31 October 2002 Nature 419, 929 - 934 (2002); doi:10.1038/nature01124 Progenitor cell maintenance requires numb and numblike during mouse neurogenesis PETUR H. PETERSEN*†, KAIYONG ZOU*†, JOSEPH K. HWANG*, YUH NUNG JAN&# 8225; & WEIMIN ZHONG* * Department of Molecular, Cellular and Developmental Biology, Yale University, New Haven, Connecticut 06520, USA ‡ Howard Hughes Medical Institute and Departments of Physiology and Bioche mistry, University of California, San Francisco, California 94143, USA † These authors contributed equally to this work Correspondence and requests for materials should be addressed to W.Z. (e-mail: w eimin.zhong@yale.edu). Neurons in most regions of the mammalian nervous system are generated over an ex tended period of time during development. Maintaining sufficient numbers of prog enitors over the course of neurogenesis is essential to ensure that neural cells are produced in correct numbers and diverse types1-3. The underlying molecular mechanisms, like those governing stem-cell self-renewal in general, remain poorl y understood. We report here that mouse numb and numblike (Nbl)4-6, two highly c onserved homologues of Drosophila numb7, 8, play redundant but critical roles in maintaining neural progenitor cells during embryogenesis, by allowing their pro genies to choose progenitor over neuronal fates. In Nbl mutant embryos also cond itionally mutant for mouse numb in the nervous system, early neurons emerge in t he expected spatial and temporal pattern, but at the expense of progenitor cells , leading to a nearly complete depletion of dividing cells shortly after the ons et of neurogenesis. Our findings show that a shared molecular mechanism, with mo use Numb and Nbl as key components, governs the self-renewal of all neural proge nitor cells, regardless of their lineage or regional identities. In Drosophila, Numb is a membrane-associated signalling protein that allows two daughter cells to adopt different fates after an asymmetric division. It does th is by localizing to only one half of the cell membrane in dividing precursor cel ls, so that it is segregated primarily to one cell7, 8. On the basis of studies in the developing mouse neocortex, we have postulated that mouse Numb segregates to, and promotes the fate of, progenitor cells in asymmetric divisions that gen erate a neuron and a daughter progenitor cell during mammalian neurogenesis4. Th is view, however, is controversial; others have postulated instead that vertebra te Numb proteins promote the neuronal fate in such divisions6, 9, 10. numb mutan t mice exhibit severe defects in cranial neural tube closure and die around embr yonic day (E) 11.5, but neurogenesis abnormalities are limited10, 11 and insuffi cient to resolve the controversy. From studies to be described elsewhere, we gen erated Nbl homozygous mutant mice, which are viable, fertile and exhibit no obvi ous phenotypes. We find low levels of Nbl expression in E8.5 embryos, including in neural progenitor cells. Moreover, embryos mutant for both mouse numb and Nbl die around E9.5 with more widespread defects than single mutants. We therefore used Cre-loxP mediated gene targeting12 to examine their function i n neurogenesis. We generated a transgenic line (NesCre8) with Cre expression con trolled by a nestin promoter13 that is active in neural progenitor cells and som ites. Within the nervous system, Cre-mediated recombination14 is readily detecta ble in a majority of the progenitor cells at E8.5 (n = 7/9) and becomes nearly c omplete by E12.5 (Fig. 1a, and data not shown). Conditional knockout (cKO) using NesCre8 and a floxed mouse numb allele11 did not cause embryonic lethality or d efects in neural tube closure. In fact, cKO mice are viable, fertile and indisti nguishable from their wild-type littermates (data not shown). As expected, immun oblots show mouse Numb protein level is already greatly reduced in E9.5 conditio nal mutant embryos (Fig. 1b). Most of the