conserved mir-8/mir-200 defines a glial niche that controls … · 2017. 2. 24. · developmental...
TRANSCRIPT
Developmental Cell
Article
Conserved miR-8/miR-200 Definesa Glial Niche that Controls NeuroepithelialExpansion and Neuroblast TransitionJavier Morante,1,* Diana M. Vallejo,1 Claude Desplan,2 and Maria Dominguez11Instituto de Neurociencias, Consejo Superior de Investigaciones Cientıficas y Universidad Miguel Hernandez, Av Santiago Ramon
y Cajal s/n, 03550 San Juan de Alicante, Spain2Center for Developmental Genetics, Department of Biology, New York University, 100 Washington Square East, New York, NY 10003, USA*Correspondence: [email protected]
http://dx.doi.org/10.1016/j.devcel.2013.09.018
SUMMARY
Neuroepithelial cell proliferation must be carefullybalanced with the transition to neuroblast (neuralstem cell) to control neurogenesis. Here, we showthat loss of the Drosophila microRNA mir-8 (thehomolog of vertebrate miR-200 family) results inboth excess proliferation and ectopic neuroblasttransition. Unexpectedly, mir-8 is expressed in asubpopulation of optic-lobe-associated cortex gliathat extend processes that ensheath the neuroepi-thelium, suggesting that glia cells communicatewith the neuroepithelium. We provide evidence thatmiR-8-positive glia express Spitz, a transforminggrowth factor a (TGF-a)-like ligand that triggersepidermal growth factor receptor (EGFR) activationto promote neuroepithelial proliferation and neuro-blast formation. Further, our experiments suggestthat miR-8 ensures both a correct glial architectureand the spatiotemporal control of Spitz protein syn-thesis via direct binding to Spitz 30 UTR. Together,these results establish glial-derived cues as key reg-ulatory elements in the control of neuroepithelial cellproliferation and the neuroblast transition.
INTRODUCTION
The correct regulation of adult brain size and function is a funda-
mental process that requires the careful regulation of neural stem
cell numbers during early neurogenesis. Inmature tissues, neural
stem cell number is precisely controlled by signal(s) from the
surrounding neural stem cell niche. However, evidence for the
occurrence and need of a niche microenvironment during early
neurogenesis has remained elusive (Knoblich, 2008). The optic
lobe of Drosophila is a well characterized model to study early
neurogenesis (Brand and Livesey, 2011) and the neural networks
underlying vision circuits (Morante and Desplan, 2008). Each
optic lobe derives from neural stem cells generated in a stereo-
typed pattern from two columnar neuroepithelia called the inner
(IPC) and outer (OPC) proliferation centers (White and Kankel,
1978) (Figures S1A–S1C available online). The OPC generates
174 Developmental Cell 27, 174–187, October 28, 2013 ª2013 Elsev
lamina and medulla neurons (Morante et al., 2011). The emer-
gence of neuroblasts in the part of the OPC neuroepithelium
that generates the medulla is dependent on the expression of
Lethal of scute [l(1)sc], a proneural basic helix-loop-helix protein
that specifies neuroblast fate (Yasugi et al., 2008). Expression of
l(1)sc progresses like a wave that resembles the morphogenetic
furrow in the fly eye imaginal disc (Treisman and Heberlein,
1998). As in the retina, the progression of neural development
associated with the expression of L(1)sc expression is positively
regulated by epidermal growth factor receptor (EGFR) signaling
and negatively regulated by Notch signaling (Egger et al., 2010;
Ngo et al., 2010; Reddy et al., 2010; Wang et al., 2011b; Yasugi
et al., 2010). Ahead of the wave that produces neuroblasts,
uncommitted neuroepithelial cells continue to divide symmetri-
cally, expanding the pool of prospective neuroblasts.
Increased signaling from the JAK/STAT and EGFR pathways
(Ngo et al., 2010; Wang et al., 2011a; Yasugi et al., 2008), or
loss of Hippo pathway activity (Kawamori et al., 2011; Reddy
et al., 2010), or of lethal (3) malignant brain tumor (Richter
et al., 2011) all cause oversized neuroepithelium and tumor
growth associated with a delay or blockade of the emergence
of neuroblasts. In contrast, loss of the Notch pathway has
the opposite consequence, advancing neuroblast progression
and causing a premature termination of neuroepithelial growth
(Egger et al., 2010; Ngo et al., 2010; Reddy et al., 2010; Wang
et al., 2011b; Yasugi et al., 2010). These observations indicate
that the EGFR pathway affects both the rate of neuroepithelial
proliferation and the rate of transition to neuroblast, reminiscent
to the reiterated use of the EGFR/Ras pathway during retinal
development (Domınguez et al., 1998; Freeman, 1996). Thus,
signaling through the EGFR/Ras pathway provides a good
candidate for coordinating proliferation with neuroblast
emergence. However, it remains obscure how the strength
and spatiotemporal activation of the EGF receptor pathway is
regulated to ensure proper brain size and pattern.
Glial cells are essential for the maintenance of nervous sys-
tem homeostasis, and the loss of glial function induces adult
neural degeneration (Kretzschmar et al., 1997). Thus far, two
distinct layers of surface glia ensheathing the Drosophila larval
brain have been described: the perineurial sheath of astro-
cyte-like cells, which divide throughout larval development;
and a layer of subperineurial glia that form septate junctions
and act as the main blood-brain barrier (Awasaki et al., 2008;
Bainton et al., 2005; Pereanu et al., 2005; Schwabe et al.,
ier Inc.
Developmental Cell
A MicroRNA-Dependent Niche for Early Neurogenesis
2005; Stork et al., 2008; reviewed in DeSalvo et al., 2011;
Edwards and Meinertzhagen, 2010; Hartenstein, 2011; Stork
et al., 2012). Here, we describe a glial cell layer ensheathing
the optic lobe neuroepithelium, ‘‘optic-lobe-associated cortex
glia,’’ which is distinct from the two previously characterized
surface-associated glia cell populations, the perineurial and
subperineurial glial cells. We show that these glia are the source
of the EGFR ligand Spitz whose production is controlled by a
microRNA (miRNA).
Here, we show that glia ensheathing the developing optic
lobe neuroepithelium are marked by miRNA mir-8, which con-
trols their ability to produce the EGFR ligand Spitz. miR-8 is
the sole fly homolog of the human mir-200 gene family of tumor
suppressors (Brabletz et al., 2011; Vallejo et al., 2011) that
control epithelial-to-mesenchymal reprogramming (Park et al.,
2008; Wellner et al., 2009) and the acquisition of stemness (Shi-
mono et al., 2009). Previous work on this miRNA in flies (Hyun
et al., 2009; Karres et al., 2007) has demonstrated effects on
the nervous system, motor activity, leg development, and over-
all body size. Our analyses unveil a dual role for miR-8 in optic-
lobe-associated cortex glia: miR-8 is required to increase the
size of the glia, thereby facilitating sprouting toward the neuro-
epithelium; it also negatively regulates Spitz levels by directly
binding to its 30 UTR, thereby preventing premature and excess
signaling in the underlying neuroepithelium. Consistently, loss
and gain of spitz in the miR-8+ cortex glia mimicked the neuro-
epithelial defects observed following mir-8 gain and loss,
respectively, whereas expression of spitz lacking its 30 UTR
rescued defects caused by mir-8 overexpression. Taken
together, these findings define a prominent role for an optic-
lobe-associated cortex glia in controlling EGFR signaling and
thus the balance between neuroepithelial proliferation and
neuroblast emergence.
RESULTS
The Conserved MicroRNA miR-8 Is Expressed in GliaLoss ofmir-8 is associated with increased apoptosis in the larval
Drosophila brain (Karres et al., 2007). To further investigate the
developmental role of miR-8 in neurogenesis, we analyzed cells
expressing mir-8 at larval stages using a mir-8-Gal4 enhancer
trap line (Karres et al., 2007) and a nuclear RFP reporter (UAS-
H2B::RFP, Figures 1A–1D).
We detected mir-8 expression in the central brain and optic
lobe using both mir-8-Gal4 and in situ hybridization (Figures 1A
and S1D–S1H). mir-8 was detected in cortex glia in the central
brain and in glia surrounding the optic lobe neuroepithelium
starting before late third (L3) stage (Figures 1A–1D). mir-8 was
not detected in neuroblasts or in neuroepithelial cells (Figures
1A–1D, S1E, and S1F).
We used a membrane-tagged GFP reporter (UAS-
mCD8::GFP, Figures 1E–1H) to more precisely define the struc-
ture of the miR-8+ glia in the optic lobe. Serial optical sectioning
revealed that miR-8 positive optic lobe glia lie on the surface
(Figure 1E) and in the boundary between the optic lobe and the
central brain and project long extensions (arrowheads, Figures
1F–1H). This notion is further confirmed by single-cell labeling
using mir-8-Gal4 driving a FLP-out construct (Figures 1I–1L).
