Endocannabinoid signaling controls pyramidal cell
specification and long-range axon patterning
Jan Muldera,b, Tania Aguadoc, Erik Keimpemab, Klaudia Baraba ´sb, Carlos J. Ballester Rosadod, Laurent Nguyene,f,
Krisztina Monoryg, Giovanni Marsicanog,h, Vincenzo Di Marzoi, Yasmin L. Hurdj, Francois Guillemote, Ken Mackiek,
Beat Lutzg, Manuel Guzma ´nc, Hui-Chen Lud, Ismael Galve-Roperhc,l, and Tibor Harkanyb,l,m
aDepartment of Neuroscience, Retzius va ¨g 8, andmDivision of Molecular Neurobiology, Department of Medical Biochemistry and Biophysics, Scheeles va ¨g 1,
Karolinska Institutet, 17177 Stockholm, Sweden;bSchool of Medical Sciences, University of Aberdeen, AB25 2ZD Aberdeen, United Kingdom;cDepartment
of Biochemistry and Molecular Biology I, School of Biology, and Centro de Investigacio ´n Biome ´dica en Red sobre Enfermedades Neurodegenerativas,
Complutense University, 28040 Madrid, Spain;dDevelopmental Neurobiology Program and Department of Pediatrics, Baylor College of Medicine,
Feigin Center, 1102 Bates Street, Houston, TX 77030;eDivision of Molecular Neurobiology, National Institute for Medical Research, The Ridgeway,
Mill Hill, London NW7 1AA, United Kingdom;gDepartment of Physiological Chemistry, Johannes Gutenberg University, 55128 Mainz, Germany;
hAVENIR INSERM, Institut National de la Sante et de la Recherche Medicale, Institute Francois Magendie, 146 rue Le ´o Saignat, 33076 Bordeaux, France;
iEndocannabinoid Research Group, Institute of Biomolecular Chemistry, Consiglio Nazionale delle Ricerche, Via Campi Flegrei 34, 80078 Pozzuoli, Italy;
jDepartments of Psychiatry, Pharmacology, and Systems Therapeutics, Mount Sinai School of Medicine, One Gustave L. Levy Place, New York, NY 10029;
andkGill Center and Department of Psychological and Brain Sciences, Indiana University, 1101 East 10th Street, Bloomington, IN 47405
Communicated by Tomas Ho ¨kfelt, Karolinska Institutet, Stockholm, Sweden, April 15, 2008 (received for review December 9, 2007)
Endocannabinoids (eCBs) have recently been identified as axon
guidance cues shaping the connectivity of local GABAergic inter-
neurons in the developing cerebrum. However, eCB functions
during pyramidal cell specification and establishment of long-
range axonal connections are unknown. Here, we show that eCB
signaling is operational in subcortical proliferative zones from
embryonic day 12 in the mouse telencephalon and controls the
proliferation of pyramidal cell progenitors and radial migration of
immature pyramidal cells. When layer patterning is accomplished,
developing pyramidal cells rely on eCB signaling to initiate the
elongation and fasciculation of their long-range axons. Accord-
ingly, CB1 cannabinoid receptor (CB1R) null and pyramidal cell-
specific conditional mutant (CB1Rf/f,NEX-Cre) mice develop deficits in
axonal pathfinding becomes impaired after in utero pharmaco-
logical blockade of CB1Rs. Overall, eCBs are fundamental devel-
opmental cues controlling pyramidal cell development during
excitation ? glutamate ? layer patterning ? neocortex ? neurogenesis
the neuronal lineage occurs in the subcortical proliferative
pyramidal cells undergo radial migration to populate the cortical
plate (CP) (1), where they acquire layer-specific neurochemical
and morphological diversity (2). Pyramidal cell positioning and
patterning of their corticofugal and intracortical axons is in part
achieved via transcriptional control acting throughout cellular
identification (2). However, epigenetic microenvironmental
cues, provided by neural progenitors, radial glia, and immature
neurons, are also fundamental in attaining cortical cell identity
with particularly robust effects on pathfinding and directional
growth of long-range axons (3).
Endocannabinoids [eCBs; anandamide (AEA) and 2-arachi-
donoylglycerol] control various forms of synaptic plasticity at
cortical glutamatergic synapses in the postnatal brain (4)
through functional CB1 cannabinoid receptors (CB1Rs) (5).
During brain development, eCBs control neuronal fate decision
(6), interneuron migration (7), and axonal specification (8).
