Journal of Neuroscience Research 69:714 722 (2002)

Developmental Regulation of the Effects of Fibroblast Growth Factor-2 and 1-Octanol on Neuronogenesis: Implications for a Hypothesis Relating to MitogenAntimitogen Opposition
T. Goto,1,2* T. Takahashi,1,2 S. Miyama,1,2 R.S. Nowakowski,3 P.G. Bhide,1 and V.S. Caviness, Jr.1
1 2

Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts Department of Pediatrics, Keio University School of Medicine, Tokyo, Japan 3 Department of Neuroscience and Cell Biology, UMDNJ-Robert Wood Johnson Medical School, Piscataway, New Jersey

Neocortical neurons arise from a pseudostratied ventricular epithelium (PVE) that lies within the ventricular zone (VZ) at the margins of the embryonic cerebral ventricles. We examined the effects of broblast growth factor-2 (FGF-2) and 1-octanol on cell output behavior of the PVE in explants of the embryonic mouse cerebral wall. FGF-2 is mitogenic and 1-octanol antimitogenic in the PVE. Whereas all postmitotic cells migrate out of the VZ in vivo, in the explants some postmitotic cells remain within the VZ. We refer to these cells as the indeterminate or I fraction, because they neither exit from the VZ nor reenter S phase as part of the proliferative (P) fraction. They are considered to be either in an extremely prolonged G1 phase, unable to pass the G1/S transition, or in the G0 state. The I fate choice is modulated by both FGF-2 and 1-octanol. FGF-2 decreased the I fraction and increased the P fraction. In contrast, 1-octanol increased the I fraction and nearly eliminated the P fraction. The effects of FGF-2 and 1-octanol were developmentally regulated, in that they were observed in the developmentally advanced lateral region of the cerebral wall but not in the medial region. 2002 Wiley-Liss, Inc. Key words: cell cycle; FGF; gap junction; neuronogenesis; neuronal differentiation

that exits the cycle (the Q fraction) and, complementarily, a decrement in the fraction that reenters S phase (the P fraction). It is a further complexity of the process that it is initiated rostrolaterally and advances in a gradient fashion along the caudomedial axis of the epithelium, the so-called transverse neurogenetic gradient (Bayer and Altman, 1991). The regularity and precision with which these cycle kinetic and cell output parameters of the proliferative process are scheduled by the 11-cycle sequence in the intact animal provide little insight into the underlying complexity of the cellular mechanisms and the molecular agents upon which the operation presumably depends. The mechanisms that ensure such stability and precision in vivo are assumed to be primarily cell internal but modulated by a balance of opposing mitogenic and antimitogenic environmental signals. We investigated the modulatory effects of mitogenic and antimitogenic signals on cell proliferation in the neocortical PVE in an organotypic explant culture system in which concentrations and duration of exposure to the signals may be specied. We exposed the PVE to broblast
Contract grant sponsor: NIH; Contract grant number: NS12005; Contract grant number: NS33433; Contract grant number: HD05515; Contract grant sponsor: NASA; Contract grant number: NAG2-1367; Contract grant sponsor: Keio University Grant-in-Aid for Encouragement of Young Medical Scientists; Contract grant sponsor: Japanese Ministry of Health/ Education; Contract grant sponsor: Pharmacia Fund for Growth and Development Research. *Correspondence to: Dr. Tomohide Goto, Department of Pediatrics, Keio University School of Medicine, 35 Shinano-machi, Shinjuku-ku, Tokyo 160-8582, Japan. E-mail: goto@chp-kiyose-tokyo.jp Received 23 February 2002; Revised 21 May 2002; Accepted 27 May 2002 Published online in Wiley InterScience (www.interscience.wiley. com). DOI: 10.1002/jnr.10361

Projection neurons of the neocortex arise from a pseudostratied ventricular epithelium (PVE) that lines the margins of the ventricles of the embryonic cerebrum (Boulder Committee, 1970). In mouse, neocortical neurons are generated in the course of 11 cell cycles (Takahashi et al., 1995; Miyama et al., 1997; Caviness et al., 2000). With each successive cycle, there is an increment in the duration of the G1 phase, but not in any other phase of the cell cycle. Moreover, with each successive cell cycle, there is an increment in the fraction of postmitotic cells
2002 Wiley-Liss, Inc.

