1) Pax7-labeled somites in the transfection side of neural tube are similar to those on the contralateral control side. (B2) Merge of B1with GFP and DAPI staining. (C) Representative whole-mount embryo showing the neural tube unilaterally cotransfected with BRE-wt plus GFP constructs, cultured, and then immunofluorescently stained for Pax7 expression. The somites adjacent to left side of the neural tube, which overexpresses BRE, are grossly larger than the somites on the contralateral side of the untransfected neural tube. Somite differentiation is indicated by Pax7 expression (white arrows) and shows that BRE overexpression in the neural tube indirectly enhanced somite differentiation (wider somite formed). N = 8/10 embryos. (D1, D2) Transverse section indicated by white dotted line D in C. (D1) Pax7 marker in the dorsal neural tube and somites. The Pax7-labeled somites in the BRE overexpressed side of neural tube differentiated faster and more extensively (dermomyotome formation indicated by white arrows) than on the contralateral control side. (D2) Merge image of D1 with GFP and DAPI staining. (E1, E2) Transverse section indicated by white dotted line E in C. (E1) Pax7-labeled dorsal neural tube and somites. The somites in E1 are less mature than in D1, but, as in D1, the somites on the BRE overexpressed side of neural tube are significantly bigger (white arrows) than on the contralateral control side. (E2) Merge image of E2 with GFP and DAPI staining. (F) Representative whole-mount embryo showing the neural tube unilaterally cotransfected with BRE-siRNA plus GFP constructs, cultured, and then immunofluorescently stained for Pax7 expression. The somites opposing the BRE-silenced side of the neural tube are significantly smaller (white arrows) than the somites on the contralateral side of the untransfected neural tube. N = 7/10 embryos. (G1, G2) Transverse section indicated by white dotted line G in F. (G1) Pax7-labeled somites on the BRE-silenced side of neural tube are smaller and differentiation retarded (shorter dermomyotome formed) compared to those on the contralateral control side. (G2) Merge of G1with GFP and DAPI staining. (H) Semiquantitative RT-PCR analysis of the neural tube transfected with control-GFP, BRE-wt, or BRE-siRNA constructs. BRE overexpression enhanced Pax7 expression, whereas silencing BRE decreased Pax7 expression compared with the control. N = 3. (I) Bar chart of the somite area ratio (transfection side vs. control side from the sections) in control, BRE-wt, and BRE-siRNA– transfected embryos (N = 10). (J) Schematic illustrations showing that the neural tube unilaterally cotransfected with control-siRNA plus GFP constructs has no effect on the differentiation of somites. However, BRE overexpression in the neural tube indirectly accelerates somite differentiation (middle), whereas silencing BRE retarded somite differentiation (bottom). Scale bar, 100 μm (A, C, F), 20 μm (B1, B2, D1–D3, E1, E2, G1, G2). Ctrl-GFP, control-GFP; NT, neural tube; So, somite.

Neural tube is required for proper somitogenesis and differentiation

Developmentally, there is a very close relationship between the somites and neural tube. It has been proposed that inhibitory cross-talk between the paraxial mesoderm and neural primordium controls the timing of neural crest delamination to match the development of a suitable mesodermal substrate for subsequent NCC migration (Sela-Donenfeld and Kalcheim, 2000). Our foregoing data show that somites developed asymmetrically when the opposing side of the neural tube overexpressed BRE. To confirm directly that there is cross-talk between the somite and neural tube, we cultured explants composed of somite and presomitic mesoderm (PSM) devoid of the neural tube or explants composed of somite and PSM plus neural tube (Figure 3) in vitro at 37°C for 48 h. Immunofluorescence staining for Pax7 clearly demonstrated that somites develop and differentiate much faster in the presence of the neural tube than in its absence (Figure 3B). This confirms that somite development and differentiation are dependent on signals emitted from ipsilateral neural tube.

FIGURE 3:

Presence of the neural tube is required for normal somite development. Somite development was compared in the presence or absence of neural tube in explant culture. A HH10 chick embryo was sliced into two parts along the rostral–caudal axis and then cultured 24 h. (A) The left explant contained somites that developed in the absence of neural tube, whereas for the right explant, the somites developed in the presence of neural tube. (B) Immunofluorescence staining for Pax7 was performed on explant cultures in A. In both bright-field and Pax7 IHC images, the somites developed normally in the presence of neural tube. In contrast, in the absence of the neural tube, somite development was retarded and abnormal, illustrating the importance of the neural tube in somitogenesis (C). Scale bar, 100 μm. NT, neural tube; So, somite.

