The Journal of Neuroscience, April 1994, 74(4)- 2020-2031

Cogeneration of Neurons with a Unique Molecular Phenotype in Layers V and VI of Widespread Lateral Neocortical Areas in the Rat

Yasuyoshi Arimatsu, ltsuko Nihonmatsu, Kanako Hirata and Keiko Takiguchi-Hayashi Laboratory of Neuromorphology, Mitsubishi Kasei Institute of Life Sciences, Tokyo 194, Japan

Monoclonal antibody PC3.1 detects a unique subpopulation neurogenesis. To facilitate a better understanding of molecular of neurons located mainly in layer VI and, to a lesser extent, and cellular mechanisms underlying the early regional spccih- in layer V within the lateral neocortical areas in the rat. In cation. it is important first to know about the time and site of an attempt to characterize these neurons, we determined generation of these neurons and then to investigate the way by the time of their generation in selected neocortical areas by Lvhich cortical progenitor cells are committed to the specific a double-labeling experiment combining quantitative long- molecular phenotype. survival 3H-thymidine autoradiography and immunohisto- Neurons in the are distributed in six laminar layers chemistry for the PC3.1 antigen. We found that the vast and, within each layer, they tend to show similar cellular iden- majority of PC3.1 -positive neurons in both layers V and VI tities (Gilbert, 1953). The time of their generation (“birthday”) were generated concurrently at embryonic day 15 in all areas has been determined by autoradiographic detection of ‘H-thy- examined, demonstrating a strict correlation between the midine administered during the embponic period. It has been molecular identity of neurons and the time of their genera- well established that neurons ofthe neocortex in most mammals tion, irrespective of their final positions along the radial and tend to originate in an inside-to-outside sequence (for review. tangential axes. In contrast, PC3.1 -negative neurons, which see McConnell, 1988; lacobson, 199 1). Thus, earlier-generated should represent more diverse phenotypic identities, were neurons are located in a deeper layer and later-generated neurons generated during a more extended period of cortical devel- in a more superficial layer. opment and tended to exhibit radial (inside-to-outside) and Although it is possible, based on the intimate relationship tangential (ventral-to-dorsal and rostral-to-caudal) neuro- bctwecn cellular identity of neurons and their laminar fate, that genetic gradients. Our findings indicate that laminar and tan- the relationship between the time ofgeneration and laminar fate gential locations of cortical neurons are not established sole- is controlled by factors that regulate cell fate determination, it ly by a combination of mechanisms for the inside-out has not yet been established to what extent linkage is present movement of newly generated neurons in each cortical area between the time of generation and a specific cell‘s fate, either and for the broad tangential neurogenetic gradients. The for morphological. connectional, or molecular phenotypes. Since results of this study suggest a distinct way of cortical de- the PC3.1 -positi\ e neurons are located in different deep layers velopment in which neurons with a common molecular phe- in the neocortex (Arimatsu et al., 1992a), it would be possible notype are generated concurrently and migrate toward their to examine whether a strict linkage is indeed present between eventual positions, which are not necessarily located in a the time of their generation and the molecular phenotype. by single lamina. In addition, our results suggest some kind of comparing the time of generation in the different laminae. tangential heterogeneity in the mechanism involved in neo- Previous ‘H-thymidine birth-dating studies have also dem- cortical histogenesis, supporting the concept of early re- onstratcd tangential neurogenetic gradients when dealing with gional specification within the neocortex. ncocortical neurons as a single population. Thus, neurons in [Key words: neurogenesis, neocortex, claustrum, endo- rostra1 and ventral regions in the neocortex tend to originate piriform nucleus, regional specification, neuronal identity, carlier than those in caudal and dorsal regions, rcspectivcly. It neoronal migration, laminar fate, 3H-thymidine aotoradiog- is not known, however. whether these general tangential gra- raphy, monoclonal antibody] dients apply to each particular neuronal phenotype. Since PC3.1- positive neurons are distributed across widespread lateral neo- We have generated a monoclonal antibody, designated X3.1, cortical regions, it would also be possible to address this question that detects a neuronal entity confined to lateral. but not dor- by examining the time of generation along the rostral-caudal somedial, neocortical areas and to some other tclenccphalic and ventral-dorsal axes. regions in the rat (Arimatsu et al.. 1992a). By using this antibody We have previously shown that PC3.1-positive neurons in as a molecular marker, it was possible to show that a molecular layer VI of the parietal cortex, area 2 (Par2; = secondary so- phenotype restricted to the lateral neocortcx is specified early. matoscnsory area) are generated around embryonic day 15 (El 5; well bcforc thalamocortical interactions and even before cortical Arimatsu et al., 1992a). The primary purpose of the present study was to expand upon the previous experiment and to in- v-cstigate the time of generation of PC3.1 -positive neurons in Received May 11, 1993; revised Sept. 9, 1993: accepted Sept. 21. 1993. different deep laminae of various lateral neocortical areas in the Correspondence should bc addressed to Dr. Y. hrimatsu, Laborator) of Neu- rat. The time ofgeneration of PC3.1-positive neurons in some romorphology, Mitsubishi Kasei lnstrtute of Life Scxnccs, 11 Mmamiooya, Ma- chida-shi. Tokyo 194. Japan. cxtrancocortical regions was also examined and compared with Copqnght 81:’ 1994 S&et) for Ncurosclence 0270-6474.94 14X20-12$05.00 0 that of neocortical neurons. The Journal of Neurosclence, April 1994, 14(4) 2021

Part of this work has been presented in an abstract form Table 1. Number of animals used in a double-labeling experiment (Arimatsu et al., 1992b). for PC3.1 immunoreactivity and ‘H-thymidine incorporation