FGF-2, Gap Junction Blockade, and Neocortical Neuronogenesis

715

growth factor-2 (FGF-2), a well-known mitogen for the PVE (Cattaneo and McKay, 1990; Kilpatrick et al., 1993; Weise et al., 1993; Eckenstein, 1994; Ghosh and Greenberg, 1995; Temple and Qian, 1995; Cavanagh et al., 1997), or 1-octanol, which mediates an antimitogenic effect through gap junction blockade (Yamamoto et al., 1991; Nadarajah et al., 1998; Bittman and LoTurco, 1999). We dene a mitogen and an antimitogen for the purposes of the present study as substances having the property of driving proliferative cells of the PVE toward the P and Q choices, respectively. We report that explantation of the PVE reveals a novel proliferative fate choice for PVE cells that is not observed in vivo and that is differentially inuenced by FGF-2 and 1-octanol. Furthermore, the effects of FGF-2 and 1-octanol on proliferative fate choice are modulated by the developmental state of the PVE. These ndings reveal a balance between mitogenic and antimitogenic environmental signals and cell-internal mechanisms whose coordinated actions set the proliferative tone during neocortical neuronogenesis.
MATERIALS AND METHODS Explant Preparation and Histological Processing CD1 mice (Charles River Laboratories, Wilmington, MA) were maintained under 12 hr light/dark conditions. Breeding pairs were set up at 5:00 PM, and the day of conception (embryonic day 0; E0) was ascertained by vaginal plug at 9:00 AM on the following day. Explants of the telencephalic wall were prepared from E13 mice with the pregnant dam under deep anesthesia (ketamine hydrochloride and xylazine, 50 mg/kg and 10 mg/kg, respectively, i.p.), as described previously (Takahashi et al., 1999a). Explants were transferred to ice-cold collagen gel medium. The explants contained both the striatal and the neopallial structures. After transfer to the collagen, the striatal portion was separated from the neopallial portion by a cut at the caudatopallial angle and discarded; only the neopallial portion was maintained in culture (Takahashi et al., 1999a; Miyama et al., 2001). The culture dish containing the neopallial explant was transferred to an incubator (37C with 5% CO2 and 95% air), and 1 hr later 2.5 ml of tissue culture medium (DMEM/F12; Gibco-BRL, Grand Island, NY) was added. Experimental Treatments For FGF-treated explants, both the collagen gel and the culture medium contained either 100 ng/ml of FGF-2 (GibcoBRL) in 0.1% bovine serum albumin (BSA; Sigma Chemical Co., St. Louis, MO) as the experimental condition or BSA alone as the control condition. The endogenous concentration of FGF-2 in the neocortex is 50 ng/g (Eckenstein et al., 1991), and 40-50 ng/ml FGF-2 was used in experiments on dissociated cells from the rodent embryonic neocortex (Ghosh and Greenberg, 1995). Because higher concentrations of the ligand were required to produce an effect in explants compared with dissociated cells in our hands, we used an FGF-2 concentration of 100 ng/ml. For 1-octanol-treated explants, both the collagen gel and the medium contained either 1.0 mM of 1-octanol (Sigma) in the vehicle dimethyl sulfoxide (DMSO; Sigma) as the exper-

Fig. 1. Experimental protocols. A: Schedule of exposure of explants to BUdR in the presence of FGF-2 BSA or BSA alone (upper series) or of 1-octanol DMSO or DMSO alone (lower series). BUdR was added to the medium 5 8 hr after explantation. In both series, specimens were xed after 0.5, 6, 12, and 16 hr of exposure to BUdR. B: Schedule of exposure of explants to 3H-TdR and BUdR. A series of three injections was delivered i.p. into the pregnant dam over a 3.5 hr interval prior to removal of embryos and preparation of explants. These were. rst, 3H-TdR, followed in 1.5 hr by the second, BUdR, and again in 2 hr more by the third, BUdR. The third injection was delivered immediately before removal of embryos. The explants were divided into two sets. One was cultured with BUdR in the medium, the other without BUdR. The explants were held in culture for 13 hr, which is long enough for cells of the P fraction to reenter S phase and to become labeled with BUdR in the explants cultured with the tracer.

imental condition or DMSO alone as the control condition. These concentrations, in accordance with those used by other investigators (Kilpatrick et al., 1993; Eckenstein, 1994; Bittman et al., 1997), were considered optimal because they were associated with maximum effect on proliferative behavior yet were without deleterious effect on the explant systems by the criteria of incidence of pyknotic cells and the architectonic integrity of the explants. For each explant, the medium was changed every 4 hr throughout the interval of incubation. The explants were cultured for 5 8 hr before S-phase labeling schedules began (see below). This range of preparatory incubation intervals prior to the addition of S-phase markers is sufcient to reestablish stable, continuous, and asynchronous progression of the cell cycle in the explants (Takahashi et al., 1999a). S-Phase Labeling Protocols Cumulative 5-bromo-2-deoxyuridine labeling. Explants from experimental and control conditions were exposed to 5-bromo-2-deoxyuridine (BUdR; Sigma; nal concentration 10 M) cumulatively and harvested at 0.5, 6, 12, or 16 hr (Fig. 1A). The explants were xed with 70% ethanol, dehydrated with graded ethanol, cleared in xylene, embedded in parafn, and cut at a thickness of 4 m in the coronal plane, i.e., orthogonally to the external cerebral surface of the explant. Sections from the middle portion of each series were processed for BUdR immunohistochemistry (anti-BUdR antibody; Beckton Dickinson, Mountain View, CA) using nickel-cobalt color

716

Goto et al.

