Cerebral Cortex Advance Access originally published online on April 20, 2005
Cerebral Cortex 2006 16(1):124-135; doi:10.1093/cercor/bhi092
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Role of the Protomap and Target-derived Signals in the Development of Intrahemispheric Connections
1 Mitsubishi Kagaku Institute of Life Sciences, 11 Minamiooya, Machida, Tokyo 194-8511, Japan and 3 Genome Information Research Center, Osaka University, 3–1 Yamadaoka, Suita, Osaka 565-0871, Japan, 2 Present address: Laboratory for Neural Architecture, RIKEN Brain Science Institute, 2–1 Hirosawa, Wako, Saitama 351-0198, Japan
Address correspondence to Y. Arimatsu, Mitsubishi Kagaku Institute of Life Sciences, 11 Minamiooya, Machida, Tokyo 194-8511, Japan. Email: Ispr{at}libra.Is.m-kagaku.co.jp.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsOrganotypic Slice CultureResultsDiscussionReferences |
|---|
Mechanisms intrinsic to the early cerebral cortex have been implicated in the establishment of cortical area identity. However, the extent to which the cortical protomap contributes to the formation of highly complex intrahemispheric connections remains obscure. Mechanisms by which postmitotic neurons establish correct corticocortical connections later in corticogenesis also remain to be elucidated. Here, we used a new transplantation method, employing donor tissue harvested from enhanced green fluorescent protein-expressing rats, to show that cortical progenitors are regionally specified for connectional potential and that this controls the development of specific intrahemispheric projections. The acquisition of connectional capacity relies on positional cues within the cortical primordium, but is independent of thalamic inputs. In addition, since cortical neurons developing in organotypic slice culture extended axons more prominently into their normal cortical target tissues than into non-target tissues, we suggest that cortical neurons respond to specific signals derived from their cortical targets.
Key Words: cortical specification • corticocortical connections • latexin • neuronal birth • rat • transplantation
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsOrganotypic Slice CultureResultsDiscussionReferences |
|---|
The mammalian cerebral cortex is composed of functionally specialized regions with distinct molecular and anatomical characteristics. The establishment of regional diversity is thought to rely on signals intrinsic to the cerebral wall, i.e. protomap (Rakic, 1988
; Rubenstein and Rakic, 1999; Muzio and Mallamaci, 2003), and also on signals released from afferent axons that project from subcortical structures (O'Leary, 1989; Sur and Leamey, 2001). Increasingly, evidence points to the importance of intrinsic mechanisms in the development of regional identity. These include gene expression patterns (Barbe and Levitt, 1991; Arimatsu et al., 1992; Cohen-Tannoudji et al., 1994; Miyashita-Lin et al., 1999; Nakagawa et al., 1999; Fukuchi-Shimogori and Grove, 2001), and projections to and from subcortical structures (Barbe and Levitt, 1992; Ebrahimi-Gaillard et al., 1994; Garnier et al., 1995; Ebrahimi-Gaillard and Roger, 1996; Frappé et al., 1999; Gaillard and Roger, 2000; Pinaudeau et al., 2000; Fukuchi-Shimogori and Grove, 2001; Leingartner et al., 2003). However, despite their functional importance, there have been very few studies on the mechanisms that dictate the shaping of intrahemispheric connections, which are inherently complex (Ebrahimi-Gaillard et al., 1994; Barbe and Levitt, 1995; Kingsbury et al., 2002; Huffman et al., 2004).
The highly organized intrahemispheric circuits play principal roles in the cognitive functions of the neocortex (Felleman and Van Essen, 1991; Salin et al., 1995). Information from the primary sensory cortices is relayed by chains of feedforward pathways through the sensory association cortices, while feedback pathways arising from higher association cortices transmit information to the primary sensory cortices. Tracing studies in the rat, for example, have shown that neurons in the secondary somatosensory cortex (S2) and the lateral part of the secondary visual cortex (V2L) project to the prefrontal cortical region including the secondary motor cortex (M2), while S2 and V2L have projections to the primary somatosensory (S1)/motor (M1) and visual (V1) cortices respectively (Van Eden et al., 1992; Coogan and Burkhalter, 1993; Shi and Cassell, 1998). Intermodal connections are generally weak, such that S2 neurons have markedly fewer projections to the V1, and V2L neurons have fewer projections to the S1/M1 region (Paperna and Malach, 1991). The feedback projection neurons located in the supra- and infragranular layers innervate distinct laminae in a given area (Coogan and Burkhalter, 1993; Cauller et al., 1998; Zhang and Deschênes, 1998). Recently, it has been shown that a subpopulation of feedback projection neurons in the infragranular layers specifically coexpresses latexin, a carboxypeptidase A inhibitor (Arimatsu et al., 1999a; Bai et al., 2004), as well as the orphan nuclear receptor Nurr1 (Arimatsu et al., 2003), indicating that functionally related feedback connections are arranged in a lamina-specific manner.
To elucidate the developmental mechanism(s) underlying the construction of complex intrahemispheric circuits, it is essential to study when and how cortical cells are regionally specified. It is also imperative to elucidate how cortical neurons acquire the capacity to detect and respond to area-specific guidance cues either generated within their target areas and/or encountered on the way to the target. Since distinct transcription factors are expressed in developing corticocortical projection neurons in the supra- and infragranular layers (Arimatsu et al., 2003; Hevner et al., 2003), it is important to address whether specific connections arising from a given layer are established through a process unique to the lamina.