Materials and Methods Number of pups Experimental animals. Rats of Wistar strain were maintained under Number of mothers used for double- constant photoperiods (12 hr light/l2 hr dark; light on at 9:00 A.M.). Day of injection injected with labeling Pregnant rats were obtained by overnight mating. The day when a of ‘H-thvmidine 3H-thvmidine exoerimenta vaginal plug was observed in the morning was designated embryonic El1 1 2 (1) day 0 (EO). Birth usually occurred late on E21 (=postnatal day 0, PO). El2 Immunohistochemistry. The rats were deeply anesthetized with ether 2 4 (4) and fixed by cardiac perfusion with 4% paraformaldehyde in 0.1 M El3 2 4 (4) sodium phosphate buffer (pH 7.4). Their brains were removed and fixed El4 2 4 (1) further by immersion in the same fixative (4”C, 7 hr). Each block of El5 3 6 (4) brain was equilibrated in 30% sucrose in phosphate-buffered saline (PBS; pH 7.4), frozen in OCT compound (Lab-Tek), and serial coronal sections El6 2 4 (2) were cut on a cryostat at 10 Frn intervals. The sections were mounted El7 1 2 (1) onto gelatin/chrome alum-subbed glass slides and incubated sequen- El8 1 2 (1) tially with 5% normal goat serum in PBS containing 0.1% Triton X- 100 u The sex of the experimental animals was chosen randomly. The number of male (SNGS/T; 20 min), PC3.1 monoclonal antibody (13 &ml in SNGS/ rats in each injection group is shown in parentheses. T, 2 hr), goat anti-mouse IgG (Cappel; 1:50 dilution in SNGS/T, 2 hr), and mouse peroxidase-antiperoxidase (PAP; Jackson; 1:500 dilution in SNGS/T, 1 hr). The bound antibody was visualized by incubating the sections with 0.02% 3,3’-diaminobenzidine and 0.01% H,O, in 50 mM Tris-HCl, pH 7.6 (20 min). After rinsing in distilled water the sections were treated with 0.0 1% 0~0, to intensify the staining. To estimate the relative laminar density of PC3.1-positive Estimation of time of neuronal generation. Timed-pregnant rats were neurons, we counted the number of PC3.1-positive neuronsin iniected intraneritoneallv at 2:00 P.M. with 2H-thvmidine (New England separatelayers of selectedneocortical areas(Table 2). The ma- Niclear; 6.7 Ci/mmol, 1 Ci= 37 GBq; 0.5 mCijlO0 gm body weight). jority of them (68-91%) were found in layer VI in all areas The delivered pups were killed at P42, and cryostat sections of their brains were processed for immunohistochemical visualization of PC3.1- examined. In Vi and Par2 more PC3.1-positive neurons were antigen with PAP as described above. The animals used in the present found in the lower part of layer VI than in the upper one, but study are listed in Table 1. For each brain, 1 in 10 sections was dehy- in other areasthis tendency was not evident (Fig. 2). Substantial drated with sequential incubation in a series of ethanol (70-lOO%), numbers of PC3.1-positive neurons,although to a lesserextent, defatted in xylene (2 hr), and rehydrated in the same ethanol series down to water. The sections were air dried, dipped into autoradiographic were found also in layer V (7-29%). The positive neuronswere emulsion NR-M2 (Konica; diluted with an equal volume of distilled found at various depths of layer V and did not show any ten- water) at 43”C, and air dried. They were exposed in a desiccating box dency to be located near the border with layer VI (see Fig. (4”C, 4 weeks), developed with Konicadol-X (2o”C, 4 min), and fixed 2B,D,E). PC3.1-positive neurons in the upper layers or in the with Konicafix (2o”C, 5 min). A series of sections adjacent to those used white matter were very low in number (~3%; Fig. 30). for the double labeling was stained with cresyl violet. To estimate the time of generation of PC3.1 -positive neurons, the Although the morphological characteristicsof PC3.1-positive autoradiographic signal was quantified by using a Vidas image analysis neuronswere not determined in detail in the presentstudy, many system (Zeiss/Kontron) equipped with a light microscope (Zeiss Axio- of them were pyramidal, multipolar, or bipolar in shape(Fig. stop) with a motor-driven scanning stage. At first, the image of a low- 3). We frequently observed PC3. l-positive neurons which re- power view of a section was transferred to the screen of a video monitor at a 5 x objective and a scanning area was outlined interactively. Then, sembled “atypically oriented” or “inverted” pyramidal cells using a 40 x objective, positions of PC3.1 -positive neurons were stored (Miller, 1988a; Bueno-L6pez et al., 199 1) with their apical den- in a data file by marking the neurons interactively. Finally, using a 100 x drite extending toward various directions (toward the white objective, the number of autoradiographic silver grains was counted matter, and parallel or oblique to the white matter; seeFig. 3) interactively over the nucleus of a PC3.1 -positive neuron on the screen. as well as “typical” pyramidal cells with their apical dendrites The scanning stage moved automatically to the next registered position after each counting. The number of silver grains over PC3.1 -negative toward the pia matter. neurons was also counted. Positions of clusters of from medium to large In addition to that of the neuronal cell bodies and dendrites, numbers of silver grains (approximately 20 or more) were identified on immunostaining was observed in someother neural structures. the screen at the 40x objective and exact grain counting was then Although most ofthe fiber tracts in the forebrain were not PC3.1 performed using the 100 x objective as described above. immunopositive, some axon fibers in the cortical white matter (Figs. 2A, 30) and the fomix (not shown) were substantially Results immunostained. The PC3.1-positive axon fiberswere especially Distribution of PC3.1 -positive structures in the forebrain evident at the rostrocaudal level near Cl where both neuronal Monoclonal antibody PC3.1 recognized a neuronal subpopu- cell bodies and the neuropil were strongly immunostained (Fig. lation located in a lateral sector of the neocortex (Arimatsu et 4). This may suggestthe presenceof PC3.1-positive claustro- al., 1992a).This included the parietal cortex, area 2 (Par2; Figs. cortical and/or corticoclaustral axon fibers, although a definite lA, 2B, 3,4-D), visceral sensory area (Vi; Fig. 2A), temporal conclusion on this point remains to be determined. In some cortex, area 1 (Tel; Fig. 2C), area 2 (Te2; Fig. 2D), and area 3 regions (e.g., the mammillary body and basolateralamygdaloid (Te3; Figs. 2E, 3E,F), and occipital cortex area2, lateral (Oc2L; nucleus), from weak to moderate staining was seenin the neu- Figs. 2F, 3G). Immunopositive neuronswere also found in the ropil. In the neocortex, weak neuropil staining was seenin layer lateral periallocortex including the agranular insular (AI; Fig. IV and other layers (seeFig. lA), even in regionswithout PC3.1- 4.4), lateral orbital (not shown), and perirhinal (not shown) ar- positive neuronalcell bodies.In most ofthe forebrain structures, eas,in the claustrum (Cl; Fig. 4A) and in the dorsal and ventral including the neocortex, hippocampus, amygdala, caudate pu- endopiriform nuclei (DEn and VEn; Fig. 4A). There were some tamen, olfactory bulb, hypothalamus, and thalamus, weakly PC3.1-positive neuronsin the presubiculumas well (not shown). stained astroglial cells were seen(Fig. 3H). 2022 Arimatsu et al. l Cogeneration of Cortical Neurons with a Molecular Phenotype

