Dual control of neurogenesis by PC3 through cell cycle inhibition and induction of Math1

Abstract

Growing evidence indicates that cell cycle arrest and neurogenesis are highly coordinated and interactive processes, governed by cell cycle genes and neural transcription factors. The gene PC3 (Tis21/BTG2) is expressed in the neuroblast throughout the neural tube and inhibits cell cycle progression at the G1 checkpoint by repressing cyclin D1 transcription. We generated inducible mouse models in which the expression of PC3 was upregulated in neuronal precursors of the neural tube and of the cerebellum. These mice exhibited a marked increase in the production of postmitotic neurons and impairment of cerebellar development. Cerebellar granule precursors of PC3 transgenic mice displayed inhibition of cyclin D1 expression and a strong increase in the expression of Math1, a transcription factor required for their differentiation. Furthermore, PC3, encoded by a recombinant adenovirus, also induced Math1 in postmitotic granule cells in vitro and stimulated the Math1 promoter activity. In contrast, PC3 expression was unaffected in the cerebellar primordium of Math1 null mice, suggesting that PC3 acts upstream to Math1. As a whole, our data suggest that cell cycle exit of cerebellar granule cell precursors and the onset of cerebellar neurogenesis are coordinated by PC3 through transcriptional control of cyclin D1 and Math1, respectively.

Figure 1.

Generation and genomic analyses of transgenic mice with regulated expression of PC3. A, TRE-PC3 transgene contains tet operator sequences (TRE) fused upstream to the minimal CMV promoter (pCMVmin), PC3 ORF, and the late SV40 gene polyadenylation site (pA). B, βACT-tTA transgene includes human β-actin promoter fused to the β-actin IVS1/splice acceptor, human CMV promoter–enhancer, tetR/VP16 transactivator (tTA), and the 3′ UTR of the human β-actin gene with pA. Mechanisms of conditional activation are outlined in Materials and Methods. C, Southern blot analysis of the genomic organization of TRE-PC3 in transgenic mice. XbaI-digested tail DNA of Tg TRE-PC3 mice was hybridized with the indicated probes. Tg TRE-PC3 A and G had clearly unique junctional fragments, indicating a single site of integration. White arrowheads indicate DNA fragments containing the regions flanking the integration site depicted in D (left). The L family element has higher size because it included 0.6 kb of pUHD10-3 vector. D, Genomic organization of the TRE-PC3 transgene. E, Analysis of expression of Tg βACT-tTA in mice lineages, measured by luciferase activity in lysates from indicated organs obtained from 4- to 5-week-old mice carrying the following: no transgene (WT); Tg TRE-Luc (L7 strain); binary Tg CMV-rtTA/TRE-Luc; binary Tg βACT-tTA (strain nACT37)/TRE-Luc; and binary Tg βACT-tTA (strain nACT75)/TRE-Luc. Luciferase units per milligram of extract protein are expressed as mean ± SEM, and the number of animals is indicated. N.D., Not determined.

Figure 2.

Bitransgenic βACT-tTA/TRE-PC3 mice display reduced body and cerebellar size and ataxic gait. A, Normal-sized Tg TRE-PC3 (left) and a binary Tg βACT-tTA/TRE-PC3 of smaller size (right). Both are 30 d old and were never exposed to doxycycline. B, Newborn TRE-PC3 mouse of normal size (left) and bitransgenic Tg βACT-tTA/TRE-PC3 mice of small size (right). C, A representative weight curve of βACT-tTA/TRE-PC3 mice, families A and G; the control is a bitransgenic treated with doxycycline that did not differ from wild type. D, Brain of mice shown in A. The cerebellar vermis is absent in Tg βACT-tTA/TRE-PC3 (right). E, Hematoxylin and eosin staining of cerebella from mice shown in A, revealing a reduced length and size of lobules in bitransgenic compared with single transgenic and a selective reduction of cerebellar size. Scale bars, 750 μm. PKJ, Purkinje cell layer; ML, molecular layer.