Developmental eCB actions are underpinned by a temporally
defined assembly of functional eCB signaling networks with
coincident expression of sn-1-diacylglycerol lipases (DAGL?/?)
(9) and N-arachidonoyl-phosphatidyl ethanolamine (NAPE)-
selective phospholipase D involved in eCB synthesis, fatty-acid
amide hydrolase (FAAH) (an enzyme preferentially degrading
AEA), and CB1Rs (8). The selective axonal targeting of CB1Rs
yramidal cell specification follows a sequential scenario in
the developing cerebrum: commitment of progenitor cells to
and DAGLs in immature neurons suggests that eCBs may
function in either cell-autonomous (6, 9) or target-derived (8)
manner to control axonal elongation and postsynaptic target
Although recent findings in both mammals (8) and nonmam-
malian vertebrates (10) suggest that eCB signaling is required for
axonal elongation and fasciculation, eCB functions instructing
distinct stages of pyramidal cell development are unknown.
Here, we show that eCB signaling is operational in the subcor-
tical VZ/SVZ and drives neural progenitor proliferation and
migration, thus contributing to defining the final positions and
densities of immature pyramidal cells. Subsequently, eCB sig-
naling in immature pyramidal cells is required for axonal polar-
ization and the formation of long-range glutamatergic axons.
Accordingly, genetic and pharmacological disruption of CB1R
functions reveals fasciculation deficits and axonal mistargeting.
In sum, our data demonstrate that eCB signaling is indispensable
for the genesis, proliferation, migration, and axonal behaviors of
neocortical pyramidal cells and support the concept that eCBs
act as a novel class of morphogens during corticogenesis.
CB1R Expression in Developing Cerebrum. We used in situ hybrid-
ization (11, 12) and immunofluorescence (8) histochemistry
combined with high-resolution confocal microscopy [see sup-
porting information (SI) Methods and Figs. S1 and S2] to define
the identity of cells expressing CB1R and DAGL?/? expression
in the developing mouse and human neocortex. CB1R mRNA
expression was restricted to telencephalic differentiation zones
at embryonic day (E)12.5 in mouse (Fig. 1 A and B). By E14.5,
significant CB1R mRNA expression was evident in immature
pyramidal cells populating the CP and hippocampal primordium
(Figs. 1C and Fig. S2 A–D), reached peak expression levels by
M.G., H.-C.L., I.G.-R., and T.H. designed research; J.M., T.A., E.K., K.B., C.J.B.R., L.N., Y.L.H.,
I.G.-R., and T.H. performed research; K. Monory, G.M., V.D.M., K. Mackie, and B.L. contrib-
uted new reagents/analytic tools; J.M., T.A., E.K., K.B., M.G., H.-C.L., I.G.-R., and T.H.
analyzed data; and J.M., T.A., K. Monory, G.M., V.D.M., Y.L.H., F.G., K. Mackie, B.L., M.G.,
H.-C.L., I.G.-R., and T.H. wrote the paper.
The authors declare no conflict of interest.
fPresent address: Centre for Cellular and Molecular Neurosciences, Developmental Neuro-
biology Unit, University of Lie `ge, 1 Avenue de l’Ho ˆpital, 4000 Lie `ge, Belgium.
lTo whom correspondence may be addressed. E-mail: firstname.lastname@example.org or t.harkany@
This article contains supporting information online at www.pnas.org/cgi/content/full/
© 2008 by The National Academy of Sciences of the USA
June 24, 2008 ?
vol. 105 ?
approximately E16.5 (Fig. 1D), and gradually declined during
pyramidal cell morphogenesis in the late gestational embryo
(Fig. S2 G–J). Considerable CB1R mRNA expression in neural
progenitors exiting the cortical SVZ was sustained throughout
corticogenesis, whereas the VZ was largely devoid of CB1R
hybridization signal (Fig. 1 E and G). These findings are not
restricted to mouse development, because a similar CB1R
mRNA expression pattern was seen in the second trimester
human fetal telencephalon, comparable to E18/postnatal day
(P)0 rodent brain (13). In particular, robust CB1R mRNA
expression was detected in the human SVZ, CP neurons, and
hippocampal pyramidal cell layers (Fig. 1 H and I).