enhancement of the diaminobenzidine reaction product and counterstained with basic fuchsin (Takahashi et al., 1992). The cerebral wall was partitioned in the sections into a lateral cerebral zone (LCZ) and a medial cerebral zone (MCZ), each with a width of 100 m. The LCZ was dened to extend from 250 to 350 m from the caudate-pallial angle, and the MCZ extended from 950 to 1,050 m from the caudate-pallial angle (Miyama et al., 2001). The progenitors in the PVE of the LCZ are encountering cell cycle number 8 of the 11-cell-cycle neurogenetic sequence on E13, whereas those in the PVE of the MCZ are encountering cell cycle number 6 (Miyama et al., 1997). Thus, the LCZ is developmentally in advance of the MCZ with respect to the neurogenetic schedule. The BUdR labeling analyses were performed in the LCZ and MCZ sectors of the PVE. Each sector was 100 m in the medial-lateral direction and 4 m (corresponding to section thickness) in the rostral-caudal direction. The radial dimension of the sector was divided into bins (Takahashi et al, 1992); each bin was 10 m high (radial dimension) and 100 m wide (medial-lateral dimension). The bins were numbered 1, 2, 3, etc., from the ventricular margin outward. The numbers of BUdR-labeled and BUdR-unlabeled nuclei were counted in each bin, and the labeling index (LI; the ratio of BUdR-labeled cells to total cells) was calculated for each bin and also for the entire sector. The LI was plotted as a function of distance from the ventricular border (i.e., bin number) for LCZ and MCZ. Four nonadjacent sections from each explant were used. An average BUdR LI was calculated for each explant and then for each experimental condition (n 4 7). Identication of proliferative fates. A double S-phase labeling paradigm was used to delineate a cohort of cells that left the S phase during a 1.5 hr interval. The cohort was then partitioned into cells that terminated proliferative activity following mitosis and migrated out of the ventricular zone (VZ) as a part of the quiescent (Q) fraction or cells that entered another S phase as part of the proliferative (P) fraction. The design is shown in Figure 1B. Pregnant dams carrying E13 mice were injected i.p. with tritiated thymidine (3H-TdR; New England Nuclear, Boston, MA; 5 Ci/g body weight), followed by two injections of BUdR (Sigma; 50 g/g body weight i.p.) administered 1.5 and 3.5 hr after the 3H-TdR injection (Fig. 1B). Explants were prepared 0.5 hr after the second BUdR injection. By this time, virtually all cells of the 3H-TdR-only-marked cohort would have completed M phase, because those cells left S phase by the time of the rst BUdR injection (2.5 hr earlier), and the combined lengths of G2 and M phases are under 2 hr in the neocortical PVE (Takahashi et al., 1993, 1994; Hayes and Nowakowski, 2000). Thus, the 3H-TdR-only-labeled cells would have entered G1 phase at the time of explantation and would be exposed to conditions of explantation and culture only after entering G1 phase (Takahashi et al., 1993, 1994; Hayes and Nowakowski, 2000). To partition the 3H-TdR-only-labeled cells into to P and Q fractions, explants were divided into two subsets (Fig. 1B). One subset was cultured continuously with BUdR for a 12 hr period. In this subset, cells of the Q fraction would be labeled only with 3H-TdR, because all cells entering S phase (i.e.,

P-fraction cells) will have become labeled with BUdR during the 12 hr labeling interval. The second subset was cultured without BUdR for 12 hr, with the result that the 3H-TdRonly-labeled cohort in this subset will not be partitioned into P and Q fractions but will represent P Q fractions (Fig. 1B). The details of this experimental design have been previously published (Takahashi et al., 1994, 1999b). At the end of the experiment, the explants were xed with 70% ethanol, dehydrated in graded series of ethanol, cleared in xylene, and embedded in parafn. The explants were sectioned at 4 m thickness and stained immunohistochemically for BUdR, as described above, but without nickel-cobalt enhancement of the diaminobenzidine reaction product. For autoradiographic detection of 3H-TdR, sections processed for BUdR immunohistochemistry were dried at 37C, coated with NTB-2 nuclear emulsion (Eastman Kodak, Rochester, NY), and stored in the dark at 4C. After 5 weeks of exposure at 4C , the autoradiograms were developed with D-19 developer (Eastman Kodak, Rochester, NY) and counterstained with basic fuchsin (Takahashi et al., 1999b). For the analysis of proliferative fates, the bin method of partitioning the cerebral wall that was described for the cumulative labeling analysis was also used. Cells labeled with 3H-TdR only under each of the experimental conditions were counted with respect to each bin and plotted with respect to bin number; six nonadjacent sections were used for each explant. The average number per bin and the average total number per entire sector were calculated for the 3H-TdR-only-labeled cells. The distribution of cells labeled with 3H-TdR only was compared between each control and experimental group using a 2 analysis separately for the MCZ and the LCZ (n 3 8 for each experimental condition).

RESULTS Cellular Architecture of the Cerebral Wall Explants We have shown previously that embryonic cerebral wall explants retain their structural integrity in vitro for up to 24 hr (Takahashi et al., 1999a; Miyama et al., 2001). The pattern of stratication of the cerebral wall in vitro resembles closely that in vivo and remains stable over the 21 hr interval required by the S-phase labeling experiments of the present study. The cerebral wall explants showed comparable cytoarchitecture and stratication pattern under the different experimental and control conditions. The incidence of pyknosis was also low (1%) under the different conditions. In our previous studies with this explant culture system also, the incidence of pyknosis was low even at 24 hr in vitro (Takahashi et al., 1999a; Miyama et al., 2001). The explants were xed with ethanol in the present study. The xation is not suitable for TUNEL analysis. Therefore, we counted pyknotic gures to estimate the incidence of cell death. Previously, it has been shown that pyknotic gure counts and TUNEL labeling analyses provide comparable estimates of cell death in the VZ (Thomaidou et al., 1997).

FGF-2, Gap Junction Blockade, and Neocortical Neuronogenesis

717

Fig. 2. Cumulative BUdR labeling indices in MCZ (A) and LCZ (B) at 0.5, 6 and 12 hr in the presence of BSA alone, FGF-2BSA, DMSO alone or DMSO 1-octanol (error bars show SEM). Indices are higher in the MCZ than the LCZ at corresponding intervals under every condition. A signicant augmentation of LI is observed with exposure to FGF-2 relative to BSA alone in LCZ at 6 hr. Signicant decrease in the LI with exposure to 1-octanol relative to DMSO alone is observed in the LCZ at 6 and 12 hr. No signicant changes in LI occur at the MCZ under the different experimental conditions.