BRE modulates NCC migration and neurite outgrowth

Normally, NCCs delaminate from the dorsal neural tube of the trunk, with one population of NCCs migrating through the rostral half of adjacent differentiating somites. This raises the possibility that the larger somites formed on the BRE-transfected neural tube side might be a consequence of NCC invasion. To address this question, we performed immunofluorescence staining against HNK-1 (a maker for migrating NCCs) after ipsilateral BRE overexpression or silencing on one side of the neural tube (Figure 4). In control embryos, we ipsilaterally transfected the neural tube with a GFP-only construct and stained the embryos with HNK-1 antibody to demonstrate the normal extent of NCC migration (Figure 4, A–D, N = 6 embryos). The results show that our electroporation technique did not physically damage the neural tube to prevent NCC migration. When we ipsilaterally overexpressed BRE in the neural tube, we found that NCC migration was significantly increased on the BRE-transfected side in 93% of embryos as compared with the contralateral control side (Figure 4, E–H and M; N = 14 embryos). Conversely, NCC migration was reduced when BRE expression was ipsilaterally silenced in the neural tube (Figure 4, I–L and M, N = 8 embryos). We then calculated the area occupied by HNK-1+ cells in the BRE-manipulated neural tube side versus the control side. The ratio GFP/control was 0.97 ± 0.03 (N = 8 sections; Figure 4N), BRE-wt/control was significantly increased at 2.01 ± 0.21 (N = 8 sections, p < 0.01; Figure 4N), and BRE-siRNA/control was significantly reduced at 0.42 ± 0.04 (N = 8 sections, p < 0.001; Figure 4N). Together the results suggest that overexpression of BRE enhances the migration of NCCs (Figure 4O), which then invade the differentiating somites, making them appear larger than normal. Besides HNK-1, we examined other neural crest–specifier genes (Snail2, MSX1, FoxD3, and Sox9) by quantitative PCR analysis (Supplemental Figure S1, D–G). We found that overexpressing BRE did not affect Snail2 and MSX1 but increased FoxD3 and reduced Sox9 expression. In situ hybridization confirmed that BRE hardly affects Snail2 expression (N = 3 embryos in each group; Supplemental Figure S2).

FIGURE 4:

Effects of overexpressing and silencing BRE in the neural tube on NCC migration. (A-D) The neural tube was unilaterally cotransfected with control-siRNA plus GFP constructs. (A) Immunofluorescence staining for HNK-1 labels the migrating NCCs on both sides of the neural tube (white arrows). (A1) Merge image of HNK-1 and GFP. (B) Left side of the neural tube expressing GFP in section. (C) Immunofluorescence staining for HNK-1 labels the migrating NCCs on both sides of the neural tube (white arrows) in section. (D) Merge image of A and B stained with DAPI. (E–H) The neural tube was unilaterally transfected with BRE-wt. (E) HNK-1+ NCCs migrating more extensively on the neural tube side overexpressing BRE than on the nontransfected side (white arrows). (E1) Merge image of HNK-1 and GFP. (F) Left side of the neural tube overexpressing BRE. (G, H) HNK-1+ NCCs migrating more extensively on the neural tube side overexpressing BRE than on the nontransfected side (white arrows) in section. (I–L) The neural tube was unilaterally cotransfected with BRE-siRNA and GFP constructs. (I) HNK-1+ NCC migration was reduced on the silenced BRE neural tube compared with NCC migration on the nontransfected side (white arrows). (I1) Merge image of HNK-1 and GFP. (J) Left side of the neural tube silenced by BRE-siRNA and expressing GFP. (K, L) HNK-1+ NCC migration was reduced on the silenced BRE neural tube compared with NCC migration on the nontransfected side (white arrows) in section. (M) Bar chart showing abnormal NCC migration (black) after BRE overexpression and silencing (BRE-wt, N = 14 embryos; BRE-siRNA, N = 8 embryos). (N) Bar chart showing the relative area in the embryo invaded by HNK-1+ NCCs in control, BRE-wt, and BRE-siRNA–transfected embryos (N = 8 sections/embryo, **p < 0.01, ***p < 0.001). (O) Schematic illustration showing that BRE promotes NCC migration. Error bars, SE. Scale bar, 100 μm (A, A1, E, E1, I, I1), 20 μm (B–D, F–H, and J–L). NT, neural tube.

Besides NCC delamination, neurites also grow out from the developing neural tube. Hence we want to establish whether altering BRE expression affects this process. Neurofilament (NF) antibody was used to label intermediate filaments within neurites, which protrude from anterior horns of the spinal cord (Supplemental Figure S3, D1 and D2). Control-GFP transfection confirmed that the electroporation did not harm the neural tube and was able to produce neurite outgrowths (Figure 5, A, A1, and D; N = 6 embryos). We found that ipsilaterally overexpressing BRE resulted in an earlier appearance of NF+ neurite outgrowth from the anterior horn in 92% of embryos as compared with the contralateral control side (Figure 5, B, B1, and E; N = 13 embryos). In contrast, silencing BRE expression resulted in a delay in the appearance of NF+ neurites in 86% of the embryos examined; the phenotype can be easily observed on the rostral side (Figure 5, C, C1, and F; N = 7 embryos). We also used the TuJ-1 neuron marker to confirm our observation. TuJ-1 identifies neuron-specific class III β-tubulin. The neural tube was unilaterally transfected with control-GFP (Figure 5, G, G1, and K; N = 6 embryos), BRE-wt, or BRE-siRNA constructs. The results were consistent with the NF-staining patterns, where TuJ-1+ neurites grew out earlier from the side of the neural tube that overexpressed BRE than the control side in 86% of the embryos (Figure 5, H, H1, and K; N = 7 embryos). Conversely, silencing BRE expression delayed the outgrowth of the TuJ-1+ neurites in 88% of the embryos examined (Figure 5, I, I1, and K; N = 8 embryos). In addition, we used semiquantitative PCR to detect Neurogenin2, a basic helix-loop-helix transcription factor that functions in neuronal differentiation (Simmons ). Neurogenin2 was up-regulated after BRE-wt transfection and down-regulated after BRE-siRNA transfection (Figure 5L). In sum, altered levels of BRE expression can affect neurite outgrowth from the neural tube (Figure 5M).