Moreover, these GFP-marked clones highlighted that surface-
Developm
associated miR-8-glia cells with large and elaborate arboriza-
tions ensheathing the developing optic lobe neuroepithelium
(Figure 1J) and central brain miR-8 cortex glia that ensheath neu-
roblast and neuronal lineages (arrowheads, Figure 1L) likely orig-
inate from distinct progenitor cells as we never recovered clones
containing both types ofmiR-8 glia.We determined that a normal
brain contains on average 137.33 ± 6.53 (mean ± SEM) miR-8+
glia within the optic lobe (Figure 1A). Of those, 25.16 ± 1.99
(mean ± SEM) glia cells lie on the surface of the optic lobe and
show the elaborated morphology illustrated in Figure 1J.
miR-8+ Surface-Associated Glia Lie Underneath theSubperineurial Glial Cell LayerPrevious studies classified the surface-associated glia surround-
ing the developing Drosophila optic lobes into two populations
(Awasaki et al., 2008; Stork et al., 2008; reviewed in DeSalvo
et al., 2011; Edwards and Meinertzhagen, 2010; Hartenstein,
2011; Stork et al., 2012): the perineurial glial cells, which are
located in the outermost superficial layer and the blood-brain-
barrier (also known as subperineurial glia), beneath the perineu-
rial layer. We found that miR-8+ glia constitute a glial layer that
lies underneath the subperineurial glial cell layer as revealed
by labeling them with the miR-8 driver (mir-8>mCD8::GFP)
and the subperineurial markers Moody (Bainton et al., 2005;
Schwabe et al., 2005) (Figures 2B and 2C) and Nrx-IV-GFP
(DeSalvo et al., 2011) (Figures S2A and S2B).
We also compared the nuclei and processes of miR-8 and
subperineurial glia labeled with an UAS-DsRed reporter under
the control of moody-Gal4 (arrows in Figures 2D–2F, S2C, and
S2D) in L3 larval optic lobes. Figures 2G and 2H show the
distinctive sprouting and locations of miR-8+ and subperineurial
glia (arrowheads and arrows point to mir-8 and moody glia,
respectively) in adult optic lobes. Therefore, miR-8 expression
defines a layer of glia underneath the subperineurial glial cell
layer in both the developing and adult optic lobe. Hereafter, we
refer to these glia as ‘‘optic-lobe-associated cortex glia’’
(Figure 2I).
Glial miR-8 Controls Neuroepithelial ExpansionThe developing optic lobe progenitor region is formed by
a growing neuroepithelial zone and a ‘‘neurogenic’’ zone where
neuroblasts are generated in a wave crossing the neuroepithe-
lium. In L3 wild-type brains, the neuroepithelium can be labeled
by staining for epithelial junctional markers Patj, DE-Cadherin
(DE-Cad), or Discs large (Dlg) (Figures 3A, S3A, and S3D) (Egger
et al., 2007). Lethal (1) of scute [L(1)sc] marks the transition zone
from neuroepithelial cells to neuroblasts (Figure S3D) (Yasugi
et al., 2008), which are labeled by the expression of Miranda
(Mira) (Figures 3A and S3A) (Egger et al., 2007; Morante et al.,
2011). The expression of Mira expands across the neuroepithe-
lium as neuroblasts are generated (Yasugi et al., 2008).
In flies lacking mir-8 (mir-8D2) (Karres et al., 2007), the neu-
roepithelium of the larva is 143% overgrown (11,565.11 ±
510.65 mm2, [mean ± SEM], n = 14 optic lobes scored) as
compared with wild-type (WT) L3 neuroepithelium (7,907.54 ±
732.1 mm2, n = 13). In the subsequent stage, ectopic and
precocious neuroblasts as visualized by the ectopic unevenly
spaced Mira- and L(1)sc-positive cells scattered throughout
the enlarged neuroepithelium (Figures 3B, 3E, and S3E). This
ental Cell 27, 174–187, October 28, 2013 ª2013 Elsevier Inc. 175
Figure 1. The Conserved MicroRNA miR-8 Is Expressed in Glia
(A–D) miR-8 labels a subset of Repo (red) glial cells. (A) Whole-mount view of a L3 brain and (B–D) details of L1, L2, and L3 brains from flies carrying mir-8-Gal4
driving a nuclear UAS-H2B::RFP (green). The neuroepithelial cells (NE) are labeled by DE-Cadherin (DE-Cad, blue).
(E–H) Magnification and serial optical sections from anterior (F) to posterior (H) showing miR-8+ cortex glia structure visualized by mCD8::GFP (green). NE is
labeled for DE-Cad (blue).
(I) Scheme to generate GFP-labeled mir-8-Gal4-positive glial clones in hsp70-FLP/+; mir-8-Gal4/UAS (FRT) CD2, y+ (FRT) mCD8::GFP.
(J–L) Examples shown are the distinct classes of FLP-out clones recovered. (J) Single cell clones (glial nuclei visualized by Repo label) of surface-associated glia
overlying the NE. (K–L) Single or few cell-clones of optic-lobe-associated cortex glia (K) and central brain cortex glia that ensheath neuronal cell bodies (K) and
neuroblast (arrowheads, L). CB, central brain; IPC, inner proliferation center; OL, optic lobe; OPC, outer proliferation center. Scale bars represent 50 mm (A, F–H)
and 20 mm (B–E, J–L).
See also Figure S1.
Developmental Cell
A MicroRNA-Dependent Niche for Early Neurogenesis
suggests that miR-8 is required both to limit the growth and neu-
roblast formation within the neuroepithelium.
Next, we investigated the effects of glial-specific overexpres-
sion ofmir-8 (UAS-mir-8) (Vallejo et al., 2011) using the pan-glial
driver repo-Gal4 (Sepp et al., 2001). mir-8 overexpression pro-
duced defects that were opposite to those seen in mir-8-defi-
cient animals, with a 41% reduction in the size of the underlying
neuroepithelium (4,713.32 ± 509.09 mm2, n = 12) and overall brain
size (Figures 3C, 3E, and S3C). This smaller neuroepithelial size
was coupled with accelerated progression of the proneural wave
marked by L(1)sc (Figure S3F).
To test the endogenous requirement of mir-8 in optic-lobe-
associated cortex glia, we attempted to generate genetic
176 Developmental Cell 27, 174–187, October 28, 2013 ª2013 Elsev
mosaicsofmir-8—gliabut only recovered rareclonesof singleglial
cell, possibly reflecting the notion that these glia cells grow by en-
doreplication without dividing during larval stages (Stork et al.,
2012 and see below). We then used amiR-8microRNA ‘‘sponge’’
(UAS-mir-8-Sponge(SP)::GFP) (Loyaet al., 2009) expressed in the
endogenous mir-8 domain using mir-8-Gal4. miR-8 glia-specific
depletion of mir-8 caused defects similar to those observed in
wholly nullmir-8mutant animals (Figure 3D). For example, expres-
sion of two copies of theUAS-miR-8-(SP)::GFP bymir-8-Gal4 led
to an expansion of the neuroepithelium and ectopic neuroblasts
(Mira-positive cells) scattered throughout the enlarged neuroepi-
thelium. Thus, glia-specific gain and loss of mir-8 impact on the
growth and the emergence of neuroblast in the neuroepithelium.
ier Inc.
Figure 2. Surface-Associated miR-8 Glia
Layer Lies Underneath the Subperineurial
Glia
(A–C) Images showmiR-8+ cells (mir-8>H2B::RFP,
A ormCD8::GFP, B-C) constitute a distinct class of
surface-associated glia from that of the sub-
perineurial glia as is visualized by Moody-b anti-
body (green, B andC). Arrowhead points tomiR-8+
glia and arrow to subperineurial (miR-8� Moody+)
glia.
(D–F) Subperineurial glial membranes are labeled
by moody-Gal4 driving UAS-DsRed (red, arrow)
and glia nuclei by Repo. Single (E) and double
channel confocal image (Repo and moo-
dy>DsRed: F) are shown.
(G) Expression pattern of mir-8>DsRed in an adult
medulla cortex (MeC). Arrowheads point cortex
glia.
(H) Expression pattern of moody>DsRed in an
adult medulla cortex. Arrows point to sub-
perineurial glia. Neuropil and glia in (G) and (H) are
visualized by DN-Cad (blue) and Repo (green).
Insets show whole-mount views of adult optic
lobes.
(I) Schematic overview of the three layers of sur-
face-associated in a L3 brain lobe: miR-8+ (this
study), subperineurial (SPG) and perineurial (PG)
glial layers as single cells and highlighted in
different colors. NE, neuroepithelial cells. Scale
bar represents 20 mm.
See also Figure S2.
Developmental Cell
A MicroRNA-Dependent Niche for Early Neurogenesis
The Size and Sprouting of Optic-Lobe-AssociatedCortex Glia Is Controlled by miR-8These observations revealed communication between glia and
the underlying neuroepithelium. They suggest that miR-8+ glia
define an anatomical niche that provides structure and/or signals
for neuroepithelial progenitors.
miR-8+ optic-lobe-associated cortex glia exhibit characteristic
large nuclear and somatic sizes (44.42 ± 2.37 mm2, mean ± SEM:
n = 91 glial cells in 10 brains) as compared with subperineurial
glia (as visualized by Moody marker) (19.29 ± 1.74 mm2, 48 cells
in six brains) and perineurial glia (12.89 ± 0.59 mm2, 100 cells in
six brains) (quantification in Figure 4A; see Experimental Proce-
dures). Overexpression ofmir-8 in all glia by repo-Gal4 increased
further the glial nuclear size and processes. See quantification in
Figure 4A and compare count in Figure 4B versus 4C in late third
3D-brain reconstructions. Thus, glial cells immediately overlying
the optic lobe neuroepithelium (Moody negative and hence pre-
sumably the endogenous miR-8 glial cell population, arrow-
heads in Figures 4E and 4G) increased their size on average
400% (184.816 ± 15.48 mm2: n = 70 cells in 10 brains) and the
Moody-positive glia on top of this layer increased their size by
�500% (109.096 ± 10.979 mm2: n = 35 cells in ten brains, arrows
in Figures 4E and 4G). See also whole-mount (Figures 4D and 4F)
Developmental Cell 27, 174–187,
and high-magnification (Figures 4E and
4G) views. Central brain cortex glia also
increased their size (Figures S4A and
S4B) in response to mir-8 overexpres-
sion. However, satellite glia, epithelial
glia, marginal glia, and medulla neuropil
glia (Hartenstein, 2011; hereafter referred to as inner glia) did
not change cell size. Thus, control inner glial size area was
11.653 ± 0.86 mm2 (n = 47 cells in six brains scored) and
13.643 ± 1.16 mm2 in repo>mir-8 brains (n = 65 cells in six brains
scored) (Figures 4D and 4F and quantification in Figure 4A).