residual mouse Numb protein probably c omes from tissues where Cre is not active. Figure 1 Conditional double knockout of mouse numb and Nbl. Full legend High resolution image and legend (172k) Mouse numb cKO in the Nbl homozygous mutant background (conditional double-knock out, or cDKO), on the other hand, results in embryonic lethality. We never recov ered cDKO mice postnatally, whereas those with other allelic combinations, in pa rticular cKO in Nbl heterozygous background, are viable and exhibit no gross mor phological or behavioural defects. cDKO embryos are indistinguishable from their littermates at E9.5 (n = 7), but become completely necrotic by E12.5 (n = 4). T hose recovered at E11.5 are considerably smaller than the littermates (n = 8), s uggesting that cDKO embryos die around this stage. At E10.5, cDKO embryos were consistently recovered (33/183, 1/4 or 1/8 expected) . They are 80–90% the size of their wild-type or single-mutant littermates, although many (21/33) are within the range of variation seen in wild-type litters. cDKO embryos are morphologically appropriate for their age with the expected somite numbers, but have significantly reduced telencephalic vesicle (Fig. 1c, right panel) and undulating spinal cord (Fig. 1e, right), the combination of which can be reliably used to identify them. Histological analysis reveals that E10.5 cDKO embryos have severe thinning of the neural tube, from the most rostral telencephalon to caudal spinal cord, including optic discs (Fig. 1d?Cf, right). In 82% of the mutants, the neuroepithelium is only one-quarter to on e-half the thickness of that in the littermates, with frequent buckling of the s urface. E10.5 wild-type neuroepithelium consists mainly of progenitor cells that make up the wider, inner ventricular zone, with a much smaller number of neurons formin g the outer mantle zone. We could detect neurons in E10.5 cDKO embryos, in a reg ion-specific pattern similar to that in control littermates, wild-type or other allelic combinations, using two general neuronal markers, anti-HuC/D (Hu; Fig. 2 a'–c', in red) or Neurofilament (NF; Fig. 2d'–g', in green) (n = 10). Dll1, a marker for newborn, migrating neurons in the ventricular zone15, 16, is also exp ressed throughout the cDKO nervous system (n = 5) (Fig. 2h'–l'), including the forebrain, which has few Hu- or NF-positive neurons (Fig. 2a', h'). In fact, the cDKO neuroepithelium frequently contains large patches of Dll1-positive cells (Fig. 2i', l'), unlike in control embryos where they are invariably discrete (Fig. 2i, l). Figure 2 Effect of mouse numb and Nbl conditional double mutation on neurogenes is. Full legend High resolution image and legend (155k) To assess the severity of the neural progenitor cell loss in E10.5 cDKO embryos, we first used an antibody against phospho-Histone H3 (P-H3) to identify mitotic cells. In wild-type embryos, neural progenitor cells within the neural tube und ergo S phase (DNA synthesis) when their nuclei are in the outer half of the vent ricular zone. The nuclei then translocate towards the ventricular surface where cells undergo mitosis. Accordingly, P-H3-positive cells form a near continuous o utline of the ventricular surface (Fig. 2a–c). In the cDKO nervous system, however, there is a dramatic reduction of P-H3-positive cells (Fig. 2a'?Cc'). We quantified the loss in the forebrain, the hindbrain (at the level of ot ic vesicle), and the spinal cord (at cervical levels), which ranges from about 8 0% to near 100% (n = 5). We next performed bromodeoxyuridine (BrdU) labelling experiments. BrdU is incorp orated into DNA by S-phase cells and, therefore, the number of cells labelled du ring a short pulse reflects the number of cells still proliferating but not in m itosis. There is little variation among control littermates in the pattern and p ercentage of neural progenitor cells labelled by BrdU (Fig. 2d–g). In forebrain, hindbrain and cervical spinal cord sections, BrdU-labelled