t

*

Figure 1. Distribution of PC3.1-im- munopositivecells in the parietalcor- tex, area2 (Par2).A, A coronalsection of a rat brainthrough Par2 stained with PC3.1monoclonal antibody. B, An ad- jacentsection stained with cresylviolet. Roman numerals in Figures1 and 2 indicatecortical layers.CPU, caudate putamen; WM, white matter. In the present study, area1 and laminar boundariesof the cortex were deter- minedaccording to Paxinosand Wat- son (1986), Zilles (1985, 1990), and Zilleset al. (1990).Scale bar, 100pm.

When assumingthat the generation time of cortical neurons is Time of generation of PC3.1-positive and -negative neurons in 11-19 hr (Waechter and Jaensch, 1972), and that the number the neocortex of silver grains in many cells should have been underestimated The time of generation was estimated separately for PC3.1- due to various factors including the timing of the ‘H-thymidine positive and -negative neuronsby a double-labelingexperiment injection during their cell cycle and the variation of nuclear size combining immunohistochemistry for PC3.1-antigen and ‘H- and the site of sectioning within the nucleus (for detailed dis- thymidine autoradiography (Fig. 5). Within the neocortex, the cussion, see Hicks and D’Amato, 1968; Sidman, 1970; Bayer time of generation was analyzed in five selectedareas, the ap- and Altman, 199la), a neuron with more than one-fourth of the proximate locations of which are shown in Figure 6. We con- maximum number of grains is most likely to have been gen- sidered a neuron as being heavily labeled when the number of erated within 24 hr following the injection of 3H-thymidine autoradiographic silver grains was 40 or greater. This criterion (Cavanagh and Parnavelas, 1988, 1990). We therefore deter- was based on the observation that the maximum number of mined the percentageof the heavily labeled cells (with at least silver grains was 159 (seeFig. 7 for relative frequency of the 40 grains) for each injection group for estimation of relative number of silver grains on a single cell). Any cells with more proportion of neurons originating within 24 hr following the than half the maximum number of silver grains can be consid- injection. Since heavily labeledneurons were found in injection ered as onesthat underwent only one round of DNA synthesis groups between El 2 and El 7 but not in animals injected at and cell division following the injection of ‘H-thymidine (Sid- either E 11 or E 18, most, if not all, of the PC3.1-positive neurons man, 1970; Rakic, 1974), and cells with more than one-fourth should have been generated between El 2 and El 7. Thus, we of the maximum should not have divided more than twice. estimatedthe proportion of PC3.1-positive neuronsoriginating The Journal of Neuroscience, April 1994, M(4) 2023

Figure 2. Distribution of PC3.1 -positive cells in various lateral neocortical areas. A, Visceral sensory area (Vi). B, Parietal cortex, area 2 (Par2). C, Temporal cortex, area 1 (Tel). D, Temporal cortex, area 2 (Te2). E, Temporal cortex, area 3 (Te3). F, Occipital cortex, area 2, lateral (0~2~). The arrowhead in A indicates PC3.1 -positive axon fibers. Scale bar, 50 pm. on a particular day by dividing the percentageof heavily labeled generatedin an inside-to-outside sequencein both Te3 and Par2 neuronsfor each injection group by the sum of the percentages (Fig. 8C,D). obtained between El 2 and El 7 (Table 3). The proportion of Time of generation of neurons in dtxerent neocortical areas. PC3.1-negative neuronsoriginating was calculated by dividing To compare the time of generation of PC3.1-positive neurons the density of heavily labeledneurons for each injection group acrosswidespread lateral neocortical areas, the percentageof by the sum ofthe densitiesobtained betweenE 12 and El 7 (Table PC3. l-positive neurons with heavy )H-thymidine labeling was 4). examined in layer VI of three additional areas (Vi, Tel, and Time ofgeneration of neurons in d$erent neocortical laminae. Oc2L) and the data were combined with thosedescribed above. In Te3 and Par2, the time of generation of PC3.1-positive neu- As shown in Figure 9, the proportion of PC3.1-positive neurons rons was compared between layers V and VI. In both areas, originating was at its highest level at El5 for all five areasex- either in layer V or in layer VI, the vast majority of PC3.1- amined, whereas PC3. l-negative neurons in the same layer positive neuronswere generatedconcurrently at E 15 (Fig. 8A,B). tended to show rostral-to-caudal and lateral-to-dorsal neuro- In contrast, PC3.1-negative neurons in layers V and VI were genetic gradients. The peak day of neurogenesisfor PC3.1-neg- 2024 Arimatsu et al. l Cogeneration of Cortical Neurons with a Molecular Phenotype

Figure 3. Higher magnification of PC3.1 -positive cells. A, Par2, upper layer VI. An arrow indicates an inverted pyramidal cell. B, Par2, upper layer VI. Arrows indicate atypically oriented pyramidal cells. C, Par2, lower layer VI. D, Par2, bottom of layer VI and the white matter. The arrow and arrowheads indicate a PC3.1 -positive neuron and PC3.1 -positive fibers, respectively, in the white matter. E, Te3, layer V. F, Te3, layer VI. G, Oc2L, layer VI. H, Caudate putamen. Arrows indicate weakly stained astrocytes: In A-G, the top of each photomicrograph corresponds to the upper side of the neocortex (side of the pia mater). Scale bar, 50 pm.

