Article

Mosaic Analysis with Doub

le Markers RevealsDistinct Sequential Functions of Lgl1 in NeuralStem Cells

Highlights

d MADM-based genetic dissection of intrinsic gene function

and community effects

d Sparse and complete Lgl1 ablation distinctly affects RGP-

mediated neurogenesis

d Lgl1 cell-autonomously controls astrocyte generation in an

EGFR-dependent manner

d Critical role for Lgl1 in postnatal NSC lineage progression and

neurogenesis

Beattie et al., 2017, Neuron 94, 517–533May 3, 2017 ª 2017 Elsevier Inc.http://dx.doi.org/10.1016/j.neuron.2017.04.012

Authors

Robert Beattie, Maria Pia Postiglione,

Laura E. Burnett, ..., Valeri Vasioukhin,

Troy H. Ghashghaei,

Simon Hippenmeyer

Correspondencesimon.hippenmeyer@ist.ac.at

In Brief

Beattie et al. determined the relative

contribution of novel intrinsic Lgl1 gene

functions and non-cell-autonomous

community effects in neural stem cell

proliferation behavior. They found

distinct but sequential Lgl1 functions

controlling embryonic neurogenesis and

postnatal astrocyte and olfactory bulb

interneuron generation.

Neuron

Article

Mosaic Analysis with Double MarkersReveals Distinct Sequential Functionsof Lgl1 in Neural Stem CellsRobert Beattie,1 Maria Pia Postiglione,1,4 Laura E. Burnett,1 Susanne Laukoter,1 Carmen Streicher,1 Florian M. Pauler,1

Guanxi Xiao,2 Olga Klezovitch,3 Valeri Vasioukhin,3 Troy H. Ghashghaei,2 and Simon Hippenmeyer1,5,*1Institute of Science and Technology Austria, Am Campus 1, 3400 Klosterneuburg, Austria2Department of Molecular Biomedical Sciences, Program in Genetics, W.M. Keck Center for Behavioral Biology, College of VeterinaryMedicine, North Carolina State University, Raleigh, NC 27607, USA3Division of Human Biology, Fred Hutchinson Cancer Research Center, Seattle, WA 98109, USA4Present address: Discovery Sciences, AstraZeneca, 43183 Molndal, Sweden5Lead Contact*Correspondence: simon.hippenmeyer@ist.ac.at

http://dx.doi.org/10.1016/j.neuron.2017.04.012

SUMMARY

The concerted production of neurons and glia byneural stem cells (NSCs) is essential for neural circuitassembly. In the developing cerebral cortex, radialglia progenitors (RGPs) generate nearly all neocor-tical neurons and certain glia lineages. RGP prolifer-ation behavior shows a high degree of non-stochas-ticity, thus a deterministic characteristic of neuronand glia production. However, the cellular and mo-lecular mechanisms controlling RGP behavior andproliferation dynamics in neurogenesis and glia gen-eration remain unknown. By using mosaic analysiswith double markers (MADM)-based genetic para-digms enabling the sparse and global knockoutwith unprecedented single-cell resolution, we identi-fied Lgl1 as a critical regulatory component. Weuncover Lgl1-dependent tissue-wide community ef-fects required for embryonic cortical neurogenesisand novel cell-autonomous Lgl1 functions control-ling RGP-mediated glia genesis and postnatal NSCbehavior. These results suggest that NSC-mediatedneuron and glia production is tightly regulatedthrough the concerted interplay of sequential Lgl1-dependent global and cell intrinsic mechanisms.

INTRODUCTION

The human cerebral cortex is the seat of our cognitive abilities

and is composed of an extraordinary number of neurons and

glia cells. The developmental programs regulating the accurate

generation of postmitotic cells, by neural stem cells (NSCs),

need to be precisely implemented and regulated. At the end of

neurulation, the early neuroepithelium is composed of neuroepi-

thelial stem cells (NESCs) from which all subsequent neural

progenitor cells and their lineages derive. NESCs initially divide

symmetrically but then transform into radial glia progenitor

(RGP) cells. RGPs have been demonstrated to be the major neu-

ral progenitors in the developing cortex responsible for produc-

ing the vast majority of cortical excitatory neurons (Lui et al.,

2011; Taverna et al., 2014).

The RGP division patterns and dynamics determine the num-

ber of neurons in the mature cortex. The mitotic RGP cell divi-

sion at the surface of the embryonic ventricular zone (VZ) can

be either symmetric or asymmetric, which is defined by the

fate of the two daughter cells (Homem et al., 2015; Lui et al.,

2011; Taverna et al., 2014). Symmetric divisions can generate

two RGPs to amplify the progenitor pool (symmetric prolifer-

ative division). In contrast, asymmetric neurogenic divisions

produce a renewing RGP and a neuron, or an intermediate

progenitor (IP) that can further divide in the subventricular

zone (SVZ) to produce neurons (Noctor et al., 2004). RGPs

may also generate SNPs (Stancik et al., 2010) and outer SVZ

radial glia progenitors (oRGs, a.k.a. basal RGs or bRGs) (Fietz

et al., 2010; Hansen et al., 2010; Wang et al., 2011). RGPs

can also produce glia cells, including astrocytes and oligoden-

drocytes, which have critical roles in the development, main-

tenance, and function of neuronal circuits (Chung et al.,

2015; Freeman and Rowitch, 2013). Although gliogenesis is

generally known to follow neurogenesis in the developing brain

(Costa et al., 2009; Magavi et al., 2012; Schmechel and Rakic,

1979; Voigt, 1989), the principles regulating glia generation,

especially at the individual RGP and successive glia progenitor

level(s), are not clear (Bayraktar et al., 2014; Molofsky and

Deneen, 2015). Shortly after birth, the embryonic neuroepithe-

lium transforms into the postnatal NSC niche in the ventricu-

lar-subventricular zone (V-SVZ) within the lateral ventricle (LV)

(Lim and Alvarez-Buylla, 2016). While certain subpopulations

of RGPs give rise to ependymal (E1) cells (Spassky et al.,

2005) postnatally, other RGPs transform into V-SVZ type B1

cells (Merkle et al., 2004). Type B1 cells function as the main

progenitors in adult neurogenesis (Doetsch et al., 1999) and

generate type C cells that in turn differentiate into type A neuro-

blasts migrating to the olfactory bulb (OB) (Lim and Alvarez-

Buylla, 2016).