cells account for about 49, 33 and 40% of the ventricular zone cells, respectively (n = 3; 1-h pulse). In contrast, only a few BrdU-positive cells are present in the cDKO nervous system, indicating a loss of over 99% of the S-phase cells (n = 14) (Fig. 2d'?Cg'). To confirm that the near-absence of proliferating cells in E10.5 cDKO embryos re flects an absence of neural progenitor cells, rather than their becoming quiesce nt or defective in cell-cycle progression, we examined the expression of Nestin protein, a commonly used neural progenitor cell marker. Nestin-positive cells ar e reduced in numbers, but not completely absent, in the E10.5 cDKO neural tube ( data not shown). As Nestin is transiently detectable in newborn neurons, probabl y inherited from progenitor cells, we would expect a more severe reduction at E1 1.5. This is indeed the case (n = 3) (Fig. 2m', n'). More importantly, loss of N estin expression is specific to the nervous system; expression in somites in the same cDKO embryos is not affected (Fig. 2n', arrow). Because the results with Nestin at E10.5 were inconclusive, we performed in situ hybridization using Hes5, which marks neural progenitor cells and is essential for keeping them undifferentiated17. As expected, there is a sharp reduction of Hes5 expression in the E10.5 cDKO neuroepithelium (n = 6). In the hindbrain, for example, Hes5-expressing cells normally occupy two-thirds the width of the neur oepithelium (Fig. 2o). In the cDKO hindbrain, however, there is only a thin laye r of scattered Hes5-expressing cells near the ventricular surface (Fig. 2o'). Si milar loss can be observed in the ventral half of the spinal cord, where Hes5 is highly expressed in wild-type embryos (Fig. 2p, p'). We point out that the near-complete absence of neural progenitor cells, as descr ibed above, is based on the analysis of, and invariably observed in, embryos wit h a neural tube that is less than half the normal thickness, a group representin g 82% (27/33) of the cDKO embryos we recovered at E10.5. Furthermore, BrdU-label led cells in many non-neural tissues are also reduced in numbers. Whether this i s a secondary effect or shows a direct requirement for mouse numb and Nbl in non -neural tissues remains to be determined. What causes the near-absence of neural progenitor cells in E10.5 cDKO embryos? A t E9.5, such embryos show no significant defects in cell proliferation (Fig. 4c' and data not shown; n = 4). Consistently, the overall neural tube thickness in E9.5 and E10 cDKO embryos is comparable to that in control littermates (n = 7 ea ch), although the mutant neuroepithelium, particularly the ventral spinal cord, is sometimes punctuated by thinner regions. At E10, however, S-phase cells in th e nervous system are reduced in numbers (Fig. 3a, b, bottom, and Fig. 4h'), part icularly in regions with more active neurogenesis, indicating that the absence o f progenitor cells at E10.5 is probably due to defects in self-renewal rather th an smaller or defective founding populations. There is, however, significant var iation in the severity of such loss among cDKO embryos (n = 5), probably due to the perdurance effect of the mouse Numb protein and variations in the onset of C re expression. Consequently, E10.25 cDKO embryos show varying degrees of neural tube thinning and reduction of BrdU-labelled cells, ranging from a near-complete loss (data not shown) to about 60% (Fig. 3c, f, bottom and inset). Figure 4 Motor neuron generation at the expense of progenitor cells. Full leg end High resolution image and legend (145k) Figure 3 Effect of mouse numb and Nbl conditional double mutation on neurogenes is before neural tube thinning. Full legend High resolution image and legend (62k) Between E10 and E10.5, mostly neurons are being generated in wild-type embryos. cDKO embryos show no ectopic expression of glia markers like Sox10 or Olig218-20 , indicating an absence of premature gliogenesis (data not