Figure 4. Distribution of PC3.1 -immunopositive cells in the claustrum (Cl), endopiriform nuclei (En), and agranular (AI). A coronal section of a rat brain stained with PC3.1 monoclonal antibody (A) and its adjacent section stained with cresyl violet (B) are shown. CPU, caudate putamen; DEn, dorsal endopiriform nucleus; Pir, piriform cortex; VEn, ventral endopiriform nucleus. PC3.1 -positive neurons in Cl and those in the lateral part of DEn at a higher magnification are shown in insets of A. Scale bars: A and B, 100 pm; inset, 50 pm.

2026 Arimatsu et al. * Cogeneration of Cortical Neurons with a Molecular Phenotype

Table 2. Laminar density of PC3.1-positive neurons in lateral neocortical areas”

Lamina Vi Par2 Tel Te3 Oc2L Layer II/III, IV 0.6 rt 0.2 (2.0%) 0.8 t 0.3 (1.3%) 0.9 + 0.3 (2.8%) 1.1 + 0.3 (2.3%) 0.0 t 0.0 (0.0%) Layer V 1.9 f 0.3 (6.6%) 5.7 f 0.9 (9.3%) 3.0 zk 0.3 (9.3%) 13.9 + 1.0 (29.4%) 1.6 * 0.2 (9.2%) Layer VI 26.5 + 1.2 (91.4%) 53.7 f 2.4 (88.1%) 28.3 + 2.2 (87.9%) 32.1 f 3.3 (67.9%) 15.7 + 1.2 (90.8%) White matter ND* 0.8 + 0.2 (1.3%) 0.0 t- 0.0 (0.0%) 0.2 + 0.1 (0.4%) 0.0 f 0.0 (0.0%) Total 29.0 + 1.3 61.0 f 2.6 32.2 + 2.3 47.3 f 3.8 17.3 + 1.2 The number of PC3.1-positive neurons was counted in a vertical strip having a width of 0.56 mm (for Vi) or 0.98 mm (for Par2, Tel, Te3, and Oc2L). Each value is normalized to represent cell number per O.S-mm-wide cortical strip (mean k SEM, tz = 12). ~1The relative laminar density is shown in the parentheses. b Since the white matter in the Vi region is very thin, the number of PC3. l-positive neurons was not determined (ND).

ative neurons was at El 3-El4 for Vi, at El4 for Par2 and Te3, and at El 5 for Te 1 and Oc2L. The total generation time for Discussion PC3.1-positive neurons was shorter than that for PC3.1-nega- In the rat neocortex, PC3. l-positive neurons have been found tive neurons. Thus, PC3. l-positive neurons in layer VI were only in lateral (but not dorsomedial) areas (Arimatsu et al., generatedwithin 4 d whereasPC3.1 -negative neurons emerged 1992a). In the present study, we quantitatively showed that throughout 5 d. About 60% (55-64%) of the PC3.1-positive while the majority (68-91%) of PC3.1-positive neurons in the neuronsin layer VI were generatedat El 5 while only about 45% neocortex were located in layer VI, substantialportions (7-29%) (42-50%) of the PC3.1-negative neurons in the same layer ap- were found in layer V. We demonstratedthat neocortical PC3.1- peared at the peak day of their generation. positive neurons were generated concurrently, irrespective of their eventual positions along the radial and tangential axes. Time of generation of PC3.1 -positive neurons outside the neocortex Cogeneration of neocortical PC3. I -positive neurons in layers V The time of generation of PC3.1-positive neuronsin Cl and of and VZ those in the endopiriform nuclei (En; =DEn + VEn) was ex- It has been establishedthat there is a fairly strong correlation amined at the level shown in Figure 6 (between bregma -0.92 between the laminar fate of neurons and the time of their gen- mm and bregma -0.32 mm; Paxinos and Watson, 1986). The eration in the neocortex of various species(Angevine and Sid- majority of PC3.1-positive neurons in these regions were gen- man, 1961; Berry and Rogers, 1965; Hicks and d’Amato, 1968; erated earlier than those in the neocortex. The proportion of Shimada and Langman, 1970; Rakic, 1974; Bruckner et al., neuronsoriginating was at its highest level at E 14 for Cl and at 1976; Bayer and Altman, 199 1a). Thus, earlier-generatedneu- E13-El4 for En (Fig. 10). rons tend to be located in a deeper layer and later-generated

Figure 5. Double labeling for PC3.1- immunoreactivity and ‘H-thymidine incorporation. This is a section through the Oc2L of a rat injected with )H-thy- midine at E15. Cells 1-3 are positive for PC3.1. Cell 1 is moderately labeled for ‘H-thymidine incorporation (num- ber of silver grains, 36). CeU2 is heavily labeled (number of silver grains, 79). Cell 2 is not labeled above the back- ground level. Arrowheads indicate clus- ters of silver grains on PC3.1 -negative neurons. Scale bar, 20 pm. The Journal of Neuroscience, April 1994, 14(4) 2027

9 El5 2 3 600 z + 3 400 a 0” 5 200 Par2 Te3 s hgur~~ 6. Schematic illustration of a lateral surface vie\% of the rat forebrain showing cortical areas containing PC3.1 -positive neurons z (shaded). Approximate locations at which neurogenesis was analyzed 0 are shown by solid circles (for neocortical arcas), a rcctat& (for Cl). 0 40 80 120 and a trmng/e (for En). Enr, : Fr, frontal cortex; LO. NUMBER OF GRAINS lateral orbital cortex: 0~1, Occipital cortex, area 1; Purl, parictal cortcx~ area 1; PRh. . Figure 7. Number of autoradiographic silver grains on single PC3.1- positive neurons. Each value represents the frequency of occurrence of neurons in a tnore superficial layer. The exceptions are the neu- PC3. l-positive neurons with the number of silver grains indicated on rons in layer I, which are generated during the earliest phase of the abscissa (O-4, 5-9, 10-14, . 155-l 59). The data were deriv-cd from 42 17 PC3.1 -positive neurons in 110 neocortical sections from a neocortical development. There is, houever, substantial radial total of six rats that had received ‘H-thymidine at E15. spread of simultaneously generated neurons. In the rat Par?; for example, the majority of neurons gencratcd at El3 are destined to settle in layer VI but others are found in layer V. while most identity rather than the laminar position in the adult. It is pos- neurons generated at E 15 are in layer V and. to a lesser extent, sible to speculate that neurons with heterogeneous birthdays in in layers IV and VI (Bayer and Altman, 1991a). The radial a particular lamina represent. at least in part, certain diverse spread has also been observed in other species (monkey, Rakic. phenotypic populations, each of which are generated concur- 1974; cat, Luskin and Shatr. 1985: ferret, Jackson et al., 1989: rentl). The radial spread of each population with a single birth- hamster, Lent et al.. 1990). The present study rev~ealed that the day might vary from one to another, thus contributing to the vast majority of PC3.1-positive neurons in either layer V or extent of the radial migratory spread in general. This view is in layer VI are generated concurrently at El 5, indicating that the accordance with the observation that. even within a single layer, birthday of neurons is linked more strictly with the phenotypic