Interval to Maximum LI We compared the cumulative BUdR labeling data from the two control conditions, namely, DMEM/F12 BSA or DMEM/F12 DMSO, with those of culture in DMEM/F12 alone from our previous studies (Takahashi et al., 1999a; Miyama et al., 2001). The advance in the LI under conditions of exposure to DMEM/F12 BSA or to DMEM/F12 DMSO was comparable to that seen in culture in DMEM/F12 alone. Thus, under all three conditions, the LI advanced with time of exposure to BUdR to reach a maximum of approximately 0.7 in the MCZ and 0.55 in the LCZ at 12 hr. We examined the effects of exposure to FGF-2 or 1-octanol on the BUdR LI in the explants. In the MCZ, neither FGF-2 nor 1-octanol altered the maximum LI achieved at 0.5, 6.0, or 12 hr of cumulative BUdR exposure (Fig. 2A). In the LCZ, by contrast, both FGF-2 and 1-octanol affected both the rate of change in BUdR LI with time and the maximum LI (Fig. 2B). FGF-2 elevated

the LI by 15% at 6 hr (t-test, P 0.02) but did not produce signicant changes at 12 hr (Fig. 2B). By contrast, 1-octanol decreased the LI by 24% at 6 hr (t-test, P 0.01) and by 16% at 12 hr (t-test, P 0.05). Neither FGF-2 nor 1-octanol produced signicant changes at 0.5 hr. The maximum LI, i.e., the growth fraction or the proportion of proliferating cells, does not reach 1.0 under any of the conditions examined. This is in contrast to the VZ in the intact embryo, in which all cells of the VZ are proliferating and the maximum LI reaches 1.0 (Takahashi et al., 1995). The proportion of nonproliferating cells in the VZ of the explants is considerable, 30% in the MCZ and 45% in the LCZ. Moreover, insofar as this is a cumulative BUdR-labeling paradigm that continuously labels cells entering S phase, the unlabeled cells, which contribute to the lower LI, must be in G2, M, or G1 phase at the onset of the experiment and do not enter the S phase in culture. The second observation concerns the inuence of maturational stage of the PVE progenitors on cell cycle kinetics. Both control (i.e., BSA or DMSO) and experimental (i.e., FGF-2 or 1-octanol) conditions alter the maximum LI in the relatively more mature LCZ but do not affect it in the relatively less mature MCZ (Fig. 2A,B). Finally, the effects of FGF-2 and 1-octanol on BUdR LI in the LCZ are in opposite directions (Fig. 2A,B). FGF-2 promotes S-phase reentry, and 1-octanol inhibits it. Proliferative Fate At each pass through the cell cycle in the intact embryo, a fraction of dividing cells reenters S phase upon completion of G1 phase (P cells). The complementary fraction of the cells becomes permanently postmitotic (Q cells) and exits the VZ, relatively quickly, i.e., by the time that the P cells reenter S phase. As a consequence of this behavior, by the end of G1 phase, all cells remaining in the VZ will be P-fraction cells and will have entered S phase. Under those conditions, the BUdR LI will be 1.0 in the VZ (Takahashi et al., 1994, 1996). However, in the explant, the maximum LI in the VZ with cumulative labeling is less than 1.0 (Fig. 2A,B; Takahashi et al., 1999a). Therefore, all cells in the VZ of the explant cannot be assigned to the P fraction. The explant VZ contains three categories of cells. Some cells fulll the criteria for assignment to a P or a Q fate, as is true in vivo. However, other cells belong neither to the P fraction (because they do not become BUdR labeled within the 12 hr interval) nor to the Q fraction (because they do not exit from the VZ during the 12 hr labeling period, although they are not BUdR labeled). We designate this third fate as the I or indeterminate fate (see Fig. 4). Cells that adopt the I fate do not fulll the criteria for assignment to either P or Q. In our nal set of studies, we determined the patterns of distribution and the relative number of cells of the P, Q, and I proliferative fates under control and experimental conditions. We used the double S-phase labeling paradigm

718

Goto et al.