Neuron 94, 517–533, May 3, 2017 ª 2017 Elsevier Inc. 517

GFP / / tdT DAPI

control-MADM wild-type

I

II-IV

V

VI

ENCXD

A

0wtwt

40

20

ns

++

+%

SAT

B2

,MA

DM

/ to

tal M

AD

M

0wtwt

6

3

ns

++

+%

CTI

P2,M

AD

M /

tota

l MA

DM

0wtwt

20

10

ns

++

+%

TB

R1

,MA

DM

/ to

tal M

AD

M

genetic mosaicLgl1-MADM

NCX I

II-IV

V

VI

F G

B

0wt

-/-Lgl1

40

20

ns

++

+%

SAT

B2

,MA

DM

/ to

tal M

AD

M

0

6

3

ns

++

+%

CTI

P2,M

AD

M /

tota

l MA

DM

wt-/-Lgl1

0

20

10

ns

++

+%

TB

R1

,MA

DM

/ to

tal M

AD

M

wt-/-Lgl1

conditional (full) knockoutcKO-Lgl1-MADM

‚Double Cortex‘

NCX

SBH

I

II-IV

V

VI

H I

C

0-/-Lgl1-/-Lgl1

40

20

ns

++

+%

SAT

B2

,MA

DM

/ to

tal M

AD

M

0

6

3

ns

++

+%

CTI

P2,M

AD

M /

tota

l MA

DM

-/-Lgl1-/-Lgl1

0

20

10

ns

++

+%

TB

R1

,MA

DM

/ to

tal M

AD

M

-/-Lgl1-/-Lgl1

SATB

2C

TIP2

TBR

1O

verv

iew

Exp.

Par

adig

m

J J’ J’’ K K’ K’’ L L’ L’’

M M’ N N’ N’’ O O’ O’’M’’

P P’ Q Q’ Q’’ R R’ R’’P’’

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

I

II-IV

V

VI

+/+

+/+ Lgl1+/+

+/+Lgl1

+/+Lgl1

+/+Lgl1+/+

Lgl1+/-

-/-Lgl1

+/+Lgl1

+/-Lgl1

-/-

+/++/-

-/-Lgl1

-/-Lgl1

-/-Lgl1

-/-Lgl1

-/--/- -/-

-/-

-/- -/--/- -/- -/--/- -/- -/--/-

-/--/-

-/--/--/--/--/-

Figure 1. Sparse and Whole Neuroepithelium Ablation of Lgl1 in RGPs Differentially Affects Cortical Neurogenesis

(A–C) Schematic illustration of experimental paradigm in control-MADM (A, wild-type), Lgl1-MADM (B, genetic mosaic), and cKO-Lgl1-MADM (C, conditional/full

knockout). In control-MADM, GFP+ (green), tdT+ (red), and GFP+/tdT+ (yellow) and unlabeled (vast majority) cells are all WT. In Lgl1-MADM, GFP+ (green) cells are

Lgl1�/�, tdT+ (red) cells are Lgl1+/+, and GFP+/tdT+ and unlabeled cells are Lgl1+/�. In cKO-Lgl1-MADM, GFP+ (green), tdT+ (red), and GFP+/tdT+ (yellow) and the

vast majority of unlabeled cortical projection neurons are all Lgl1�/�.

(legend continued on next page)

518 Neuron 94, 517–533, May 3, 2017

While previous studies provided a rough framework of cortical

neurogenesis and glia production, the cellular and molecular

mechanisms dictating the quantitative neuron and glia output

at the individual NSC level remain to be resolved. Recent mosaic

analysis with double markers (MADM)-based lineage tracing,

however, indicates that the proliferation behavior of RGPs is

remarkably coherent and predictable (Gao et al., 2014). RGPs

initially undergo symmetric amplification division with a defined

proliferation potential before transiting to asymmetric neuro-

genic division. Because MADM affords single-cell resolution,

and thus a quantitative assessment of the neurogenic potential

(Hippenmeyer, 2013; Postiglione and Hippenmeyer, 2014;

Zong et al., 2005), RGPs in their neurogenic phase were shown

to follow a defined nonrandom program of cell-cycle exit, result-

ing in a unitary output of approximately eight to nine neurons per

individual RGP. Upon completion of neurogenesis approxi-

mately one in six neurogenic RGPs proceed to produce glia

(Gao et al., 2014). Although the above MADM-based clonal anal-

ysis provided a quantitative model of NSC proliferative behavior,

the cellular and molecular mechanisms controlling the size of

pre-programmed RGP output through neurogenesis and glio-

genesis are currently unknown.

Key regulators orchestrating the RGP mode of cell division

include the signaling protein LGL1 (a.k.a. Llgl1, lethal giant larvae

homolog 1 [Drosophila]), which regulates intracellular polarity in a

variety of cellular contexts (Betschinger et al., 2003; Klezovitch

et al., 2004; Yamanaka et al., 2003). Although Lgl1 is predicted

to regulate embryonic RGPproliferation behavior, the cell-auton-

omous function of Lgl1 in vivo is not clear. The possible require-

ment of Lgl1 at postnatal stages during NSC lineage progression

is unknown due to lethality of Lgl1 knockout mice at birth. The

analysis of RGP-mediated neurogenesis in Lgl1 mutant mice is

further compromised due to the progressive disruption of the

VZ resulting in disorganization and tumor-like growth of RGPs

in the form of rosettes (Klezovitch et al., 2004). This raises the

possibility that substantial aspects of the phenotype in whole

tissue Lgl1 knockout could be the result of a combination of

both cell-autonomous and non-cell-autonomous and/or com-

munity effects. In this study, we address the above questions

and determine the relative contribution of cell-autonomous

Lgl1 signaling and non-cell-autonomous mechanisms in RGP

proliferation behavior in neurogenesis and glia production. By

capitalizing on the MADM system, we established genetic para-

digms to either ablate Lgl1 in very sparse mosaic and single RGP

clones or neuroepithelium-wide in all RGPs, both coupled with

single-cell labeling enabling high resolution quantitative pheno-

typic analysis. Our MADM-based functional analysis led to the

identification of previously unknown cell-autonomous Lgl1 func-

tions in RGPs and suggests that a concerted interplay of cell-

(D–I) Overview of MADM-labeling pattern in somatosensory cortex in control-MA

magnification of boxed areas in (D), (F), and (H) illustrating the CPwith emerging la

‘‘Double Cortex’’). Scale bars, 500 (D, F, and H) and 60 mm (E, G, and I).