Table 4 Neurogenesis of PC3.1-negative neurons in Te3 Table 3. Neurogenesis of PC3.1-positive neurons in Te3 Number of Number of Embry- Number of heal-ill labeled % Cells Embryonic sections o/o Heavily % Cells onic sections cells per 0.1 origi- dav Layer observed labeled cells originating dav Layer observed mm”” natinn

El1 V 4 0.0 (107) 0.0 El1 V 4 0.0 (0.94) 0.0 VI 3 0.0 (250) 0.0 VI 4 0.0 (1.10) 0.0 El2 V 12 0.0 (210) 0.0 El2 V 8 0.0 (1.81) 0.0 VI 8 0.0 (395) 0.0 VI 8 0.8 (2.09) 1.9 El3 V 11 0.0 (23 1) 0.0 El3 V 8 1.1 (1.80) 4.0 VI 8 1.6 (371) 4.6 VI 8 10.7 (1.97) 25.4 El4 V 12 2.5 (283) 1.2 El4 V 8 7.4 (1.58) 26.1 VI 8 3.1 (360) 8.8 VI 8 20.9 (1.93) 49.1 El5 V 19 24.1 (514) 70.3 El5 V 12 15.3 (2.77) 53.8 VI 13 2 I .O (685) 60.2 VI 13 7.7 (3.18) 18.4 El6 V 12 7.7 (285) 22.5 El6 V 8 4.5 (1.73) 15.7 VI 8 9.2 (445) 26.4 VI 8 2.0 (2.14) 4.7 El7 V 6 0.0 (156) 0.0 El7 V 4 0.1 (0.90) 0.4 VI 3 0.0 (253) 0.0 VI 4 0.0 (1.12) 0.0 El8 V 4 0.0 (111) 0.0 El8 V 4 0.0 (0.94) 0.0 VI 4 0.0 (299) 0.0 VI 4 O.O(l.13) 0.0

The number of PC3.1.positive neurons wth heav) ‘H-th>midine labeling was The number of PC.l-negative cells with heavy ‘H-thymidme labeling was ex- examined in 0.98.mm-nide vertical strips of Te3. The percentage of cells ongi- amined in an area of0. I SO.28 mm’ per section. The percentage ofcells originating natlng at each embryonic day nas calculated by dlvldmg percentage of heavily at each embryonic da) nab calculated by dividing the number of heavily labeled labeled cells m each inJectloo group b> the sum of percentages obtamed between cells m each mJectmn group by the sum of the number obtained between El 2 and El2 and El7 (34.3% for la)er V, 34.9% for layer VI). El 7 (28.4 for layer V, 42. I for layer VI). ‘ The total number of PC? I -poslti\ e cells counted is shown in the parentheses. 1 The total area (mm’) examined is shown in the parentheses 2028 Arimatsu et al. * Cogeneration of Cortical Neurons with a Molecular Phenotype

+ LAYER Q 80 a’ B --*-- LAYER “I

fjo PC3.1+ :I’ : Par2 : : : ( 40

20 : ‘$3 ,/’ 0 h 11 12 13 14 15 16 17 16 11 12 13 14 15 16 17 16

& LAYER Q + LAYER ” C --a-- LAYER “I “1 D --*-- LAYER Q PC3.1-

11 12 13 14 15 16 17 18 11 12 13 14 15 16 17 16 DAY OF ORIGIN DAY OF ORIGIN

Figure8. Time ofgeneration of PC3. I -positive (A and B) and -negative (C and D) neurons in layers V and VI. A and C, Te3. B and D, Par2. Source data for Te3 are shown in Tables 3 and 4. The proportion of neurons originating in Par2 was calculated in a similar way as that for Te3. Since the density of PC3.1 -positive neurons in layer V of Par2 was low (see Table 2), about twice as many sections were examined for Par2 as for Te3 in each injection group.