FIGURE 5:

BRE affects the initiation of motor neuron outgrowth from the neural tube anterior horns. BRE was overexpressed or silenced on the left side of HH10 neural tube. The embryos were cultured, and the motor neurons started to develop and were stained with NF (A–F) and TuJ-1 (G–I) antibodies. (A–C) Whole-mount immunohistochemistry against NF to label the neurites. (A1–C1) Merge images of GFP and NF staining. (D–F) Transverse sections. (A, A1, D) Left side of the neural tube transfected with control-GFP constructs, showing that NF expression is symmetrical on both sides of the neural tube. (B, B1, E) Left side of the neural tube cotransfected with BRE-wt and GFP constructs, showing more NF+ cells on the BRE-overexpressed side than on the untransfected side, as indicated by white arrows. (C, C1, F) Left side of the neural tube cotransfected with BRE-siRNA and GFP constructs, showing fewer NF+ cells on the BRE-silenced side than on the untransfected side, as indicated by the white arrows. (G, G1) Left side of the neural tube transfected with control-GFP construct, showing that TuJ-1+ cells are symmetrically distributed on both sides of the neural tube (white arrows). (H, H1) Left side of the neural tube cotransfected with BRE-wt and GFP constructs, showing that there are more TuJ-1+ cells on the side overexpressing BRE than on the nontransfected side (white arrows). (I, I1) Left side of the neural tube cotransfected with BRE-siRNA and GFP constructs. There are fewer TuJ-1+ cells on the BRE-silenced side than on the nontransfected side. (J) Bar chart showing that NF+ neurite outgrowth is affected (black) by BRE misexpression (control-GFP, 6 embryos; BRE-wt, 7 embryos; BRE-siRNA, 8 embryos). (K) Bar chart showing that TuJ-1+ neurite outgrowth is affected (black) by BRE misexpression (control-GFP, 6 embryos; BRE-wt, 13 embryos; BRE-siRNA, 7 embryos). (L) Semiquantitative RT-PCR analysis showing that BRE overexpression in the neural tubes increased Neurogenin2 expression, whereas silencing BRE reduced Neurogenin2 expression (N = 3). (M) Schematic drawing illustrating that BRE promotes motor neurite outgrowth from anterior horns. Scale bar, 100 μm (A–C), 50 μm (D, E1, F, F1, G, H1, I, I1). NT, neural tube.

BRE modulates cell cycle progression and survival in the developing neural tube

We investigated whether there was any association between BRE expression and cell cycle, which would help explain why NCC migration and neurite outgrowth were abnormal. We established that 68.01 ± 1.60% of BRE-overexpressing (GFP+) cells were bromodeoxyuridine positive (BrdU+), whereas 49.03 ± 1.56% of control GFP+ were BrdU+. This implies that that BRE overexpression significantly accelerated cells into S phase compared with control cells in the neural tube (p < 0.001; Figure 6, A–A2, B–B2, and J). In contrast, there were significantly fewer BRE-silenced (GFP+) cells that were BrdU+ cells (43.06 ± 1.11%) than the control (p < 0.05; Figure 6, C–C2 and J). In the BRE-overexpressed group, we also found GFP+ and BrdU+ colocalization in NCCs, as indicated by arrows (Figure 6, B–B2). In addition, we found that BrdU+ cells in the BRE-wt–transfected side were greater than on the contralateral control side in somites (control-GFP, 0.96 ± 0.03, N = 6; BRE-wt, 1.72 ± 0.06, N = 12; ***p < 0.001; Supplemental Figure S4, A–D and G). Conversely, the BrdU+ cells in the BRE-siRNA–transfected side of somites were fewer (BRE-siRNA, 0.70 ± 0.02, N = 12; ***p < 0.01; Supplemental Figure S4, A–B and E–G). These results imply that BRE is also involved in neural tube, NCC, and somite development by modulating the cell cycle.

FIGURE 6:

Ability of BRE to modulate cell cycle progression and survival in the developing neural tube. The neural tubes were unilaterally transfected with control-GFP (A), BRE-wt plus GFP (B), or BRE-siRNA plus GFP (C) constructs. (A1–C1) Transverse sections of the neural tube immunofluorescently stained for BrdU after transfection. (A2–C2) Merge images of BrdU incorporation and GFP marker expression. There are significantly more BrdU+/GFP+ cells in the BRE-wt–overexpressed samples (B2) than in the control-GFP samples (A2). Cyclin D1 expression after unilateral transfection of the neural tube (left side) with control-GFP (D), BRE-wt plus GFP (E), and BRE-siRNA plus GFP (F) constructs. In situ hybridization shows overexpressing BRE increased cyclin D1 expression (E, E1) as compared with the control (D, D1). Conversely, silencing BRE expression decreased cyclin D1 expression (F, F1). (G–I) TUNEL assay showing that overexpressing BRE did not increase apoptosis (H) compared with the control (G). However, silencing BRE increased the incidence of apoptosis (the transfected side of the neural tube is marked by a green asterisk). (J) Bar chart showing the ratio of BrdU+/GFP+ cells in neural tube transfected with control-GFP, BRE-wt, and BRE-siRNA constructs. *p < 0.05, ***p < 0.001. (K) Bar chart displaying the incidence of normal and abnormal phenotypes resulting from BRE overexpression and silencing. Control-GFP (N = 6), BRE-wt (N = 5), and BRE-siRNA (N = 6) transfection. (L) Semiquantitative RT-PCR analysis showing that BRE overexpression in the neural tubes increased cyclin D1 expression, whereas silencing BRE reduced cyclin D1 expression (N = 3). (M) Bar chart showing that silencing BRE expression significantly increased apoptosis. Ratio of TUNEL+ cells in the transfected side of the neural tube vs. the contralateral control side (N = 6, ***p < 0.001). Error bars, SE. Scale bars, 20 μm (A–C, G–I, D1–F1), 100 μm (D–F). NT, neural tube.