(J–R) Analysis of cortical marker expression and laminar distribution. Expression

MADM+ cells, J00–L00), CTIP2 (white, M–O and M0–O0; % CTIP2+, MADM+/total MA

MADM+ cells, P00–R00) in MADM-labeled cells in control-MADM, Lgl1-MADM, and

Nuclei were stained using DAPI (blue). Cortical layers are indicated (roman nume

See also Figures S1 and S2 and Table S1.

intrinsic mechanisms coupled with stem cell niche interactions

are essential for faithful NSC proliferation behavior.

RESULTS

MADM-Based Experimental Paradigm for Sparse andComplete Lgl1 AblationIn order to determine the cell-autonomous function of Lgl1 and to

assess the relative contribution of non-cell-autonomous mecha-

nisms in RGP lineage progression, we developed a quantitative

MADM-based genetic strategy (Figures1A–1C and S1; Table

S1). The main assay consists of subtractive RGP phenotypic

analysis at single cell resolution of (1) genetic Lgl1 mosaic with

only sparse Lgl1 deletion in just a few RGPs, within heterozygous

and/or wild-type (WT) genetic background; and (2) conditional

Lgl1 knockout with global Lgl1 deletion in all RGPswithin Lgl1�/�

mutant background. In other words, individual Lgl1�/�RGP cells

are either surrounded by a local microenvironment with ‘‘normal’’

heterozygous and WT cells (abbreviated Lgl1-MADM), or by

Lgl1�/� mutant cells (abbreviated cKO-Lgl1-MADM). In both

Lgl1-MADM and cKO-Lgl1-MADM, Cre recombinase expressed

from the Emx1 locus (Gorski et al., 2002) is used to induce sparse

MADM labeling and/or geneticmanipulation specifically in dorsal

telencephalic RGPs (Figures 1 and S1). In order to generate Lgl1-

MADM animals, we genetically linked the Lgl1-flox allele (Klezo-

vitch et al., 2004) via meiotic recombination to the MADM-TG

cassette on chromosome (chr) 11 (MADM-11) (Hippenmeyer

et al., 2010); in parallel, the MADM-GT cassette is linked to the

corresponding WT allele (Figure S1A). By using a conditional

Lgl1-flox allele, Emx1-Cre mediates interchromosomal trans-

recombination between MADM cassettes, as well as cis-recom-

bination between the LoxP sites of the Lgl1-flox allele (rendering

the Lgl1-flox allele into Lgl1-D mutant allele). Thus, MADM-

labeled green cells are homozygous Lgl1�/� while red cells are

homozygous Lgl1+/+ (Figures 1B and S1A). Yellow cells are

also generated by certain MADM events (for details, see Fig-

ure S1A) (Hippenmeyer et al., 2010, 2013) and are Lgl1+/�. Forthe generation of cKO-Lgl1-MADM, we genetically linked the

Lgl1-flox allele to both TG and GT cassettes on chr11, respec-

tively (Figure S1B). In these experimental cKO-Lgl1-MADM

mice, Emx1-Cre-mediated cis-recombination renders both

Lgl1-flox alleles into Lgl1-D alleles globally in all Emx1 expressing

dorsal telencephalic RGPs (i.e., conditional Lgl1 knockout). In

addition Emx1-Cre-mediated trans-recombination leads to

sparse MADM labeling with green, red, yellow, and unlabeled

cells in the background, all carrying the Lgl1�/� homozygous

mutation (Figures 1C and S1B). Because individual cells can

be traced at high resolution, these Lgl1-MADM and cKO-Lgl1-

MADM paradigms offer an unprecedented quantitative platform

DM (D), Lgl1-MADM (F), and cKO-Lgl1-MADM (H) at P0. (E, G, and I) Higher

yers. Cyan broken line (H) outlines the subcortical band heterotopia (SBH, a.k.a.

and quantification of SATB2 (white, J–L and J0–L0; % SATB2+, MADM+/total

DM+ cells, M00–O00), and TBR1 (white, P–R and P0–R0; % TBR1+, MADM+/total

cKO-Lgl1-MADM at P0. Scale bars, 60 mm (J–R0).rals). NCX, neocortex; ns, nonsignificant. Values represent mean ± SEM.

Neuron 94, 517–533, May 3, 2017 519

to assay for the influence of the local microenvironment on

Lgl1�/� cells. In Lgl1-MADM, the surrounding cells are WT or

Lgl1+/� and thus could potentially exert positive influence on

sparse Lgl1�/� cells (green cells in Figures 1 and S1A).

In contrast, in cKO-Lgl1-MADM, all cells are Lgl1�/� and there-

fore sparsely labeled green Lgl1�/� mutant cells are encom-

passed by a community of mutant cells that could exert

influence by collective effects with the potential to possibly rem-

edy or worsen the individual cell-autonomous loss-of-function

phenotype.