In the present study, we initially used a new transplantation technique to determine when the connectional specificity of corticocortical projection neurons is established during corticogenesis. We examined projections from S2 and V2L grafts, taken from enhanced green fluorescent protein-expressing (EGFP+) rat embryos at embryonic day (E) 12–E17, following their transplantation into the S2 region of newborn, wild-type, host animals. We also performed organotypic slice culture experiments to elucidate whether signals from the targets contribute to the construction of region-specific corticocortical projections. In both the in vivo and in vitro experiments, latexin was used to monitor the growth and maintenance of infragranular corticocortical neurons. We examined whether latexin+ infragranular and latexin– supragranular neurons behave differently in cortical explants in vivo and in vitro. Our findings demonstrate that proliferating progenitor cells in the cortical primordium are regionally specified during a limited time window to produce both supra- and infragranular layer neurons that project to specific cortical targets. We also provide evidence suggesting that target-derived signals contribute to region-specific axon targeting from both supra- and infragranular layers.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsOrganotypic Slice CultureResultsDiscussionReferences |
|---|
Animals
EGFP-transgenic and wild-type Sprague–Dawley rats were used. The EGFP transgenic line TgN(acro/act-EGFP)4Osb, which expresses EGFP under the promoter of ß-actin, has been described previously (Tashiro et al., 2001). Timed-pregnant wild-type and transgenic rats were supplied by Japan-SLC (Hamamatsu, Japan). The days on which the vaginal plug was present was designated as E0. Birth occurred at E21 or E22 and the 24 h period after birth was designated P0. All animal experiments were performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals.
Transplantation
Donor cortical grafts were obtained from heterozygous EGFP-transgenic rat embryos at E12, E13, E14 and E17 from dams deeply anesthetized with ether. Coronal forebrain slices of 400 µm (E12, E17) or 600 µm (E13, E14) thickness were made on a vibrating microtome. A pair of anterior lateral or posterior lateral cortical fragments was dissected out of a slice obtained from each embryo, except for E17 anterior lateral cortical fragments which were prepared from two consecutive forebrain slices (Fig. 1). It was previously shown that anterior lateral and posterior lateral parts of the cerebral cortex from E12–E16 embryos develop many latexin+ neurons in slice culture, whereas anterior dorsal and posterior dorsal cortices produce a lot fewer latexin+ neurons (Arimatsu et al., 1992; Arimatsu and Ishida, 1998). Correlating the capacity of specific embryonic cortical regions to generate latexin+ neurons in vitro with the known distribution pattern of latexin+ neurons in the adult cortex (Arimatsu et al., 1999c), we had determined regions that were likely to contain presumptive S2 or V2L. The ‘presumptive S2 graft’ corresponded to the lateral neocortex at the level of the anterior part of the lateral ganglionic eminence and may have contained not only S2 but also a part of S1 and/or granular and agranular insular cortices. The ‘presumptive V2L graft’ corresponded to the upper part of the lateral neocortex at the level of the posterior ganglionic eminence and may have contained V2L, and a part of V1 and/or primary and associative auditory cortices. We used these grafts as donors in the transplantation experiments.
|
Newborn host rats (P0) were anesthetized by hypothermia and the scalp was opened. A flap was created in the temporal bone and a small hole, reaching the white matter, was made in S2 by aspiration (
600 µm in diameter). A thick piece of donor tissue comprising all cortical layers down to the white matter or to the ventricular surface was aspirated into a glass cannula and injected into the host lesion cavity. We did not attempt to orient the graft in the host cortex. Retrograde Labeling in Vivo
Fluoro-Gold (Fluorochrome, Denver, CO; 2% in H2O, 0.2 µl per injection) was pressure injected through a glass micropipette (outer diameter, 150–200 µm) into V1 (two points: Bregma –6.3 and –7.3 mm, lateral 3.5 mm) or M1 (two points: Bregma +2.7 and +1.7 mm, lateral 3.0 mm) regions of 6–10 week old rats with transplants (body weight: 200–300 g), under deep anesthesia with sodium pentobarbital (50 mg/kg). Following 4 days of survival, rats were perfused with 4% paraformaldehyde, and the brains were postfixed and cryosectioned (10 µm), as described previously (Arimatsu et al., 1999a). Every fifth section was mounted onto silane-coated slides (Matsunami Glass Ind., Osaka, Japan), and stained for Fluoro-Gold and latexin immunofluorescence using a rabbit anti-Fluoro-Gold antibody (Chemicon, Temecula, CA; 1:1000 dilution) and a mouse monoclonal anti-latexin antibody (PC3.1, Arimatsu et al., 1992; 10 µg/ml) respectively. The secondary antibodies used were Alexa Fluor 350 anti-rabbit IgG (Molecular Probes, Eugene, OR; 1:200 dilution) and Alexa Fluor 594 anti-mouse IgG (Molecular Probes; 1:200 dilution). In the EGFP+ population, the number of Fluoro-Gold-filled neurons with and without coincident latexin immunoreactivity were counted using an Axioskop fluorescence light microscope (Carl Zeiss, Oberkochen, Germany) equipped with a C5810 color-chilled 3CCD camera (Hamamatsu Photonics, Hamamatsu, Japan). To trace axon fibers from transplants, sections from animals with and without the tracer injection were stained with an anti-GFP mouse monoclonal antibody (Molecular Probes; 10 µg/ml) followed by Alexa Fluor 488 anti-mouse IgG (Molecular Probes; 1:200 dilution) to intensify the EGFP signal. Cortical areas were defined according to the cytoarchitecture and the positional information linked with anatomical landmarks, e.g. the forceps minor for M1 and M2, and the forceps major for V1.