shown). We therefore examined neuron production in E10 and E10.25 cDKO embryos to ascertain whether t he inability of neural progenitor cells to self-renew results from their progeni es all adopting neuronal fates, which would cause an initial overproduction of n eurons and, more importantly, a significant increase of their percentage within the neuroepithelium owing to a depletion of progenitor cells. Other causes, such as progenitor cells becoming quiescent or defective in cell-cycle progression, or undergoing programmed cell death (apoptosis), should affect neuron production negatively, resulting in fewer neurons, both in absolute numbers and as a perce ntage of the neuroepithelium. We analysed E10.25 cDKO embryos in which neural tube thinning is not yet apparen t (n = 2). Three lines of evidence show unambiguously that a diminished self-ren ewal capability among cDKO neural progenitor cells indeed results from over-diff erentiation of their progenies. First, reductions in the number of BrdU-labelled cells are accompanied by a similar decrease in cells expressing progenitor mark er Hes5 (Fig. 3c, f, bottom). Second, within progenitor domains marked by Hes5, a higher percentage of cells express Dll1, a marker for newborn neurons15, 16 (F ig. 3d, g, bottom; sections are adjacent to those in Fig. 3c, f, bottom, respect ively). Third and most important, there is a significant expansion of, proportio nal to the decrease of progenitor domains, cells expressing neuronal marker Hu ( Fig. 3e, h, bottom, and Fig. 5). In these cDKO embryos, there are patches of Hes 5-positive progenitor cells at the most lateral positions of the neuroepithelium (Fig. 3c, f, bottom). It remains to be determined whether they had migrated abe rrantly, or whether their nuclei were physically prevented from translocating me dially owing to neuron overproduction near the ventricular surface. Figure 5 Neuron overproduction and death in mouse numb and Nbl conditional doub le-mutant embryos. Full legend High resolution image and legend (319k) We analysed in more detail neurogenesis in the ventral spinal cord, where motor neurons emerge shortly after E9 and are continuously generated until E13.5. At E 9.5 (Fig. 4a–c, a'–c'), cDKO and control embryos (n = 4) show little differenc e in Olig2 and Isl1 expression, which marks motor progenitors and motor neurons, respectively21-24. By E10, less than two cell cycles later, there are dramatic differences. At cervical levels in control embryos (Fig. 4d–f), wild-type or other allelic combinations, which are indistinguishable, Isl1-positive neurons colonize the lateral half of the motor domain, whereas Olig2-positive progenitor cells occupy the medial half21-24. On average, only 25.6% (n = 6 sections from three embryos) of the Olig2-expressing cells are doubly positive for Isl1, representing newborn motor neurons. In contrast, Isl1-expressing cells frequently span the entire width of the cDKO neuroepithelium (n = 4/5) (Fig. 4d'?Cf', right). As expected, more Olig2-positive cells in the mutant co-express Isl 1, ranging from 40.6% to as high as 68.7% (Fig. 4f', right). Even in regions whe re motor neurons appear to have only colonized the lateral positions, Olig2-posi tive cells are also absent from the ventricular zone and many, consistent with t heir position, co-express Isl1 (Fig. 4d'–f', left). Neurogenesis in the spinal cord proceeds in a rostral to caudal gradient. More c audally, there are more Olig2 single-positive cells than that at cervical levels in E10 cDKO embryos. Although the overall reduction of BrdU-labelled cells with in the spinal cord is limited, it is much more severe among Olig2-expressing cel ls, which sometimes show no BrdU incorporation at all (Fig. 4g'–h', right). Accordingly, in E10.25 cDKO embryos (n = 2) at similar caudal levels, Isl1-positive motor neurons span the entire width of the neuroepithelium and most of the remaining Olig2-positive cells co-express Isl1 (Fig. 