The pattern of the vertebrate brain is formed as brain vesicles form just after the closure of the neural tube (Fig. 1). Primary brain vesicles such as the prosencephalon, mesencephalon and metencephalon are made and subdivided into secondary brain vesicles: the telencephalon, diencephalon, mesencephalon, metencephalon and myelencephalon (Fig. 1). The fate of the brain vesicles is determined by the combination of transcription factors. Signals from the secondary organizer, the anterior neural ridge and the mesencephalon–metencephalon boundary (isthmus), regulate expression of the transcription factors in the adjacent region and organize its fate (Fig. 1).
The vertebrate central nervous system is elaborated from a simple neural tube. Brain vesicles formation is the first sign of regionalization. Classical transplantation using quail and chick embryos revealed that the mesencephalon–metencephalon boundary (isthmus) functions as an organizer of the mesencephalon and metencephalon. Fgf8 is accepted as a main organizing molecule of the isthmus. Strong Fgf8 signal activates the Ras-ERK signaling pathway to differentiate the cerebellum. In this review, the historical background of the means of identifying the isthmus organizer and the molecular mechanisms of signal transduction for tectum and cerebellum differentiation is reviewed.
Fate determination of the brain vesicle
Heterotopic transplantation studies of the brain vesicles between quail and chick embryos were carried out to examine whether or not the fate of the brain vesicles are determined from the beginning (Alvarado-Mallart & Sotelo 1984; Nakamura et al. 1986, 1988, 1991; Nakamura 1990; Nakamura & Itasaki 1992). Quail and chick cells are easily distinguished following a Feulgen-Rossenbeck reaction because heterochromatin is condensed and attached to the nucleolus in the quail cells, while in the chick cells heterochromatin is dispersed in the nucleus (Le Douarin 1969, 1973). The mesencephalon and metencephalon kept their original fate at the ectopic place (Alvarado-Mallart & Sotelo 1984; Nakamura 1990; Nakamura et al. 1986, 1988). The alar plate of the mesencephalon differentiated into the tectum in the prosencephalon and metencephalon. The alar plate of the metencephalon differentiated into the cerebellum in the prosencephalon and in the mesencephalon (Martinez et al. 1991; Bally-Cuif et al. 1992; Bally-Cuif & Wassef 1994). It was shown that the ectopically differentiated optic tectum could receive retinal fibers (Alvarado-Mallart & Sotelo 1984). Moreover, although the fate of the mesencephalon is determined around stage 10, its rostro-caudal polarity is not determined. Thus, the ectopic tectum differentiated in the diencephalic region was used for the study of axis formation of the tectum (Itasaki et al. 1991; Itasaki & Nakamura 1992, 1996).
It was a surprise that the alar plate of the diencephalon changed its fate, and differentiated to the tectum when transplanted into the mesencephalon around stage 10 (Nakamura et al. 1986, 1988, 1991). The diencephalon changed its fate to the tectum when it was transplanted to the posterior part of the mesencephalon and integrated into the host (Nakamura et al. 1986; Nakamura & Itasaki 1992). Fate change did not occur when the diencephalon was transplanted to the anterior part of the tectum, even when the transplant was integrated into the host (Nakamura & Itasaki 1992). When the diencephalic alar plate was transplanted to the posterior part of the mesencephalon, the transplant expressed En2 and differentiated to the tectum, which was shown to receive retinal fibers (Nakamura et al. 1991; Nakamura & Itasaki 1992). More surprisingly, the isthmus induced En2 expression in the host diencephalon when the isthmus was transplanted ectopically to the diencephalon, and the tectum was induced near the transplant (Martinez et al. 1991; Bally-Cuif et al. 1992; Bally-Cuif & Wassef 1994). Furthermore, the isthmus induced an ectopic cerebellum when transplanted to the hindbrain (Martínez et al. 1995). From these studies, it was suggested that the isthmus functions as an organizer for the tectum and cerebellum.
What is the organizing molecule?
Following these results, the identity of the organizing molecules was pursued. Crossley et al. (1996) paid attention to Fgf8 because of its expression in the isthmus. They implanted a Fgf8-soaked bead into the diencephalon, and showed that Fgf8 could mimic the isthmic activity. Fgf8 induced En and Wnt1 expression in the presumptive diencephalon, which was transformed into the tectum. The results suggested that Fgf8 is the main isthmus organizing molecule. Subsequent gain-of-function studies of Fgf8 in chick and mice have all suggested that Fgf8 is a signaling molecule emanating from the isthmus (Crossley et al. 1996; Liu et al. 1999; Martinez et al. 1999; Shamim et al. 1999). Fgf8 mutant zebrafish and mice showed deletion of midbrain and cerebellum (Brand et al. 1996; Meyers et al. 1998; Reifers et al. 1998).