|
Organotypic Slice Culture |
|---|
|
TopAbstractIntroductionMaterials and MethodsOrganotypic Slice CultureResultsDiscussionReferences |
|---|
Wild-type and EGFP-transgenic rat embryos were removed at E20 or E21 as described above. To label neurons in the supragranular layers, a pregnant dam with E17 embryos was administered 5-bromo-2'-deoxyuridine (BrdU) intraperitoneally (50 mg/kg, four times at 4 h intervals), and the embryos were removed at E20 or E21. Coronal slices of the forebrain were cut (400 µm) and portions of the cerebral wall corresponding to S1, S2, V1, V2L, and the piriform cortex (Pir) were dissected out on the basis of maps of the developing rat brain (Paxinos et al., 1994
). From each cortical hemisphere, two pieces of S1, two pieces of S2, two pieces of Pir, two pieces of V1 and two pieces of V2L were obtained. Where necessary, the white matter and subplate, together with adjacent lower layer 6a, were surgically removed from cortical slices. In a separate experiment, we were able to identify the position of subplate layer at E20 by the specific expression of Nurr1 (Arimatsu et al., 2003). An S2 or V2L slice was co-cultured on a collagen-coated membrane between two slices from other cortical regions, as described previously (Arimatsu et al., 1999a; Arimatsu and Ishida, 2002). The S2 (or V2L) slice was placed immediately adjacent to other cortical slices so that the cortical radial axis became mutually parallel and the white matter surface was linearly oriented on the same side. At 5 (S1-S2-Pir) or 12 (S1-S2-V1, V1-V2L-S1) days in vitro, small crystals of either of the fixable fluorescent dextran amines (fluoro-ruby, 3000 MW; micro-emerald, 3000 MW; both from Molecular Probes) were placed at 10 sites within the S1 (S1-S2-Pir, S1-S2-V1, V1-V2L-S1), Pir (S1-S2-Pir) and V1 (V1-V2L-S1) slices with the tip of a glass micropipette. At 18–20 h after the tracer injection, the co-cultures were fixed with 4% paraformaldehyde (2 h, 4°C). After removal of the tracer-injected portions from the co-cultures, the remaining S2 or V2L slices were cryosectioned serially (10 µm) and thaw-mounted onto slides. The sections were treated with anti-latexin (10 µg/ml) or anti-BrdU (DAKO, Denmark; 1:20 dilution) monoclonal antibodies and then with Alexa Fluor 350 goat anti-mouse IgG (Molecular Probes; 1:200). Micro-emerald fluorescence was intensified by treating the sections with Alexa 488 streptavidin (Molecular Probes; 1:100 dilution). The number of fluoro-ruby- and micro-emerald-labeled neurons, with and without latexin-immunoreactivity, was counted in every second section. In the alternating sections, the number of tracer-filled neurons, with and without BrdU-labeling, was counted under the fluorescence microscope. Individual neurons labeled with both tracers were doubly counted, once for each tracer. Statistics
For statistical analyses, we performed Mann–Whitney U-test in the transplantation experiment, and t-test for paired samples in the slice culture experiment, by using the GraphPad Prism version 3 software (GraphPad Software Inc., San Diego, CA).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsOrganotypic Slice CultureResultsDiscussionReferences |
|---|
Transplantation of EGFP+ Cells to Assess Connectional Potential
We studied the patterns of long-distance corticocortical projections from embryonic cortical grafts containing presumptive S2 or V2L regions (Fig. 1) transplanted homotopically and heterotopically into the cortex of newborn host rats. Since we used cortical grafts from EGFP+ rats as a source of donor tissue, cells and axons of donor origin fluoresced green and were readily identifiable in the host brain.
Neurons in homotopic cortical grafts (S2-to-S2) taken from E14 embryos exhibited basically the same projection patterns as those of normal rats, even though they were transplanted heterochronically into newborn hosts. Prominent EGFP+ axon fibers originating from the donor tissue extended into the white matter and grew into the contralateral hemisphere through the corpus callosum (Fig. 2A). Within the ipsilateral cortex, axon fibers also projected through the bottom of layer 6, and some axons from these projections grew into the upper layers (Fig. 2B). Dense axonal arborizations were observed in the normal targets of S2 neurons (Vaudano et al., 1991; Van Eden et al., 1992; Shi and Cassell, 1998; Zhang and Deschênes, 1998): layers 1–6 of ipsilateral S1 (Fig. 2C–G), M1 (Fig. 2H, I), M2, granular and agranular insular and perirhinal cortices as well as contralateral S2 (not shown). In contrast, virtually no axonal projections were observed in V1 (Fig. 2J), consistent with observations in normal rats (Paperna and Malach, 1991). In addition, we observed normal corticothalamic projections from the homotopic S2 transplants. Prominent axonal arborizations were found in the ventrobasal and posterior thalamic nuclei (Fig. 2K,L), but not in the visual thalamus, i.e. the lateral geniculate and lateral posterior thalamic nuclei (not shown).
|
Cortical Cells Are Regionally Specified for Connectional Potential Prior to Neuronal Birth
We assessed the capacity of E14 cells in presumptive V2L grafts to establish specific projections to their cortical targets. When posterior lateral cortical grafts containing presumptive V2L were heterotopically transplanted into S2 of newborn hosts, axons originating from the V2L-to-S2 transplants grew extensively into V1 (Fig. 3A–C), the perirhinal cortex (Fig. 3E) and M2 (Fig. 3F,G), the normal targets of V2L (Vaudano et al., 1991; Van Eden et al., 1992; Coogan and Burkhalter, 1993). In these three areas, dense axonal arborizations were seen from layer 1 to layer 6. The prominent axonal projections into V1 from the V2L-to-S2 transplant contrast with the paucity of axonal ingrowths into V1 from the S2-to-S2 transplant (Fig. 2J). In contralateral S2, callosal projections from the heterotopic transplant were observed running in deep layers, but they were much sparser than those from the homotopic transplant (not shown). These observations strongly suggest that cortical cells are regionally specified by E14 for connectional potential. Since most cells that later differentiate into cortical projection neurons are still in a proliferative phase at E14 (Bayer and Altman, 1991), cortical cells must be specified prior to neuronal birth (i.e. when they exit the cell cycle). Given that thalamic afferents grow into the cerebral wall at E16–E18 (Catalano, 1991), this commitment must occur in the absence of instructive signals from the thalamic inputs. We therefore conclude that development of region-specific intrahemispheric connections relies on the protomap of the proliferative cortical primordium.
|
In addition to normal cortical targets, the V2L-to-S2 transplant sent prominent axons into S1 and to granular and agranular insular cortices not normally connected with V2L (Fig. 3F). Although graft axons rarely grew into the piriform cortex beyond the insular cortices, they invaded the M1 region crossing the boundary with S1. These observations suggest that E14 cortical cells generally have the capacity to send axons into neighboring cortical tissue, even when it is not their normal target.