Analysis of Corticogenesis upon Sparse and CompleteLgl1 Deletion in RGPsWe first analyzed cortical neuron production and overall cortical

morphogenesis in control-MADM, Lgl1-MADM, and cKO-Lgl1-

MADM at postnatal day (P) 0 (Figures 1D–1I). In our genetic

paradigms, we employ constitutive Emx1-Cre, and thus

MADM events occur stochastically at any given time in a

random subset of dividing Emx1+ RGPs. Still, if progenitor

cell division were symmetric, the number of red and green cells

within an individual clone would be close to identical (Gao

et al., 2014; Hippenmeyer et al., 2010). Even if a large number

of RGP divisions were asymmetric, such that red and green

progeny numbers were different in individual clones, the by-

chance distribution of colors in asymmetric clones would

ensure that the number of red and green progeny were equal

overall. The green/red (g/r) ratio of all MADM-labeled cells in

the somatosensory area in control-MADM, Lgl1-MADM, and

cKO-Lgl1-MADM was �1 when analyzed at P0 (Figures 1D–

1I; Table S2). Previous studies assessing ventral telencephalic

RGPs in Lgl1 full knockout suggest that Lgl1�/� RGPs show

a lack of differentiation (Klezovitch et al., 2004). We thus

assessed the cell fate specification of Lgl1�/� cells in control-

MADM, Lgl1-MADM, and cKO-Lgl1-MADM. We evaluated the

expression of TBR1 (layer VI), CTIP2 (layer V), and SATB2

(layer IV-II) at P0 (Figures 1J–1R). The coarse laminar architec-

ture appeared similar in all three experimental paradigms, and

there was no significant difference in the relative numbers

of TBR1+, CTIP2+, and SATB2+ neurons in the neocortex in

control-MADM, Lgl1-MADM, and cKO-Lgl1-MADM mice (Fig-

ures 1J–1R).

Lgl1 Is Not Cell-Autonomously Required inRGP-Mediated NeurogenesisA g/r ratio of 1 at P0 in control-MADM and cKO-Lgl1-MADM

was expected due to the identical genotypes of red and green

cells, WT in control-MADM and Lgl1�/� in cKO-Lgl1-MADM,

respectively. In contrast a g/r ratio of 1 in Lgl1-MADM was

against our expectations because LGL1 has been suggested

to control the switch from symmetric to asymmetric RGP divi-

sion mode, and loss of Lgl1 in global Lgl1 knockout leads to

exuberant RGP proliferation (Klezovitch et al., 2004). While the

full Lgl1 knockout is lethal at birth (Klezovitch et al., 2004),

Lgl1-MADM mice survive beyond 1 year of age. We therefore

determined the g/r ratio of mutant to WT neurons at postnatal

stages P21, 3 months, and 12 months. We found that the g/r

ratio was consistently �1 at all postnatal stages analyzed (Table

S2). To more precisely assess the consequences of Lgl1 loss-

520 Neuron 94, 517–533, May 3, 2017

of-function at the individual progenitor level, we conducted

MADM-based clonal analysis (Gao et al., 2014; Hippenmeyer

et al., 2010; Zong et al., 2005). To this end, we utilized tamoxifen

(TM)-inducible Emx1-CreER (Kessaris et al., 2006) to induce

RGP clones at E11 in control-MADM and Lgl1-MADM. We

analyzed the size and composition of these MADM clones at

E13 and E16 (Figures S2A–S2H). We observed no significant

difference in the total clone size and zonal distribution of clonally

related WT and Lgl1�/� cells indicating that RGP-mediated neu-

rogenesis is not dependent on cell-autonomous Lgl1 function.

Next, we analyzed the unitary neuronal output of RGPs (Gao

et al., 2014) in the absence of Lgl1. TM was injected in Lgl1-

MADM at E12 when RGPs transit from symmetric proliferative

division to asymmetric neurogenic division. As expected, we

observed asymmetric clones with a majority population in

one color and a minority population in the other color (Figures

S2I–S2K), whereby the two colors correspond to different geno-

types (i.e., WT and Lgl1�/�). We assessed the size of the major-

ity and minority populations within individual clones but

observed no significant difference, and the overall neuronal

unit size was unchanged regardless of the color and thus Lgl1

genotype (Figures S2L, S2M, and S2P). We next assessed the

relative number of upper versus lower layer neurons, and similar

to the total unit size, there was no significant difference (Figures

S2N and S2O). We conclude from these clonal analyses that

Lgl1 is not cell-autonomously required for RGP-mediated

neurogenesis.

Increased RGPProliferation Results in Subcortical BandHeterotopia in cKO-Lgl1-MADMAlthough the overall g/r ratio in cKO-Lgl1-MADMwas�1, a large

heterotopic cell mass was present beneath the neocortex (Fig-

ures 1H and S3). The ectopic cell mass resembled subcortical

band heterotopia (SBH) or ‘‘double cortex syndrome’’ in human

(Gleeson et al., 1998). While green Lgl1�/� RGPs have an iden-

tical genotype in both Lgl1-MADM and cKO-Lgl1-MADM, their

surrounding cells have distinct genotypes and thus properties

that could differentially influence the proliferation dynamics of

individual Lgl1�/� RGPs. To test whether distinct RGP prolifera-

tion could be the origin of SBH in cKO-Lgl1-MADM, we injected

bromodeoxyuridine (BrdU) at E13 and analyzed its incorporation

after 1 hr. We noticed a significant increase in the number of

BrdU+ cells in cKO-Lgl1-MADM when compared to control-

MADM (Figure S3). Rosette-like structures that stained positively

for PAX6 and included BrdU+ proliferating RGPs were also

frequently observed throughout the VZ and cortical plate (CP)

in cKO-Lgl1-MADM but not control-MADM (Figure S3). Although

we detected a significant increase in apoptosis, the number of

proliferating cells remained significantly higher at P0 in cKO-

Lgl1-MADM when compared to control-MADM (Figure S3). We

did not detect any signs of tumors, and the size of the SBH

was stable throughout adult stages from P21 onward and up

to 12 months of age (Figure S3). Because the sparse ablation

of Lgl1 in Lgl1-MADM did not lead to the formation of an SBH,

we conclude that in cKO-Lgl1-MADM, community effects

emerging in an environment where all RGPs lack Lgl1 expression

influence their mutual RGP proliferation behavior and/or dy-

namics resulting in SBH.