11 12 13 14 15 16 17 18 PC3.1-positive neuronsare cogeneratedduring a more restricted period than are PC3.1-negative neurons,which should represent DAY OF ORIGIN more diverse phenotypic identities in that layer. Figure 9. Time of generation of PC3. l-positive (A) and -negative (B) Several lines of evidence suggestthere being an intimate cor- neurons in layer VI of five different lateral neocortical areas. Data were relation between the time of generation of cortical neuronsand obtained in a similar way as that for Te3 (Tables 3,4) except that PC3. l- their morphological identities. (1) In the reeler mutant mouse, positive neurons in Vi were examined in vertical strips 0.56 mm wide. although migration of young cortical neuronsis disrupted and thus the postmigratory positions of neuronsare abnormal, cor- ent study, we have provided convincing evidence for a strict respondingcell classesare still generatedin the same sequence linkage between the time of generation of cortical neuronsand (Caviness, 1982). (2) The neuronal migration is abnormal in their certain phenotypic identity by usingthe specificmolecular rats exposed to x-rays during their fetal stage. Even in these marker PC3.1. Although physiological functions of the PC3.1 animals,at least certain numbers of corticospinal neuronswere antigen are not yet known, there is evidence that it is a 29 kDa generated appropriately (Jensenand Killackey, 1984). (3) Ac- protein with an amino acid sequencethat has not been reported cording to recent reports, albeit from separatelaboratories, the thus far (Arimatsu et al., 1992a; Y. Hatanaka, Y. Uratani, K. corticocortical and corticotectal neuronsin the rat visual cortex Takaguchi-Hayashi, A. Omori, K. Satoh, M. Miyamoto, and may be generatedduring rather different embryonic days (Miller Y. Arimatsu, unpublished observations). 1985, 1988b; Contamina and Boada, 1992). Despite this cir- Although PC3.1-positive neurons in layer V represent only cumstantial support for a strict phenotype-birthday correlation, about 10% of the total PC3. l-positive neurons in Par2, those however, it is not yet fully known which kinds of morphological in Te3 representabout 30% of the total. Thus, it seemsunlikely and molecular phenotypes are actually correlated with the time that the apparent cogeneration of PC3.1-positive neurons in of generation. Previous studieshave revealed that various neu- layer V with those in layer VI was derived from various kinds rotransmitter phenotypes of cortical neurons are not related to of “fluctuation” or “inaccuracy” in the mechanismregulating their time ofgeneration (Cavanagh and Pamavelas, 1988, 1989, the formation of the inside-out laminar structure. It is, however, 1990; Miller, 1988b, 1992; Peduzzi, 1988). Neocortical GA- not known what is the way by which fractions of a single pop- BAergic neuronsin the rat, for example, are generatedthrough- ulation that are generatedconcurrently are delivered to separate out the entire period ofcortical neurogenesis(Miller, 1985,1988b; layers. Fairen et al., 1986). None ofthe subtypesofGABAergic neurons classifiedaccording to various morphologicalcharacteristics (e.g., Cogenerationof neocortical PC3.1-positive neuronsacross shapeand size of cell body, pattern of dendritic arborization) widespreadareas have been correlated with distinct birthdays (Cobasand Fairen, Another striking finding of the present study is the cogeneration 1988). Furthermore, it has been demonstratedthat both typical of PC3.1-positive neurons in tangentially widespread cortical and atypically oriented pyramidal neurons are generated con- regions. Previous ‘H-thymidine birth-dating studies have re- currently in an inside-out sequence(Miller, 1988a). In the pres- vealed broad neurogeneticgradients along rostral-to-caudal and The Journal of Neuroscience, April 1994, 14(4) 2029 ventral-to-dorsal axes in the rodent neocortex. Thus, neurons ao- in a particular lamina tend to be generated earlier in rostra1 and i2 PC3.1+ - Cl ventral regions than in caudal and dorsal ones, respectively - - En (Gardette et al., 1982; Smart and Smart, 1982; Miller, 1987; Bayer and Altman, 199 la). However, the tangential neuroge- netic gradients should not be complete enough to include all the neocortical neurons. We show in the present study that the vast majority of PC3.1 -positive neurons in layer VI are cogenerated at E 15, irrespective of their tangential locations. This is in con- trast with the observation that PC3.1 -negative neurons, repre- senting the major portion of neurons in the same layer, showed tangential neurogenetic gradients. According to a widely accepted assumption, most if not all young cortical neurons migrate radially, either strictly or through ii 12 13 14 15 16 17 ia a modified way, from the ventricular zone to their final desti- nation (Rakic, 1988; Misson et al., 1991). If this is applicable DAY OF ORIGIN for both PC3.1 -positive and -negative neurons, the vast majority Figure 10. Time of generation of PC3.1-positive neurons in Cl and of PC3.1 -positive neurons in layer VI of the rostra1 and ventral En. PC3. I -positive neurons in Cl and En (=DEn + VEn) were examined for heavy ‘H-thymidine labeling in four (E 11, E I 7) or eight (E 12-E 16) areas (Par2, Te3) are likely to be cogenerated at El5 together sections from each injection group between bregma -0.92 mm and with the majority of PC3.1-negative neurons in layer V, and - 1.30 mm (Paxinos and Watson, 1986). At this level (approximate thus the migratory movement of these two populations should location is shown in Fig. 6), 102.7 i 4.4 and 88.8 i 3.6 (both mean be different along the radial axis. On the other hand, PC3.1- t SEM, n = 48) PC3.1 -positive neurons were found in Cl and En, respectively, in one side of each section. Values for Par2 are also shown positive neurons in layer VI of the more caudal and dorsal areas as a reference. (Te l,Oc2L) are likely to be cogenerated at El 5 with the majority of PC3.1-negative neurons in the same layer and might comi- grate toward their final positions. Thus, there might be certain of thosein En are in good agreementwith the reported birthdays kinds of tangential heterogeneity that are involved in the mi- of neuronstreated asa singlecell population (Bayer and Altman, gration of PC3.1 -positive neurons. 1991 b). The time of generation of the PC3.1-positive neurons It might also be possible that most of the presumptive PC3.1- in Cl and En being earlier than that in the neocortex indicates positive layer VI neurons are generated in a restricted sector of that theseextraneocortical neuronsare developmentally distinct the ventricular zone and migrate more extensively in tangential from the neocortical ones. directions than the presumptive PC3.1 -negative neurons. The Previously, we reported that PC3.1-positive neurons were results of recent lineage tracing experiments using retroviral presentjust outside DEn (Arimatsu et al., 1992a).In the present vectors or an embryonic stem cell line as a tracer showed that study, we examined the exact localization of PC3.1-positive some immature neurons indeed migrate across widespread cor- neuronsin the portion deep to the piriform cortex by comparing tical areas (Austin and Cepko, 1990; Nakatsuji et al., 199 1; side by side the adjacent sectionsstained with PC3.1 and cresyl Walsh and Cepko, 1992, 1993). With sequential 3H-thymidine violet, respectively, and confirmed that the central core of DEn autoradiography, Bayer et al. (199 1) suggested the presence of was mostly devoid of PC3.1-positive neurons. We found, how- a prominent migratory path, that is, lateral cortical stream, fol- ever, that the majority of PC3.1-positive neuronswere located lowed by neurons migrating to the lateral and ventrolateral cor- within the lateral portion of DEn in addition to those outside tical plate. Furthermore, tangential dispersion of fluorescently the nucleus corresponding to VEn (deolmes et al., 1985; see labeled cells was directly observed in an explant of developing Fig. 4A,B). It is thus evident that DEn can be divided into two using time-lapse confocal microscopy, although distinct subregionswith or without PC3.1-positive neurons. In their exact cell type (e.g., neuronal or glial) was not known this connection, it may be important to note that DEn was (O’Rourke et al., 1992; Fishell et al., 1993). When considering reported to be a heterogeneousnucleus with respectto its affe- the recent report that both radial mosaicism and tangential cell rent projections (Luskin and Price, 1983). A recent study on migration contribute to neocortical development (Tan and Breen, the distribution of phosphatase inhibitor- 1-immunoreactive 1993) it would be of great interest to examine whether or not neurons has also demonstratedthat the lateral and medial sec- PC3.1 -positive neurons represent a specific population that mi- tors of the nucleus were distinct (Gustafson et al., 1991). It is grate extensively toward tangential directions. Whether or not of great interest that the distribution pattern of PC3.1-positive extensive tangential migration of presumptive