Certain E14 V2L cells seem to be specified for corticothalamic connections as evidenced by the fact that neurons in V2L-to-S2 transplants (E14) projected to the lateral posterior, lateral geniculate and posterior thalamic nuclei — all normal targets of V2L neurons (Fig. 3D). However, we did not observe extensive projections from the heterotopic transplants into the ventrobasal thalamic nucleus.
Signals from the Early Cortical Primordium Are Necessary for Full Development of Connections
Both S2 and V2L cortical grafts taken at E12 and E13, rather than E14, developed less prominent axonal projections to their normal targets in the dorsal neocortex. Although some axons from S2-to-S2 transplants were observed to extend toward S1 and M1 through layer 6, their growth into the upper layers was very sparse (Fig. 4A) compared with that from E14 transplants. Similarly, axons from E12 and E13 V2L-to-S2 transplants grew poorly into V1 (Fig. 4B), in contrast to those from E14 heterotopic transplants. We found no axonal growth from E12 or E13 S2-to-S2 transplants into host V1, as was observed for E14 homotopic transplants. Callosal projections from E12 and E13 transplants were also poor compared to those developed from E14 transplants (not shown). Furthermore, only very sparse axonal arborizations were found in the ventrobasal and posterior thalamic nuclei of host rats with either E12 or E13 S2-to-S2 transplants (Fig. 4C). These observations support the suggestion that certain cortical cells require signals from the E13–E14 cortical primordium to fully develop intrahemispheric, callosal and thalamic projections. However, because V2L-to-S2 E12 transplants distributed prominent axonal projections to the perirhinal cortex, it appears that different populations of cortical projection neurons acquire their fates at different times (Fig. 4D).
|
Retrograde Labeling of M1- and V1-projecting Neurons in Transplants
To quantitatively assess axonal growth from the transplant into M1 and V1, the retrograde axonal tracer Fluoro-Gold was injected into the M1 and V1 regions of host rats (Table 1). We evaluated axonal projections from latexin+ (infragranular) and latexin– (supragranular and infragranular) neurons by triple-color fluorescence for EGFP, latexin and Fluoro-Gold (Fig. 5). The number of retrogradely labeled neurons in the transplant was in agreement with the qualitative observation described above (Fig. 6). In S2-to-S2 grafts taken from E14 and E17 embryos, both latexin+ and latexin– neurons were prominently labeled following Fluoro-Gold injections into M1 (Figs 5A–C, 6A,B). Labeled neurons were extremely rare in the homotopic S2 transplants following the tracer injection into V1 (Fig. 6C,D). In contrast, significantly more V1-projecting neurons, both latexin+ and latexin–, were found in E14 and E17 heterotopic V2L-to-S2 transplants (Figs 5D–I, 6C,D). In E12 and E13 transplants, M1- and V1-projecting neurons were generally sparse (Fig. 6A–D). Taken together, these retrograde labeling experiments confirm the assumption that the connectional potential of cortical cells is regionally specified by E14. Furthermore, these experiments indicate that this connectional potential has a similar temporal emergence in latexin+ and latexin– neuronal populations.
|
|
|
Target-derived Signals Play a Role in Inter-regional Connections
In the transplantation experiments, the V2L-to-S2 transplants, but not the S2-to-S2 transplants, from E14 and E17 donors grew extensively into V1. Since axons from the V2L-to-S2 transplants must have reached the V1 target through an aberrant pathway distinct from the route that V2L axons follow in normal rats, it is likely that the corticocortical axon targeting depends, at least in part, on signals derived from the target. To test this possibility, we performed an organotypic slice culture experiment, in which a cortical slice was co-cultured between slices from target and non-target areas. When an S2 slice from an E20 or E21 EGFP rat embryo was co-cultured between slices of wild-type S1 and piriform cortex (Pir), or between slices of wild-type S1 and V1, S2 neurons appeared to project preferentially to the normal S1 target rather than the non-target areas (Fig. 7A,B). To quantitatively evaluate the number of S2 neurons projecting to target and non-target areas, we performed similar experiments using S2 slices from wild-type animals. Two different fluorescent tracers, micro-emerald and fluoro-ruby, were injected into S1 and Pir slices of the S1-S2-Pir co-culture (5 days in vitro). Notably, more neurons were retrogradely labeled in the S2 slice from tracer injections into the S1 slice (373 ± 40, mean ± SEM) than from injections into the Pir slice (113 ± 30) (Fig. 7C; P < 0.001, t-test for paired samples, number of co-cultures = 10). Similarly, following tracer injections into S1 and V1 slices of the S1-S2-V1 co-culture (12 days in vitro), more neurons were labeled in the S2 slice from injections into the S1 slice (865 ± 86) than from injections into the V1 slice (405 ± 55) (Fig. 7D; P < 0.001, t-test for paired samples, number of co-cultures = 18). Furthermore, when a V2L slice (E21) was co-cultured between slices from V1 and S1, neurons in the V2L slice projected more to V1, the normal target of V2L, than to the non-target area, S1. Consequently, more neurons were retrogradely labeled from injections of micro-emerald into the V1 slice (695 ± 127) than from injections of fluoro-ruby into the S1 slice (299 ± 70) (12 days in vitro; P < 0.001, t-test for paired samples, number of co-cultures = 10). Overall, these experiments revealed that even in the absence of potential guidance cues normally available to corticocortical axons in vivo, developing cortical neurons still project more prominently to their normal targets than to non-target regions in vitro. These results support the idea that target-derived signals play a role in establishing and/or maintaining specific corticocortical connections.
