Glutamatergic vs. GABAergic neuron subtypes
Although there are thousands of types of neurons in the central nervous system, they can largely be classified into two groups; excitatory and inhibitory neurons, which are mainly represented by glutamatergic and GABAergic neurons, respectively. In the rodent telencephalon, glutamatergic neurons are produced from the ventricular zone (VZ) of the dorsal telencephalon and migrate radially, while GABAergic neurons emerge from the ganglionic eminence in the ventral telencephalon and migrate tangentially (1). It has been suggested that several bHLH transcription factors participate in glutamatergic vs. GABAergic neuronal subtype specification in the telencephalic neuroepithelium; Neurogenins 1/2 are expressed in the dorsal telencephalon and involved in glutamatergic neuron generation, while MASH1 is expressed in the ventral telencephalon and participates in producing GABAergic neurons (2-4). These bHLH factors are expressed in neuroepithelial cells and not in mature neurons, implying that glutamatergic/GABAergic neuronal subtypes are determined prior to differentiation. However, the actual molecular machinery governing glutamatergic vs. GABAergic neuronal subtype specification seems to be more complex, as suggested by previous reports (5-7).
The cerebellum consists of three parts; cerebellar cortex, cerebellar white matter and deep cerebellar nuclei (DCN). The major types of neurons found in the cerebellar cortex are Purkinje, Golgi, basket, stellate and granule cells, while the DCN contains large and small DCN neurons (8). Granule cells and large DCN neurons are glutamatergic, while the other neurons are all GABAergic. These neurons regulate each other's activity to achieve proper cerebellar function. During cerebellar development, it has long been believed that only granule cells are generated from the rhombic lip, while the remaining neurons are produced from the cerebellar VZ (9). Although the genes and/or proteins involved in granule cell generation from the rhombic lip were described (10, 11), the molecular mechanisms underlying neuronal production and/or subtype specification of the other neuronal subtypes had not been clarified until 2005.
Our recent study on a novel cerebellar mutant and its responsible gene revealed a molecular mechanism that specifies the GABAergic lineage within the cerebellar neuroepithelium (12). Moreover, two papers, which were published just after our paper, on the lineage analyses of cerebellar rhombic lip-derived cells conflict in part with the classic model (13, 14). Consequently, these three reports present new insights into cerebellar development(15). Here, I will introduce the role of Pancreatic transcription factor 1a (Ptf1a) in producing cerebellar GABAergic neurons and propose a new model regarding the patterning or regionalization processes during cerebellar development.
cerebelless, a novel mouse mutant which lacks the cerebellar cortex
During the generation of some transgenic mice, we obtained a mutant line in which homozygotes exhibited uncoordinated locomotion, ataxic gait and tremors. Surprisingly, they were found to lack the entire cerebellar cortex at adult stages (Fig. 1A-D) and therefore they were named "cerebelless". Despite the absence of the entire cerebellar cortex, these mice survive as long as wild type mice (about 2 years). Although several lines of mutant mice lacking the entire cerebellar cortex have been reported, most of them die during development. cerebelless is the first mutant known to survive and live a normal life span.
Ptf1a is involved in generating cerebellar GABAergic neurons
Fluorescence in situ hybridization revealed that the transgene was inserted in a locus on the second chromosome, around regions 2A3-2B. The genetic linkage study localized the transgene between two markers that were 4.5 Mb apart. Because the process used for generation of transgenic mice is known to sometimes cause a deletion of genomic sequences near the insertion site (25), we searched for possible genomic deletions, and identified a 313 kb deletion in the cerebelless genome.
According to the NCBI mouse genome database, the Ptf1a gene is located 60 kb from one end of the deletion. Ptf1a encodes a bHLH transcription factor which was originally reported as pancreatic determiner that drives undifferentiated cells in the foregut endoderm to differentiate into a pancreatic lineage. Targeted disruption of this gene resulted in pancreatic malformation and dysfunction as well as perinatal death (26, 27). Because transheterozygotes of cerebelless and a Ptf1a null allele (Ptf1acre) (28) share the abnormal cerebellar morphology of cerebelless, we concluded that Ptf1a was likely the responsible gene for cerebelless.
In cerebelless mutants, Ptf1a expressionwas absent in the cerebellum but maintained in the pancreas during development. The mutant mice possess a normal pancreas, which may underlie the viability of these animals. This suggests that a cerebellum-specific enhancer element of this gene is defective in the cerebelless allele (Ptf1acbll), while the pancreas-specific element is intact. in situ hybridization revealed that Ptf1a is expressed in the cerebellar neuroepithelial cells in the VZ during development. In contrast, Ptf1a expression was not observed in neuroepithelial cells of the rhombic lip nor postmitotic cerebellar neurons.