Loss of RGPCell Polarity Correlateswith theEmergenceof SBH in cKO-Lgl1-MADMIn order to evaluate whether the emergence of SBH in cKO-Lgl1-

MADM results from a disruption of the junctional adhesion belt,

we analyzed cKO-Lgl1-MADM at early embryonic stages (Fig-

ure 2). Indeed, already at E12, small ectopic cell formations

were apparent in cKO-Lgl1-MADM, but not in control-MADM

or Lgl1-MADM, respectively (Figures 2A–2C). We next assessed

the expression of a set of proteins that localize at basolateral

(CDH2 and b-catenin) and apical (CD133, Pals1, and g-tubulin)

sites in RGPs by immunohistochemistry at E13 when the ectopic

cell masses reached sizes of hundreds of cells. While in control-

MADM and Lgl1-MADM, the ventricular adhesion belt was uni-

formly stained for all basolateral and apical proteins indicated

above, all the components were either not expressed or mislo-

calized at sites of ventricular heterotopia (Figure 2). These results

indicate that excessive proliferation of Lgl1�/� RGPs (Figure S3)

in cKO-Lgl1-MADM is accompanied by the disruption of the

ventricular zone and mislocalization or loss of expression of

basolateral and apical components of the junctional adhesion

belt in RGPs.

Sparse and Complete Lgl1 Deletion DifferentiallyAffects Postnatal NeurogenesisShortly after birth, the embryonic neuroepithelium transforms

into the postnatal NSC niche in the V-SVZ within the LV (Lim

and Alvarez-Buylla, 2016). Because RGPs are lineally related

to progenitors in the adult stem cell niche, we assessed the

constitution and neurogenic properties of the V-SVZ in Lgl1-

MADM, cKO-Lgl1-MADM, and control-MADM. Because we

used Emx1-Cre, MADM-based labeling and Lgl1 ablation only

occurred in the dorsal wall (DW) of the V-SVZ but not in the

lateral and medial walls (LW andMW) of the LV. Given the strong

phenotype in the early embryonic neuroepithelium and VZ in

cKO-Lgl1-MADM, we first analyzed the ependymal cell layer

integrity in the DW and LW in cKO-Lgl1-MADM at P21. To this

end, we stained for CD133, which is strongly expressed in the

highly ciliated E1 ependymal cells. While the ependymal layers

in the DW in control-MADM and Lgl1-MADM appeared indistin-

guishable, the DW in cKO-Lgl1-MADM was completely devoid

of CD133 expression (Figures 3A–3C). The LW was not affected

as expected due to absence of Cre recombinase expression

(Figures 3D–3F). Previous studies have demonstrated that an

intact ependymal cell layer is critical for postnatal neurogenesis

and thus olfactory neuron production (Jacquet et al., 2011;

Paez-Gonzalez et al., 2011). We therefore quantified MADM-

labeled cells in the OB in cKO-Lgl1-MADM, Lgl1-MADM, and

control-MADM. Typically, we detected�80–100 MADM-labeled

red and green cells per mm2 on a representative mid-rostrocau-

dal OB section in control-MADM, and the g/r ratio was �1 (Fig-

ures 3M, 3P, and 3S) (Hippenmeyer et al., 2010). In contrast,

we detected less than five MADM-labeled cells in cKO-Lgl1-

MADM OB per mm2 indicating a dramatic defect in postnatal

neurogenesis (Figures 3O, 3R, and 3U). We did not detect any

MADM-labeled cells along the entire rostral migratory stream

(RMS) reflecting a true lack of DW postnatal neurogenesis rather

than a migration deficit in cKO-Lgl1-MADM (Figures 3I and

3L). These results indicate that global embryonic loss of Lgl1

function in all RGPs in cKO-Lgl1-MADM not only disrupts the

embryonic neuroepithelium, resulting in SBH, but subsequently

impedes the establishment of the DW ependymal cell layer and

thus postnatal neurogenesis. The Lgl1-MADM DW ependymal

cell layer integrity appeared indistinguishable when compared

to control-MADM (Figures 3A and 3B). However, we observed

a significant reduction of Lgl1�/� migrating neuroblasts along

the RMS (Figures 3H and 3K) and a decrease of OB neurons

in Lgl1-MADM. The g/r ratio of Lgl1�/� to WT olfactory granule

and periglomerular cells was �0.5 in Lgl1-MADM (Figures 3N,

3Q, and 3T). These data suggest an important cell-autonomous

function for Lgl1 in the V-SVZ stem cell niche and postnatal

neurogenesis.

Lgl1 Cell-Autonomously Controls Postnatal NSCLineage Progression in V-SVZTo identify the crucial Lgl1-dependent step in the establishment

of the adult stem cell niche and/or the role of Lgl1 in postnatal

neurogenesis, we traced the development of the dorsal V-SVZ

in Lgl1-MADM in a time course (Figures 4A–4J). While we

observed no gross abnormalities at P0, a large fraction of

Lgl1�/� mutant cells appeared in multicellular clusters in the

V-SVZ from P7 onward. We determined the frequency at which

multicellular clusters of distinct sizes (number of cells/cluster)

occurred in the V-SVZ. In control-MADM, the vast majority

(�90%) of both green and red MADM-labeled WT cells ap-

peared as single cells (Figure 4K), two cell clusters were

detected at a low rate, and clusters containing more than two

cells were only very rarely detected. In contrast, in Lgl1-

MADM the majority (>50%) of green Lgl1�/� cells were found

in clusters with more than two cells and larger clusters with 15

or more cells were detected frequently (Figure 4L). Next, we

quantified the g/r ratio of Lgl1�/� to WT cells and detected a

slight but significant �2-fold increase of Lgl1�/� at P0 indicating

that Lgl1 is cell-autonomously required in the emerging V-SVZ

immediately after birth (Figure 4M). The g/r ratio increased

to �8 at P21 and remained stable up until at least 12 months.