|
Target-derived Signals Are Important for Connections from Both the Supra- and Infragranular Layers
To further explore whether neuronal populations in both the supra- and infragranular layers respond to target-derived signals, we performed an additional co-culture experiment. Since a large number of the supragranular layer neurons are generated at E17 and later (Bayer and Altman, 1991), we were able to identify them in culture by prior administration of the thymidine analog BrdU to E17 embryos in vivo. In the S1-S2-V1 co-culture system, both BrdU-labeled (supragranular) neurons and latexin+ (infragranular) neurons in the S2 slice were retrogradely labeled approximately three times more frequently from the S1 slice than from the V1 slice [Fig. 8: BrdU-labeled (52.3 ± 6.3 versus 16.5 ± 3.3), latexin+ (66.8 ± 11.6 versus 22.3 ± 3.8); Ps < 0.001, t-test for paired samples, number of co-cultures = 23; data from two independent experiments]. This suggests that connections from both the supra- and infragranular layers are dependent on target-derived signals.
|
Inter-regional Connections in the Absence of Subplate Neurons
Subplate neurons in the embryonic cortex are implicated in the development of cortical afferents and efferents (McConnell et al., 1989; Ghosh and Shatz, 1993; Molnár and Blakemore, 1995). Furthermore, subplate neurons in adult and early postnatal rats display extensive intrahemispheric projections (Clancy and Cauller, 1999; Arimatsu et al., 2003), raising the question of whether developing corticocortical axons respond to specific guidance cues generated by subplate neurons. To examine the potential involvement of subplate neurons in the process of target selection, we performed a slice co-culture experiment, in which an S2 slice without the subplate was co-cultured between S1 and V1 slices, both of which were also lacking subplate neurons. The axons from the S2 slice grew more extensively into the S1 slice than into the V1 slice, as was observed in co-cultures with subplate neurons (Table 2). Thus, more neurons were retrogradely labeled from injections of micro-emerald (or fluoro-ruby) into the S1 slice than from injections of fluoro-ruby (or micro-emerald) into the V1 slice. The absence of the subplate and adjacent lower layer 6a in the culture was verified by the marked reduction in latexin+ neuron frequency observed in the S2 slice, since these cells are mainly located in layer 6a. Our findings suggest that signals generated within layers 1–6a are sufficient to mediate area-specific axon targeting.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsOrganotypic Slice CultureResultsDiscussionReferences |
|---|
Technical Considerations
Previously in conventional transplantation studies, it was observed that grafts of E14–E17 neocortical tissue tend to develop the capacity to receive callosal and thalamic inputs appropriate to their embryonic site of origin (Barbe and Levitt, 1992; Frappé et al., 1999; Gaillard and Roger, 2000). Moreover, it has been shown that certain neocortical cells are already specified to project appropriately to intrahemispheric targets at E16 (Ebrahimi-Gaillard et al., 1994) and to subcortical targets at E13–E17 (Ebrahimi-Gaillard et al., 1994; Ebrahimi-Gaillard and Roger, 1996). In line with these observations, we used transplants from EGFP transgenic rats to show that at E14, presumptive S2 and V2L cells differ in their potential to develop specific intrahemispheric projections. Although cell mixing of host and donor cells and tracer leakage outside the transplant from the injection site are unavoidable problems in conventional transplantation experiments, the present study is free from such caveats. Our finding that homotopic transplants project to their normal cortical and thalamic targets indicates that this new transplantation method, in which embryonic EGFP transgenic rats are used as donors and newborn wild-type rats as hosts, is suitable for assessing the potential for developing cortical cells to establish specific corticocortical and corticothalamic projections.
In transplantation experiments that address cortical regional specification, it is critical to appropriately dissect the young cerebral wall in order to generate donor grafts. We prepared cortical grafts containing presumptive S2 and V2L from anterior lateral and posterior lateral parts of the embryonic cortex, respectively. Although the donor tissue from the anterior lateral cortex may have contained a small part of presumptive S1 and/or granular and agranular insular cortices, especially when derived from younger embryos, these regions are not known to have substantial projections to the V1 region in the adult rat (Shi and Cassell, 1998). This makes it appropriate to use the anterior lateral cortical graft as presumptive S2 in the present transplantation experiments. On the contrary, it is possible that posterior lateral cortical grafts may have contained not only presumptive V2L, but also a part of presumptive V1 and the primary and association auditory cortices. Whereas adult V1 and auditory cortices have only sparse connections with either M1 or M2 (Van Eden et al., 1992; Mascagni et al., 1993; Shi and Cassell, 1997), V2L regions have rich projections to M2 (Miller and Vogt, 1984; McDonald and Mascagni, 1996) passing through deep layers in M1 (Van Eden et al., 1992). In fact, in a study by Reep et al. (1994), a large number of Oc2L (V2L) neurons were retrogradely labeled following Fluoro-Gold injection into M1. This would have presented a serious problem if we quantitatively assessed M1-projecting neurons in the transplant by the conventional means of injecting Fluoro-Gold into M1, as M2-projecting neurons may also have been labeled. However, here we did not evaluate the number of M1-projecting neurons in the posterior lateral cortical transplant (V2L-to-S2), nor did we attempt to compare the capacity of homo- and heterotopic transplant cells in projecting to M1.