eCB Signaling During Corticogenesis. DAGL?-like immunoreactiv-
ity (i.r.) was present in the cortical VZ/SVZ from E12.5 until
birth (Fig. 1 J and K and Fig. S3A) (10), suggesting the coinci-
dence of local eCB synthesis with neuronal progenitor prolifer-
ation. CB1R i.r. was evident in intermediate progenitor cells
(Tbr2?), known to differentiate into pyramidal cells (2), that had
engaged in radial migration toward the CP (Fig. 1 L and M and
Fig. S3). CB1R mRNA expression by immature pyramidal cells
concurred with the targeting of CB1Rs to developing long-range
axons between E13.5 and P0 (Fig. 1 N–R). Coexistence of
DAGL? and CB1Rs in developing glutamatergic axons (Fig. S2
E and F) reinforces the hypothesis (9) that eCB signaling is
required for pyramidal cell development and functional
eCBs Control SVZ Progenitor Proliferation. We tested whether eCBs
control neural progenitor proliferation (6) in subcortical VZ/
BrdU on E14.5. Lack of CB1Rs significantly decreased neural
progenitor proliferation (Fig. 2A). Conversely, FAAH?/?(14)
increased proliferation of VZ/SVZ progenitors (Fig. 2B). Here,
genetic manipulation of FAAH activity was used to elevate eCB
levels; however, the particular eCB mediating these phenomena
was not identified.
Arrows in C and E and G denote CB1R mRNA hybridization signal in pyramidal cells and SVZ progenitors, respectively. (H and I) Distribution of CB1R mRNA in human
fetal brain. Nissl/AChE histochemistry reveals territorial boundaries. (J–M) DAGL? and CB1R expression in mouse VZ/SVZ. Cortical Tbr2?projection neurons migrating
axons committed to the fibria. (Scale bars: E–G and K–M, 30 ?m; N–R, 85 ?m; A–D, H, and I, 100 ?m.) See SI Text for abbreviations.
(A) CB1R deletion significantly reduces the rate of neural progenitor pro-
liferation, as defined by the density of BrdU?cells in the VZ/SVZ. (B)
Conversely, elevated eCB levels in FAAH?/?mice (14) significantly increase
neural progenitor proliferation. (C) Conditional CB1R deletion in subcor-
tical SVZ progenitors (arrows) (16) decreases the rate of Ki67?progenitor
proliferation (n ? 3 per genotype).**, P ? 0.01, compared with wild-type
littermates. (Scale bar, 100 ?m.)
eCBs regulate neural progenitor proliferation in cortical VZ/SVZ.
Mulder et al.
June 24, 2008 ?
vol. 105 ?
no. 25 ?
eCB functions specifically underpinning pyramidal cell pro-
genitor proliferation were elucidated in mice with conditional
control of regulatory sequences of NEX, a neuronal basic
helix–loop–helix protein (15) (CB1Rf/f,NEX-Cre). Prominent Cre
activity is observed by approximately E11.5 in cortical progen-
itors in NEX-Cre mice (15), ensuring the lack of CB1Rs in
pyramidal cells at all developmental stages studied here. In these
mutants (16), VZ/SVZ progenitor proliferation was significantly
impaired, as indicated by reduced Ki67 and GOLGA5 cell
density in proliferative zones (Fig. 2C and Fig. S4A). Because
NEX is expressed by pyramidal progenitors in SVZ but not in
VZ (16), a coincident proliferation deficit in CB1Rf/f,NEX-Creand
CB1R?/?mice suggests that eCBs exert differential control on
subcortical progenitor pools. Therefore, we analyzed BrdU?
progenitor proliferation rates separately in SVZ and VZ of
CB1R?/?(SVZ, 55%; VZ, 77% of control) and FAAH?/?mice
(SVZ, 226%; VZ, 150% of control). Thus, progenitor prolifer-
ation in SVZ may rely more on autocrine eCB regulation,
whereas non-cell-autonomous signaling predominates in VZ.