We next examined the effects of altered BRE expression on cyclin D1 expression, since the latter is indispensable for G1/S cell cycle transition. Using in situ hybridization, we showed that cyclin D1 expression was little affected in the control group (Figure 6, D, D1, and K; N = 6). When BRE was overexpressed, cyclin D1 expression was correspondingly increased in the BRE-wt–transfected side of the neural tube in 60% of the embryos (Figure 6, E, E1, and K; N = 5). Conversely, silencing of BRE expression reduced cyclin D1 expression in 100% of the embryos examined (Figure 6, F, F1, and K; N = 6). Furthermore, the extent of cyclin D1 expression in the neural tube was validated by semiquantitative RT-PCR analysis after electroporation with BRE-wt or BRE-siRNA (Figure 6L). These data suggest that BRE promotes G1-S transition. Aberrant cell cycle reactivation in postmitotic neurons could lead to apoptosis (Becker and Bonni, 2005), and cyclin D1 is essential for regulating neuronal cell death (Kranenburg ). Hence we examined whether cell survival was altered after BRE-wt and BRE-siRNA transfection. Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay revealed that there was no significant difference in the ratio (transfection side vs. control side) of TUNEL+ between the control-GFP group and the BRE-wt group (Figure 6, G, H, and M; control, 1.15 ± 0.07, and BRE-wt, 1.24 ± 0.09; p > 0.05). However, there are more apoptotic cells present in the BRE-silenced side of the neural tube than the control side (Figure 6, I and M; BRE-siRNA, 1.93 ± 0.15; p < 0.001). We also found that p53 and BRCA1, which can regulate the cell cycle (Agarwal ; Deng, 2006), were significantly decreased (Supplemental Figure S1H) or increased (Supplemental Figure S1I), respectively, after overexpressed BRE.

BRE regulates gene expression pattern along the dorsoventral neural tube

BMP4 and Shh are two important genes expressed in the dorsoventral neural tube, and they also play crucial roles in somite differentiation (Murtaugh ; Anderson ; Resende ; Kalcheim, 2011; Van Ho ) and neuron differentiation in the ventral neural tube (Wilson and Maden, 2005; Ribes and Briscoe, 2009). Hence we investigated whether influence of BRE on the neural tube, NCC, and somite development was mediated by BMP4 and Shh signaling. We performed in situ hybridization to examine BMP4 and Shh expression after BRE overexpression and silencing in the neural tube. We found that ipsilateral transfection of the neural tube with our GFP construct (control) did not affect BMP4 expression in the dorsal neural tube or Shh expression in notochord and ventral neural tube. Both genes were expressed symmetrically (Figure 7, A, A1, B, and B1, N = 3/3). After BRE overexpression, we found BMP4 expression correspondingly up-regulated at the affected ipsilateral neural tube side (Figure 7, C and C1; N = 3/4). When BRE expression was silenced, we determined that BMP4 expression was reduced in the neural tube (Figure 7, E and E1; N = 3/5). In contrast, Shh expression in the ventral neural tube was inhibited after BRE overexpression (Figure 7, D and D1; N = 4/6), and enhanced after BRE silencing (Figure 7, F and F1; N = 4/6). We validated these observations by semiquantitative RT-PCR analysis (Figure 7G). The results imply that BRE promotes BMP4 expression in the dorsal neural tube and inhibits Shh expression in the ventral neural tube (Figure 7H). BMP and Wnt family members are known to antagonize Shh expression and play an important role in the patterning of the dorsal axis (Patten and Placzek, 2002; Ulloa and Marti, 2010). Thus we also detected whether Wnt signaling was affected after BRE transfection. The results showed that BRE overexpression could up-regulate β-catenin, whereas BRE silencing could down-regulate β-catenin (Supplemental Figure S5).

FIGURE 7:

BRE affects BMP4 and Shh expression in the chick neural tube. HH10 neural tubes were unilaterally (green asterisks) transfected with control-GFP (A, B), BRE-wt plus GFP (C, D), and BRE-siRNA plus GFP (E, F) constructs. After 20 h, in situ hybridization was performed to establish BMP4 and Shh expression. (A1, B1) Normal expression pattern for BMP4 (black arrows in A1) and Shh on both sides of the neural tube. Shh expression is symmetrically distributed in the neural tube, as indicated by the dotted white line in B1. (C1, D1) BRE overexpression enhanced BMP4 expression (black arrows) in the dorsal neural tube compared with the untransfected side. In contrast, Shh expression was reduced in the BRE-overexpressed side of the neural tube, as indicated by the dotted white line in D1. (E1, F1) Silencing BRE represses BMP4 expression (black arrow in E1) compared with the contralateral, untransfected side of the neural tube. In contrast, Shh expression slightly increased compared to the untransfected side (dotted line in F1). (G) Neural tubes were collected for RT-PCR analysis after BE overexpression and silencing. Overexpressing BRE increased BMP4 expression, whereas silencing BRE suppressed BMP4, compared with the control (N = 3). (H) Schematic illustration showing BRE promoting BMP4 expression and suppressing Shh expression in the chick neural tube. Scale bar, 20 μm (A–F). NT, neural tube