In the mesencephalon and metencephalon, Fgf17 and Fgf18 are also expressed (Maruoka et al. 1998; Ohuchi et al. 2000; Xu et al. 2000). Their expression covers wider area and expression commences later than that of Fgf8. Studies of gain and loss of function of these molecules suggest that Fgf8, 17 and 18 coordinately function as organizing molecules, although Fgf8 is likely to play more crucial role (Sato et al. 2004).
Wnt1 is expressed in the most caudal part of the mesencephalon abutting the expression of Fgf8 (McMahon & Bradley 1990), and is one of candidate molecules for the isthmus organizing activity. Wnt1 knock out mice show defects in the midbrain and hindbrain (McMahon & Bradley 1990; Thomas & Capecchi 1990; McMahon et al. 1992). But misexpression of Wnt1 in chick embryos exerted weak effects: increase in uptake of BrdU, and weak induction of Fgf8 and En2 (Sugiyama et al. 1998; Matsunaga et al. 2002; Canning et al. 2007). Limx1b is expressed in the mesencephalon overlapping with Wnt1, and can induce Wnt1 expression (Adams et al. 2000; Matsunaga et al. 2002). Limx1b repressed Fgf8 expression cell autonomously, but induced Fgf8 non cell autonomously via Wnt1. Thus, it was suggested that interaction among Fgf8, Limx1b and Wnt1 contributes to patterning the Fgf8 expression (Fig. 1, Matsunaga et al. 2002).
Fgf8 as an organizing molecule
We have been analyzing the isthmus organizer, focusing on Fgf8. Among several splice isoforms of Fgf8 (Crossley & Martin 1995; MacArthur et al. 1995a), Fgf8a and Fgf8b were shown by reverse transcription–polymerase chain reaction (RT–PCR) to be expressed in the isthmic region (Sato et al. 2001). The difference between Fgf8a and Fgf8b is very subtle: Fgf8b contains 33 additional base pairs (Crossley & Martin 1995; MacArthur et al. 1995a). But they exert very different effects on mesencephalon–metencephalon development.
Misexpression of Fgf8a by in ovo electroporation (Muramatsu et al. 1997; Funahashi et al. 1999; Nakamura et al. 2000, 2004) did not affect expression of most of the mesencephalon–metencephalon-related genes such as Irx2, Pax2, Otx2 and Gbx2, but induced En1 and En2 in the diencephalons, that is, En1 and En2 expression extended anteriorly to the diencephalic region, where Pax6 expression was repressed. As indicated by Matsunaga et al. (2000), that diencephalon–mesencephalon boundary is settled by repressive interaction between Pax6 and En1, the diencephalic region where En expression was induced after Fgf8a misexpression changed the fate to the mesencephalon. The alar plate differentiated to the tectum, and the oculomotor nucleus, from which additional oculomotor nerve trunks came out, differentiated in the basal plate. Misexpression in chick by electroporation, and in mice under Wnt1 regulation, resulted in similar effects (Lee et al. 1997; Sato et al. 2001).
Misexpression of Fgf8b in chick embryos by in ovo electroporation exerted drastic effects. Presumptive mesencephalon changed its fate, and acquired the characteristics of the metencephalon. Alar plate differentiated as the cerebellum: external granular layer, and Purkinje cells were differentiated (Fig. 2). The oculomotor nucleus disappeared, and its nerve trunk was not formed. These results indicated that the presumptive mesencephalon was completely changed to the metencephalon (Sato et al. 2001). Fgf8b repressed Otx2 expression, and Gbx2 and Irx2 expression extended rostrally up to the presumptive diencephalic region. Thus the region where Gbx2 and Irx2 expression is induced may have changed its fate to differentiate into the cerebellum. En2 was so sensitive to Fgf8, and was induced in the diencephalon by either Fgf8a or Fgf8b, which may explain the transformation of the fate of the presumptive diencephalon to the mesencephalon. Misexpression of Fgf8b in transgeneic mice in which Fgf8b is misexpressed under Wnt1 regulation exerted similar effects on downstream gene expression (Liu et al. 1999).
Although Fgf8a and Fgf8b exert different effects, quantitative experiments indicated that the type difference could be attributable to the difference in the strength of the signal (Sato et al. 2001). Electroporation with a 1/100 concentration of a Fgf8b expression vector exerted Fgf8a-type effects. It was also shown in vitro that Fgf8b has stronger transformation activity (MacArthur et al. 1995b). Structural analysis by surface plasmon resonance showed that Fgf8b binds more intimately to the c isoforms of the Fgf receptor 1-3 (FGFR1-3) according to the additional 11 amino acids in Fgf8b, phenylalanine 32 (F32) being most important (Olsen et al. 2006). Mutation of F32 to alanine (F32A) reduced the affinity to FGFR, and the mutation functionally converted Fgf8b to Fgf8a; misexpression of this mutant Fgf8F32A exerted similar effects to Fgf8a (Olsen et al. 2006). These results, together with that of the Fgf8b-bead implantation in the diencephalon experiements, in which cerebellar structure differentiated in the center and the tectum in the periphery (Martinez et al. 1999), suggest that the region which is exposed to strong Fgf8 signal may differentiate into the cerebellum, and the region where Fgf8 signal is weak may differentiate into the tectum.