Our genetic data were further validated by HU-210 (synthetic
CB1R agonist) and URB597 (FAAH inhibitor) treatment of
E14.5 organotypic slices revealing increased VZ/SVZ progenitor
proliferation upon sustained cannabinoid receptor stimulation
eCBs Regulate Pyramidal Precursor Migration. eCB effects on radial
migration of immature pyramidal cells was tested by injecting
pregnant mice (E14.5) with BrdU and allowing the offspring to
develop until P2.5, when layer-specific distribution of BrdU?
cells was determined. Profuse BrdU incorporation was found in
the brains of wild-type mice with labeled cells enriched in
neocortical layers (L)2 and 3 (bins 2–4 in Fig. 3A; see also Table
S1). In contrast, BrdU?cells accumulated in L4–6 (correspond-
ing to bins 6–10; Fig. 3A) in CB1R?/?mice. Likewise, cell
of BrdU?cells in L3 and L4 (bins 5–8 in Fig. 3B; see also Table
S1). In vivo evidence of a role for eCBs during radial cell
migration was confirmed by showing that HU-210 or URB597
application enhanced (Fig. 3C), whereas FAAH overexpression
inhibited (Fig. S4C) radial cell migration from the VZ/SVZ to
superficial cortical layers in organotypic cultures.
Pyramidal Cell Specification Relies on eCB Signaling. Exogenous
CB1R agonists have been identified (7, 8) as morphogens and
chemotropic guidance cues for cortical interneurons. However,
it is unknown whether eCBs affect pyramidal cell development.
Therefore, we exposed pyramidal cells isolated from E14.5
neocortex to NGF (100 ng/ml), a differentiation promoting
neurotrophin (17), AEA (200 nM) (8), or AM251 (1 ?M), a
CB1R inverse agonist. NGF increased axonal arbors of
VGLUT1?pyramidal cells (Fig. 4 A and B). In contrast, AEA
induced the elongation of a leading axon while inhibiting axon
branching. AM251 effects were reminiscent of those of NGF: an
expansion of axonal arbors and reversal of AEA effects (Fig. 4
A and B). Disrupting CB1R signaling by coapplication of AEA
and AM251 significantly inhibited neurochemical differentia-
tion by decreasing the density of VGLUT1?neurons (Fig. 4B).
Our data also support that pyramidal cells require an intrinsic
‘‘eCB tone’’ to initiate axonal polarization and neurochemical
differentiation. Both DAGLs were expressed by pyramidal cells
(Fig. 4C and Fig. S5A) and targeted to the axon (Fig. 4 D and E)
and navigating growth cones (Fig. 4F). Notably, DAGL? was
detected in elongating axon shafts and growth cones (including
filopodia) and exhibited proximal localization to CB1R (Fig.
S5A), whereas DAGL?-like i.r. was limited to axonal shafts.
DAGL? levels inversely correlated with axon development;
initially a random DAGL? distribution was seen in quiescent
axons, followed by DAGL? concentrating in axon varicosities
(18) (Fig. S5C). These data reveal that axonal DAGL? levels
remain high in developing axons and undergo redistribution
during axon maturation and synapse specification (Fig. S5C).
DAGL inhibition by O-3841 (19) significantly reduced VGLUT1
expression in pyramidal cells, thus supporting the involvement of
eCB signaling during the acquisition of a glutamatergic pheno-
type (Fig. 4G). Moreover, O-3841-induced increased synapto-
genesis in vitro (Fig. S5D) suggests that impaired temporal and
spatial integrity of eCB signaling in pyramidal cells may disrupt
postsynaptic targeting of glutamatergic axons.
CB1R Deletion Reveals Fasciculation Deficits. We assessed the in vivo
significance of our findings in CB1Rf/f,NEX-Cremice (15). In
new-born CB1Rf/f,NEX-Cre(Fig. 5 A and B?) but not CB1Rf/for
NEXCre/?mice (Fig. S6), L1 neuronal cell adhesion molecule
(L1-NCAM)?pyramidal axons formed bundles with aberrant
trajectories in the corpus callosum, failed to invade the dorsal
striatum, and exhibited a change in their striatal paths. In
CB1R?/?(null) mice, axonal fasciculation deficits were pro-
nounced at E14–16 (Fig. 5C) with a significant degree of
compensation by birth. Nevertheless, corticofugal axons invari-
ably failed to invade the dorsal striatum in CB1R?/?neonates
In Utero CB1R Blockade Disrupts Neurodevelopment. CB1R function
in long-range axon fasciculation was tested also by i.c.v. injection
Cortical distribution of neurons at P2.5 whose progenitors were BrdU labeled
at E14.5. Note the migration arrest of a population of neural progenitors in
deep cortical layers of CB1R?/?mice. (B) Cortical progenitor distribution in
FAAH?/?mice was assessed as above. Cell counts were performed in grouped
cortical layers defined as equal binned areas. (C) Distribution of GFP?cells in
brain slices from E14.5 mouse embryos after ex vivo electroporation of VZ
progenitors with pCIG2-GFP. Slices were maintained for 48 h in the presence
of HU-210 (1 ?M) or URB597 (1 ?M) in vitro. Cumulative cell counts were
obtained in CP, IZ, and VZ/SVZs.**, P ? 0.01;*, P ? 0.05, compared with
analysis). (Scale bars: A and B, 75 ?m; C, 35 ?m.)