FLP-out clones overexpressing mir-8 (Figures S4C and S4D)
recovered contained a single cell, suggesting that mir-8 overex-
pressing glial cells do not proliferate (Stork et al., 2012; Vallejo
et al., 2011). These single-cell FLP-out mir-8 expressing clones
increased glial cell size cell-autonomously. Additionally, the
number of non-miR-8 surface-associated glial cells was reduced
from 78.91 ± 2.86 perineurial (i.e., Moody-negative) glia and
17.75 ± 1.48 subperineurial (i.e., Moody+) glia in WT (n = 21 optic
lobes scored) to 9.27 ± 0.91 glia in repo>mir-8 optic lobes (n =
18), most of them being Moody+ glia (Figures 4B and 4C). The
reduction of glial cell number may result in part from migrating
defects throughout the optic stalk (see Figures S4F and S4G)
and/or inhibition of cell proliferation. The role of miR-8 in cell
growth is evolutionarily conserved because overexpression of
human mir-200c by repo-Gal4 similarly increased cell-autono-
mously surface glial size (Figure S4E).
Cell size is often proportional to the amount of nuclear DNA
and many differentiating cells increase in size by undergoing
October 28, 2013 ª2013 Elsevier Inc. 177
Figure 3. Glial miR-8 Expression Controls
Neuroepithelial Expansion
(A–C) Whole-mount views of L3 brains in WT (A),
mir-8D2 (B) and animals overexpressing mir-8 by
repo-Gal4 (C). The NE are visualized by DE-Cad
(blue) and Patj (green) and the location of neuro-
blasts by Mira (red). (B) Larger neuroepithelium in
mir-8D2 brains (arrowheads) and precocious neu-
roblast transitions (arrows). (C) Image illustrates
the smaller neuroepithelial size (as measured in E)
in animals overexpressing mir-8 by repo-Gal4.
(D) Glia-specific depletion of mir-8 by expressing
two copies of UAS-mir-8-SP::GFP with mir-8-
Gal4.
(E) Histogram shows NE area of the indicated ge-
notypes measured using ImageJ (n = 13 [repo >],
n = 14 [mir-8D2] and n = 12 [repo>mir-8] brain lobes
scored). Shown are mean ± SEM and p values
were calculated by two-tailed unpaired t test.
Scale bar represents 50 mm.
See also Figure S3.
Developmental Cell
A MicroRNA-Dependent Niche for Early Neurogenesis
endoreplication without cell division (Park and Asano, 2008;
Zielke et al., 2011). Indeed, key markers of endoreplication,
including the APCFzr/Cdh1–lacZ enhancer trap line (Zielke et al.,
2008), were specifically only activated in central brain and op-
tic-lobe-associated cortex glial cells positive for mir-8 and spitz
in L3 larvae (Figures 4H and S4H–S4J; see also Figure S5),
consistent with the lack of mitotic figures in these cells (Figures
4I and S4K–S4M). Indeed, in L3 brains, we could only seemitosis
in neuroepithelial cells and neuroblasts (Figure S4N).
Similar to the increased cell size observed by overexpressing
mir-8 (Figures 4F and 4G), we observed increase glial cell size
when the fly homolog of the essential endoreplication factor
cdt1 (called dup in Drosophila) (Whittaker et al., 2000) was over-
expressed by repo-Gal4 (Figure S4O), albeit the effect was less
strong than when mir-8 was overexpressed. Conversely, RNA
interference (RNAi) silencing of dup/Cdt1 or of components of
the prereplication complex (Table S1) (Park and Asano, 2008)
in glial cells reduced cell number and size of the surface glia
(14 ± 1.89 mm2 in repo>dup-IR brains; n = 30 cells in 10 brains,
Figure 4A). In these repo>dup-IR brains, the only remaining sur-
face glia (10 ± 0.81 cells, n = 10) displayed short processes as
visualized by mCD8::GFP (arrowheads in Figure 4J and 4K and
compare with Figures 4D–4G) and were positive for the subper-
ineurial marker Moody. These alterations in size and sprouting
were fully reversed by expressing mir-8 in a repo>dup-IR back-
ground (Figures 4L and 4M). Thus, miR-8 seems to influence glial
size via a mechanism that might involve the depletion of a nega-
tive regulator of endoreplication, whose identity is unknown.
The underlying neuroepithelia in brains with smaller glia in re-
po>dup-IR and repo>orc1-IR were highly disorganized as visu-
alized by Patj staining (Figures 4N and 4O) with intermingled,
ectopic neuroepithelial-neuroblast transitions, as evident by
L(1)sc and Mira labeling (Figures S4P–S4S). This nonautono-
178 Developmental Cell 27, 174–187, October 28, 2013 ª2013 Elsevier Inc.
mous effect is specific for miR-8 glial
cells because selective expression of
mir-8 transgene in the endoreplicating
subperineurial glia cells (Unhavaithaya
and Orr-Weaver, 2012) usingmoody-Gal4 resulted in overgrown
glial cells (Figures S4T–S4Y), but the growth of the underlying
neuroepithelium and the timing of neuroblast transition were
not unaffected. Thus, we concluded that the proper structure
and signaling of miR-8 glia is necessary for proper neuroepithe-
lium development. We speculate that the nonautonomous influ-
ence of miR-8 glia likely involve direct regulation of a secreted
factor(s) by the microRNA.
Optic-Lobe-Associated Cortex Glia Is a Source of theEGFR Ligand SpitzThe brain defects observed in mir-8-deficient flies resembled
those seen in argos (aos) mutants (Figures 5C and 5D), where
we observed an enlarged neuroepithelium and then in later stage
a disruption of the orderly specification of neuroblasts, with
ectopic formation of neuroblasts visualized with Mira (compare
Figure 5D with Figure S3E). The aos gene encodes a negative
regulator of EGF receptor signaling that sequesters the EGFR
ligand TGF-a-like called Spitz (Freeman et al., 1992; Klein
et al., 2004). Spitz/TGF-a-like ligand is produced as an inactive
precursor that is converted to the active form when it is pro-
cessed by the intramembrane protease Rhomboid (Rho) (Sturte-
vant et al., 1993).
In the L3 brain, the aos-lacZ (Figure 5E) and rho-lacZ
(Figure 5F) enhancer traps were specifically expressed in the
optic-lobe-associated cortex glia. Rho is generally produced
by the same cells that express the ligand Spitz and its negative
regulator Aos. Consistently, spitz expression was detected in
miR-8+ cortex glial cells as visualized using two independent
enhancer traps (Figures 5G and S5A–S5C and compare with Fig-
ures 1A and 1D) spi-lacZ and spi-Gal4 (Jiang and Edgar, 2009).
No expression of the other EGFR ligands vein/Neuregulin and
keren was observed and downregulation of their expression by
Figure 4. Structure of the Surface Glia and Their Effects on the Neuroepithelium
(A) Histogram represents the means ± SEM of the glial nuclear area of wild-type perineurial (red bar), Moody+ subperineurial (green), miR-8+ surface cortex glia
(dark blue), and inner glia (gray); the glial nuclear area of non-miR-8 surface (green), miR-8 surface cortex glia (dark blue) and inner (gray) glia upon mir-8
overexpression, and the glial nuclear area of Moody-subperineurial glia (green) and inner (gray) glia upon dup/cdt1 depletion by RNAi (dup-IR) by repo-Gal4.
(B and C) Illustrative examples of 3D reconstructions of whole-mount WT (B) and repo>mir-8 (C) L3 brains.
(D–G) Changed size in glia cells of control WT (D-E) and repo>mir-8>CD8::GFP (F-G) L3 brain lobes. Glia nuclei were visualized with Repo (D, F, red) and
neuroepithelia stained with DE-Cad (D, F) and Dlg (E, blue). Arrowheads in (E) and (G) point to miR-8+ glia (green). Subperineurial (SPG) glia are visualized by
Moody (red, arrow in E and G).
(H) APCFzr/Cdh1-lacZ (green) labels a subset of glial cells that by position agrees with miR-8+ glia.
(I) miR-8+ glia (mir-8>H2B::RFP, green) are always negative for pH3 mitotic marker (red).
(J and K) Depletion of dup by RNAi (dup-IR) in repo cells. Arrowheads in (J) point to disrupted surface glia, which are all positive for Moody (red; arrow in K).
(L–O) The glial undergrowth defect of repo>dup-IR is rescued bymir-8 overexpression (repo>dup-IR>mir-8). Glial nuclei were visualized by Repo (red; L, M), the
NE by Patj (green, N, O) and DE-Cad (blue, L–O) and neuroblasts by Mira (red, N and O). CB, central brain; IPC, inner proliferation center; OPC, outer proliferation
center; VNC, ventral nerve cord. Scale bars represent 100 mm (B and C), 50 mm (D, F, H–J, L), and 20 mm (E, G, K, M–O).