shown in Figure 1B. In this paradigm, when the explants are cultured without BUdR, cells of all the three proliferative fates, i.e., P, I, and Q, are labeled only with 3 H-TdR. When the explants are cultured with BUdR, cells of the I and Q fractions are labeled only with 3HTdR, and cells of the P fraction are labeled with 3H-TdR and BUdR. Thus, we can distinguish cells of each of the three fractions using data from the two labeling paradigms. When the position of cells of the I and Q fractions is plotted, a bimodal distribution is seen across the full width of the cerebral wall (Fig. 3A,B). The outer and smaller of the two populations, distributed through the subventricular zone and lower part of the intermediate zone (bins 10 15) corresponds to the Q fraction, because it has exited from the VZ, as in vivo (Takahashi et al., 1994). The inner and larger of the two populations is distributed throughout the full width of the VZ, with a peak in the outer half of the VZ. A small proportion of Q fraction cells is found at the outer border of the VZ in vivo (Takahashi et al., 1994, 1996). However, the inner population in the present study is considerably larger than its counterpart in the in vivo studies. In addition, our BUdR LI data show that some postmitotic cells (i.e., non-BUdR-labeled cells) remain in the VZ in the explants, and those cells belong to the I fraction and contribute to the inner population of cells. Therefore, we consider the inner population to consist of a mixture of Q- and I-fraction cells. FGF-2 and 1-octanol produce different effects on the distribution of Q and I populations (i.e., the inner population), but only in LCZ (Fig. 3A,B). FGF-2 decreases the height of the inner peak (open arrowheads in Fig. 3A), whereas 1-octanol increases it (solid arrowheads in Fig. 3B). There is no signicant effect of either FGF-2 or 1-octanol in the MCZ (Fig. 3 A,B; 2, P 0.05). The relative numbers of P, Q, and I cells were compared under the different conditions; the relative number of cells in each fraction equals the number of cells in the respective fraction/total number of cells in the cohort (Table I). FGF-2 decreased the relative number of cells in the I fraction in the LCZ from 0.593 to 0.343 compared with the control condition (Table I). The decrease may have contributed to an increase in the number of cells in the P fraction. 1-Octanol decreased the relative number of cells in the P fraction (by almost 100%). This decrease may have contributed to an increase in the number of cells in the I fraction. DISCUSSION Our principal nding is that explantation of the cerebral wall reveals a proliferative fate choice for PVE cells that is not observed in vivo and that is differentially inuenced by FGF-2 and 1-octanol. Furthermore, the effects of FGF-2 and 1-octanol on proliferative fate are modulated by the developmental state of the PVE, insofar as the effects are elicited only in the developmentally more advanced LCZ. We advance, based on these observations, a hypothesis for the role of mitogenic and antimitogenic substances in the regulation of neocortical neuronogenesis.

Fig. 3. Bimodal distribution of Q and I fractions across the full width of the cerebral wall (error bars show SEM). The inner and larger of the two populations is distributed in the VZ (bins 17; shaded area), with a peak in the outer half of the VZ. This population consists of a mixture of Q- and I-fraction cells. The outer and smaller of the two populations corresponds to the Q fraction and extends through the subventricular zone and lower part of the intermediate zone (bins 10 15). A: Under conditions of culture with FGF-2, there is a signicant decrease in Q I-fraction cells in LCZ (open arrowhead) but not in MCZ. B: Under conditions of culture with 1-octanol in LCZ, there is a signicant enlargement of Q I fractions (solid arrowhead). This effect of 1-octanol is not seen in MCZ.

A Novel Proliferative Fate for PVE Cells in the Explants: The I Fraction The most salient consequence of explantation is redirection of the proliferative fates of cells of the PVE. In vivo, there are two options for cells in G1 phase, either to exit the cycle and migrate out of the VZ or to reenter S phase and continue to cycle within the VZ. In the explants, by contrast, a substantial proportion makes neither of those choices but instead persists in the VZ in an indeterminate state (I fraction). The I population appears to be a fate choice, or state, entered by cells in G1 phase (Fig. 4). The magnitude of the I population is sufcient to account for the reduction in the growth fraction (or maximum BUdR LI) in the explants compared with that in vivo. At least three conditions could give rise to the I fate. First, I-fraction cells could be cells fated to be Q that failed to exit the VZ, perhaps because the cytoskeletal infrastructure necessary for the migratory state was not assembled

FGF-2, Gap Junction Blockade, and Neocortical Neuronogenesis

TABLE I. Proliferative Fates of PVE Cells in E13 Cerebral Wall Explants P fraction Number of cells MCZ BSA alone BSA FGF DMSO alone DMSO OCT LCZ BSA alone BSA FGF DMSO alone DMSO OCT 3.39 4.11 2.85 4.28 0.94 3.94 3.88 0.07* Relative number 0.216 0.215 0.192 0.270 0.053 0.192 0.206 0.004 Q fraction Number of cells 4.47 6.33 4.58 4.51 6.33 9.53 5.76 6.64 Relative number 0.286 0.330 0.308 0.284 0.354 0.465 0.307 0.350 I fraction Number of cells 7.81 8.71 7.46 7.09 10.6 7.03** 9.14 12.2

719

Relative number 0.498 0.455 0.501 0.446 0.593 0.343 0.487 0.646

The average number and relative number assigned to the P, Q, or I fates are shown for the MCZ and the LCZ under the different control and experimental conditions. *Difference between DMSO alone and DMSO OCT signicant at P 0.001. **Difference between BSA alone and BSAFGF signicant at P 0.001.

Fig. 4. Proliferative fates of neocortical progenitors in vitro in MCZ and LCZ. There are two choices for PVE progenitors in vivo, namely, Q or P. Our data reveal a novel fate choice, namely, the indeterminate, or I, fate choice in vitro.

properly (Iacopetti et al., 1999). A second possibility is that the I cells are still in G1 phase but that G1 phase is greatly, perhaps indenitely, prolonged by explantation and culture conditions. The progression from G1 phase to S phase is contingent, inter alia, upon a stoichiometric dominance of cyclin E over unbound p27Kip1 in proliferating cells (Cheng et al., 1998; Geng et al., 1999; Sherr and Roberts, 1999). Because their abundance and state of activity are time and condition dependent, the I population may be the cells that are approaching but have not reached the threshold to trigger S-phase entry. A third possibility is that the I cells have entered a G0 state. The G0 state is generally considered not to be a fate option for neuronal precursors of the PVE (Murray and Hunt, 1993). However, under the circumstances of explant culture, this proliferative fate must also be considered because G0 is essentially an adaptive state of proliferative eukaryotic cells in response to nutrient or trophic factor deprivation. It is possible that cells stressed by the minimal culture conditions might have resorted to this state as well (Sherr, 1994; Zetterberg et al., 1995). The I fraction is prominent in both MCZ and LCZ. However, it is modulated, in opposing directions and to different degrees, by FGF-2 and 1-octanol in LCZ but not