eCBs control radial migration of pyramidal cell progenitors. (A)
www.pnas.org?cgi?doi?10.1073?pnas.0803545105 Mulder et al.
of SR141716 (10 mM in 1 ?l) in utero at E13.5, as a proof-of-
concept for CB1R antagonist actions. Analysis of corticofugal
connectivity at P8 revealed enlarged lateral ventricles (Fig. 5E)
in conjunction with impaired progenitor proliferation in the
subcortical VZ/SVZ (DMSO, 39 ? 3%; SR141716, 27 ? 3%
BrdU?cells; P ? 0.05; n ? 4), resulting in cortical delamination.
Expansion of the corpus callosum was due to a deficit in axonal
pathfinding, as indicated by a reduced commitment of pyramidal
cell axons to the developing callosal trajectory (Fig. 5 F and G).
Overall, data obtained in genetic models and after in utero
SR141716 treatment confirm that proliferation, migration, and
axonal pathfinding decisions of pyramidal cells are reliant on
eCB signaling through CB1Rs.
The molecular mechanisms of eCB actions in the adult brain are
well appreciated. However, a dearth of knowledge exists on the
cellular specificity of eCB actions in the developing brain (20).
In this article, we show the onset and spatial restriction of eCB
signaling during embryogenesis and demonstrate the cellular
roles of eCB signaling on neurochemically identified cortical
pyramidal cells and their progenitors. These data, together with
prior results on eCB control of neuronal progenitor cell fate and
GABA interneuron development, define the complexity of eCB
signaling in the developing cerebrum and identify eCBs as
developmental signals indispensable for cortical neuron speci-
fication and connectivity patterning (Fig. S7).
eCB Effects on VZ/SVZ Neural Progenitor Commitment. Recent stud-
ies demonstrate that eCBs are developmental signals helping to
determine neuronal identity at both the cellular and neuronal
network levels (6–8, 10, 20). This article significantly expands
existing knowledge on neuronal fate decisions during cortico-
genesis by showing that acquiring a glutamatergic neuronal
phenotype, in addition to eCB actions instructing interneuron
development (7, 8), is influenced by eCB signaling. First, the
coincident expression of eCB synthesis enzymes and receptors in
cortical progenitor zones demonstrates the existence of opera-
tional eCB signaling networks in neurogenic niches of the
developing brain, where eCBs tune the rate of neural progenitor
proliferation and the number of immature neurons committed to
radial migration. Our findings indicate that eCB signaling is
instrumental for embryonic neurogenesis, as also reported in
adult brain (21), and data from knockout mice support a role for
CB1Rs in committing SVZ progenitors to a pyramidal cell fate
(10). Our in vitro morphometric and biochemical data also
extend previous findings showing that both DAGL?/? are
recruited to elongating glutamatergic axons (9) with DAGL?
being the primary enzyme concentrating in axonal growth cones.
Concurrent with the recent hypothesis that DAGL functions are
required for axonal polarization and growth (9), we define a
negative correlation between DAGL?/? levels in axons and the
complexity of excitatory axon morphology (18). Also, we show
that inhibition of eCB synthesis attenuates neurochemical pyra-
midal cell differentiation by suppressing VGLUT1 expression, a
functional marker of glutamatergic synapses (22), and triggers
premature synapse formation.