Overexpressing BRE in the neural tube activated BMP signaling

We examined the expression of p-Smad1/5/8 after we overexpressed BRE. In control embryos, we ipsilaterally transfected the neural tube with a GFP-only construct and stained the embryos with p-Smad1/5/8 antibody to demonstrate that our electroporation technique did not cause any nonspecific changes (N = 3/3; Figure 8, A and A1). When we ipsilaterally overexpressed BRE in the neural tube, we found that p-Smad1/5/8 was significantly increased on the BRE transfected–side compared with the contralateral control side (N = 4/4; Figure 8, B and B1). Next we used in vitro experiments and added LDN-193189, which can inhibit the p-Smad1/5/8 (Yu ), to rescue BRE effects by BMP modulation. The neural tubes were cotransfected with control-GFP plus control-siRNA or BRE-wt or control-GFP plus BRE-siRNA and then explanted into culture dishes at 37°C with 5% CO2 for 48 h. These cultured neural tube explants were checked for extent of cell migration and HNK-1 expression in migratory cells. The results revealed that BRE promoted NCC migration, as the area covered by migrating cells increased when BRE was up-regulated (1.85 ± 0.28 mm2, N = 5, p < 0. 01). In contrast, when BRE was down-regulated, the area covered by migrating cells was significantly reduced (0.56 ± 0.19 mm2, N = 4, p < 0.05) when compared with control GFP explants (1.09 ± 0.17 mm2, N = 4; Figure 8, C–F and K). In addition, there were significantly more HNK-1+ cells in the BRE-overexpressing explants (78.88 ± 5.22%; p < 0.001) and significantly fewer HNK-1+ cells in the BRE-silenced explants (31.88 ± 5.05%; p < 0.01) compared with the control explants (39.75 ± 3.45%, Figure 8, G–J1 and L). However, NCC migration of the BRE-wt–transfected tube was significantly inhibited in the LDN-193189 culture group (area, 0.81 ± 0.44 mm2; HNK-1+ cell percentage, 44.16 ± 5.16%) compared with the BRE-wt–transfected tube in control culture (p < 0.001).

FIGURE 8:

BRE modulates NCC migration through activation of BMP signaling. (A, A1) Transverse sections of neural tube immunofluorescently stained for p-Smad1/5/8 after transfection of the control-GFP. (B, B1Transverse sections of neural tube immunofluorescently stained for p-Smad1/5/8 after transfection of the BRE-wt. (C–F) Cultured neural tube explants transfected with control-GFP + control-siRNA (C), BRE-wt (D), control-GFP + BRE-siRNA (E), or BRE-wt with LDN-193189 added (F), immunofluorescently stained for HNK-1. (G–J) Higher-magnification images of C–F (top left corner shows the edge of the neural tube). (G1–J1) Merged images (DAPI, HNK-1, and GFP). (K) Area occupied by HNK+ migratory cells. Image-Pro Plus 6.0 software was used to analyze the data. (L) Ratio of HNK-1+ to total number of cells in each sample. *p < 0.05, **p < 0.01, and ***p < 0.001 indicate significant difference between experimental and control-GFP groups. ###p < 0.001 indicates significant difference between BRE-wt and BRE-wt+LDN-193189 embryos. Scale bars, 50 μm (A–B1), 500 μm (C–F), 100 μm (G–J1). Error bars, SD.

DISCUSSION

We recently reported that BRE is important for maintaining stemness in human umbilical cord perivascular mesenchymal stem cells. We found that silencing BRE expression in these cells could dramatically accelerate osteogenic induction and differentiation (Chen ). This observation suggests that BRE may potentially play a role in development, but the function of this gene in embryogenesis had never been investigated. The early chick embryo is a good model for determining the multifunctional nature of the BRE gene during development. In adult mice, BRE is mainly expressed in the nervous system and its precursors (Li ; Miao ), and consistent with this, we found BRE expressed in the neural plate of primitive streak–stage embryos and in the neural tube of older embryos. We also detected BRE expression in the cranial NCCs. In this context, we decided to manipulate BRE gene expression in the neural tube of all experiments since the tissue normally express high levels of BRE during early embryonic development. Of interest, either overexpression or knockdown of BRE expression in the neural tube affected somites as the embryo developed. Our data suggest that this is due to the effect of BRE on expression of morphogenes involved in dorsoventral neural tube patterning and on NCCs.

In ovo electroporation, a widely used method in the chick embryo (Itasaki ), allowed us to transfect half of the neural tube with BRE constructs while the contralateral half was left untransfected and served as control. Surprisingly, we found that overexpressing and silencing BRE in the chick neural tube did affect neural tube patterning genes BMP4 and Shh but affected the morphology of the flanking somites. When BRE was overexpressed, we found that the somites opposing the side that overexpressed BRE were significantly larger than those on the control side. The inverse effect was elicited when BRE expression in the neural tube was silenced. Because Pax7 is expressed in dorsal neural tube and upper-lateral somites (Otto ; Galli ), it was used to mark the process of somite differentiation after transfection of BRE in the neural tube (Figure 2). This suggests that misexpression of BRE in the neural tube can indirectly affect somitogenesis. It is now well established that the neural tube, somites, and notochord all interact intimately to regulate each other's development (Patten and Placzek, 2002; Ulloa and Marti, 2010).