See also Figure S4 and Table S1.
Developmental Cell
A MicroRNA-Dependent Niche for Early Neurogenesis
Developmental Cell 27, 174–187, October 28, 2013 ª2013 Elsevier Inc. 179
Figure 5. Optic-Lobe-Associated Cortex
Glia Is a Source of the EGFR Ligand Spitz
(A–D) Whole brain (A and C) and a magnification of
an area (B and D) from a L3 WT (A and B) and
homozygous aosD7 animals (C and D). The NE is
visualized by DE-Cad (blue) and neuroblasts by
Mira (red).
(E–G) Enhancer-traps showing the transcription of
aos-lacZ (E), rho-lacZ (F) (green staining) occurs in
glia (Repo+, red). Coexpression of spi-lacZ (green)
and mir-8>H2B::RFP (G) in optic lobe-associated
cortex glia. Scale bars represent 50 mm (A and C)
and 20 mm (B, D, E–G).
See also Figure S5.
Developmental Cell
A MicroRNA-Dependent Niche for Early Neurogenesis
RNA interference had no effect on the neuroepithelial phenotype
(data not shown).
We therefore investigated the defects in the neuroepithelium
following specific interference with spi and aos genes in the
miR-8+ glia. We first used the biologically active processed
form of Spitz tagged with GFP (UAS-sSpi::GFP) (Miura et al.,
2006) to activate the EGF receptor, while allowing us to track
localization of the Spi protein. This construct was expressed us-
ing mir-8-Gal4 (Figure 6). The observed effect was comparable
with that seen using repo-Gal4, which drives expression in all
glia (Figures S6A and S6B).
180 Developmental Cell 27, 174–187, October 28, 2013 ª2013 Elsevier Inc.
The presence of overexpressed
sSpi::GFP in miR-8 surface cortex glial
cells was evident in puncta overlying the
cells of the neuroepithelium and resulted
in a dramatic expansion of the neuroepi-
thelium coupled with precocious and
ectopic neuroblasts (Figures 6A and 6B),
which resembled the defects seen inmir-
8deficient flies.Due to the lethality caused
by early expression of sSpi::GFP, we used
a temperature-sensitive Gal80 construct
(tub-Gal80ts: McGuire et al., 2004) and
grew larvae at the permissive temperature
of 18�C until early L3 stage. The tempera-
ture was then increased to 31�C to inacti-
vate Gal80 and release Gal4 repression
for 24 hr in order to characterize the ef-
fects of sSpi::GFP overexpression on the
underlying neuroepithelium.
Endogenous depletion of spitz by RNA
interference (UAS-spi-IR) (Dietzl et al.,
2007) using mir-8-Gal4 (mir-8>UAS-
spi-IR) resulted in a dramatic reduction
in neuroepithelial size (Figure 6C).
These defects were less severe when
we used repo-Gal4 (Figure S6C). More-
over, expression of the spi-IR transgene
by moody-Gal4 had no effect on the
morphogenesis of the underlying neuroe-
pithelium (Figures S6G–S6J). Therefore,
the miR-8+ optic-lobe-associated cortex
glia represents an endogenous source
of Spitz that is required to control neuroe-
pithelial proliferation and neuroblast formation. In agreement
with the secreted nature of Spi protein, small clones of glial cells
null for spi (spi1) recovered using theMARCM technique (Lee and
Luo, 1999) revealed noneffect in the underlying larval neuroepi-
thelium and in the final optic lobe size, likely due to nonautono-
mous rescue by the adjacent spi+ glial cells (Figures S6K–S6M).
Spitz Is a Direct Target of miR-8Gain and loss of the EGF receptor ligand Spitz in miR-8 positive
glial cells, through loss of aos or Spitz depletion by RNAi
increased or reduced the expansion of the neuroepithelium,
Figure 6. Surface Cortex Glia-Derived Spitz
Promotes Neuroepithelial Expansion and
Neuroblast Generation
(A and B) L3 brains shown are of larvae raised at
18�C (A), or shifted to 31�C during the early to late
L3 larval stage (B) to drive expression of sSpi::GFP
by mir-8-Gal4.
(C) Inhibition of endogenous spi in miR-8 glia by
UAS-spi-IR. NE, neuroepithelial cells; Nbs, neu-
roblasts. Scale bar represents 50 mm.
See also Figure S6.
Developmental Cell
A MicroRNA-Dependent Niche for Early Neurogenesis
respectively, and disrupted the orderly emergence of neuroblast
transition. Because these effects were also seen in loss and gain
of mir-8 animals, we investigated whether Spitz may be directly
downregulated by miR-8. The 30 UTR of spitz mRNA contains
only one potential binding site for miR-8 that is highly conserved
across Drosophila species (Figure 7A). We validated (Figures
7B–7D) this binding site using tubulin-luciferase reporters that
contained either the full-length WT 30 UTR of spitz or a mutated
version that abolished microRNA binding. The luciferase activity
of these constructs was measured in Drosophila Schneider cells
(S2) that were cotransfected with a tub-mir-8 plasmid to express
the microRNA (Figure 7B). Luciferase activity from the reporter
containing the WT spitz-30UTR was reduced by 77% when mir-
8 was expressed. This effect was reversed when the conserved
miR-8 binding site wasmutated (Figure 7B). We also generated a
tub-eGFP sensor containing the WT or the mutated 30 UTR of
Spitz and overexpressed the mir-8 transgenes in vivo in cells
of the posterior wing compartment using engrailed (en)-Gal4
(en>UAS-mir-8>DsRed). We observed a specific decrease in
the expression of a tub-eGFP-sensor containing the WT 30
UTR of spi, whereas we found no change in the expression of
a mutated version of the sensor containing the same three point
mutations in the miR-8 binding motif that prevented downregu-
lation in S2 cells (Figures 7C and 7D).
Endogenous levels of spimRNAs were elevated 1.5-fold in the
larval brain of flies lacking mir-8 (Figure 7E), as determined by
quantitative RT-PCR (qRT-PCR) analysis, consistently with spi
being a biologically relevant target of the microRNA. Expression
of spitz transgene without its 30 UTR by repo-Gal4, and therefore
incapable of being downregulated by miR-8, fully reverted the
defects caused by mir-8 overexpression (Figures 7F and 7G).
Moreover, the growth defect in brain overexpressing mir-8 was
enhanced by further reducing the expression levels of its target
Spitz (repo>mir-8>spi-IR, Figures 7H and 7I). In addition, some
of these brains (n = 13 out of 53 brains) exhibited areas of cell
death, which lacked the typical features of apoptosis and auto-
phagy, and hence seemed to be caused by necrotic cell death.
The occurrence of these cell deaths suggest that an acute depri-
vation of spi (gain of mir-8 and downregulation by spi-IR) may
cause brain trauma or some forms of neurodegeneration. Further
comprehension of this process could have important implica-
tions in Spi-regulated processes.
Intrinsic Activation of EGFR Mimics mir-8 DefectsThese results suggest that there is communication between
miR-8 glia and the neuroepithelial progenitors likely through
Developm
the EGF pathway that sustains the growth of the neuroepithelium
and the location and extent of neuroblast emergence. Directly
altering EGFR activity in the neuroepithelium mimicked the
defects observed in mir-8 and glial-specific spi mutants (Fig-
ure 6). We used a Gal4 line that drives expression exclusively
in neuroepithelial cells (c855a-Gal4) (Egger et al., 2007) of a
ligand-independent constitutively active form of the EGFR
(c855a>UAS-EGFRAct) (Domınguez et al., 1998), or of an acti-
vated form of Ras (c855a>UAS-RasV12). As in mir-8 flies (Fig-
ure 3B), or upon glial-specific increase in Spitz expression
(through loss of aos or Spitz overexpression; Figures 5C, 5D,
and 6B), the neuroepithelium expressing the activated form of
EGFR (Figure 8B) or RasV12 (data not shown) was increased in
size and ectopic neuroblasts were found throughout the neuro-
epithelium (Figures S7A–S7C). These phenotypes are reminis-
cent to those of Spitz overexpression in glial cells and of the brain
mutant for aos or mir-8.
We confirmed that both the overgrowth and ectopic neuroepi-
thelial-neuroblast transitions were cell-autonomous responses
to the activation of the EGFR signaling cascade. We used
MARCM to generate and mark clones of cells that express
constitutively active RasV12. MARCM clones of RasV12 cells pro-
duced cell-autonomous outgrowths and ectopic neuroblasts
(Figure 8D), supporting the view that intrinsic activation of the
EGFR receptor pathway is directly involved in autonomously
stimulating growth and generating neuroblasts. Inactivation of
the EGFR pathway by the expression of dominant negative
forms of the receptor inhibited neuroepithelial growth and fewer
neuroblasts formed (Figure 8E), demonstrating that the endoge-
nous EGFRpathway is required for neuroepithelial expansion. As
shown in Figure S7G, inactivation of EGFR by the expression of
dominant negative led to a 52% smaller size of the optic lobe
(1,714.695 ± 156.499 mm2 in c855a>EGFRDN: 5 hr after puparium
formation, APF5, n = 8 brains scored) as compared withWT con-
trol (3,523.1 ± 305.18 mm2: APF5, n = 7 brains scored). These
phenotypes of overactivation of EGFR/Ras pathway resembled
disc overgrowth and formation of ectopic morphogenetic furrow
in eye imaginal discs (Domınguez et al., 1998).