These results raise the question whether cluster formation

would include and/or affect type B1 cells in their lineage

progression. To this end, we performed high resolution morpho-

logical analysis of the postnatal V-SVZ stem cell niche in con-

trol-MADM and Lgl1-MADM. While at P0, RGPs (transforming

into type B1 cells) with their cell bodies located in the VZ could

be readily detected by virtue of their basal process (Figures 4N

and 4O), no morphological abnormalities could be observed. In

contrast, at P7 when many large clusters of green Lgl1�/� cells

were apparent (Figure 4D), high resolution morphological anal-

ysis revealed unusually long cellular extensions in Lgl1�/� cells

(Figures 4P and 4Q). Morphologically abnormal and clustered

Lgl1�/� cells expressed high levels of glial fibrillary acidic protein

(GFAP) indicating that these cells represent type B1 cells (Fig-

ure S4). At later postnatal stages, Lgl1�/� cells display even

more dramatic morphological abnormalities including syncy-

tium-like cell assemblies with long cytoplasmic extensions be-

tween individual cells (Figures 4R and 4S). These data indicate

that Lgl1 is required for the maintenance of type B1 cellular

morphology and their integrity within the postnatal V-SVZ

stem cell niche.

Neuron 94, 517–533, May 3, 2017 521

A CB

E FD

G H I

J K L

M N O

P Q R

Figure 2. Subcortical Band Heterotopia in

cKO-Lgl1-MADM IsAssociatedwithDisrup-

tion of Ventricular Apical-Basal Polarity

(A–C) Analysis of MADM-labeling pattern and

ventricular neuroepithelium integrity in control-

MADM (A), Lgl1-MADM (B), and cKO-Lgl1-MADM

(C) at E12. White arrows in (C) mark ectopic apical

cells accumulating into the ventricle. Scale

bars, 20 mm.

(D–R) Analysis of CDH2 (white, D–F), b-catenin

(white, G–I), Pals1 (white, J–L), g-tubulin (white,

M–O), and CD133 (white, P–R) expression pattern

in control-MADM (D, G, J, M, and P), Lgl1-MADM

(E, H, K, N, and Q), and cKO-Lgl1-MADM (F, I,

L, O, and R) at E13. White arrows mark the mis-

expression of apical and basolateral components

in RGPs in ectopic cell mass accumulating into the

ventricle in cKO-Lgl1-MADM. Scale bars, 45 (D–F)

and 30 mm (G–R).

See also Figure S3.

522 Neuron 94, 517–533, May 3, 2017

Lgl1 Cell-Autonomously Controls Cortical AstrocyteGenerationWhile Lgl1 appears not to be required cell-autonomously

for RGP-mediated cortical neurogenesis, we observed an

exuberant large number of Lgl1�/� cells with apparent astrocyte

morphology in Lgl1-MADM at P21 (Figures 5A–5D). To confirm

the cell fate of these MADM-labeled cells, we performed immu-

nohistochemistry for the mature astrocyte markers (Figure S5).

Next, we assessed the size and morphology of MADM-labeled

astrocytes in Lgl1-MADM and control-MADM. We quantified

the size of the cell body, total cell volume, and branching pattern

using Sholl analysis (Sholl, 1953) but could not detect any signif-

icant difference when Lgl1�/� astrocytes were compared to WT

(Figure S5). We quantified the g/r ratio of green Lgl1�/� to redWT

cortical astrocytes and noticed an increase of Lgl1�/� astrocytes

by a factor of �10 at P21 and throughout adulthood (Figure 5G).

In order to assess the cell-autonomous Lgl1 function in astrocyte

production at the single RGP level, we carried out MADM-based

clonal analysis using Emx1-CreER (Kessaris et al., 2006). MADM

clones were induced at E11 and analyzed at P21 (Figures 5E and

5F). Because astrocytes are not always present in both red and

green subclones (Gao et al., 2014), we quantified the number of

clonally related astrocytes in distinctly colored (and thus pre-

senting with different genotype) subclones independently (Fig-

ure 5H). We found that the total number of astrocytes per red

subclone, WT in control-MADM, and WT in Lgl1-MADM was

not significantly different. In contrast, the number of astrocytes

per green Lgl1�/� subclone in Lgl1-MADM was significantly

increased when compared to the green WT subclone in con-

trol-MADM (Figure 5H). The clonal analysis corroborates our re-

sults obtained from the population analysis in Lgl1-MADM.

Excessive astrocyte production was also observed in Lgl1-

MADM where we recombined the Lgl1-D allele (generated with

Hprt-Cre germline deleter [Tang et al., 2002]) instead of the

Lgl1-flox allele via meiotic recombination to the MADM-TG

cassette on chr11 (Figure S6). No difference in the astrocyte

phenotype was observed when Lgl1�/� cells exhibit either

maternal or paternal uniparental chromosome disomy, respec-

tively (Figure S6), and identical defects in astrocyte production

were seen when Lgl1�/� cells were labeled with tdT (red) instead

of GFP (green) (Figure S6). Increased astrocyte production was

also observed in the hippocampus in Lgl1-MADM (Figures 5B,

S6, and S7) and in the striatum in Lgl1-MADM induced by

Nestin-Cre (data not shown). Altogether, these data indicate an

important general cell-autonomous function of Lgl1 in astrocyte

production in the neocortex and diverse forebrain structures.

Increased Numbers of Early Postnatal AstrocyteProgenitors Precede Excess Astrocyte ProductionIn order to identify the critical Lgl1-dependent step in RGP line-

age progression and/or the transition to astrocyte production,

we traced the developmental origin of increased cortical astro-

cyte production. Once neurogenesis is completed, RGPs adopt

a gliogenic potential and can either proliferate locally to produce

intermediate astrocyte progenitor cells (aIPCs) or directly trans-

form into aIPCs and/or astrocytes (Kriegstein and Alvarez-

Buylla, 2009). We thus reasoned that astrocytes in Lgl1-MADM

could be emerging precociously or as a result of increased pro-

liferation capacity in aIPCs. To address these questions, we first

evaluated late embryonic Lgl1-MADM and early postnatal

stages for the emergence of astrocytes that could be unambig-

uously defined on the basis of their characteristic morphology.