Figure 3.

Regulated PC3 transgene expression in Tg βACT-tTA/TRE-PC3 mice and selective targeting to neuroepithelia in Tg nestin-tTA/TRE-PC3. A, B, Semiquantitative RT-PCR of exogenous and endogenous PC3 mRNA was performed on RNA from heads of E13 embryos, obtained by breeding Tg βACT-tTA with Tg TRE-PC3 family A (A) or family G (B). Equal amounts of RT-PCR amplified products were electrophoresed, blotted on filters, and hybridized to [³²P]-labeled probes for transgenic PC3, endogenous PC3, and 18 S mRNA. RT + or – refers to the products of amplification performed in parallel on aliquots of each RNA sample, preincubated or not with RT, as controls for the presence of DNA contamination. C, RT-PCR analysis of PC3 mRNA in various tissues from 40-d-old bitransgenic βACT-tTA/TRE-PC3 mice exposed to doxycycline (250 μg/ml of drinking water) from 1 week before fertilization until birth, to allow only postnatal expression. D, Whole-mount ISH of E12 embryos from Tg βACT-tTA/TRE-PC3 and Tg nestin-tTA/TRE-PC3, without or with doxycycline, respectively, or control. A transgenic PC3-specific antisense riboprobe was used. Te, Telencephalon; Me, mesencephalon; Rh, rhomboencephalon; Sc, spinal cord; DRG, dorsal root ganglia; V, viscera. Arrow indicates the retinal primordium.

Figure 4.

Increase of postmitotic newborn neurons after overexpression of PC3 in the neural tube at E12.5. Transverse sections of the neural tube at cervical level (Kaufman, 1999, corresponding to his plate 27b, sections a,b) from E12.5 embryos of Tg nestin-rtTA/TRE-PC3, either activated by doxycycline since conception or control, were analyzed by ISH for expression of exogenous PC3 (A, F), the indicated neuronal markers (B, C, G, H), BrdU incorporation (D, I), and apoptosis by the TUNEL assay (E, J). Sections were spaced ∼10 μm. K, L, Confocal microscopy of a VZ region of the neural tube defined by a box in B and G, performed in an adjacent section double labeled for PC3/βIII tubulin. Activation of the PC3 transgene (K) is associated with an increase of βIII tubulin⁺ cells, in which most are yellow–green (unlike in the control, L), indicating coexpression of PC3. M, Quantitative analysis of differentiation (percentage ratio of βIII tubulin⁺ neurons to the total number of cells and percentage ratio of βIII tubulin ⁺PC3 ⁺ neurons to the total number of cells), apoptosis (percentage ratio of TUNEL⁺ neurons to the total number of cells), total cell number per area, and total area of the neural tube. The total number of cells was obtained by counting nuclei stained by Hoechst 33258. Values were calculated within the neural tube area from transverse sections localized and spaced as indicated above (3 transversal adjacent sections for each of 3 mice, for a total of 9 sections for either the activated or the control Tg; the fields analyzed covered the entire neural tube area). *p < 0.001 versus control (Student's t test). nt, Neural tube; drg, dorsal root ganglia; cng, cranial nerve ganglia, indicated by a black asterisk; t, tongue. Scale bars: A, F, 1100 μm; D, I, 240 μm.

Figure 5.