The NCCs normally delaminate from the region between the dorsal neural tube and overlying ectoderm and migrate out toward the periphery of the embryo. Therefore we investigated whether these cells were directly affected when BRE was misexpressed in the neural tube. We used HNK-1 as the marker for migrating NCCs (Tucker ) and found that overexpressing BRE increased HNK-1+ NCC emigration. It was reported that transcription factor FoxD3 induced NCC delamination, and Sox9 was down-regulated after NCCs initiated their migration (McKeown ). This is consistent with our result that overexpressing BRE accelerates NCC migration. Sela-Donenfeld and Kalcheim (2000) proposed that the timing of NCC delamination was regulated by developing somites and that they serve as substrates for NCC migration. Because overexpressing BRE in the neural tube causes larger somites to develop, this may feed back to further enhance NCC migration.

The segmented peripheral nervous system in the trunk is generated by the ventral migration of NCCs, which invade the anterior sclerotome and then differentiate into metameric dorsal root and sympathetic ganglia (Kuo and Erickson, 2010). The ventral spinal motor axons also project through the somites in a segmental manner (Roffers-Agarwal and Gammill, 2009). We explored whether neurite outgrowth from the ventral neural tube was affected by BRE misexpression. We used NF and TuJ-1 as neuronal markers, since NF is an intermediate filament and TuJ-1 is a neuron-specific, class III β-tubulin. We found that overexpressing BRE enhanced neurite outgrowth, whereas silencing BRE inhibited outgrowth. It has been reported that the developing motor and sensory axons are intimately associated with their surrounding tissues, which help to direct and guide axon growth (Tannahill ). Motor axons turn or branch away after encountering the posterior sclerotome cells, whereas the anterior sclerotome stimulates axon motility (Oakley and Tosney, 1993). This suggests that neurite outgrowth across somites might also act as a medium between neural tube signaling and somite development.

For neural tube to develop normally, it relies on a correct balance between cell proliferation and apoptosis in the neuroepithelium (Wei ). Cell proliferation is an essential process found in every aspect of embryo development, especially during the early developmental stages (Fu ). This is why early embryos are susceptible to changes in the external microenvironment. We performed BrdU-incorporation assays and determined that overexpressing BRE accelerated NCCs and neuroepithelial cells into S phase, whereas silencing BRE inhibited this process. We also examined other cell cycle–related genes, such as cyclin D1 (Figure 6, D–F, and L), p53, and BRCA1 (Supplemental Figure S1, H and I), to further establish an association between BRE and cell cycle. We found that BRE regulated NCC migration and cell cycle by modulating transcription of cyclin D1, which is similar to what was previously reported: the cytoplasmic N-cadherin fragment translocates into the nucleus, stimulates cyclin D1 transcription and crest delamination, and enhances transcription of β-catenin (Shoval ). It was also reported that p53 can coordinate NCC growth and EMT delamination by altering the expression of cell cycle–related genes (Rinon ). We found that overexpressing BRE down-regulated p53 expression. The BRCA1 protein complex forms from an assembly of proteins, RAP80/CCDC98/BRCC36/BRE/MERIT40/BRCA1, at the site of DNA breakage and plays a pivotal role in DNA repair and maintenance of genomic integrity. Increased apoptosis is caused by BRCA1 depletion, and this involves the p53 pathway (Pulvers and Huttner, 2009). This suggests that the cell survival effects seen in the neuroepithelium after BRE transfection might be attributed to the disturbance of p53 pathway.

The neurites are initially formed as outgrowths from the anterior horns of the spinal cord, in which BRE might be involved in regulating the neurogenesis through promoting Neurogenin2 expression (Figure 5L). BMP4 also regulates Neurogenin2 expression (Ota and Ito, 2006). The neurite outgrowth process is associated with Shh and BMP4 signaling from the notochord and the dorsal neural tube, respectively. Hence we investigated the expression patterns of BMP4 and Shh in the neural tube after BRE misexpression. We found BMP4 expression was significantly increased in the dorsal neural tube when we overexpressed BRE. This may explain why when we overexpressed BRE in the neural tube, oversized somites were produced, and there also was increased NCC migration. In support, we observed the inverse effect when BRE expression was silenced in the neural tube. Because BMP4 antagonizes Shh, we found that silencing BRE correspondingly increased Shh expression in the ventral neural tube because BMP4 expression was reduced dorsally by BRE knockdown.

BMPs achieve their inductive effects both locally, via direct cell–cell communication, and over a long range, via BMP-binding proteins, which establish diffusible BMP gradients (Hegarty ). The concentration of active BMP proteins (which deceases ventrally from the roof plate) is crucial for the dorsal/ventral patterning (Liu and Niswander, 2005). The transcription factors Smad1/5/8 are the pivotal intracellular effectors of the BMP family of proteins (Hegarty ). The results of p-Smad1/5/8 suggested that BMP signaling was activated after BRE overexpression. In early neurogenesis, BMP-Smad could act upstream of Wnt-β-catenin (Muller ), and BMP-Smad cooperates with Wnt-β-catenin signaling to control the neurogenesis (Muller ), with BMP-Smad important in the specification of neural fates and Wnt-β-catenin signaling functioning in appropriate proliferation (Hegarty ). These results suggested that the BRE affected neuroepithelial cell proliferation, differentiation, apoptosis, and NCC migration though both BMP-Smad and Wnt-β-catenin signaling.