Transduction of Fgf8 signal for mes/metencephalon development
Next, one asks the question of how the Fgf8 signal is transduced to organize cerebellar differentiation. Fgf signal is received by a receptor tyrosine kinase, the signal from which is mainly transduced through the Ras-ERK signaling pathway (Fig. 2; reviewed by Katz & McCormick 1997; Rommel & Hafen 1998). Indeed, ERK is activated around the isthmus where Fgf8 is expressed, which was shown immunohistochemically with an anti-diphosphorylated ERK (dpERK) antibody (Sato & Nakamura 2004). So, disrupting Ras-ERK signaling pathway by misexpression of a dominant negative form of Ras (DN-Ras) was carried out. Misexpression of DN-Ras changed the property of the metencephalon to that of the mesencephalon, Gbx2 expression in the metencephalon was repressed, and Otx2 expression was induced in the metencephalon. Finally a tectum differentiated in place of the cerebellum (Fig. 2; Sato & Nakamura 2004). It was suggested that Irx2 is phosphorylated by ERK, and be involved in cerebellar differentiation (Matsumoto et al. 2004).
In conclusion, strong Fgf8 signal activates Ras-ERK signaling pathway to organize cerebellar differentiation (Fig. 3).
Negative regulators of the Ras-ERK pathway
Fgf8 induces negative regulators of Ras-ERK signaling pathway such as Sprouty2, Sef and MKP3 in the isthmus (Fig. 2; Hacohen et al. 1998; Casci et al. 1999; Kramer et al. 1999; Chambers & Mason 2000; Chambers et al. 2000; Zhang et al. 2001; Fürthauer et al. 2002; Lin et al. 2002; Tsang et al. 2002; Kawakami et al. 2003; Echevarria et al. 2005; Smith et al. 2005; Suzuki-Hirano et al. 2005); reviewed in (Mason et al. 2006). Expression of Sprouty2 is always overlapping to Fgf8 in the isthmus and in the anterior neural ridge (Casci et al. 1999; Kramer et al. 1999; Chambers & Mason 2000; Chambers et al. 2000; Echevarria et al. 2005; Suzuki-Hirano et al. 2005). Misexpression of Fgf8 rapidly induced Sprouty2, which in turn downregulated ERK phosphorylation.
Misexpression of Sprouty2 repressed Gbx2 expression and induced Otx2 expression in the metencephalon, and the tectum differentiated in place of the cerebellum (Suzuki-Hirano et al. 2005). On the other hand, misexpression of dominant negative form of Sprouty2 (DN-Sprouty2) kept ERK activated around the isthmus. After misexpression of DN-Sprouty2, the posterior limit of the tectum shifted anteriorly (Suzuki-Hirano et al. 2005). The results indicate that Ras-ERK signaling should be regulated strictly for the proper mesencephalic and metencephalic regionalization. If the signal flows too much, the mes–metencephalic boundary shifts anteriorly, and if the signal flows less, the mes–metencephalic boundary shifts posteriorly. In another words, when the signal flows too much, fate change of the mesencephalon to the metencephalon occurs, and vice versa.
Regional identity of the brain vesicles is determined by the combination of transcription factors expressed. Regionalization proceeds gradually, and earlier expression of transcription factors prepatterns the region. As for the mes–metencephalic region, Otx2 is expressed rostral to the isthmus, and caudal to it Gbx2 is expressed from a very early stage of development (Acampora et al. 1995; Matsuo et al. 1995; Ang et al. 1996). Their expression is overlapping at first, but becomes abutting by their mutual repressive interaction at the isthmus (Fig. 1). At the boundary of their expression, Fgf8 expression is induced overlapping to Gbx2 expression (Katahira et al. 2000). The mesencephalon and metencephalon are already prepatterned by Otx2 and Gbx2, and strong Fgf8 signal activates the Ras-ERK signaling pathway to differentiate cerebellum. The posterior border of the mesencephalon may receive strong Fgf8 signal but it differentiates into the tectum because of Otx2 expression, which raises the threshold of sensitivity to Fgf8 signal toward cerebellar differentiation (Sato et al. 2001).
Transcription factors for tectal development have been well studied; the region of Otx2, En1 and Pax2 expression may acquire the mesencephalic property (Fig. 1, Nakamura 2001a,b), and the addition of Pax3/7 expression may lead to tectal differentiation (Fig. 1, Matsunaga et al. 2001). The optic tectum and cerebellum differentiate responding to the isthmic organizing signal. The signal transduction pathway for cerebellar differentiation has been well elucidated as discussed in this review (Fig. 3), but the signal transduction pathway for the tectum development is not well elucidated. Further study to unravel the transcription factors that are activated downstream of ERK to organize cerebellar differentiation is necessary.
Conflict of Interest
No conflict of interest has been declared by H. Nakamura, T. Sato or A. Suzuki-Hirano.