in cortical neuron specification depends on the molecular iden-
tity of neurons. Pyramidal cells express CB1Rs as postmitotic
progenitors in the subcortical VZ/SVZ. However, GABAergic
progenitors in ganglionic eminences lack this receptor, and
immature GABAergic interneurons express CB1Rs only during
and after intracortical (radial) migration (8). Postmitotic, pyra-
midal cell lineage-committed neurons harbor the capacity of
eCB synthesis throughout their morphological and functional
except for sporadic NAPE-selective phospholipase D expres-
sion, eCB synthetic enzymes. Functional implications of these
findings include the following. First, eCB signaling is a key
determinant of the number of pyramidal cells destined to
particular cortical laminae. Second, eCB control of axonal
elongation in pyramidal cells is a primarily cell-autonomous
mechanism, with DAGLs and CB1Rs being in close proximity
predominantly along the axon stem allowing direct ligand-
receptor coupling. Conversely, CB1Rs concentrate in axonal
growth cones of GABAergic interneurons (8), suggesting that
treatment in vitro. Bracketed numbers indicate population sizes (see also Table S3). (C–E) DAGL? and ? undergo axonal targeting and exhibit differential
is targeted to filopodia (arrows), whereas DAGL? concentrates in the axon stem with a clear demarcation from the growth cones (arrowheads). (G) DAGL
inhibition by O-3841 significantly reduces VGLUT1 expression in pyramidal cells by 6 days in vitro. O-3841 effects were prevented by exogenous application of
AEA.*, P ? 0.05, compared with control. (Scale bars: A, 25 ?m; C, 15 ?m; D–F, 3 ?m.)
Mulder et al.
June 24, 2008 ?
vol. 105 ?
no. 25 ?
GABA cells use target-derived eCBs as microenvironmental
directional cues for growth cone steering decisions. Third, local
GABAergic and long-range glutamatergic axons are inherently
different with regards to the dynamics of their development,
arborization, and mechanisms of postsynaptic target selection.
Because axonal arbors of cortical interneurons exhibit a high
degree of complexity within volumetrically limited cellular mi-
crodomains, an autocrine eCB tone (9) would likely restrict
formation of ramified, complex axonal arbors by cortical inter-
neurons. In contrast, CB1R agonists trigger the elongation of a
single pyramidal cell axon with concomitant inhibition of col-
lateral formation in a manner reversed by CB1R inverse agonists
and NGF (data not shown). Consequently, our present genetic
and pharmacological findings revealing perturbed fasciculation
of corticofugal glutamatergic axons endowed with CB1Rs tar-
geted to the surface of their plasmalemma (8) and high focal
DAGL activity (Fig. S5) support the hypothesis that eCB
signaling has simultaneous roles in long-range axon develop-
ment: Autocrine eCB signaling (9) facilitates the formation of
axonal pathways (10), where an eCB gradient along individual
axons propels their extension, with focal elevation of eCB
concentrations among developing axons being critical to restrict
premature axonal dispersion. Thus, disrupting the temporal and
spatial dynamics of eCB signaling in the developing telenceph-
alon ultimately leads to aberrant axonal behaviors.
CB1R Knock-Outs Develop Axonal Deficits. Significant axon fascic-
ulation deficits were observed in neonatal CB1Rf/f,NEX-Cremice.
Notably, CB1R?/?mice exhibited disrupted axon development
at approximately E15 that was reminiscent of the phenotype in
conditional mice; however, in CB1R?/?mice the deficit normal-
the eCB system at the receptor level and suggest that the
expression of non-CB1Rs on neurons or redundancy of signaling
pathways (e.g., overt expression of other chemotropic/repulsive
guidance cues) may compensate for the complete loss of CB1R-
mediated signaling in CB1R?/?mice. The finding that pyramidal
cell CB1Rs are only abundantly expressed during the restricted
period of their morphological and functional specification
(E14.5–E18.5) argues that targeted disruption of CB1R signaling
during this period will selectively disrupt axon elongation and
targeting and impose permanent deficits to a restricted set of
neocortical pyramidal cells in CB1Rf/f,NEX-Creconditional
Overall, our present and previous (6–8) data define discrete
spatial and temporal niches for eCB action and identify the
cellular basis of their dichotomy on glutamatergic and GABAer-
gic cortical neurons. Although the identity of eCB(s) mediating
particular cellular actions has not been elucidated, developmen-
levels suggest that both ligands may control pyramidal cell
specification through their promiscuity at the CB1Rs. Consid-
ering the high degree of phylogenetic conservation of eCB
signaling, both at enzyme and receptor levels, across vertebrate
species (23), data described herein on mammalian neurons (8)
together with those on zebrafish and chick neurons (10) formu-
late the unifying concept that eCBs are key developmental cues
establishing neuronal diversity and synaptic connectivity in the
Animals. The generation and genotyping of glutamate decarboxylase 67-GFP
(?neo) (GAD67gfp/?) (8), CB1R?/?(11), FAAH?/?(14), CB1Rf/f,NEX-Cre(11), and
respective littermate controls has been reported elsewhere. Mouse and
Sprague–Dawley rat embryos and tissues were obtained from timed matings.