Overexpression of PC3 inhibits the proliferation and the cyclin D1 expression of GCPs in cerebellum at P1, in Tg βACT-tTA/TRE-PC3 (A–L) and in Tg nestin-rtTA/TRE-PC3 (M–X). A, G, M, S, Hematoxylin and eosin staining indicates reduced size of lobules. –doxy corresponds to active Tg βACT-tTA/TRE-PC3, whereas +doxy corresponds to active Tg nestin-rtTA/TRE-PC3. Endogenous (end.; B, H, N, T) and exogenous (ex.; C, I, O, U) PC3 mRNA expression is shown as detected by ISH using specific biotin-labeled probes (boxes in A, G, M, S). ML, Molecular layer; PL, Purkinje layer; Calb., calbindin. D, J, P, V, Confocal microscopy analysis of folia indicated by arrows in A, G, M, and S for PC3 protein expression, which increases in the outer EGL and, to a lower extent, in the IGL, and for calbindin (which marks Purkinje neurons). The markers analyzed are indicated. Purkinje neurons are organized in clusters rather than in layers as in controls. E, K, Q, W, BrdU incorporation is reduced; F, L, R, X, cyclin D1 expression decreases when transgenes are active. Scale bars, 140 μm. Adjacent sagittal sections were spaced 10 μm.

Figure 6.

Overexpression of PC3 strongly increases the differentiation of granule cells and the expression of Math1 in cerebellum at P1, in Tg βACT-tTA/TRE-PC3 (A–J) and in Tg nestin-rtTA/TRE-PC3 (K–T). A, F, K, P, Increase of NF 160 expression in cerebella of mice with active transgene. The width of the molecular layer (ML) is increased, and NF⁺ cells are present also in outer EGL (arrow in A). B, G, L, Q, Math1 expression increases in EGL and also in IGL. The regions analyzed in L and Q are defined by boxes in K and P. The arrow in L indicates the presence of NF⁺ cells within the inner EGL in the activated Tg, never observed in control mice, as also shown by confocal microscopy in M. C, H, M, R, Confocal microscopy magnification from adjacent sections (the region analyzed in C and H is defined by boxes in B and G). D, I, N, S, Analysis of NeuroD/NF160 expression by confocal microscopy in adjacent sections (in regions indicated by arrowheads in A, F, K, P), showing increase of NeuroD⁺ cells in the outer EGL and decrease in the ML-IGL. E, J, O, T, Representative TUNEL analyses of EGL fields. WM, White matter; PL, plexiform layer. Scale bars, 160 μm. Adjacent sagittal sections were spaced 10 μm.

Figure 8.

Analysis of cell cycle, differentiation, and apoptosis in the cerebellum of Tg βACT-tTA/TRE-PC3 (A–F) and Tg nestin-tTA/TRE-PC3 (A′–F′) at P1 and P5. A, A′, BrdULI within the EGL. B, B′, Cyclin D1⁺/total cells within the EGL. C, C′, IGL area positive for NF160. D, D′, Math1⁺/total cells within the IGL. E, E′, TUNEL⁺/total cells in EGL and IGL. F, F′, Cell number/EGL area. Sections were double labeled with Hoechst 33258 and with BrdU, cyclin D1, NF160, Math1 (revealed by FITC- or TRITC-conjugated goat anti-mouse secondary antibodies), or TUNEL assay. Single cells positive for BrdU, cyclin D1, Math1, or TUNEL were identified by merging the two digital images (see Materials and Methods). BrdULI and the other markers analyzed are represented as mean ± SEM percentage ratio between number of labeled cells and total number of cells (or between mean NF160⁺ area and total area). For each binary Tg, the cerebella of six mice were analyzed (3 with active transgenes and 3 controls). Cells or areas positive for each labeling were counted within three fields covering the entire EGL or IGL extension (as specified) of three sagittal sections per mouse adjacent to the midline, spaced ∼10 μm. Black bars, Active transgene; white bars, control. *p < 0.05 versus control (Student's t test).

Figure 7.