In summary, BRE is mainly expressed in neural tube, NCCs, somites, and neurites during early embryo development. We altered BRE expression specifically in the neural tube, since it is capable of determining the development of NCCs, somites, and neurites. We found that overexpressing BRE could simultaneously enhance BMP4 and also inhibit Shh expression, dorsoventrally, in the neural tube. This may explain how BRE was able to enhance directly NCC migration and neurite outgrowth and also indirectly somite development. We schematically illustrate the role of BRE during early embryogenesis in Figure 9.

FIGURE 9:

Model depicting how BRE misexpression in the neural tube can modulate somite development. Overexpression of BRE enhances BMP4 and inhibits Shh expression (possibly indirectly) in the neural tube. This would in turn affect the cell cycle through the activation of cell cycle–related genes. NCC migration and neurite outgrowth (neuron differentiation) are directly or indirectly enhanced by BMP4, Shh, and cell cycle. Finally, the enhanced NCC migration and neurite outgrowth might promote somite development.

MATERIALS AND METHODS

Chick embryos

Fertilized leghorn eggs were acquired from the Avian Farm of South China Agriculture University (Guangzhou, China). The eggs were incubated in a humidified incubator (Yiheng Instruments, Shanghai, China) set at 38°C and 70% humidity until the embryos reached the desired developmental stage.

Gene transfection of chick embryos

pEGFP-N3 vector was purchased from Clontech, and BRE-siRNA was purchased from Guangzhou Ribobio. Full-length human BRE (BRE-wt) cDNA was ligated into the pEGFP-N3 vector. The plasmid DNAs were prepared and concentrated to 2 μg/μl using a Tiangen DP107-02 kit. The following strategy was designed for transfecting the different constructs into different stages of chick embryos according to experimental requirements. For in ovo electroporation, plasmid DNA were microinjected into the lumen of the neural tube of HH10-stage chick embryos. The electroporation parameters used for archiving maximum transfection efficiency were as previously described (Yang ; Wang ). The transfected chick embryos were then incubated at 38°C with 70% humidity until the embryos reached the HH13 stage. Some of these embryos were exposed to BrdU (10 μg/ml; Sigma-Aldrich, St. Louis, MO) for 2 h in Early Chick (EC) culture to determine the cell rate in S phase (Chapman ). The somatic level from the eighth to 13th pairs of somites (rostral to caudal) was used for analysis.

Immunohistochemistry

Immunohistological staining was performed on whole-mount chick embryos to determine Pax7, HNK-1, TuJ-1, phospho-Smad1/5/8 (p-Smad1/5/8), β-catenin, and NF expression, as previously described (Yang ; Yue ). All embryos were fixed in 4% paraformaldehyde (PFA) at 4°C overnight and then washed with phosphate-buffered saline (PBS). The embryos were then blocked with 2% bovine serum albumin plus 1% Triton-X plus 1% Tween 20 in PBS for 2 h at room temperature to prevent unspecific immunoreaction. After several washes in PBS, the embryos were incubated with primary monoclonal antibody specific for Pax7 (1:100; Developmental Studies Hybridoma Bank, Iowa City, IA), HNK-1 (1:200; Sigma-Aldrich), TuJ-1 (1:200; Neuromics, Edina, MN), GFP (1:1000; Novus Biologicals, Littleton, CO), p-Smad1/5/8 (1:100, Cell Signaling, Boston, MA), β-catenin (1:100, Epitomics, Burlingame, CA), and NF (1:200; Invitrogen, Carlsbad, CA) overnight at 4°C on a rocker. After extensive washing, the embryos were incubated with anti-mouse or anti-rabbit Alexa Fluor 555 antibody (2 μg/ml; Invitrogen) overnight at 4°C on a rocker to reveal the primary antibody staining. Some embryos were pretreated with BrdU (10 μg/ml; Sigma-Aldrich) for 2 h and then immunohistological staining with BrdU antibody (1:200; BD Biosciences, San Jose, CA). All embryos were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; 5 μg/ml; Invitrogen) for 1 h at room temperature. Subsequently the embryos were sectioned on a cryostat microtome (CM1900; Solms, Germany). The sections were mounted onto glass slides using Mowiol 4-88 mounting solution (Sigma-Aldrich) and sealed with coverslips.

In situ hybridization

BRE antisense probes were synthesized from a 1.2-kb BRE fragment inserted into a pGEMT vector. The BRE-pGEMT plasmid was cut with Xho1 and linearized to generate digoxigenin-labeled antisense BRE probes using T6 polymerase (Boehringer Mannheim Biochemica). Antisense probes that are specific for cyclin D1 (Shoval ), FGF8 (Yang ), BMP4 (Somi ), Snail2 (Nieto ), and Shh (Diez del Corral ) were also synthesized as previously described. Whole-mount in situ hybridization of chick embryos was performed according to a standard in situ hybridization protocol (Henrique ). Briefly the embryos were fixed with 4% PFA overnight at 4°C, dehydrated in a graded series of methanol, and stored at −20°C (overnight). The next day, the embryos were hybridized with BRE digoxigenin-labeled antisense probe overnight at 65°C. After hybridization, the bound RNA probe was visualized by incubation with alkaline phosphatase–conjugated anti-digoxigenin antibodies, and the color was developed in NBT/BCIP (Roche, Basel, Switzerland). The whole-mount stained embryos were photographed and then prepared for cryosectioning to a thickness of 15–20 μm on a cryostat microtome (Leica CM1900).