Neuroanatomy. In situ hybridization in mouse and human fetal brains was
performed as described (8, 11, 12). Multiple immunofluorescence labeling,
acetylcholinesterase histochemistry, Hoechst 33,258 (Sigma) and Nissl stains
were performed with quality-controlled immunoreagents (6, 8) (Fig. S1) with
n ? 2–5 brains per group analyzed by laser-scanning microscopy.
Cell Proliferation, Migration, and ex Vivo Electroporation. Cell proliferation and
migration was determined after i.p. BrdU injection (100 ?g/g) of pregnant
females at E14.5. For proliferation assays, embryos were harvested 2 h after
labeling. For migration experiments, pups were killed at P2 [n ? 6/6 (wild
type/CB1R?/?); n ? 7/7 (wild type/FAAH?/?), from three litters]. Cortical layers
were identified by their discrete cell densities as visualized by Hoechst 33528
(Sigma) and ?-III-tubulin (TuJ1; Promega) counterstaining. Ex vivo electropora-
tion was performed by using a pCIG2-GFP vector (24) and slice cultures were
maintained in semidry conditions in wells containing neurobasal medium (1%
rofloxacine). pCIG2-GFP vector was also used to overexpress FAAH in an IRES-
EGFP cassette under the control of a CMV enhancer and chicken ?-actin pro-
ulation deficits. (A) Conditional deletion of CB1Rs in pyramidal cells evokes
fasciculation deficits in L1-NCAM?long-range axons. Arrows point to en-
larged fascicles. Asterisk in open boxes denotes the general localization of
reminiscent of those seen in neonatal conditional mutants. (D) Impaired
axonal targeting toward the striatum persists in CB1R?/?until P2. (E) In utero
SR141716 infusion induces ventricle enlargement, migration arrest, and cor-
tical delamination. Arrowheads indicate clusters of progenitor cells in the
ventricular zone. Open boxes denote the position of insets. (F) Neurofilament
M staining reveals fasciculation deficits in the subventricular corpus callosum
after SR141716 infusion. (G) SR141716 decreases the commitment of pyrami-
dal cell axons to descending projections. (Scale bars: A, C, and E, 250 ?m; A
Inset and D, 100 ?m; B? and G, 30 ?m; F, 50 ?m.)
Genetic and pharmacological ablation of CB1Rs leads to axon fascic-
www.pnas.org?cgi?doi?10.1073?pnas.0803545105 Mulder et al.
moter. For proliferation studies, brain slices (n ? 6 per condition) were cultured Download full-text
acid 3?-carbamoyl-biphenyl-3-yl ester (URB597) was obtained from Cayman
Chemical. A minimum of six slices per treatment were analyzed (6).
Pyramidal Cell Cultures. Rat cortices were isolated at E14.5, cells were disso-
ciated by trypsin digestion (0.1%; 5 min) and plated at a density of 25,000 or
200,000 cells per well for morphometry or biochemical analysis, respectively.
Cultured neurons were maintained in DMEM/F12, supplemented with 1%
B27/1% glutamine/1% penicillin/streptomycin for 3–6 days (7). Ligands were
first added 12 h after cell seeding and replenished every other day (7), except
for O-3841 (1 ?M) (19), which was used between days 2–6 in vitro. Density of
glutamatergic neurons was defined as a ratio of VGLUT1?/TuJ1?cells from
?10 randomly selected view fields per coverslip. Morphometric analysis was
performed as described (7, 8), with also defining the longest axon segment.
Western blotting of SNAP25 and DAGL? was used to verify morphometric
In Utero SR141716 Treatment. The uteruses of anesthetized pregnant mice
(E13.5) were externalized, and a glass micropipette filled with 1 ?l SR141716
(10 mM) or DMSO (vehicle) was advanced through the uterine wall and into
the other served as an internal control. After drug infusion, the uterus was
gently repositioned and the abdominal wall was sutured. Mice were transcar-
dially perfused on P8 and processed (Fig. S1).
Statistics. Data were expressed as means ? SEM. Statistical analysis was
performed by either ANOVA with Student–Neuman–Keuls post hoc tests or
two-tailed unpaired Student’s t test on independent samples. P ? 0.05 was
considered statistically significant.
ACKNOWLEDGMENTS. We thank C. Ljungberg, E. Resel, and B. Julien for
technical assistance; Sanofi–Aventis (Montpellier, France) for SR141716; R.