Throughout the present study, latexin was used as a molecular tag to monitor the growth and maintenance of infragranular corticocortical neurons in the transplant. This is based on our previous finding that latexin expression in cortical cells is dependent on their embryonic site of origin, even under various environmental conditions in vitro (Arimatsu et al., 1992, 1999b). Indeed, latexin+ neurons developed well in cortical grafts even though they had been transplanted heterochronically and heterotopically. We selected transplants for analysis on the basis of latexin expression, such that rats with a graft containing a substantial number of latexin+ neurons were used for retrograde labeling of M1- or V1-projecting neurons. The number of latexin+ neurons was then used to normalize the number of M1- or V1-projecting neurons in the transplant.
Regionalization and Neuronal Birth
Given that latexin+ corticocortical projection neurons in the infragranular layers and latexin– neurons in the supragranular layers are generated at E15 (Arimatsu et al., 1994) and E17–E19 (Bayer and Altman, 1991) respectively, the present findings indicate that regional diversification for intrahemispheric connections occurs at least 1 day prior to neuronal birth for latexin+ infragranular neurons and 3–5 days prior to neuronal birth for latexin– supragranular neurons. The apparent lack of correlation of neuronal birth date with regionalization is in contrast to the strong correlation that is observed between birth and phenotype in various cortical populations. During normal development, neurons that will populate a more superficial layer tend to be born later than those that populate deeper layers (Bayer and Altman, 1991). Neuronal birth date is linked even more closely with certain molecular and connectional fates. For example, latexin+ neurons are born simultaneously even though they are distributed in both layers 5 and 6 (Arimatsu et al., 1994), and (latexin+) corticocortical neurons in layer 6 are born at a different time than corticothalamic neurons in the same layer (Arimatsu and Ishida, 2002). With regard to laminar fate, it has been postulated that early cortical progenitors, which exist at the time when layer 6 neurons are being generated (E14–E15, in rats), have the capacity to adopt a supragranular layer fate when exposed to an appropriate extracellular environment (McConnell and Kaznowski, 1991). Given the pluripotency of progenitors at or before their birth, the laminar fate of individual neurons must be determined after cortical regionalization. Molecular (latexin+ versus latexin–) and connectional (corticocortical versus corticothalamic) fates may also be determined after cortical regionalization. The exact timing of commitment remains to be determined since a fraction of early (E12, E13) progenitors can produce latexin+ neurons, even in an in vitro system with minimum cell-to-cell interactions (Arimatsu et al., 1999b; Takiguchi-Hayashi, 2001).
Connectional Fate Potential of Early Progenitors
One of the unexpected findings in the present study is that E12 and E13 cortical cells in presumptive S2 and V2L regions developed very sparse projections to their normal neocortical targets when transplanted in the cortex of P0 rats. It might be that progenitors in the E12 and E13 transplants proliferated less and/or regressed more than those in E14 transplants, leading to very sparse axonal projections from the transplants. Although the cell growth and maintenance were not quantitatively estimated, it seems unlikely that the sparse innervations of all neocortical targets from E12 and E13 transplants were due to a general paucity of cell growth, because the number of latexin+ neurons in E12 and E13 transplants was very comparable to that in E14 transplants. It is also possible that E12/E13 cortical grafts did not send long-distance axonal projections because younger progenitors required longer time periods to develop their responsiveness to guidance cues in the host cortex, which might become less permissive for axonal growth in later postnatal periods. However, this is unlikely since it was shown that cortical tissue was appropriately and massively innervated by axons from a homotopic embryonic cortical graft transplanted into the adult cortex (Gaillard et al., 2004). Thus, the less prominent axonal growth from E12/E13 cortical grafts was likely due to incomplete development of cortical progenitors for axonal projections rather than insufficient availability of trophic and guidance signals in the host cortex. Taken together, we conclude that signals generated by E13–E14 cortical primordium are essential for cortical progenitors to fully develop intrahemispheric connections.
Previously, it was observed that E12 cortical grafts removed from perirhinal and visual cortices, and heterotopically transplanted into newborn S1 cortex, received callosal and thalamic inputs respectively, that were appropriate for the host locus where they developed (Ebrahimi-Gaillard et al., 1994; Barbe and Levitt, 1995). Moreover, some E12 heterotopic cortical transplants sent axons to subcortical targets appropriate for their new environment (Pinaudeau et al., 2000; Gaillard et al., 2003). These observations led the authors to propose that early (E12) cortical progenitors are multipotent, and retain the capacity to respond to local cues in the newborn host cortex and to change their connectional fate. However, in our study, E12 and E13 progenitor cells did not fully develop intrahemispheric projections without exposure to the cortical primordium at E13–E14. Therefore, the E13–E14 cortical primordium and P0 cerebral wall should be distinct in their capacity to instruct corticocortical connections. On the other hand, it seems premature to draw a definite conclusion with respect to the precise timing of regional specification of early progenitors. While it is possible that E12/E13 progenitors are multipotent and would be specified with positional information from E13–E14 cortical primordium, it is also possible that they are already specified but could be compromised in their ability to send out projections without appropriate signals from the early cortical primordium.
Patterning Signals for Intrahemispheric Connections
Given that cortical progenitors are regionally specified for intrahemispheric connections at E13–E14 or before and that they are not competent to respond to positional cues in the P0 cortex, it is possible that the cortical area map is established by signaling molecules that are highly expressed at E13–E14 or before and downregulated by P0. Recent molecular genetic studies have provided evidence of a role of certain transcription factors, e.g. Emx2 and Pax6, and signaling molecules, e.g. FGF8, in patterning events within the cerebral cortex (O'Leary and Nakagawa, 2002; Garel et al., 2003; Grove and Fukuchi-Shimogori, 2003; Muzio and Mallamaci, 2003). For example, FGF8 is expressed in the anterior pole of the mouse cortical primordium at E9.5–E12.5 and is downregulated at E14.5 (Heikinheimo et al., 1994; Crossley and Martin, 1995). When endogenous FGF8 action was augmented or suppressed in the mouse cortical primordium at E11.5 (corresponding to E13 in the rat), the area boundary was shifted along the anterior–posterior axis, as assessed by gene expression patterns (Fukuchi-Shimogori and Grove, 2001). Ectopic expression of FGF8 even caused duplication of the cytoarchitecturally distinct S1 barrel field. Thus, it is tempting to speculate that signaling molecules such as FGF8 also pattern intrahemispheric connections along the anterior–posterior axis. Consistent with this idea, a very recent study has demonstrated that Fgf8-deficient mice display an altered pattern of intrahemispheric projections (Huffman et al., 2004).