4i', j'). Although neurons are initially overproduced in cDKO embryos, there is significan t increase in apoptosis among mutant neurons (Fig. 5, in red), indicating that m any die shortly after birth. This is consistent with the absence of axon tracts underneath the floor plate (Fig. 2e', g', arrow), and suggests a role for mouse numb and Nbl in later events of neural development, as we have postulated5, alth ough neuronal death may be a secondary effect. It remains possible that cDKO progenitor cells adopt an abnormal developmental p athway and differentiate en masse before the onset of neurogenesis. However, two lines of evidence strongly suggest that they are rapidly depleted as neurogenes is progresses because their daughter cells all adopt neuronal fates instead of s elf-renewing after division. First, in cDKO embryos where BrdU-labelled S-phase cells are significantly reduced in numbers but not totally absent, the number of mitotic cells is comparable to that in control littermates (Fig. 3b, bottom, ar rowheads, Fig. 5, P-H3, in blue). Second, whereas neuron overproduction is obser ved throughout the cDKO nervous system, it is more pronounced in regions where n eurogenesis has been more active. In the control E10.25 forebrain, for example, only a few scattered Hu-positive neurons are present (Fig. 5a, left), indicating that neurogenesis was just underway. Neurons are overproduced in the cDKO foreb rain, but represent only a small fraction of the neuroepithelium (Fig. 5a, right ). On the other hand, ventral spinal cord in control littermates contains large numbers of neurons (Fig. 5e, left). Accordingly, Hu-positive neurons not only ar e overproduced in the cDKO but also span nearly the entire width of the neuroepi thelium (Fig. 5e, right). Similar differences in neuron overproduction can be ob served in other regions (Fig. 5b–d, right) and along the dorsoventral neural axis (Fig. 3h, bottom). The findings reported here are consistent with our earlier hypothesis4, 5, and d emonstrate unequivocally that the main function of numb homologues in mouse neur ogenesis is to maintain progenitor cells, not promoting neuronal fates, during t he initial progenitor versus neuronal fate decision. This is also supported by g ain-of-function studies in chick showing an increase in progenitors among neuroe pithelial cells over-expressing chick Numb9. A near-absence of motor and sensory neurons in mouse numb single mutants at E10.5 due to impaired differentiation h as been reported10. This phenotype is not seen in our single or cDKO mutants and , therefore, is unlikely to be due to a loss of mouse numb function. Our findings are, to our knowledge, the first direct evidence of a pan-neural pr ogram for precursor cells, regardless of their lineage or regional identities, t o choose between proliferation and differentiation. There is growing evidence fr om direct imaging experiments that asymmetric cell division occurs during mammal ian neurogenesis25-27. Mouse Numb is asymmetrically localized to the apical memb rane in dividing progenitor cells4, 5, 28, but how Nbl protein is distributed in these cells is unknown, owing to low levels of expression. Therefore, although an effect on asymmetric division can account for pan-neural progenitor depletion in cDKO embryos (Fig. 5f, left), further studies are necessary to ascertain thi s, in particular the specific effects on multipotential neural stem cells and pr ogenitors with more limited developmental potentials. Similarly, to ascertain if mouse Numb and Nbl act by inhibiting Notch activity like Drosophila Numb, it is necessary to determine first whether Notch signalling plays a generic role in r egulating cell fate choices between proliferation and differentiation—as seen w ith mouse Numb and Nbl—or acts differently in different progenitor populations. Finally, the widespread defects exhibited by mouse numb and Nbl constitutive dou ble mutants at E9.5 raise an interesting possibility