in MCZ; i.e., it is modulated only at the more advanced region of the neurogenetic gradient (Fig. 5). This differential modulation is interesting and presumably reects differences in the maturational state of the LCZ vs. the MCZ. We have shown previously that the LCZ is upstream of the cell cycle gradient in the PVE compared with the MCZ (Miyama et al., 1997). Thus, at the time of explantation, the progenitors in the PVE of the LCZ are encountering cell cycle number 8 of the 11-cell-cycle neurogenetic sequence on E13, whereas those in the PVE of the MCZ are encountering cell cycle number 6 (Miyama et al., 1997). The expression of p27Kip1 transcripts is higher in the PVE at the LCZ compared with the MCZ on E13 (Delalle et al., 1999), illustrating a putative molecular basis for the cell cycle gradient in the PVE. This indicates that, in the present study, the explant conditions have apparently unmasked relatively selectively mechanisms regulatory to the proliferative fates of cells of the PVE, expressed only at the relatively advanced region in the developing neocortex. Exposure to FGF-2 in the LCZ was associated with a redirection of the I population to both the Q and the P fates (Fig. 5). This suggests that at least some of the I population corresponds to cells that in vivo would have been destined for both P and Q fates. We emphasize that the I population represents a large fraction of the fate choices in the explants. It is greater than the Q fraction in vivo in either the MCZ or the LCZ at E13 (Miyama et al., 1997). That is, the I population is sufciently large to include a substantial fraction of cells diverted from both of the normal fate choices made in vivo by cells in G1 phase at the time of explantation. The fact that, with respect to the in vivo condition, the P choice is greatly underrepresented in the VZ of both the MCZ and the LCZ is consistent with this. Implications for Histogenetic Regulation In mouse, the Q fraction follows an ascending path from 0.0 to 1.0, and the P fraction follows a complemen-

720

Goto et al.

Fig. 5. Schematic representation of proliferative fates in MCZ and LCZ in the intact embryo, in vivo and under conditions of exposure to FGF-2 and 1-octanol in vitro. The width of the arrows indicating a given fate corresponds approximately to the proportion of daughter cells that will follow that fate. There are only the two fates, P and Q, in vivo. At E13, the assignment to P greatly outweighs that to Q in the MCZ, but the two fates are approximately balanced in the LCZ. There are three fate choices in vitro, P, Q, or I. In the MCZ, in vitro under control conditions, again P outweighs Q, but there is an assignment of much of the P fraction to the I fraction. In the LCZ, there is an even more dramatic assignment to the I fate, particularly at the expense of P but also of Q. FGF-2 decreases the I fraction and appears to increase both P and Q fractions at the LCZ in vitro. 1-Octanol increases the I fraction and virtually completely obliterates the P fraction.

tary descending path from 1.0 to 0.0 in the course of 11 cell cycles (Takahashi et al., 1996). This in vivo pattern of regulation is quite unlike the high-amplitude shift of cells away from P and Q observed in the explant. What is the mechanism that ensures such stability and precision in vivo and that is not reproduced in vitro despite preservation of both vitality of cells and their epithelial integrity? What is the biological signicance of the I population that appeared only under culture conditions and its modication by 1-octanol and FGF-2? To address these questions, we draw attention to two paradoxes. The rst is the contrast in the potential effect of an antimitogen, 1-octanol, on cell cycle kinetics and proliferative fate in the MCZ and the LCZ in the explant and that in vivo. There are, for example, no comparable local variations across the expanse of the murine PVE in the pattern of advance of Q fraction with cell cycle such as might be ascribable to local variations in the inuence of antimitogens (Miyama et al., 1997). The second is that, whereas the in vivo PVE is tonically bathed in a host of

substances with antimitogenic properties but of diverse pharmacological class, there appear to be only a few mitogenic trophic factors that have been established denitely (Cattaneo and McKay, 1990; DiCicco-Bloom et al., 1990; Ghosh and Greenberg, 1995; Temple and Qian, 1995; Nadarajah et al., 1996; Bittman et al., 1997). We suggest for consideration the hypothesis that the critical histogenetic role of the antimitogenic factors relates not so much to their antimitogenic properties but to their diversity. Thus, each antimitogenic substance thus far recognized also has a known role in the histogenetic development or functional properties of terminally postmitotic neurons. Thus, glutamate, -aminobutyric acid (GABA), and catecholamines are neurotransmitters with antimitogenic properties, and neurotrophin-3 (NT3) is a neurotrophic factor and an antimitogen (Maisonpierre et al., 1990; LoTurco et al., 1995; Ghosh and Greenberg, 1995; Temple and Qian, 1995; Bittman et al., 1997). We suggest, then, that collectively these substances of the cellular milieu contribute nonspecically to a tonic upward pressure on Q fate but that the specicity of their effects is reserved for the histogenetic fate of the terminally postmitotic neuron. It is possible that a few kinds of mitogens, for example, the FGF family, may serve proliferative roles on PVE in vivo and be sufcient, possibly with gap junctional connection of PVE, to offset the tonic upward drive on Q fate exerted by the antimitogens. Indeed, in the FGF-2 knockout mouse (Vaccarino et al., 1995, 1999; Raballo et al., 2000), the neuronal decit is 20 50%, corresponding to a relatively small fraction of the normal output per cycle. This result indicates either that there are other mitogens, for example, epidermal growth factor (EGF; Tropepe et al., 1999; Martens et al., 2000), or that there are signicant cell autonomous factors, for example, cell-cycle-regulatory proteins, such as cyclin E and p27Kip1, that control P and Q fraction. The latter possibility could be supported by the fact that the explant system shows that the MCZ is relatively insensitive to mitogen/antimitogen treatment, although the progenitors in the region express FGF receptors (Raballo et al., 2000), and that the rst several cell cycles have high cyclin E and low p27Kip1 levels (Delalle et al., 1999), suggesting intrinsic, autonomous mechanisms that relate to the degree of developmental maturity of PVE. This hypothesis shifts the emphasis for principal regulation of the pace of neuronogenesis away from cellexternal and toward cell-internal mechanisms, with the implication that the latter operate with regular, precisely calibrated step advances in the length of each G1 phase and Q fraction as scheduled with successive cell cycles (Caviness et al., 2000). The amplitude of advance is invariably correlated with cell cycle number, independently of region of the PVE. The plausibility of this view is supported by the observation here and elsewhere that the determination of P and Q fates is essentially unresponsive to cell-external factors, be they mitogens or antimitogens, until the neuronogenetic interval is relatively advanced (LoTurco et al., 1995; Bittman et al., 1997). Moreover,