Because the Ptf1acre allele was generated by the replacement of the Ptf1a protein-encoding genomic region with a Cre recombinase-encoding cDNA (28), lineage tracing can be done by crossing these mice with Gt(ROSA)26Sortm1sor (R26R) mice which carry a modified lacZ gene driven by the cell type-independent ROSA26 promoter (29). Lineage tracing in the heterozygous background (Ptf1acre/+; R26R) revealed that all types of cerebellar GABAergic neurons, including Purkinje, Golgi, basket, stellate cells and small DCN neurons, are generated from Ptf1a-expressing neuroepithelial cells in the cerebellar VZ, while glutamatargic neurons such as granule cells and large DCN neurons are derived from Ptf1a-nonexpressing cells. Since the loss of Ptf1a expression in cerebelless results in inhibition of GABAergic neuron production, Ptf1a likely plays an essential role in generating cerebellar GABAergic neurons from the neuroepithelial cells in the VZ.
Furthermore, lineage tracing analyses in the mutant background (Ptf1acre/cbll; R26R) revealed that neuroepithelial cells in the cerebellar VZ produce neurons even in the absence of Ptf1a expression, however the resultant neurons cannot differentiate into GABAergic neurons. This seems to indicate that Ptf1a is not a "proneural gene" that is involved in generating neurons from neuroepithelium, although the possibility exists that other proneural genes may compensate for the proneural function of Ptf1a in cerebelless, resembling the case in neurogenin 2-deficient mice (3).
Two bHLH factors, PTF1A and Math1 may be involved in regionalization of the cerebellar neuroepithelium
The Math1 gene, which encodes a bHLH transcription factor, is known to be expressed in neuropithelial cells of the rhombic lip as well as proliferating granule cell precursors in the EGL. About a decade ago, it was shown that targeted disruption of this gene resulted in complete loss of the granule cell lineage, suggesting an essential role of Math1 in cerebellar granule cell generation (10). Last October, Zoghbi's and Fishell's groups revealed a molecular fate map of the derivatives of Math1-expressing neuroepithelial cells in the cerebellar rhombic lip (13, 14). They showed that not only granule cells but also, at least in part, DCN neurons are derived from the rhombic lip, although they did not discriminate between glutamatergic and GABAergic DCN neurons. In their studies, development of rhombic lip-derived DCN neurons was shown to be disrupted in the Math1 mutants.
As described above, we reported that GABAergic but not glutamatergic DCN neurons are derived from Ptf1a-expressing neuroepithelial cells in the VZ (12). Together these findings suggest that GABAergic and glutamatergic DCN neurons originate from distinct neuroepithelial regions, the cerebellar VZ and the rhombic lip, respectively, and therefore imply that all cerebellar GABAergic neurons are generated from the VZ and all cerebellar glutamatergic neurons from the rhombic lip. Furthermore, normal development of these neurons requires the activity of Ptf1a or Math1.
These facts suggest a model in which the cerebellar neuroepithelium is regionalized into two distinct regions, the VZ and the rhombic lip, by the two bHLH transcription factors, PTF1A and Math1 (Figure 5). During embryonic development, the ventral part of the cerebellar neuroepithelium expresses PTF1A, leading to the acquirement of cerebellar VZ characteristics including the generation of GABAergic neurons, while the dorsal part of cerebellar neuroepithelium expresses Math1 and becomes the cerebellar rhombic lip that produces glutamatergic neurons. In the telencephalon, a similar regionalization by bHLH transcription factors occurs; the glutamatergic neurons are produced by dorsal neuroepithelium expressing Neurogenins 1/2 and GABAergic neurons are generated by ventral neuroepithelium expressing Mash1 (1).
We ectopically introduced Ptf1a into the dorsal telencephalic neuroepithelium (12) by means of in utero electroporation (30, 31). This experiment showed that neurons produced from Ptf1a-introduced dorsal neuroepithelium were GABA-positive and exhibited a morphology similar to that of GABAergic neurons in the cerebral cortex. Furthermore, time-lapse microscopy of slice cultures showed that these neurons tended to migrate tangentially, resembling the GABAergic neurons in the cerebral cortex. These findings suggest that Ptf1a confers GABAergic characteristics on the progeny neurons, in terms of not only the neurotransmitter subtype but also their morphology and migratory behavior. Therefore Ptf1a appears to act as a neuronal subtype determiner that instructs neuroepithelial cells to differentiate into GABAergic neurons, rather than just as a simple activator of GABA synthesis, such as GAD65 and GAD67. Moreover, these results can be interpreted as that the ectopic expression of Ptf1a converts the regional identity of the dorsal neuroepithelium to that of ventral neuroepithelium.