that these two genes are in volved in stem-cell self-renewal in other tissues. If mouse Numb and Nbl protein s indeed act like their Drosophila counterpart, they would provide an attractive molecular mechanism to integrate lineage or cell-extrinsic cues for progenies o f various stem cells to choose between self-renewal and adopting appropriate dif ferentiated fates. Methods Generation of mutant mice The Nbl null (-) allele, lacking 90% of the coding reg ion, was generated using homologous recombination in ES cells and will be descri bed elsewhere. The floxed (f) and null (also named 5,6) alleles of mouse numb we re generated previously11. The NesCre8 transgene was constructed by inserting a Cre complementary DNA with a nuclear localization signal and chick -actin poly-a denylation site (a gift from M. Lewandoski) immediately 3' to the nestin transcr iption start site in a vector with 5.8 kilobases (kb) of upstream promoter and 3 .2 kb of downstream sequence, including the first two introns13 (a gift from U. Lendahl). Transgenic mice were produced as previously described29. cDKO embryos were generated by mating numbf/f;Nbl-/- males with numb+/-;Nbl+/-;NesCre8 female s. The latter were either hemizygous or homozygous for the Cre transgene. Histology and embryo staining X-gal staining, immunoblot, immunostaining, BrdU l abelling, and in situ hybridization were as previously described with minor modi fications4, 5, 11, 29. Anti-Actin, BrdU (Sigma), HuC/D (Molecular Probes), Isl1 and Nestin (Developmental Studies Hybridoma Bank) are mouse monoclonal, whereas anti-NF (mr 150K, Chemicon), Olig2 (a gift from H. Takebayashi) and P-H3 (Upstat e Biotechnology) are rabbit polyclonal. Apoptosis was detected using a kit (Roch e). Probes for Dll1 and Hes5 are based on mouse and rat cDNAs, respectively. Received 18 June 2002;accepted 23 August 2002 References 1. Anderson, D. A. Stem cells and pattern formation in the nervous sy stem: the possible versus the actual. Neuron 30, 19-35 (2001) | PubMed | 2. Edlund, T. & Jessell, T. M. Progression from extrinsic to intrinsic signaling in cell fate specification: a view from the nervous system. Cell 96, 211-224 (1 999) | PubMed | 3. McConnell, S. K. Constructing the cerebral cortex: neurogenesis and fate dete rmination. Neuron 15, 761-768 (1995) | PubMed | 4. Zhong, W., Feder, J. N., Jiang, M.-M., Jan, L. Y. & Jan, Y. N. Asymmetric loc alization of a mammalian Numb homolog during mouse cortical neurogenesis. Neuron 17, 43-53 (1996) | PubMed | 5. Zhong, W., Jiang, M.-M., Weinmaster, G., Jan, L. Y. & Jan, Y. N. Differential expression of mammalian Numb, Numblike and Notch1 suggests distinct roles durin g mouse cortical neurogenesis. Development 124, 1887-1897 (1997) | PubMed | 6. Verdi, J. M. et al. Mammalian NUMB is an evolutionarily conserved signaling a dapter protein that specifies cell fate. Curr. Biol. 6, 113-145 (1996) 7. Rhyu, M. S., Jan, L. Y. & Jan, Y. N. Asymmetric distribution of numb protein during division of the sensory organ precursor cell confers distinct fates to da ughter cells. Cell 76, 477-491 (1994) | PubMed | 8. Spana, E. P., Kopczynski, C., Goodman, C. S. & Doe, C. Q. Asymmetric localiza tion of numb autonomously determines sibling neuron identity in the Drosophila C NS. Development 121, 3489-3494 (1995) | PubMed | 9. Wakamatsu, Y., Maynard, T. M., Jones, S. U. & Weston, J. A. NUMB localizes in the basal cortex of mitotic avian neuroepithelial cells and modulates neuronal differentiation by binding to NOTCH-1. Neuron 23, 71-81 (1999) | PubMed | 10. Zilian, O. et al. Multiple roles of mouse numb in tuning developmental cell fates. Curr. Biol. 11, 494-501 (2001) | Article | PubMed | 11. Zhong, W. et al. Mouse numb is an essential gene involved in cortical neurog enesis. Proc. Natl Acad. Sci. USA 97, 6844-6849 (2000) | Article | PubMed | 12. Gu, M., Marth, J. D., Orban, P. C., Mossmann, H. & Rajewsky, K. Deletion of a DNA polymerase gene segment in T cells using cell type-specific gene targetin g. Science 265, 103-106 (1994) | PubMed | 13. Zimmerman, L. et al. Independent regulatory elements in the nestin gene dire ct transgene expression to neural stem cells or muscle precursors. Neuron 12, 11 -24 (1994) | PubMed | 14. Soriano, P. Generalized lacZ expression with the ROSA26 Cre reporter strain. Nature Genet. 