The effects of inappropriate activation of the EGFR pathway
within the neuroepithelium differed from previous studies that
suggested a role of EGFR only in the step of neuroblast transition
(Yasugi et al., 2010). It has been proposed that Notch signaling
negatively controls neuroblast transition by antagonizing EGFR
signaling and hence the effects of EGFR and Notch gain and
loss of function are opposed (Egger et al., 2010; Ngo et al.,
2010; Reddy et al., 2010; Yasugi et al., 2010). However, we
ental Cell 27, 174–187, October 28, 2013 ª2013 Elsevier Inc. 181
BA
control mir-8 Δ2 mir-8>sSpi;tub-Gal80 ts
UTR C CACGC A U3 GACA GCC AGUAUUA CUGU UGG UCAUAAU
miR-8 3 AGAAA ACUG
5
305
spitz 3´UTR (394bp)
AAAAAA
0
0,2
0,4
0,6
0,8
1
1,2
tub-emptytub-mir-8
E
C D
en-Gal4>DsRed>mir-8
L3
tub-eGFP::spi-3’UTR wt tub-eGFP::spi-3’UTR mut
0
0,5
1
1,5
2
Rel
ativ
e lu
cife
rase
act
ivity
spi-UTR wt spi-UTR mut
F G IH
repo-Gal4>sSpi::GFP>mir-8; tub-Gal80 repo-Gal4>spi-IR>mir-8controlts
Mira DE-Cad Repo Patj DE-Cad DAPI
Rel
ativ
e sp
itz m
RN
A
Figure 7. Spitz Is a Direct Target of miR-8
(A) Computer prediction of miR-8-binding site in the spi 30 UTR (seed sequence in red).
(B) Luciferase assay in S2 cells cotransfected with tub-mir-8 or the tub-empty vector, together with a firefly luciferase vector containing the spi 30 UTR(tub-luc::spi-30UTRwt) or a luciferase vector with mutations (tub-luc::spi-30UTRmut) in the seed sequence (underlined bases in A).
(C and D) Wing discs showing overexpression of UAS-mir-8 and UAS-DsRed by en-Gal4 (red) and of GFP in a tub-eGFP::spi-30UTRwt (C) and mutated
tub-eGFP::spi-30UTRmut (D) sensors.
(E) Histogram shows the relative mRNA levels of spi determined by qRT-PCR in three independent experiments (mean + SEM) in the brains of controls, mir-8D2
mutants and flies expressing mir-8>sSpi::GFP and tub-Gal80ts.
(F and G) Rescue of the effects ofmir-8 overexpression by a Spi transgene lacking its 30 UTR. Flies reared at 18�C (F) or shifted to 31�C during the early to late L3
larval stage (G).
(H and I) NE and neuroblast transitions in controls (H) and in larvae expressing repo>mir-8>spi-IR (I). Glial cells (Repo+, green) and NE cells [(DE-Cad+, blue) or
(Patj+, yellow)]. DAPI counterstaining highlights all nuclei. Scale bar represents 20 mm (F–I).
Developmental Cell
A MicroRNA-Dependent Niche for Early Neurogenesis
observed that the phenotypes were not identical. As shown in
Figure 8F, in Notch loss-of-function mutations, the wave of L(1)
sc expression that marks the neurepithelial-neuroblast transition
zone moved faster whereas activation of the EGFR pathway
182 Developmental Cell 27, 174–187, October 28, 2013 ª2013 Elsev
(c855a>UAS-EGFRAct or c855a>UAS-RasV12) gave rise to
ectopic neuroepithelial-neuroblast transition zones across the
length of the neuroepithelium (Figure 8B) and enhanced neuroe-
pithelial growth, a phenotype not observed in Notch mutants
ier Inc.
Figure 8. Intrinsic Alteration of EGFR Signaling Mimics the Defects
Induced by Altered miR-8 Expression
(A and B) Control (A) and c855a-Gal4>EGFRAct (B) neuroepithelia stained with
Dlg (blue)andwith theneuroblast transitionmarkersL(1)sc (red)andMira (green).
Developmental Cell
A MicroRNA-Dependent Niche for Early Neurogenesis
Developm
(Figures S7A–S7D). This inappropriate activation of the EGFR
pathway ultimately exerted profound effects on optic lobe
patterning and size (Figures S7E–S7H).
Collectively, these findings indicate that the expression of the
conserved microRNA miR-8 in a population of optic-lobe-asso-
ciated cortex glia is required for the development of the underly-
ing neuroepithelium. Moreover, our results demonstrate that
expression of Spitz in the miR-8 glia is necessary and sufficient
to sustain neuroepithelial expansion as well as the generation
of neuroblasts. We demonstrate that Spitz is a direct, biologically
relevant target of miR-8, molecularly linking proliferation and the
progression of neuroblast transition. Based on these findings,
we propose that optic-lobe-associated miR-8-Spitz cortex glia
represent an anatomical niche that provides the signals and sup-
port required for neuroepithelial cell expansion and neuroblast
formation and that the production of signals from this niche
must be tightly regulated. It is unclear whether Spi exerts these
influences directly, by activating EGFR in the neuroepithelial
cells, or via secondary signals (Figure 8G).
DISCUSSION
Our analysis reveals that the production by an optic-lobe-asso-
ciated cortex glia of the EGFR ligand Spitz is critical for coordina-
tion of neuroepithelial proliferation and the spatiotemporal
emergence of neuroblasts. External signaling from the surround-
ingmicroenvironment is a commonmechanism for the regulation
of stem cell number and behavior in mature tissues, but the
need of a niche microenvironment during early neurogenesis
was unknown.
Signaling through Surface-Associated miR-8-Spitz-Positive Cortex GliaFlies deficient for the microRNAmir-8 exhibit brain degeneration
and behavioral defects. Strikingly, the microRNA is expressed in
cortex glial cells lying underneath the blood-brain-barrier (sub-
perineurial) glial layer (Figures 1 and 2). These glia are of large
size and produce long protrusions that ensheath the developing
optic lobe neuroepithelium and can be distinguished by their
(C and D) c855a-Gal4 MARCM GFP (green) control clones (C) and those
overexpressing UAS-RasV12 (D). Neuroblasts are visualized by Mira (red) and
NE by DE-Cad (blue).
(E) Intrinsic inactivation of EGFR signaling via EGFRDN.
(F) Brains in which Notch receptor is inactivated via RNAi by c855a-Gal4 and
stained with Dlg (blue), L(1)sc (red), and Mira (green).
(G) Model of miR-8-Spitz regulatory interactions and glial-derived Spitz
in coordinating neuroepithelial expansion and neuroblast transition. miR-8
increases, in an autocrine manner, cortex glial cellular size that facilitates the
formation of long cellular processes that ensheath the underlying neuro-
epithelium. miR-8 also negatively regulates, in a paracrine manner, neuro-
epithelial growth, and neuroblast transition via direct inhibition of Spitz protein
translation. Blocking endoreplication and/or loss of miR-8 in cortex glia
releases miR-8 inhibition of Spitz levels, thereby resulting in inappropriate or
enhanced EGFR signaling. Feedback signaling via EGFR may also stimulate
the expression of rho and aos in miR-8+ positive cortex glia, thereby contrib-
uting to the fine-tuning of Spitz protein activation and secretion. Glial-to-
neuroepithelial cells signaling via secondary signal (X) is not ruled out. Scale
bar represents 20 mm.
See also Figure S7.
ental Cell 27, 174–187, October 28, 2013 ª2013 Elsevier Inc. 183
Developmental Cell
A MicroRNA-Dependent Niche for Early Neurogenesis
selective expression of the EGFR ligand spitz as well as its mod-
ulators aos and rho.
Genetic manipulation ofmir-8, spi, or aos, in this glial cell pop-
ulation unveiled cell nonautonomous influences of these glia on
the development of the underlying neuroepithelium. Similar glial
signaling to neuroblasts in the larval Drosophila brain has been
demonstrated for a class of neuroblasts that remain quiescent
until nutrient-responsive satellite and cortical glia reactivate their
proliferation (Chell and Brand, 2010; Sousa-Nunes et al., 2011).
Our study identifies another population of glia cells that sustain
growth of neuroepithelial cells and the neuroepithelial-neuro-
blast transition via a miR-8-Spitz axis.
Glial-mediated regulation of the neuroepithelium is reminis-
cent to the roles of mammalian astrocytes that are known
to potently stimulate neurogenesis in cell culture (Lim and
Alvarez-Buylla, 1999; Song et al., 2002) and a component of
endogenous neural stem cell niche in adult mammalian neuro-
genesis (Nern and Momma, 2006). Additionally, EGFR is also
implicated in glial cell proliferation in Drosophila (Read et al.,
2009; Reddy and Irvine, 2013) and human glioma often exhibits
elevated EGFR signaling.