La Jolla, CA) for FAAH?/?mice; and Y. Yanagawa (Gunma University, Mae-
bashi City, Japan) for GAD67gfp/?mice. This work was supported by the
Alzheimer’s Research Trust (J.M.), Alzheimer’s Association (K. Mackie and
T.H.), Swedish Medical Research Council (T.H.), European Molecular Biology
Organization Young Investigator Programme (T.H.), Hja ¨rnfoden (T.H.), Scot-
tish Universities Life Science Alliance (T.H.), and Deutsche Forschungsgemein-
schaft (B.L.). This work was also supported by National Institutes of Health
Grants DA11322, DA15916, and DA21696 (to K. Mackie); R01DA023214 (to
Y.L.H. and T.H.); NS048884 (to H.-C.L.); and ES07332 (to C.J.B.R.). I.G.-R. was
supported by Santander Complutense Grant PR27/05-13988, and M.G. was
supported by Comunidad de Madrid Grants S-SAL/0261/2006 and 950344.
1. Rakic P (2006) A century of progress in corticoneurogenesis: From silver impregnation
to genetic engineering. Cereb Cortex 16(Suppl 1):i3–i17.
2. Fishell G, Hanashima C (2008) Pyramidal neurons grow up and change their mind.
3. Lopez-Bendito G, et al. (2006) Tangential neuronal migration controls axon guidance:
a role for neuregulin-1 in thalamocortical axon navigation. Cell 125:127–142.
4. Domenici MR, et al. (2006) Cannabinoid receptor type 1 located on presynaptic
terminals of principal neurons in the forebrain controls glutamatergic synaptic trans-
mission. J Neurosci 26:5794–5799.
5. Kawamura Y, et al. (2006) The CB1 cannabinoid receptor is the major cannabinoid
6. Aguado T, et al. (2006) The endocannabinoid system promotes astroglial differentia-
tion by acting on neural progenitor cells. J Neurosci 26:1551–1561.
7. Berghuis P, et al. (2005) Endocannabinoids regulate interneuron migration and mor-
phogenesis by transactivating the TrkB receptor. Proc Natl Acad Sci USA 102:19115–
8. Berghuis P, et al. (2007) Hardwiring the brain: Endocannabinoids shape neuronal
connectivity. Science 316:1212–1216.
9. Bisogno T, et al. (2003) Cloning of the first sn1-DAG lipases points to the spatial and
temporal regulation of endocannabinoid signaling in the brain. J Cell Biol 163:463–468.
10. Watson S, Chambers D, Hobbs C, Doherty P, Graham A (2008) The cannabinoid
receptor, CB1, is required for normal axonal growth and fasciculation. Mol Cell
in the hippocampus. Neuron 51:455–466.
cannabinoid receptor mRNA in the human fetal brain. Neuroscience 118:681–694.
14. Cravatt BF, et al. (2001) Supersensitivity to anandamide and enhanced endogenous
cannabinoid signaling in mice lacking fatty acid amide hydrolase. Proc Natl Acad Sci
15. Goebbels S, et al. (2006) Genetic targeting of principal neurons in neocortex and
hippocampus of NEX-Cre mice. Genesis 44:611–621.
16. Wu SX, et al. (2005) Pyramidal neurons of upper cortical layers generated by NEX-
17. Patapoutian A, Reichardt LF (2001) Trk receptors: Mediators of neurotrophin action.
Curr Opin Neurobiol 11:272–280.
18. Dotti CG, Sullivan CA, Banker GA (1988) The establishment of polarity by hippocampal
neurons in culture. J Neurosci 8:1454–1468.
19. Bisogno T, et al. (2006) Development of the first potent and specific inhibitors of
endocannabinoid biosynthesis. Biochim Biophys Acta 1761:205–212.
20. Harkany T, et al. (2007) The emerging functions of endocannabinoid signaling during
CNS development. Trends Pharmacol Sci 28:83–92.
mice. Mol Pharmacol 66:204–208.
two classes of excitatory synapse. Neuron 31:247–260.
23. McPartland JM (2004) Phylogenomic and chemotaxonomic analysis of the endocan-
nabinoid system. Brain Res Brain Res Rev 45:18–29.
24. Nguyen L, et al. (2006) p27kip1 independently promotes neuronal differentiation and
migration in the cerebral cortex. Genes Dev 20:1511–1524.
Mulder et al.
June 24, 2008 ?
vol. 105 ?
no. 25 ?