PC3.1 -positive neurons was strikingly similar to that of phosphataseinhibitor- neurons occurs, we would still conclude that the developing I-immunoreactive neurons in the regions of the lateral neo- cortical anlage is not homogeneous along the tangential axes in cortex, Cl and En. the mechanism of neocor-tical histogenesis. Implication fbr neocortical development Generation of PC3. I -positive neuronsin Cl and En Recently, we demonstrated the selective appearanceof PC3.1- The present study demonstratesthat the majority of PC3.1- positive neurons in cultures derived from a lateral, but not positive neurons in Cl and those in En (DEn + VEn) were dorsal, sector of neocortical tissue at El2 or later, suggesting generatedat El 4 and E 13-E 14, respectively, at least at the level that somekind of neocortical regional specification occurs very examined. This is in contrast with the cogeneration at El5 of early along the dorsal-ventral axis (Arimatsu et al., 1992a). The neocortical PC3.1-positive neurons across widespread loca- tangential heterogeneity of the developing neocortex predicted tions. The time of origin of PC3.1-positive neurons in Cl and from the cogeneration of PC3.1-positive neurons further sup- 2030 Anmatsu et al. * Cogeneration of Cortical Neurons wth a Molecular Phenotype ports the concept of early regional specification within the neo- projection neurons in the rat visual cortex. J Comp Neurol 323:570- cortex. 576. deOlmes J, Alhcide GF, Beltramino CA (1985) Amygdala. In: The Although it may be premature to discuss the timing and fac- rat ner\ ous system. Vol I, forebrain and midbrain (Paxinos G, ed). tors that contribute to the cell fate determination of PC3.1- pp 223-334. Sydney: Academic. positive neurons, it is important to note that it is now possible Falren A, Cobas A. Fonseca M (1986) Times ofgeneration of glutamic to test, both in vivo and in vitro, whether some of the newly acid decarboxylase immunoreactive neurons in mouse somatosensory postmitotic neurons at E 15 are already committed to the specific cortex. J (‘amp Neurol 25 1:67-83. Fishell G. Mason CA. Hatten ME (1993) Dispersion of neural pro- molecular phenotype. LJsing a heterochronic transplantation genitors within the germinal zones of the forebrain. Nature 362:636- technique. hlcConnel1 and Kaznowski (199 1) suggested that cor- 638. tical progenitor cells in the fcrrct are initially multipotent but Gardette K, Courtoris M, Bisconte JC (1982) Prenatal development are committed to a specific laminar fate by environmental fac- of mouse central nervous structures: time of origin and gradients of tors acting at around their fmal round of cell division. Based on neuronal production. A radioautographic study. J Himforsch 23:4 l5- 431. the strict birthday cell fate relationship dcmonstratcd in the Gilbert CD (1983) Microcircuitry of the visual cortex. Annu Re\ present study, it is tempting to speculate that the commitment Neurosci 612 17-247. of cells to the PC3.1 phenotype occurs around El 5, depending Gustafson EL, Girault J-A, Hemmings HC Jr, Naim AC, Greengard P on their interactions with their environments. Howe\.er, it is (I 99 1) Immunocq tochemical localization of phosphatase inhibitor- 1 in rat brain. J Comp Neurol 310:17-188. also possible that such a commitment occurs earlier in some Hicks SP, D’Amato CJ (1968) Cell migration to the isocortex in the progenitor cells in the lateral sector of the ventricular zone. rat. Anat Ret 160:6 19-634. These two alternative scenarios can be examined experimen- Jackson CA, Peduzzi JD, Hickey TL (1989) Visual cortex development tally, and the resultant information would shed new light on the in the ferret. I. Genesis and migration of visual cortical neurons. J mechanism of neocortical development. Neurosci 9:1242-1253. Jacobson M (199 1) DC\ elopmental neurobiology. 3d ed. New York: Plenum. References Jensen KF, Killackey HP (1984) Subcortical projections from ectopic Angevine JB. Sidman RL (1961) Autoradiographic study of cell mi- neocortical neurons. Proc Nat1 Acad Sci USA 8 1:964-968. gration during histoeenesis of cerebral cortex of the mouse. Nature Lent R, Hedin-Pereira c‘, Menezes JRL. Jhaveri S (1990) Neurogcnesis i92:766-168.- _ and development ofcallosal and intracortical connections in the ham- Arimatsu Y, Miyamoto M, Nihonmatsu I, Hirata K, Uratani Y, Ha- ster. Neuroscience 38:21-37. tanaka Y. Takiguchi-Hayashi K (I 992a) Early regional specification Luskin MB, Price JL (1983) The topographic organization of associ- for a molecula; neuronal phenotype in the ra; ncocortex. Proc Nat1 ational fibers of the olfactory system in the rat, including centrifugal Acad Scl USA 89:8879-8883. fibers to the olfactor) bulb. J Comp Neurol 2 16:264-29 1. Arimatsu Y. Nihonmatsu 1, Hirata K, Takiguchi-Hayashi K. Milamoto Luskin MB. Shatz CJ (1985) Neurogenesis of the cat’s primar) visual M (1992b) Generation of neocortical layer 6 neurons with region- cortex. J Comp Neural 242:6 1 l-63 1. specific PC3.1 antigen in the rat. Neurosci Res [Suppl] 17:s 16 1. McConnell SK (1988) Dcvclopmcnt and decision-making in the mam- Austin CP, Cepko CL (1990) Cellular migration patterns in the de- malian cerebral cortex. Brain Res Kev 13: l-23. veloping mouse cerebral cortex. Development 1 lo:71 3-732. McConnell SK. Kaznowski CE (199 1) Cell cycle dependence of lam- Bayer SA. Altman J (1991a) Neocortical development. New York: inar determination in developing neocortex. Science 254:282-285. Raven. hliller MW (1985) Cogeneration of retrogradely labelled corticocorti- Baver SA. Altman J (199 1 b) Development of the endopiriform nucleus cal projection and GABA-immunoreactive local circuit neurons in and the claustrum in the rat brain. Neuroscience 45:391-412. cerebral cortex. Dcv Brain Res 23: 187-l 92. Baver SA. Altman J. Russo RJ. Dai X. Simmons JA (1991) Cell Miller MW (I 987) Effect of prenatal exposure to ethanol on the dis- kigration in the rat embryonic neocortex. J Comp Neurol 307:499- tribution and time of origin of corticospinal neurons in the rat. J 516. Comp Ncurol 257:372-3X2. Berry M, Rogers AW (1965) The migration of neuroblasts in the Miller M W (1988a) Maturation of rat visual cortex. IV. The genera- developing cerebral cortex. J Anat 99:691-709. tion, migration. morphogenesis and connectivity of atypically ori- Bruckner G. Mares V, Biesold D (1976) Neurogenesis in the visual ented pyramidal neurons. J Comp Ncurol 273:387405. system of the rat. An autoradiographic in\-estigation. J Comp Neurol Miller Gw (1988b) Development of projection and local circuit neu- 166~245-256. rons in ncocortex. In: Cerebral cortex. Vol 7 (Jones EG, Peters A, Bueno-Lbpez JL, Reblet C, Lbpez-Mcdina A, G6mez-llrquijo SM, eds), pp 133-l 75. New York: Plenum. Grandes P, Gondra J, Hennequct L (1991) Targets and laminar Miller MW (1992) Migration of peptidc-immunoreactive local circuit distribution of projection neurons with ‘inverted’ morphology in rab- neurons to rat cingulate cortex. Cereb Cortex 2:444455. bit cortex. Eur J Neurosci 3:415430. Misson J-P, Austin CP. Takahashi T. Cepko CL, Cavmess VS Jr (199 1) Cavanagh ME. Pamavelas JG (1988) Development of somatostatin The alignment of migrating neural cells in relation to the murine immunoreactive neurons in the rat occipital cortex: a combined im- neopallial radial glial fiber system. Ccreb Cortex 1:221-229. munocytochemical-autoradiographic study. J Comp Neural 268: l- Nakatsuji N, Kadokawa Y, Suemori H (199 1) Radial columnar patch- 12. es in the chimeric cerebral cortex visualized by use of mouse cmbry- Cavanagh ME. Pamavelas JG (1989) Development of vasoactlve in- onic stem cells expressing P-galactosidase. Dev Growth Differ 33: testinal-polypeptide immunoreacti\c neurons in the rat occipital cor- 571-578. tex. A combined immunocytochemical-autoradiographic study. J O’Rourke NA, Dailey ME, Smith SJ, McConnell SK (1992) Diverse Comp Neurol 284:637-645. migratory pathways in the developing cerebral cortex. Science 258: Cavanagh ME, Pamavelas JG (1990) Development of neuropcptide 299-302. Y (NPY) immunoreactive neurons in the rat occipital cortex: a com- Paxinos G, Watson C (1986) The rat brain in stereotaxic coordinates, bined immunocytochemical-autoradiographic study. J Comp Neural 2d ed. North Rydc: Academic. 297~553-563. Peduzzi JD (1988) Genesis of GABA-immunorcactive neurons in the Caviness VS Jr (1982) Neocortical histogenesis in normal and reeler ferret visual cortex. J Neurosci X:920-93 1. mice: a developmental study based on [‘Hlthymidine autoradiogra- Rakic P (1974) Neurons in rhesus monkey visual cortex: sy-stematic phy. Dcv Brain Res 4:293-302. relation between time of origin and eventual disposition. Science 183: Cobas A, Fairen A (1988) GABAergic neurons of different morpho- 425-427. logical classes are cogenerated in the mouse barrel cortex. J Neurocytol Rakic P (1988) Spccificatlon of cerebral cortical areas. Science 241: 17:5l l-519. 170-176. Contamina P, Boada E (1992) Time of origin of cortico-collicular Shimada M. Langman J (1970) Cell proliferation, migration and dif- The Journal of Neuroscience. April 1994, 14(4) 2031