Analysis of programmed cell death

The extent of cell death in embryonic tissues was established by TUNEL analysis, using an In Situ Cell Death Detection Kit (Roche). The staining was performed according to the protocol provided by the manufacturer, which we adapted for whole-mount chick embryo labeling. TUNEL-positive cells were counted using Image Analysis Software (Olympus, Tokyo, Japan).

siRNA interference experiments

The BRE-siRNA corresponding to TCTGGCTGCACATCATTGA was custom designed to specifically target BRE mRNA; Ctrl-siRNA sequence was AAGCCUCGAAAUAUCUCCU (Tang ; Chen ). Both were purchased from Guangzhou Ribobio. The siRNA was diluted to a concentration of 1 mM in 20 mM KCl, 6 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (pH 7.5), and 200 mM MgCl2. To track the siRNA transfection, we cotransfected BRE-siRNA or Ctrl-siRNA along with a pEGFP-N3 marker (1:1) in the half-side neural tube of HH10 chick embryos. The electroporation was performed using the same condition as described earlier.

RNA isolation, semiquantitative RT-PCR, and quantitative PCR

Transfected chick embryos (control-GFP, BRE-wt, and BRE-siRNA; N = 10 embryos for each group) were incubated at 38°C with 70% humidity until the embryos reached the HH13 stage. Total RNA was then isolated from the embryonic neural tube and somites using TRIzol (Invitrogen) according to the manufacturer's instructions. First-strand cDNA was synthesized at a final volume of 25 μl using a SuperScript III First-Strand kit (Invitrogen). After reverse transcription, RT-PCR amplification of the cDNA was performed using specific primers for chick BRE (5′-GCAGCCTTCCTGAGTCACTT-3′ and 5′-TGCTCTCTTGGCCATTTCGT-3′), Neurogenin2 (5′-TCTCCGTGATTTACGAGCGG-3′ and 5′-CGCTGTAATGTCCCGTGTCT-3′), β-catenin (5′-AGTTCTGGGAGCACAGCAAG-3′ and 5′-TGAACCAT­AACCGCAGCCTT-3′), Pax7 (5′-GCTTACTGAAGAGGTCCGACTGTG-3′ and 5′-ACAAGTTGATGCGAGGTGGAAGG-3′), cyclin D1 (5′-TCGGTGTCCTACTTCAAGTG-3′ and 5′ -GGAGTTGTCGGTGTAAATGC-3′), chick BMP4 (5′-AGGAGTGGCAGAAGTAG-3′ and 5′-CGGCTAATCCTGACGTGTTT-3′), and chick glyceraldehyde-3-phosphate dehydrogenase (5′-GTCAACGGATTTGGCCGTAT-3′ and 5′-AATGCCAAAGTTGTCATGGATG-3′) as previously described (Huber ; Endo ; Scott-Drechsel ). PCR was performed in a Bio-Rad S1000 thermal cycler (Bio-Rad, Hercules, CA). cDNAs were amplified for 30 cycles. One round of amplification was performed at 98°C for 10 s, at 60°C for 15 s, and at 72°C for 30 s (Takara, Japan). The PCR products (20 μl) were resolved on 1% agarose gels (Biowest, Spain) in 1× TAE buffer (0.04 M Tris-acetate and 0.001 M EDTA) and GeneGreen Nucleic Acid Dye (Tiangen, Beijing, China). The reaction products were visualized using a transilluminator (Syngene, Cambridge, United Kingdom) and a computer-assisted gel documentation system (Syngene). Quantitative PCR was also performed using SYBR Premix Ex Tag (Takara, Dalian, China) using a 7900HT Fast Real-Time PCR system (Applied Biosystems, Foster City, CA). The sets of primers used for quantitative PCR are provided in Supplemental Figure S6. Each of these experiments was replicated at least three times.

Primary explant culture

The neural tubes either alone or in the presence of somites of chick embryos (at stage HH10) were incubated in DMEM-F12 culture medium (Gibco, Grand Island, NY) or 200 nM LDN-193189 (SML0559; Sigma-Aldrich) DMEM-F12 inside an incubator (Galaxy S; RS Biotech, Scotland, UK) at 37°C and 5% CO2 for 48 h. For analyzing the neural crest cell migration in vitro, the neural tubes cotransfected with BRE-siRNA, Ctrl-siRNA along with pEGFP-N3, or BRE-wt-GFP were excised from the embryos at the 1st–10th somite levels and then incubated in DMEM-F12 culture medium for the required time. The cultured explants were used for photography and immunofluorescence staining.

Photography

After immunohistological staining, the whole-mount embryos were photographed using a fluorescence stereomicroscope (MVX10; Olympus, Osaka, Japan) and imaging software (Image-Pro Plus 7.0). Sections of the stained embryos were photographed using an epifluorescence microscope (Olympus IX51, Leica DM 4000B) at 200× or 400× magnification using the Olympus software package (Leica CW4000 FISH).

Image acquisition and analysis

For the quantification in the BrdU experiments, we manually counted BrdU+ GFP+ cells versus BrdU+ cells of the neural tube in the transfected side (Cayuso ); the HNK-1+ area was quantified with Image-Pro Plus 6.0; for the primary explant culture experiments, cell migration was quantified by measuring HNK-1+ staining with Image-Pro Plus 6.0; HNK-1+ cells were manually counted on DAPI and HNK-1 merged images. The data are presented as mean ± SE or mean ± SD. Statistical analysis for the experimental data were performed using a SPSS 13.0 statistical package program for Windows. Statistical significance was established using one-way analysis of variance. p < 0.05 was considered to be significantly different.