21, 70-71 (1999) | Article | PubMed | 15. Dunwoodie, S. L., Henrique, D., Harrison, S. M. & Beddington, R. S. P. Mouse Dll3: a novel divergent Delta gene which may complement the function of other D elta homologues during early pattern formation in the mouse embryo. Development 124, 3065-3076 (1997) | PubMed | 16. Campos, L. S., Duarte, A. J., Branco, T. & Herique, D. mDll1 and mDll3 expre ssion in the developing mouse brain: role in the establishment of the early cort ex. J. Neurosci. Res. 64, 590-598 (2001) | Article | PubMed | 17. Ohtsuka, T. et al. Hes1 and Hes5 as Notch effectors in mammalian neuronal di fferentiation. EMBO J. 18, 2196-2207 (1999) | Article | PubMed | 18. Kuhlbrodt, K., Herbarth, B., Sock, E., Hermans-Borgmeyer, I. & Wegner, M. So x10, a novel transcriptional modulator in glial cells. J. Neurosci. 18, 237-250 (1998) | PubMed | 19. Lu, Q. R. et al. Sonic hedgehog-regulated oligodendrocyte lineage genes enco ding bHLH proteins in the mammalian central nervous system. Neuron 25, 317-329 ( 2000) | PubMed | 20. Zhou, Q., Wang, S. & Anderson, D. J. Identification of a novel family of oli godendrocyte lineage-specific basic helix-loop-helix transcription factors. Neur on 25, 331-343 (2000) | PubMed | 21. Takebayashi, H. et al. Dynamic expression of basic helix-loop-helix Olig fam ily members: implication of Olig2 in neuron and oligodendrocyte differentiation and identification of a new member, Olig3. Mech. Dev. 99, 143-148 (2000) | Artic le | PubMed | 22. Mizuguchi, R. et al. Combinatorial roles of olig2 and neurogenin2 in the coo rdinated induction of pan-neuronal and subtype-specific properties of motoneuron s. Neuron 31, 757-771 (2001) | PubMed | 23. Novitch, B. G., Chen, A. I. & Jessell, T. M. Coordinate regulation of motor neuron subtype identity and pan-neuronal properties by the bHLH repressor Olig2. Neuron 31, 773-789 (2001) | PubMed | 24. Pfaff, S. L., Mendelsohn, M., Stewart, C. L., Edlund, T. & Jessell, T. M. Re quirement for LIM homeobox gene Isl1 in motor neuron generation reveals a motor neuron-dependent step in interneuron differentiation. Cell 84, 309-320 (1996) | PubMed | 25. Chenn, A. & McConnell, S. K. Cleavage orientation and the asymmetric inherit ance of Notch1 immunoreactivity in mammalian neurogenesis. Cell 82, 631-641 (199 5) | PubMed | 26. Qian, X., Goderie, S. K., Shen, Q., Stern, J. H. & Temple, S. Intrinsic prog rams of patterned cell lineages in isolated vertebrate CNS ventricular zone cell s. Development 125, 3143-3152 (1998) | PubMed | 27. Noctor, S. C., Flint, A. C., Weissman, T. A., Dammerman, R. S. & Kriegstein, A. R. Neurons derived from radial glial cells establish radial units in neocort ex. Nature 409, 714-720 (2001) | Article | PubMed | 28. Cayouette, M., Whitmore, A. V., Jeffery, G. & Raff, M. Asymmetric segregatio n of Numb in retinal development and the influence of the pigmented epithelium. J. Neurosci. 21, 5643-5651 (2001) | PubMed | 29. Zhong, W., Mirkovitch, J. & Darnell, J. E. Jr Tissue specific regulation of mouse hepatocyte nuclear factor 4 expression. Mol. Cell. Biol. 14, 7276-7284 (19 94) | PubMed | Acknowledgements. We thank members of the Zhong laboratory for discussions, C. J acobs and H. Keshishian for comments, K. Jakobsdottir for support (P.H.P.), J. Z hang for technical assistance, Developmental Studies Hybridoma Bank, D. Anderson , R. Kageyama, U. Lendahl, M. Lewandoski, M. Nakafuku, H. Takebayashi and M. Weg ner for reagents. K.Z. is supported by a Yale Anderson fellowship. Y.N.J. is a H HMI investigator. This work was supported by a Yale Hellman Fa -- ※ 来源:．生命玄机站 bbs.cst.sh.cn.