We did not detect any precocious astrocytes at late embryonic

stages throughout the CP (Figure 6). From P7 onward, nascent

(and mature) astrocytes could be clearly discerned and the g/r

rate of Lgl1�/� to WT astrocytes was significantly increased (Fig-

ure 6M). To increase the temporal resolution and pinpoint the

critical Lgl1-dependent stages in RGPs during aIPC and/or

astrocyte production, we stained Lgl1-MADM in a time course

for brain lipid-binding protein (BLBP) (Feng et al., 1994), which

labels RGPs during embryonic development, early postnatal

aIPC, and mature astrocytes. To quantify aIPCs in the devel-

oping CP, we have only included BLBP+ cells that do not contain

apical and/or basal processes that would be indicative for RGPs

integrated either in the ventricular zone or translocating toward

the pia, rather than aIPCs. We quantified the percentage of

BLBP+/GFP+ double-positive Lgl1�/� cells and BLBP+/tdT+ WT

cells in comparison to all MADM-labeled cells. While at E16,

virtually no BLBP+/GFP+ and BLBP+/tdT+ cells could be de-

tected, BLPB+ Lgl1�/� cells are significantly overrepresented

at E17, E18, and P0 when compared to BLBP+ WT MADM-

labeled cells (Figures 6A–6K). Next, we determined the fraction

of proliferating cells that co-labeled for the cell-cycle marker

Ki67. At E17, when neurogenesis has largely ceased, the fraction

of Ki67+ Lgl1�/� cells in the CP was significantly elevated when

compared to Ki67+ WT MADM-labeled cells. Increased prolifer-

ation was not observed before E17 but through E18, P0, and P7

(Figure 6L). Together, these results suggest that proliferating

Lgl1�/� aIPCs are increased in late embryonic Lgl1-MADM, re-

sulting in a larger overall number of mature astrocytes at post-

natal stages.

We next quantified the relative fraction of astrocytes within the

total of MADM-labeled cells (neurons plus astrocytes) (Fig-

ure 6M). At P7, the fraction of mutant astrocytes (�25%) in

Lgl1-MADM was increased already by 2-fold when compared

to the fraction (�13%) of WT astrocytes in control-MADM. While

the astrocyte fraction in control-MADM reached a plateau of

�20% at P21, the relative fraction of Lgl1�/� mutant astrocytes

in Lgl1-MADM was nearly 50% of all MADM-labeled cells. The

relative number of mutant astrocytes remained stable in more

mature mice and no signs of astrocytoma were observed

(Figure S7). These data indicate that the overproduction of

astrocytes upon loss of Lgl1 still obeys tissue homeostasis

mechanisms. Interestingly, the relative fraction of astrocytes in

cKO-Lgl1-MADM was also increased and reached �50% (Fig-

ure 6M). Thus, the loss of Lgl1�/� in aIPCs results in dramatic

increase in mature astrocytes regardless of the state of the envi-

ronment, i.e., normal in Lgl1-MADM or Lgl1�/� in cKO-Lgl1-

MADM. Altogether, these results suggest that the control of

astrocyte production at normal numbers represents a critical

cell-autonomous Lgl1 function.

Genetic Interaction of Lgl1with Egfr Reveals FunctionalRelationship in Cortical Astrocyte GenerationOn the cellular level, loss of Lgl1 function leads to an exuberant

number of aIPCs and mature cortical astrocytes. To obtain

Neuron 94, 517–533, May 3, 2017 523

CD133 / / / GFP tdT DAPI

V

DW LW

cont

rol-M

AD

MLg

l1-M

AD

McK

O-L

gl1-

MA

DM

A

C

E

D

B

F

Olfactory Bulb

0

1

0.5

0

1

0.5

Rat

io o

lf. g

ranu

le c

ells

0Lgl1-/-Lgl1-/-

100

50

2#

olf.

gran

ule

cells

/ m

m

P S

T

U

Q

R

G

M

GCL

G

M

GCL

G

M

GCL

ns

0wt / wt

1

0.5

0-/-Lgl1 / wt

1

0.5

0Lgl1-/-Lgl1-/-

100

50

2#

RM

S ne

uron

s / m

m

RMS

H

I

G J

K

L

M

N

O

Rat

io R

MS

neur

ons

GCL

G

M

RMS

LW

NCXDW

MG

GCL

ERMS

Rat

io R

MS

neur

ons

wt / wt

-/-Lgl1 / wt

Rat

io o

lf. g

ranu

le c

ells

Figure 3. Sparse and Whole Neuroepithelium Ablation of Lgl1 Differentially Affects Postnatal Neurogenesis

(A–F) Analysis of CD133 (white) expression and integrity of postnatal V-SVZ in dorsal (DW, A–C) and lateral (LW, D–F) walls of lateral ventricle (LV) in control-

MADM (A and D), Lgl1-MADM (B and E), and cKO-Lgl1-MADM (C and F) at P21. Note the absence of CD133 expression and irregular arrangement of V-SVZ cells

in DW in cKO-Lgl1-MADM (C). Scale bars, 10 mm.

(legend continued on next page)

524 Neuron 94, 517–533, May 3, 2017

mechanistic insights at the molecular level, we pursued a

candidate gene approach and conceived genetic interaction ex-

periments in a MADM context. We focused on Egfr because pre-

vious studies have suggested a crucial and dose-dependent

regulatory function of Egfr signaling in gliogenesis (Burrows

et al., 1997; Sibilia et al., 1998). We first tested the hypothesis

that exuberant astrocyte generation upon loss of Lgl1 could be

dependent on functional Egfr. Because Egfr is located on

chr11, it was possible to genetically link an Egfr-flox allele (Lee

and Threadgill, 2009) with the MADM cassette and Lgl1-D

allele to generate Egfr�/�,Lgl1�/� double mutant cells in a

mosaic context (MADM-11GT/TG,Egfr-flox,Lgl1-D;Emx1Cre/+ [Egfr-

Lgl1-MADM; red cells, Egfr+/+,Lgl1+/+; green cells, Egfr�/�,Lgl1�/�; background, Egfr+/�,Lgl1+/�]). Whereas in control-