FGF-2, Gap Junction Blockade, and Neocortical Neuronogenesis

721

the appearance of responsiveness correlates with the appearance of receptors or gap junction components, again an appearance that must be the expression of cell-internal mechanisms (LoTurco et al., 1995; Bittman et al., 1997; Nadarajah et al., 1997, 1998). At present, there are no substantial hypotheses relating to the nature of these cellinternal mechanisms or their modes of regulation. ACKNOWLEDGMENT T.T. was supported by a fellowship from The Medical Foundation, Inc., Charles A. King Trust (Boston, MA). REFERENCES
Bayer SA, Altman J. 1991. Neocortical development. New York: Raven Press. Bittman KS, LoTurco JJ. 1999. Differential regulation of connexin 26 and 43 in murine neocortical precursors. Cereb Cortex 9:188 195. Bittman K, Owens DF, Kriegstein AR, LoTurco JJ. 1997. Cell coupling and uncoupling in the ventricular zone of developing neocortex. J Neurosci 17:70377044. Boulder Committee. 1970. Embryonic vertebrate nervous system: revised terminology. Anat Rec 166:257262. Cattaneo E, McKay R. 1990. Proliferation and differentiation of neuronal stem cells regulated by nerve growth factor. Nature 347:762763. Cavanagh JF, Mione MC, Pappas IS, Parnavelas JG. 1997. Basic broblast growth factor prolongs the proliferation of rat cortical progenitor cells in vitro without altering their cell cycle parameters. Cereb Cortex 7:293 302. Caviness VS Jr, Takahashi T, Nowakowski RS. 2000. Neocortical malformation as consequence of nonadaptive regulation of neuronogenetic sequence. Ment Retard Dev Disabil Res Rev 6:2233. Cheng M, Sexl V, Sherr CJ, Roussel MF. 1998. Assembly of cyclin D-dependent kinase and titration of p27Kip1 regulated by mitogenactivated protein kinase kinase (MEK1). Proc Natl Acad Sci USA 95: 10911096. Delalle I, Takahashi T, Nowakowski RS, Tsai LH, Caviness VS Jr. 1999. Cyclin E-p27 opposition and regulation of the G1 phase of the cell cycle in the murine neocortical PVE: a quantitative analysis of mRNA in situ hybridization. Cereb Cortex 9:824 832. DiCicco-Bloom E, Townes-Anderson E, Black IB. 1990. Neuroblast mitosis in dissociated culture: regulation and relationship to differentiation. J Cell Biol 110:20732086. Eckenstein FP. 1994. Fibroblast growth factors in the nervous system. J Neurobiol 25:14671480. Eckenstein FP, Shipley GD, Nishi R. 1991. Acidic and basic broblast growth factors in the nervous system: distribution and differential alteration of levels after injury of central versus peripheral nerve. J Neurosci 11:412 419. Geng Y, Whoriskey W, Park MY, Bronson RT, Medema RH, Li T, Weinberg RA, Sicinski P. 1999. Rescue of cyclin D1 deciency by knockin cyclin E. Cell 97:767777. Ghosh A, Greenberg ME. 1995. Distinct roles for bFGF and NT-3 in the regulation of cortical neurogenesis. Neuron 15:89 103. Hayes NL, Nowakowski RS. 2000. Exploiting the dynamics of S-phase tracers in developing brain: interkinetic nuclear migration for cells entering versus leaving the S-phase. Dev Neurosci 22:44 55. Iacopetti P, Michelini M, Stuckmann I, Oback B, Aaku-Saraste E, Huttner WB. 1999. Expression of the antiproliferative gene TIS21 at the onset of neurogenesis identies single neuroepithelial cells that switch from proliferative to neuron-generating division. Proc Natl Acad Sci USA 96: 4639 4644.