The continued increase at P5 of granule cell differentiation and of Math1 expression is dependent on the presence of exogenous PC3. A–L, Continued overexpression of PC3 in the EGL and molecular layer of Tg βACT-tTA/TRE-PC3 at P5 maintains the increase of granule cell differentiation (as shown in C, F) and of Math1 expression (as shown in G, J), as well as the arrest of proliferation in GCPs (for quantitative analyses of BrdULI and of NF and Math1 expression, see Fig. 8 A-F). The various markers and Tg treatments are as in Figures 5 and 6. Boxes in A and D mark the region analyzed by ISH for exogenous (ex.) PC3 mRNA in B and E; arrows mark G and J fields, and arrowheads indicate the region analyzed by confocal microscopy in H and K. I, L, The radial organization of parallel fibers in the EGL, evident in P5 control mice (+doxy), is absent in Tg βACT-tTA/TRE-PC3, and Bergmann glial cells (representatively indicated by arrowheads) are misplaced throughout the EGL. p, Pia; pf, parallel fibers. M–T, Physiological arrest of nestin promoter-driven expression of exogenous PC3 in Tg nestin-rtTA/TRE-PC3 at P5 is associated with the disappearance of the cerebellar phenotype. Arrows in M and O mark the lobe analyzed in R and T for Math1 expression. Scale bars: A, D, 190 μm; M, O, 200 μm. Adjacent sagittal sections were spaced 10 μm.

Figure 9.

Dissection of the molecular pathway of Math1 induction by PC3. A, Recombinant adenovirus encoding PC3 (Adeno-PC3) induces Math1 mRNA in primary cultures of cerebellar granule cells, as indicated by semiquantitative RT-PCR analysis, although no expression is seen in β-gal adenovirus-transduced cells. mRNAs of other proneural genes are analyzed, as indicated. Rev.T.ase(+/–), Amplifications with or without reverse transcriptase as controls for genomic DNA contamination. One representative experiment out of four is shown. B, PC3, ectopically expressed by transfection in primary cultures of cerebellar granule cells from P2 rats, and in chromaffin PC12 cells differentiated into sympathetic neurons (after 48 hr treatment with NGF, 50 ng/ml), stimulates the activity of the Math1 promoter region. This comprised 1200 nt 5′ to the putative transcription start (Akazawa et al., 1995), placed upstream to a luciferase reporter (construct pGL3-Math1-pr/-1200, indicated by green bars in the luciferase assays. The green line in the schematic representation of the Math1 mouse gene and of the constructs shows the promoter region analyzed). Math1 gene enhancer sequence (Helms et al., 2000) was not transcriptionally stimulated by PC3 [construct pGL3-Math1-enh (Helms et al., 2001, corresponding to their construct #9)] (see Materials and Methods) (indicated by red bars in the luciferase assays; the red boxes in the scheme below show the enhancer region analyzed). Min. βg pr., Minimal βglobin promoter. In PC12 cells, NGF treatment was started immediately after transfection of pSCT-PC3 or pSCT empty expression vector (Corrente et al., 2002). The average ± SEM fold increase in luciferase activity of four experiments is shown for each type of cells relative to the level of control samples. Luciferase activities were measured in luciferase units per microgram of protein normalized to the activity of the coreporter pRL-TK (containing the herpes simplex virus thymidine kinase promoter region upstream of the Renilla luciferase gene; Promega) present in each extract, as a measure of the efficiency of transfection. The activity of Math1 promoter and enhancer (constructs pGL3-Math1/-1200 and pGL3-Math1-enh) resulted severalfold above the background, represented by the activity of the empty pGL3 basic vector. The expression of transfected PC3 was verified by Western blot (data not shown). *p < 0.05 versus control (Student's t test on raw data). C, In mice with ablation of Math1, the levels of PC3/Tis21 (Tis21, i.e., the mouse homolog of PC3) were not affected, confirming that PC3 acts upstream. Real-time RT-PCR analysis of PC3/Tis21 and Math1 mRNA levels in Math1^–/– and wild-type E14 cerebella: the average ± SEM values are shown from four replicates. D, A schematic model proposed for PC3 activity in cerebellar granule cells (see Discussion). Math1 is depicted upstream of NeuroD1, consistent with previous findings (Miyata et al., 1999).