Surface-Associated miR-8-Spitz-Positive GlialArchitecture and SignalingThe notion that miR-8-Spitz-positive cortex glia constitute an
anatomically and functionally distinct population of surface-
associated glia cells is strongly supported by the finding that
RNAi knockdown of spi in subperineurial glia using moody-
Gal4 has no effect on neuroepithelial development. Spitz protein
is converted to its active form by the Rhomboid protease, which
is also expressed by miR-8 glia. The extracellular factor Aos
limits Spitz spreading and signaling level that may influence
the effects of Spitz-EGFR in the responding neuroepithelium
(e.g., sustaining proliferation and preventing premature or
ectopic neuroblast formation). The posttranscriptional silencing
of spi mRNA by the miR-8 binding to a sequence in its 30 UTRprovides another layer of regulation to fine-tune the timing, local-
ization and/or amount of Spitz protein translation. Moreover,
the distinctive architecture of miR-8-Spitz-positive cortex glia
appears to be regulated by endoreplication regulators dup/
Cdt1 and themicroRNAmiR-8. Importantly, expression of a spitz
transgene lacking its 30 UTR (and hence unable of regulating by
miR-8) fully rescued the undergrowth defect caused by mir-8
overexpression in the glia cells. Therefore, we suggest that cor-
tex glia employ a coordinated strategy that is mediated bymiR-8
to ensure that: (1) the glia establish a correct architecture to pro-
vide a continuous layer of cortex glia cells that extend long pro-
cesses to the neuroepithelium; and (2) correct local (or temporal)
control of Spitz protein synthesis. Given the expression of rho
and aos is directly induced by EGFR signaling in other context,
a feed-back signaling via EGFR may also occur in miR-8-Spitz
positive glial cells, thereby contributing to the fine-tuning of Spitz
protein activation and secretion.
Do Optic-Lobe-Associated miR-8-Spitz Cortex GliaRepresent an Anatomical Niche?A niche typically refers to a confined anatomical location where
adult stem cells reside and provides the signals required to
sustain stem cell function and number. Niches are usually
184 Developmental Cell 27, 174–187, October 28, 2013 ª2013 Elsev
composed of supporting cells that make physical contact with
the stem cells and act locally (Morrison and Spradling, 2008).
The optic-lobe-associated miR-8-Spitz-positive cortex glia
appear to represent a niche that contributes signals for the
growth andmorphogenesis of the neuroepithelium and that con-
stitutes a functionally distinct population of that of the blood-
brain-barrier glial cells.
Extrinsic signaling in the coordination of neuroepithelial prolif-
eration in the developing mammalian forebrain of Foxc1 mutant
mice has also been suggested. In these mice, the meninges are
reduced or absent, resulting in an expansion of the neuroepithe-
lium due to the predominance of symmetric divisions (Sie-
genthaler et al., 2009). The meninges are a source of the retinoic
acid required for the transition of neuroepithelial cells into radial
glia and neurons (Siegenthaler et al., 2009). Furthermore, menin-
geal cells secrete and organize the pial basement membrane, a
thin sheet of extracellular matrix that covers the brain and that is
enriched in a variety of growth factors. Rupture of the basement
membrane in the developing brain causes type II lissencephaly,
generating ectopic precursor clusters and cortical heterotopias
due to impaired attachment of the radial glia to the basement
membrane, resulting in a general laminar disorganization
(Haubst et al., 2006). These defects are reminiscent to those of
disrupted surface glia cells described here.
In summary our findings suggest that neuroepithelial prolifera-
tion and the onset of neuroblasts in the developing optic lobe
neuroepithelium are largely influenced by extrinsic cues via a
miR-8-dependent mechanism in the overlying glia. The reprog-
ramming of neuroepithelial cells into neural stem cells (neuro-
blasts) is associated with dramatic morphological and molecular
changes, including the loss of epithelial determinants DE-
Cadherin, Crumbs and PatJ (Egger et al., 2007; Ngo et al.,
2010; Wang et al., 2011a), and enhanced expression of the
Snail-family zinc-finger transcriptional repressor Worniu (Ashraf
et al., 2004). These changes are strikingly reminiscent of the
events that drive the epithelial-to-mesenchymal transition
(EMT), which confers a stem-like character inmammalian epithe-
lial cells and in cancer cells (Mani et al., 2008) andwhich are regu-
lated by the miR-200 family. Indeed, downregulation of human
mir-200 genes in epithelial normal and cancer cells promotes
EMTand the acquisition of ‘‘stemness’’ (Park et al., 2008;Wellner
et al., 2009) effects that are presumed to be cell-autonomous.
Our findings demonstrate that miR-8 can exert its effects non-
cell-autonomously, opening the possibility that microRNA of
the miR-200 family may play similar roles in stem cell fate niches
and/or microenvironmental regulation of metastasis.
EXPERIMENTAL PROCEDURES
Drosophila Husbandry
The lines used in this study are described in FlyBase: c855a-Gal4, en-Gal4,
mir-8NP5427-Gal4, moodyXA12-Gal4, repo-Gal4, spiNP0261-Gal4, aprk568-lacZ,
aosD7, aosw11-lacZ, APCFzr/Cdh1-lacZ, mir-8D2, rhoX81-lacZ, Nrx-IV-GFP
(Buszczak et al., 2007), spis3547-lacZ, tub-Gal80ts, UAS-mCD8::GFP, UAS-
DsRed, UAS-H2B::RFP, UAS-lacZ, UAS-aos, UAS-Dcr2, UAS-dup (gift from
B. Calvi), UAS-EGFRAct, UAS-EGFRDN1-7, UAS-keren-IR (VDRC KK104299),
UAS-mir-200c, UAS-mir-8, UAS-mir-8-SP::GFP, UAS-Notch-IR (VDRC
KK100002), UAS-spi-IR (VDRC GD3922), UAS-RasV12, and UAS-sSpi::GFP.
RNAi lines in Table S1 are described in the Supplemental Experimental
Procedures.
ier Inc.
Developmental Cell
A MicroRNA-Dependent Niche for Early Neurogenesis
FLP-Out Clonal Analysis with mir8-Gal4
Larvae of the genotype yw hsp70-FLP122; UAS>CD2,y+ stop>mCD8::GFP/
mir-8-Gal4 were heat shocked for 20 min at 37�C at 72 hr after egg-laying
(AEL) to fuse the UAS sequences to mCD8::GFP coding sequences by
removing the CD2, y+ stop cassette flanked by the FRT (denoted as > )
sequences, specifically in a subset of miR-8+ cells. FLP-out clones marked
by GFP were analyzed at 24 hr postinduction (96 hr AEL).
MARCM Clones
MARCM clones were generated in glial cells using repo-FLP by crossing
female virgins tub-Gal80 FRT40A; repo-Flp1C, repo-Gal4 UAS-mCD8::GFP/
TM6 (gift of C. Klambt) to males FRT40A spi1/CyO and brains were dissected
at L3 stage (6 clones in optic lobe glia in n = 69 larval brains analyzed and 13
clones in the optic lobe in n = 85 adult brains).
MARCM RasV12-expressing clones were induced by 20 min heat-shock at
37�C at 72 hr AEL and brains were dissected 24 hr postinduction (96 hr AEL)
in animals of the genotype: yw hsp70-FLP122; UAS-mCD8::GFP; FRT42D/
FRT42D tubP-Gal80; c855a-Gal4/UAS-RasV12. Control MARCM clones were
of the genotype: yw hsp70-FLP122 UAS-mCD8::GFP; FRT42D/FRT42D
tubP-Gal80; c855a-Gal4/TM2.
Antibodies and In Situ Hybridization
Brains were dissected out in cold PBS and fixed in 4%PFA for 20min (Morante
and Desplan, 2011). The following primary antibodies were used: mouse anti-
Dlg (1/100, DSHB), mouse anti-Elav (1/50, DSHB), mouse anti-Mira (1/50)
(Ohshiro et al., 2000), mouse anti-Repo (1/50, DSHB), rabbit anti-b-Gal (1/
20,000, Cappel), rabbit anti-GFP (1/1,000, Molecular Probes), rabbit anti-
Mira (1/2,000) (Ikeshima-Kataoka et al., 1997), rabbit anti-Patj (1/2,000) (Bhat
et al., 1999), rabbit anti-pH3 (1/2,000, Upstate), rat anti-DE-Cad (1/50,
DSHB), rat anti-DN-Cad (1/25, DSHB), rat anti-L(1)sc (1/800, gift from A. Car-
mena), and rat anti-Moody-b (1/200, gift from U. Gaul). Secondary antibodies
were purchased from Invitrogen and Jackson ImmunoResearch. DAPI (Invitro-
gen) counterstaining was done using a solution (0.3 mg/ml). Brains were
mounted in Vectashield (Vector Labs), maintaining their 3D configuration
and images were obtained using a Leica TCS SP2 Confocal microscope.
Locked nucleic acid (LNA)-probe in situ hybridization was performed as in
Vallejo et al. (2011) using amiR-8 probe (Exiqon) end-labeled with a DIG oligo-
nucleotide 30-end labeling kit (Roche) and hybridized at 42�C.
Glial and Neuroepithelial Area Measurements
Nuclear size of surface glia and neuroepithelial areas were measured as
described in the Supplemental Experimental Procedures.
miRNA-mRNA 30 UTR Alignment and Transgenic Constructs
The binding site for miR-8 in the spi 30 UTR was identified using the BiBiServ
server (Rehmsmeier et al., 2004). TheUAS-mir-8 construct was described pre-
viously (Vallejo et al., 2011). The tub-eGFP::spi-UTR and tub-eGFP::spi-UTR-
mut constructs were generated by cloning the full-length 30 UTR of the
Drosophila spi gene, or a mutated sequence, into the 30 end of the tub-eGFP
reporter vector (a gift from S.M. Cohen).