Target-derived Signals for Corticocortical Connections
Following the patterning events in the cortical primordium, young, post-mitotic neurons in a given area must find targets to innervate. Although the exact timing of when individual cortical cells are specified to project subcortically or contribute to particular intrahemispheric connections remains to be determined, corticothalamic and corticocortical neuron populations in layer 6 are already distinct at the perinatal period (Arimatsu and Ishida, 2002). It has been shown that perinatal cortical efferent neurons are able to detect and respond to local cues emanating from their intermediate or final targets (Bolz et al., 1990; Kuang et al., 1994; Sato et al., 1994; Métin et al., 1997; Richards et al., 1997). Although previous experiments failed to find regional specificity in the thalamocortical or corticothalamic connections formed in vitro (Molnár and Blakemore, 1991), guidance cues within the cortex have recently been identified in vivo for afferent fibers (Frappé et al., 1999; McQuillen et al., 2002; Dufour et al., 2003). Correspondingly, the present in vivo and in vitro findings provide strong evidence in favor of target-derived guidance cues for the construction of corticocortical connections. Notably, while at least some regional information resides on subplate neurons (layer 6b) for thalamocortical connections (McQuillen et al., 2002), it seems that the corticocortical projections from S2 to S1 in vitro depend on signals derived from layers 1–6a. Since the present co-culture experiments were performed using E20/E21 cortical grafts, they do not address target-derived signals and potential role(s) of subplate neurons during earlier stages of cortical development. However, since the long-range corticocortical axonal growth from S2 appears to occur mainly during early postnatal periods (Arimatsu and Ishida, 2002), it is likely that many neurons in the co-culture responded to the target-derived signals when growing de novo rather than regrowing to appropriate targets. A recent in vivo ablation study has also provided evidence suggesting that subplate neurons are not required for intra-cortical connectivity within the visual cortex (Kanold et al., 2003). Since we failed to find any differences between latexin+ infragranular and latexin– supragranular neurons in terms of their axonal behavior in cortical explants, these neuronal populations might respond to common signals from the target area. Finally, it should be mentioned that the targeting of S2 or V2L neurons was not as accurate in the co-culture system we employed as in normal rats in vivo. Thus, one-third to one-half as many cells projected to the wrong cortical targets as projected to the correct targets. This is in agreement with the present observation in vivo that the V2L-to-S2 transplant (E14) could send axons into wrong targets (S1, granular and agranular insular cortices) when placed in their neighborhood. Moreover, it suggests that some major cues generated on the way to the target, and missing in the culture system, are as important as those generated within the target. The precise molecular nature of axon guidance cues for corticocortical connections remains to be resolved.
In summary, we have shown that the development of cortical area identity with regard to intrahemispheric connections relies on a cortex-autonomous mechanism. Thus, cortical regionalization represents a sequential process that includes the initial setting of a protomap in the proliferative cortical primordium, not only for dictating molecular features of the early cortex, but also for intracortical projection patterns of the mature cortex. Cortical cells later acquire various capacities to form area-specific connections, including receptivity to local cues presented within, or emanating from, cortical targets.
|
Acknowledgments |
|---|
We thank K. Rockland, Y. Hatanaka and N. Hashimoto for their comments on the manuscript.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsOrganotypic Slice CultureResultsDiscussionReferences |
|---|
Arimatsu Y, Ishida M (1998) Early patterning of the rat cerebral wall for regional organization of a neuronal population expressing latexin. Dev Brain Res 106:71–78.[CrossRef][Medline]
Arimatsu Y, Ishida M (2002) Distinct neuronal populations specified to form corticocortical and corticothalamic projections from layer VI of developing cerebral cortex. Neuroscience 114:1033–1045.[ISI][Medline]
Arimatsu Y, Miyamoto M, Nihonmatsu I, Hirata K, Uratani Y, Hatanaka Y, Takiguchi-Hayashi K (1992) Early regional specification for a molecular neuronal phenotype in the rat neocortex. Proc Natl Acad Sci USA 89:8879–8883.
Arimatsu Y, Nihonmatsu I, Hirata K, Takiguchi-Hayashi K (1994) Cogeneration of neurons with a unique molecular phenotype in layers V and VI of widespread lateral neocortical areas in the rat. J Neurosci 14:2020–2031[Abstract]
Arimatsu Y, Ishida M, Sato M, Kojima M (1999a) Corticocortical associative neurons expressing latexin: specific cortical connectivity formed in vivo and in vitro. Cereb Cortex 9:569–576.
Arimatsu Y, Ishida M, Takiguchi-Hayashi K, Uratani Y (1999b) Cerebral cortical specification by early potential restriction of progenitor cells and later phenotype control of postmitotic neurons. Development 126:629–638.[Abstract]
Arimatsu Y, Kojima M, Ishida M (1999c) Area- and lamina-specific organization of a neuronal subpopulation defined by expression of latexin in the rat cerebral cortex. Neuroscience 88:93–105.[CrossRef][ISI][Medline]
Arimatsu Y, Ishida M, Kaneko T, Ichinose S, Omori A (2003) Organization and Development of corticocortical associative neurons expressing the orphan nuclear receptor Nurr1. J Comp Neurol 166:180–196.