The studies on the cerebelless mutant and its responsible gene have revealed that a bHLH transcription factor, PTF1A plays an essential role in the generation of cerebellar GABAergic neurons. Furthermore, molecular fate map analyses of Ptf1a and Math1 have suggested a new model for the regionalization of cerebellar neuroepithelium. The strategy for glutamatergic vs. GABAergic neuronal subtype specification might be similar between the telencephalon and cerebellum in terms of bHLH transcription factor involvement in the specification of the neuroepithelial identities. Recently, it was shown that Ptf1a is also required for GABAergic neuron development in the dorsal horn of the spinal cord(32). However, Ptf1a expression was not found in mitotic neuroepithelial cells but rather in postmitotic cells in the spinal cord. Therefore, Ptf1a may not be involved in neuroepithelial regionalization in the spinal cord.
Although the molecular machinery governing GABAergic neuron specification in the cerebellar VZ has been clarified as described above, molecular mechanisms to specify each member of GABAergic subtype (e.g. Purkinje, Golgi, basket, stellate cells and small DCN neurons) remain unclear. Many transcription factors, including bHLH proteins, may be involved in those events in neuroepithelial cells and/or immature neurons in the developing cerebellum. Accumulation of such knowledge will make the cerebellum an excellent model system to study neuronal subtype specification, as it contains limited number of well characterized neurons.
1. Wilson SW, Rubenstein JL. Neuron 2000;28(3):641-51.
2. Casarosa S, et al. Development 1999;126(3):525-34.
3. Fode C, et al., Genes Dev 2000;14(1):67-80.
4. Ross SE, et al. Neuron 2003;39(1):13-25.
5. Cheng L, et al. Tlx3 and Nat Neurosci 2004;7(5):510-7.
6. Parras CM, et al. Genes Dev 2002;16(3):324-38.
7. Schuurmans C, et al. EMBO J 2004;23(14):2892-902.
8. Palay SL, Chan-Palay V. CEREBELLAR CORTEX- CYTOLOGY AND ORGANIZATION, Springer-Verlag. 1974.
9. Wang VY, Zoghbi HY. Nat Rev Neurosci 2001;2(7):484-91.
10. Ben-Arie N, et al. Nature 1997;390(6656):169-72.
11. Aruga J, et al. J Neurosci 1998;18(1):284-93.
12. Hoshino M, et al. Neuron 2005;47(2):201-13.
13. Machold R, Fishell G. Neuron 2005;48(1):17-24.
14. Wang VY, et al. Neuron 2005;48(1):31-43.
15. Wingate R. Neuron 2005;48(1):1-4.
16. Heckroth JA. J Comp Neurol 1994;343(1):173-82.
17. Collins P. GRAY'S ANATOMY 35TH EDITION, ChAPTER 3 EMBRYOLOGY AND DEVELOPMENT, Churchill Livingstone (New York). 1995.
18. Dahmane N, Ruiz-i-Altaba A. Development 1999;126(14):3089-100.
19. Lewis PM, et al. Dev Biol 2004;270(2):393-410.
20. Wallace VA. Curr Biol 1999;9(8):445-8.
21. Wechsler-Reya RJ, Scott MP. Neuron 1999;22(1):103-14.
22. Bennet MR, et al. 50 years on. Auton Neurosci 2002;95(1-2):1-23.
23. Chu T, et al. J Neurobiol 2000;43(1):18-30.
24. Herrup K, et al, Brain Res Dev Brain Res 1996;96(1-2):28-35.
25. Kuro-o M, et al. Nature 1997;390(6655):45-51.
26. Kawauchi T, et al. Biochem. Biophys. Res. Comm. 2005;331:50-55.
27. Krapp A, et al. Genes Dev 1998;12(23):3752-63.
28. Kawaguchi Y, et al. Nat Genet 2002;32(1):128-34.
29. Soriano P. Nat Genet 1999;21(1):70-1.
30. Kawauchi T, et al. EMBO J 2003;22(16):4190-201.
31. Nikolic M. Trends Cell Biol 2004;14(1):1-5.
32. Glasgow SM, et al. Development 2005;132(24):5461-9.
33. Sellick GS, et al. Nat Genet 2004;36(12):1301-1305.