Luciferase Reporter Assays
For Drosophila S2 cell luciferase assays, cells were cotransfected in 24-
well plates with the tub-mir-8 plasmid (250 ng) (Karres et al., 2007), the
luciferase::spi-UTR or luciferase::spi-UTR-mut constructs (25 ng), and
the Renilla luciferase plasmid (25 ng) for normalization. Target sites in the
mutant construct were generated using the QuickChange Site-Directed
Mutagenesis kit (Stratagene). The relative luciferase activity was measured
60 hr posttransfection using the Dual-Glo Luciferase Assay system
(Promega).
qRT-PCR
Total RNAwas extracted from L3 brains using the RNAeasy-Mini Kit (QIAGEN),
and cDNA was synthesized using an oligo-dT primer and the SuperScript RT-
III Kit (Invitrogen). Quantitative real-time PCR was performed using the ABI
SYBR green system using gene-specific primers (Supplemental Experimental
Procedures) designed with Universal Probe Library Assay Design Center
(ROCHE; see https://qpcr1.probefinder.com). The dGAPDH and dUba2
Developm
primers were used to normalize the mRNA levels. The assays were performed
in an ABI7500 apparatus and analyzed using the Applied Biosystem software.
All qPCRs were performed in triplicate and the relative expression was calcu-
lated using the comparative CT method.
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental Procedures,
seven figures, and one table and can be found with this article online at
http://dx.doi.org/10.1016/j.devcel.2013.09.018.
ACKNOWLEDGMENTS
We are very grateful to M. Asano, N. Baker, H. Bellen, E. Bier, R. Cagan, B.
Calvi, A. Carmena, S. Cohen, B. Edgar, M. Freeman, U. Gaul, C. Klambt, F.
Matsuzaki, R. Mann, R. Lehmann, L. Luo, P. Simpson, B. Shilo, J. Treisman,
D. van Vactor, M. Vidal, the Bloomington Stock Center, the NIG-Fly Stock Cen-
ter, the Drosophila Genetic Resource Center at Kyoto Institute of Technology,
the VDRC, and the Developmental Studies Hybridoma Bank for reagents. We
thank the TRiP at Harvard Medical School (NIH/NIGMS R01-GM084947) for
providing transgenic RNAi fly stocks used in this study. We also thank Andrea
Brand, Boris Egger, Marion Silies, Tetsuya Tabata, and members of the Des-
plan and Dominguez laboratories for in-depth discussions and suggestions
and Giovanna Exposito for helping with Imaris software. This work was sup-
ported by a Ramon y Cajal grant (RyC-2010-07155) and a Ministerio de Econ-
omia y Competitividad grant (SAF2012-31467) to J.M., a National Institutes of
Health grant (R01 EY017916) to C.D., and Spanish National grants (BFU2009-
09074 and SAF2012-35181), a Generalitat Valenciana grant (PROMETEO II/
2013/001), and Fundacion Botin to M.D.
Received: February 4, 2013
Revised: July 22, 2013
Accepted: September 19, 2013
Published: October 17, 2013
REFERENCES
Ashraf, S.I., Ganguly, A., Roote, J., and Ip, Y.T. (2004). Worniu, a Snail family
zinc-finger protein, is required for brain development in Drosophila. Dev.
Dyn. 231, 379–386.
Awasaki, T., Lai, S.L., Ito, K., and Lee, T. (2008). Organization and postembry-
onic development of glial cells in the adult central brain of Drosophila.
J. Neurosci. 28, 13742–13753.
Bainton, R.J., Tsai, L.T., Schwabe, T., DeSalvo, M., Gaul, U., and Heberlein, U.
(2005). moody encodes two GPCRs that regulate cocaine behaviors and
blood-brain barrier permeability in Drosophila. Cell 123, 145–156.
Bhat, M.A., Izaddoost, S., Lu, Y., Cho, K.O., Choi, K.W., and Bellen, H.J.
(1999). Discs Lost, a novel multi-PDZ domain protein, establishes and main-
tains epithelial polarity. Cell 96, 833–845.
Brabletz, S., Bajdak, K., Meidhof, S., Burk, U., Niedermann, G., Firat, E.,
Wellner, U., Dimmler, A., Faller, G., Schubert, J., and Brabletz, T. (2011). The
ZEB1/miR-200 feedback loop controls Notch signalling in cancer cells.
EMBO J. 30, 770–782.
Brand, A.H., and Livesey, F.J. (2011). Neural stem cell biology in vertebrates
and invertebrates: more alike than different? Neuron 70, 719–729.
Buszczak, M., Paterno, S., Lighthouse, D., Bachman, J., Planck, J., Owen, S.,
Skora, A.D., Nystul, T.G., Ohlstein, B., Allen, A., et al. (2007). The Carnegie pro-
tein trap library: a versatile tool for Drosophila developmental studies. Genetics
175, 1505–1531.
Chell, J.M., and Brand, A.H. (2010). Nutrition-responsive glia control exit of
neural stem cells from quiescence. Cell 143, 1161–1173.
DeSalvo, M.K., Mayer, N., Mayer, F., and Bainton, R.J. (2011). Physiologic and
anatomic characterization of the brain surface glia barrier of Drosophila. Glia
59, 1322–1340.
Dietzl, G., Chen, D., Schnorrer, F., Su, K.C., Barinova, Y., Fellner, M., Gasser,
B., Kinsey, K., Oppel, S., Scheiblauer, S., et al. (2007). A genome-wide
ental Cell 27, 174–187, October 28, 2013 ª2013 Elsevier Inc. 185
Developmental Cell
A MicroRNA-Dependent Niche for Early Neurogenesis
transgenic RNAi library for conditional gene inactivation in Drosophila. Nature
448, 151–156.
Domınguez, M., Wasserman, J.D., and Freeman, M. (1998). Multiple functions
of the EGF receptor in Drosophila eye development. Curr. Biol. 8, 1039–1048.
Edwards, T.N., and Meinertzhagen, I.A. (2010). The functional organisation of
glia in the adult brain of Drosophila and other insects. Prog. Neurobiol. 90,
471–497.
Egger, B., Boone, J.Q., Stevens, N.R., Brand, A.H., and Doe, C.Q. (2007).
Regulation of spindle orientation and neural stem cell fate in the Drosophila
optic lobe. Neural Dev. 2, 1.
Egger, B., Gold, K.S., and Brand, A.H. (2010). Notch regulates the switch from
symmetric to asymmetric neural stem cell division in the Drosophila optic lobe.
Development 137, 2981–2987.
Freeman, M. (1996). Reiterative use of the EGF receptor triggers differentiation
of all cell types in the Drosophila eye. Cell 87, 651–660.
Freeman, M., Klambt, C., Goodman, C.S., and Rubin, G.M. (1992). The argos
gene encodes a diffusible factor that regulates cell fate decisions in the
Drosophila eye. Cell 69, 963–975.
Hartenstein, V. (2011). Morphological diversity and development of glia in
Drosophila. Glia 59, 1237–1252.
Haubst, N., Georges-Labouesse, E., De Arcangelis, A., Mayer, U., and Gotz,
M. (2006). Basement membrane attachment is dispensable for radial glial
cell fate and for proliferation, but affects positioning of neuronal subtypes.
Development 133, 3245–3254.
Hyun, S., Lee, J.H., Jin, H., Nam, J., Namkoong, B., Lee, G., Chung, J., and
Kim, V.N. (2009). Conserved MicroRNA miR-8/miR-200 and its target USH/
FOG2 control growth by regulating PI3K. Cell 139, 1096–1108.
Ikeshima-Kataoka, H., Skeath, J.B., Nabeshima, Y., Doe, C.Q., andMatsuzaki,
F. (1997). Miranda directs Prospero to a daughter cell during Drosophila
asymmetric divisions. Nature 390, 625–629.
Jiang, H., and Edgar, B.A. (2009). EGFR signaling regulates the proliferation of
Drosophila adult midgut progenitors. Development 136, 483–493.
Karres, J.S., Hilgers, V., Carrera, I., Treisman, J., and Cohen, S.M. (2007). The
conserved microRNA miR-8 tunes atrophin levels to prevent neurodegenera-
tion in Drosophila. Cell 131, 136–145.
Kawamori, H., Tai, M., Sato, M., Yasugi, T., and Tabata, T. (2011). Fat/Hippo
pathway regulates the progress of neural differentiation signaling in the
Drosophila optic lobe. Dev. Growth Differ. 53, 653–667.
Klein, D.E., Nappi, V.M., Reeves, G.T., Shvartsman, S.Y., and Lemmon, M.A.
(2004). Argos inhibits epidermal growth factor receptor signalling by ligand
sequestration. Nature 430, 1040–1044.
Knoblich, J.A. (2008). Mechanisms of asymmetric stem cell division. Cell 132,
583–597.
Kretzschmar, D., Hasan, G., Sharma, S., Heisenberg, M., and Benzer, S.
(1997). The swiss cheese mutant causes glial hyperwrapping and brain
degeneration in Drosophila. J. Neurosci. 17, 7425–7432.
Lee, T., and Luo, L. (1999). Mosaic analysis with a repressible cell marker for
studies of gene function in neuronal morphogenesis. Neuron 22, 451–461.
Lim, D.A., and Alvarez-Buylla, A. (1999). Interaction between astrocytes and
adult subventricular zone precursors stimulates neurogenesis. Proc. Natl.
Acad. Sci. USA 96, 7526–7531.