Bai W-Z, Ishida M, Arimatsu Y (2004) Chemically defined feedback connections from infragranular layers of sensory association cortices in the rat. Neuroscience 123:257–267.[CrossRef][ISI][Medline]
Barbe MF, Levitt P (1991) The early commitment of fetal neurons to the limbic cortex. J Neurosci 11:519–533.[Abstract]
Barbe MF, Levitt P (1992) Attraction of specific thalamic input by cerebral grafts depends on the molecular identity of the implant. Proc Natl Acad Sci USA 89:3706–3710.
Barbe MF, Levitt P (1995) Age-dependent specification of the corticocortical connections of cerebral grafts. J Neurosci 15:1819–1834.[Abstract]
Bayer SA, Altman J (1991) Neocortical development. New York: Raven Press.
Bolz J, Novak N, Götz M, Bonhoeffer T (1990) Formation of target specific neuronal projections in organotypic slice cultures from rat visual cortex. Nature 346:359–362.[CrossRef][Medline]
Catalano SM, Robertson RT, Killackey HP (1991) Individual axon morphology and thalamocortical topography in developing rat somatosensory cortex. J Comp Neurol 367:36–53.
Cauller LJ, Cancy B, Connors BW (1998) Backward cortical projections to primary somatosensory cortex in rats extend long horizontal axons in layer I. J Comp Neurol 390:297–310.[CrossRef][ISI][Medline]
Clancy B, Cauller LJ (1999) Widespread projections from subgriseal neurons (layer VII) to layer I in adult rat cortex. J Comp Neurol 407:275–286.[CrossRef][ISI][Medline]
Cohen-Tannoudji M, Babinet C, Wassef M (1994) Early determination of a mouse somatosensory cortex marker. Nature 368:460–463.[CrossRef][Medline]
Coogan TA, Burkhalter A (1993) Hierarchical organization of areas in rat visual cortex. J Neurosci 13:3749–3772.[Abstract]
Crossley PH, Martin GR (1995) The mouse Fgf8 gene encodes a family of polypeptides and is expressed in regions that direct outgrowth and patterning in the developing embryo. Development 121:439–451.[Abstract]
Dufour A, Seibt J, Passante L, Depaepe V, Ciossek T, Frisen J, Kullander K, Flanagan JG, Polleux F, Vanderhaeghen P (2003) Area specificity and topography of thalamocortical projections are controlled by ephrin/Eph genes. Neuron 39:453–465.[CrossRef][ISI][Medline]
Ebrahimi-Gaillard A, Roger M (1996) Development of spinal cord projections from neocortical transplants heterotopically placed in the neocortex of newborn hosts is highly dependent on the embryonic locus of origin of the graft. J Comp Neurol 365:129–140.[CrossRef][ISI][Medline]
Ebrahimi-Gaillard A, Guitet J, Garnier C, Roger M (1994) Topographic distribution of efferent fibers originating from homotopic or heterotopic transplants: heterotopically transplanted neurons retain some of the developmental characteristics corresponding to their site of origin. Dev Brain Res 77:271–283.[CrossRef][Medline]
Felleman DJ, Van Essen DC (1991) Distributed hierarchical processing in the primate cerebral cortex. Cereb Cortex 1:1–47.
Frappé I, Roger M, Gaillard A (1999) Transplants of fetal frontal cortex grafted into the occipital cortex of newborn rats receive a substantial thalamic input from nuclei normally projecting to the frontal cortex. Neuroscience 89:409–421.[CrossRef][ISI][Medline]
Fukuchi-Shimogori T, Grove E (2001) Neocortex patterning by the secreted signaling molecule FGF8. Science 294:1071–1074.
Gaillard A, Roger M (2000) Early commitment of embryonic neocortical cells to develop area-specific thalamic connections. Cereb Cortex 10:443–453.
Gaillard A, Nasarre C, Roger, M (2003) Early (E12) cortical progenitors can change their fate upon heterotopic transplantation. Eur J Neurosci 17:1375–1383.[CrossRef][ISI][Medline]
Gaillard F, Domballe L, Gaillard A (2004) Fetal cortical allografts project massively through the adult cortex. Neuroscience 126:631–637.[CrossRef][ISI][Medline]
Garel S, Huffman KJ, Rubenstein JL (2003) Molecular regionalization of the neocortex is disrupted in Fgf8 hypomorphic mutants. Development 130:1903–1914.
Garnier C, Arnault P, Ebrahimi-Gaillard A, Létang J, Roger M (1995) The topographic distribution of the efferents from neocortical neurons is not only dependent upon where in the neocortex the cells develop. A transplantation study within one single neocortical region. Dev Brain Res 89:1–10.[CrossRef][Medline]
Ghosh A, Shatz CJ (1993) A role for subplate neurons in the patterning of connections from thalamus to neocortex. Development 117:1031–1047.[Abstract]
Grove EA, Fukuchi-Shimogori T (2003) Generating the cerebral cortical area map. Annu Rev Neurosci 26:355–380.[CrossRef][ISI][Medline]
Heikinheimo M, Lawshe A, Shackleford GM, Wilson DB, MacArthur CA (1994) Fgf-8 expression in the post-gastrulation mouse suggests roles in the development of the face, limbs and central nervous system. Mech Dev 48:129–138.[CrossRef][ISI][Medline]
Hevner RF, Daza RA, Rubenstein JL, Stunnenberg H, Olavarria JF, Englund C (2003) Beyond laminar fate: toward a molecular classification of cortical projection/pyramidal neurons. Dev Neurosci 25:139–151.[CrossRef][ISI][Medline]
Huffman KJ, Garel S, Rubenstein JL (2004) Fgf8 regulates the development of intra-neocortical projections. J Neurosci 24:8917–8923.
Kanold PO, Kara P, Reid RC, Shatz CJ (2003) Role of subplate neurons in functional maturation of visual cortical columns. Science 301:521–525.