Kilpatrick TJ, Talman PS, Bartlett PF. 1993. The differentiation and survival of murine neurons in vitro is promoted by soluble factors produced by an astrocytic cell line. J Neurosci Res 35:147161. LoTurco JJ, Owens DF, Heath MJS, Davis MBE, Kriegstein AR. 1995. GABA and glutamate depolarize cortical progenitor cells and inhibit DNA synthesis. Neuron 15:12871298. Maisonpierre PC, Belluscio L, Friedman B, Alderson RF, Wiegand SJ, Furth ME, Lindsay RM, Yancopoulos GD. 1990. NT-3, BDNF, and NGF in the developing rat nervous system: parallel as well as reciprocal patterns of expression. Neuron 5:501509. Martens DJ, Tropepe V, Van der Kooy D. 2000. Separate proliferation kinetics of broblast growth factor-responsive and epidermal growth factor-responsive neural stem cells within the embryonic forebrain germinal zone. J Neurosci 20:10851095. Miyama S, Takahashi T, Nowakowski RS, Caviness VS Jr. 1997. A gradient in the duration of the G1 phase in the murine neocortical proliferative epithelium. Cereb Cortex 7:678 689. Miyama S, Takahashi T, Goto T, Bhide PG, Caviness VS Jr. 2001. Continuity with ganglionic eminence modulates interkinetic nuclear migration in the neocortical pseudostratied ventricular epithelium. Exp Neurol 169:486 495. Murray A, Hunt T. 1993. The cell cycle. New York: W.H. Freeman. Nadarajah B, Thomaidou D, Evans WH, Parnavelas JG. 1996. Gap junctions in the adult cerebral cortex: regional differences in their distribution and cellular expression of connexins. J Comp Neurol 376:326 342. Nadarajah B, Jones AM, Evans WH, Parnavelas JG. 1997. Differential expression of connexins during neocortical development and neuronal circuit formation. J Neurosci 17:3096 3111. Nadarajah B, Makarenkova H, Becker DL, Evans WH, Parnavelas JG. 1998. Basic FGF increases communication between cells of the developing neocortex. J Neurosci 18:78817890. Raballo R, Rhee J, Lyn-Cook R, Leckman JF, Schwartz ML, Vaccarino FM. 2000. Basic broblast growth factor (Fgf 2) is necessary for cell proliferation and neurogenesis in the developing cerebral cortex. J Neurosci 20:50125023. Sherr CJ. 1994. G1 phase progression: cycling on cue. Cell 79:551555. Sherr CJ, Roberts JM. 1999. CDK inhibitors: positive and negative regulators of G1-phase progression. Genes Dev 13:15011512. Takahashi T, Nowakowski RS, Caviness VS Jr. 1992. BUdR as an S-phase marker for quantitative studies of cytokinetic behaviour in the murine cerebral ventricular zone. J Neurocytol 21:185197. Takahashi T, Nowakowski RS, Caviness VS Jr. 1993. Cell cycle parameters and patterns of nuclear movement in the neocortical proliferative zone of the fetal mouse. J Neurosci 13:820 833. Takahashi T, Nowakowski RS, Caviness VS Jr. 1994. Mode of cell proliferation in the developing mouse neocortex. Proc Natl Acad Sci USA 91:375379. Takahashi T, Nowakowski RS, Caviness VS Jr. 1995. The cell cycle of the pseudostratied ventricular epithelium of the embryonic murine cerebral wall. J Neurosci 15:6046 6057. Takahashi T, Nowakowski RS, Caviness VS Jr. 1996. The leaving or Q fraction of the murine cerebral proliferative epithelium: a general model of neocortical neueonogenesis. J Neurosci 16:6183 6196. Takahashi T, Bhide PG, Goto T, Miyama S, Caviness VS Jr. 1999a. Proliferative behavior of the murine cerebral wall in tissue culture: cell cycle kinetics and checkpoints. Exp Neurol 156:407 417. Takahashi T, Goto T, Miyama S, Nowakowski RS, Caviness VS Jr. 1999b. Sequence of neuron origin and neocortical laminar fate: relation to cell cycle of origin in the developing murine cerebral wall. J Neurosci 19:1035710371. Temple S, Qian X. 1995. bFGF, neurotrophins, and the control of cortical neurogenesis. Neuron 15:249 252.

722

Goto et al.
Doetschman T, Cofn JD, Wyland JJ, Hung YT. 1999. Changes in cerebral cortex size are governed by broblast growth factor during embryogenesis. Nat Neurosci 2:246 253. Weise B, Janet T, Grothe C. 1993. Localization of bFGF and FGF-receptor in the developing nervous system of the embryonic and newborn rat. J Neurosci Res 34:442 453. Yamamoto T, Kardami E, Nagy JI. 1991. Basic broblast growth factor in rat brain: localization to glial gap junctions correlates with connexin43 distribution. Brain Res 554:336 343. Zetterberg A, Larsson O, Wiman KG. 1995. What is the restriction point? Curr Opin Cell Biol 7:835 842.

Thomaidou D, Mione MC, Cavanagh JFR, Parnavelas JG. 1997. Apoptosis and its relation to the cell cycle in the developing cerebral cortex. J Neurosci 17:10751085. Tropepe V, Sibilia M, Ciruna BG, Rossant T, Wagner EF, Van der Kooy D. 1999. Distinct neural stem cells proliferate in response to EGF and FGF in the developing mouse telencephalon. Dev Biol 208:166 188. Vaccarino FM, Schwartz ML, Hartigan D, Leckman JF. 1995. Basic broblast growth factor increases the number of excitatory neurons containing glutamate in the cerebral cortex. Cereb Cortex 5:64 78. Vaccarino FM, Schwartz ML, Raballo R, Nilsen J, Rhee J, Zhou M,