Loya, C.M., Lu, C.S., Van Vactor, D., and Fulga, T.A. (2009). Transgenic
microRNA inhibition with spatiotemporal specificity in intact organisms. Nat.
Methods 6, 897–903.
Mani, S.A., Guo, W., Liao, M.J., Eaton, E.N., Ayyanan, A., Zhou, A.Y., Brooks,
M., Reinhard, F., Zhang, C.C., Shipitsin, M., et al. (2008). The epithelial-mesen-
chymal transition generates cells with properties of stem cells. Cell 133,
704–715.
McGuire, S.E., Mao, Z., and Davis, R.L. (2004). Spatiotemporal gene expres-
sion targeting with the TARGET and gene-switch systems in Drosophila. Sci.
STKE 2004, pl6.
186 Developmental Cell 27, 174–187, October 28, 2013 ª2013 Elsev
Miura, G.I., Buglino, J., Alvarado, D., Lemmon, M.A., Resh, M.D., and
Treisman, J.E. (2006). Palmitoylation of the EGFR ligand Spitz by Rasp in-
creases Spitz activity by restricting its diffusion. Dev. Cell 10, 167–176.
Morante, J., and Desplan, C. (2008). The color-vision circuit in the medulla of
Drosophila. Curr. Biol. 18, 553–565.
Morante, J., and Desplan, C. (2011). Dissection and staining of Drosophila
optic lobes at different stages of development. Cold Spring Harb. Protoc.
2011, 652–656.
Morante, J., Erclik, T., and Desplan, C. (2011). Cell migration in Drosophila
optic lobe neurons is controlled by eyeless/Pax6. Development 138, 687–693.
Morrison, S.J., and Spradling, A.C. (2008). Stem cells and niches:mechanisms
that promote stem cell maintenance throughout life. Cell 132, 598–611.
Nern, C., and Momma, S. (2006). The realized niche of adult neural stem cells.
Stem Cell Rev. 2, 233–240.
Ngo, K.T., Wang, J., Junker, M., Kriz, S., Vo, G., Asem, B., Olson, J.M.,
Banerjee, U., and Hartenstein, V. (2010). Concomitant requirement for Notch
and Jak/Stat signaling during neuro-epithelial differentiation in the
Drosophila optic lobe. Dev. Biol. 346, 284–295.
Ohshiro, T., Yagami, T., Zhang, C., and Matsuzaki, F. (2000). Role of cortical
tumour-suppressor proteins in asymmetric division of Drosophila neuroblast.
Nature 408, 593–596.
Park, S.Y., and Asano, M. (2008). The origin recognition complex is dispens-
able for endoreplication in Drosophila. Proc. Natl. Acad. Sci. USA 105,
12343–12348.
Park, S.M., Gaur, A.B., Lengyel, E., and Peter, M.E. (2008). ThemiR-200 family
determines the epithelial phenotype of cancer cells by targeting the E-cadherin
repressors ZEB1 and ZEB2. Genes Dev. 22, 894–907.
Pereanu, W., Shy, D., and Hartenstein, V. (2005). Morphogenesis and prolifer-
ation of the larval brain glia in Drosophila. Dev. Biol. 283, 191–203.
Read, R.D., Cavenee, W.K., Furnari, F.B., and Thomas, J.B. (2009). A
Drosophila model for EGFR-Ras and PI3K-dependent human glioma. PLoS
Genet. 5, e1000374.
Reddy, B.V., and Irvine, K.D. (2013). Regulation of Hippo signaling by EGFR-
MAPK signaling through Ajuba family proteins. Dev. Cell 24, 459–471.
Reddy, B.V., Rauskolb, C., and Irvine, K.D. (2010). Influence of fat-hippo and
notch signaling on the proliferation and differentiation of Drosophila optic
neuroepithelia. Development 137, 2397–2408.
Rehmsmeier, M., Steffen, P., Hochsmann, M., and Giegerich, R. (2004). Fast
and effective prediction of microRNA/target duplexes. RNA 10, 1507–1517.
Richter, C., Oktaba, K., Steinmann, J., Muller, J., and Knoblich, J.A. (2011). The
tumour suppressor L(3)mbt inhibits neuroepithelial proliferation and acts on
insulator elements. Nat. Cell Biol. 13, 1029–1039.
Schwabe, T., Bainton, R.J., Fetter, R.D., Heberlein, U., and Gaul, U. (2005).
GPCR signaling is required for blood-brain barrier formation in Drosophila.
Cell 123, 133–144.
Sepp, K.J., Schulte, J., and Auld, V.J. (2001). Peripheral glia direct axon
guidance across the CNS/PNS transition zone. Dev. Biol. 238, 47–63.
Shimono, Y., Zabala, M., Cho, R.W., Lobo, N., Dalerba, P., Qian, D., Diehn, M.,
Liu, H., Panula, S.P., Chiao, E., et al. (2009). Downregulation of miRNA-200c
links breast cancer stem cells with normal stem cells. Cell 138, 592–603.
Siegenthaler, J.A., Ashique, A.M., Zarbalis, K., Patterson, K.P., Hecht, J.H.,
Kane, M.A., Folias, A.E., Choe, Y., May, S.R., Kume, T., et al. (2009).
Retinoic acid from the meninges regulates cortical neuron generation. Cell
139, 597–609.
Song, H., Stevens, C.F., and Gage, F.H. (2002). Astroglia induce neurogenesis
from adult neural stem cells. Nature 417, 39–44.
Sousa-Nunes, R., Yee, L.L., and Gould, A.P. (2011). Fat cells reactivate
quiescent neuroblasts via TOR and glial insulin relays in Drosophila. Nature
471, 508–512.
Stork, T., Engelen, D., Krudewig, A., Silies, M., Bainton, R.J., and Klambt, C.
(2008). Organization and function of the blood-brain barrier in Drosophila.
J. Neurosci. 28, 587–597.
ier Inc.
Developmental Cell
A MicroRNA-Dependent Niche for Early Neurogenesis
Stork, T., Bernardos, R., and Freeman, M.R. (2012). Analysis of glial cell devel-
opment and function in Drosophila. Cold Spring Harb. Protoc. 2012, 1–17.
Sturtevant, M.A., Roark, M., and Bier, E. (1993). The Drosophila rhomboid
gene mediates the localized formation of wing veins and interacts genetically
with components of the EGF-R signaling pathway. Genes Dev. 7, 961–973.
Treisman, J.E., and Heberlein, U. (1998). Eye development in Drosophila:
formation of the eye field and control of differentiation. Curr. Top. Dev. Biol.
39, 119–158.
Unhavaithaya, Y., andOrr-Weaver, T.L. (2012). Polyploidization of glia in neural
development links tissue growth to blood-brain barrier integrity. Genes Dev.
26, 31–36.
Vallejo, D.M., Caparros, E., and Dominguez, M. (2011). Targeting Notch signal-
ling by the conserved miR-8/200 microRNA family in development and cancer
cells. EMBO J. 30, 756–769.
Wang, W., Li, Y., Zhou, L., Yue, H., and Luo, H. (2011a). Role of JAK/STAT
signaling in neuroepithelial stem cell maintenance and proliferation in the
Drosophila optic lobe. Biochem. Biophys. Res. Commun. 410, 714–720.
Wang, W., Liu, W., Wang, Y., Zhou, L., Tang, X., and Luo, H. (2011b). Notch
signaling regulates neuroepithelial stem cell maintenance and neuroblast
formation in Drosophila optic lobe development. Dev. Biol. 350, 414–428.
Wellner, U., Schubert, J., Burk, U.C., Schmalhofer, O., Zhu, F., Sonntag, A.,
Waldvogel, B., Vannier, C., Darling, D., zur Hausen, A., et al. (2009). The
Developm
EMT-activator ZEB1 promotes tumorigenicity by repressing stemness-inhibit-
ing microRNAs. Nat. Cell Biol. 11, 1487–1495.
White, K., and Kankel, D.R. (1978). Patterns of cell division and cell movement
in the formation of the imaginal nervous system in Drosophila melanogaster.
Dev. Biol. 65, 296–321.
Whittaker, A.J., Royzman, I., and Orr-Weaver, T.L. (2000). Drosophila double
parked: a conserved, essential replication protein that colocalizes with the
origin recognition complex and links DNA replication with mitosis and the
down-regulation of S phase transcripts. Genes Dev. 14, 1765–1776.
Yasugi, T., Umetsu, D., Murakami, S., Sato, M., and Tabata, T. (2008).
Drosophila optic lobe neuroblasts triggered by a wave of proneural gene
expression that is negatively regulated by JAK/STAT. Development 135,
1471–1480.
Yasugi, T., Sugie, A., Umetsu, D., and Tabata, T. (2010). Coordinated sequen-
tial action of EGFR and Notch signaling pathways regulates proneural wave
progression in the Drosophila optic lobe. Development 137, 3193–3203.
Zielke, N., Querings, S., Rottig, C., Lehner, C., and Sprenger, F. (2008). The
anaphase-promoting complex/cyclosome (APC/C) is required for rereplication
control in endoreplication cycles. Genes Dev. 22, 1690–1703.
Zielke, N., Kim, K.J., Tran, V., Shibutani, S.T., Bravo, M.J., Nagarajan, S., van
Straaten, M., Woods, B., von Dassow, G., Rottig, C., et al. (2011). Control of
Drosophila endocycles by E2F and CRL4(CDT2). Nature 480, 123–127.
ental Cell 27, 174–187, October 28, 2013 ª2013 Elsevier Inc. 187