ferentiation in the cerebral cortex of the golden hamster. J Comp Walsh C, Cepko CL (1992) Widespread dispersion of neuronal clones Neural 139:227-244. across functional regions of the cerebra1 cortex. Science 255:434-440. Sidman RL (I 970) Autoradiographic methods and principles for study Walsh C, Cepko CL (1993) Clonal dispersion in proliferative layers of the nervous system with thymidine-H’. In: Contemporary research of developing cerebra1 cortex. Nature 362:632-635. methods in neuroanatomy (Nauta WJH, Ebbesson SOE, eds), pp 252- Zilles K (1985) The cortex ofthe rat. A stereotaxic atlas. Berlin: Spring- 274. New York: Springer. er. Smart IHM, Smart M (1982) Growth patterns in the lateral wall of Zilles K (1990) Anatomy of the neocortex: cytoarchitecture and mye- the mouse telencephalon. I. Autoradiographic studies of the histo- loarchitecture. In: The cerebral cortex of the rat (Kolb B, Tees RC, genesis of the isocortex and adjacent areas. J Anat 134:273-298. eds), pp 77-l 12. Cambridge, MA: MIT Press. Tan S-S, Breen S (1993) Radial mosaicism and tangential cell dis- Zilles K, Wree A, Dausch N-D (1990) Anatomy of the ncocortex: persion both contribute to mouse neocortical development. Nature neurochemical organization. In: The cerebra1 cortex of the rat (Kolb 362:638-640. B, Tees RC, eds), pp 113-l 50. Cambridge, MA: MIT Press. Waechter RV, Jaensch B (1972) Generation times of the matrix cells during embryonic brain development: an autoradiographic study in rats. Brain Res 46:235-250.