MADM, the relative fractions of redWT and greenWT astrocytes

were each �20% (Figures 7A–7C), the fraction of green

Egfr�/�,Lgl1�/� double mutant astrocytes in Egfr-Lgl1-MADM

was reduced close to zero, and the fraction of red Egfr+/+,Lgl1+/+

astrocytes was increased to �40% (Figures 7J–7L). These data

show that the emergence of the cell-autonomous Lgl1 loss

of function (LOF) phenotype (i.e., excessive astrocyte produc-

tion) is dependent on functional Egfr. Moreover, Egfr+/+

astrocytes in an Egfr+/� background seem to exhibit a competi-

tive dose-sensitive advantage when compared to control-

MADM (all cells Egfr+/+). In order to scrutinize such concepts,

we conditionally ablated Egfr from all Emx1+ progenitor

lineages on top of Lgl1-MADM mosaic (Figures 7P–7R;

MADM-11GT,Egfr-flox/TG,Egfr-flox,Lgl1-D;Emx1Cre/+ [cKO-Egfr-Lgl1-

MADM; red cells, Egfr�/�,Lgl1+/+; green cells, Egfr�/�,Lgl1�/�;background, Egfr�/�,Lgl1+/�]). Strikingly, in such experimental

conditions, we detected almost no green and red astrocytes at

all, indicating not only a loss of dosage sensitivity (all cells are

Egfr�/�) but also confirming the requirement of Egfr signaling

for astrocyte genesis regardless of Lgl1 genotype (green cells

are Egfr�/�,Lgl1�/�). Next, we tested the possibility that loss of

Lgl1 in combination with Egfr dosage-sensitive competitive

advantagemight exhibit synergistic effects on astrocyte produc-

tion. We thus generated MADM-11GT,Lgl1-D/TG,Egfr-flox;Emx1Cre/+

(Figures 7M–7O; Egfr-Lgl1-MADM; red cells, Egfr+/+,Lgl1�/�;green cells, Egfr�/�,Lgl1+/+; background, Egfr+/�,Lgl1+/�) andquantified the relative levels of red and green astrocytes. While

green Egfr�/�,Lgl1+/+ cells were almost completely absent

(similar as in the above conditions), the relative fraction of red

Egfr+/+,Lgl1�/� astrocytes was �70% (Figure 7O). This number

(G–I) Analysis of MADM-labeling pattern in RMS in control-MADM (G), Lgl1-MAD

(J–L) Quantification of green/red ratio in control-MADM (J) and Lgl1-MADM (

neurons/mm2) in cKO-Lgl1-MADM (L) in RMS at P21.

(M–R) Analysis of MADM-labeling pattern in olfactory bulb (OB) in control-MADM

boxed areas in (M–O) illustrating the distribution of OB interneurons (oINs) acros

the drastic decrease of red and green Lgl1�/� MADM-labeled cells in cKO-Lgl1

40 mm (P–R).

(S–U) Quantification of green/red ratio in control-MADM (S) and Lgl1-MADM (T

granule cells/mm2) in cKO-Lgl1-MADM (U) in GCL at P21.

(V) Schematic illustrating the V-SVZ adult stem cell niches in DW (orange) and LW

migrate along the RMS to the OB. NCX, neocortex; RMS, rostral migratory stream

layer (extension of RMS) within the OB.

Values represent mean ± SEM. ns, nonsignificant.

See also Figure S8.

is much higher than in separate conditions with relative astrocyte

fractions of green Lgl1�/� in Lgl1-MADM (Figures 7D–7F) or red

Egfr+/+ in Egfr-MADM (Figures 7G–7I) reaching�40% each, indi-

cating a synergistic effect in Egfr+/+,Lgl1�/� astrocytes in an

Egfr+/� background. Altogether, our Lgl1/Egfr genetic interaction

experiments suggest a critical functional relationship in cortical

astrocyte generation.

DISCUSSION

In the developing cerebral cortex, NSCs are in charge of gener-

ating cell-type diversity. However, the underlying cellular and

molecular mechanisms controlling NSC proliferation behavior

and lineage progression are poorly defined. In our study, by

using quantitative MADM-based experimental paradigms at sin-

gle cell resolution, we found that Lgl1 is required at distinct

sequential stages in cortical NSCs to control quantitative neuron

and glia output (Figure 8). At early embryonic stages, Lgl1 func-

tion is required at the global neuroepithelium tissue level during

cortical neurogenesis. The loss of Lgl1 in all RGPs triggers a dy-

namic community effect that results in the overproduction of

cortical projection neurons and the formation of an SBH. In

contrast, Lgl1 is cell-autonomously required in aIPCs at later

stages to control astrocyte production and in postnatal V-SVZ

neurogenic niche for lineage progression of NSCs in the LV.

Collectively, our results define distinct sequential non-cell-

autonomous and intrinsic cell-autonomous Lgl1 functions con-

trolling cortical neuron and glia genesis and postnatal stem cell

behavior. We discuss our findings with emphasis on the interplay

of cell-autonomous gene function with non-cell-autonomous

and/or community effects and in the context of the general prin-

ciples of NSC proliferation behavior.

Genetic Dissection of Cell-Autonomous Gene Functionand Environmental Community Effects at Single-CellResolutionWhile genetic loss of function can reveal cell-autonomous gene

functions, the contribution of non-cell-autonomous gene func-

tions and/or community effects often remain poorly defined.

Non-cell-autonomous gene functions may involve directed

cell-to-cell communication either via contact-mediated or

secreted signaling cues (Greenman et al., 2015; Hippenmeyer,

2014). For instance, excess activation of AKT3 in just a small

population of cells is associated with human focal malformations

M (H), and cKO-Lgl1-MADM (I) at P21. Scale bars, 30 mm.

K); and absolute numbers (i.e., almost complete absence; number of RMS

(G), Lgl1-MADM (H), and cKO-Lgl1-MADM (I) at P21. Higher magnification of

s the granule cell layer (GCL) and in periglomerular areas in the P21 OB. Note

-MADM (O and R). Nuclei are stained using DAPI. Scale bars, 200 (M–O) and

), and absolute numbers (i.e., almost complete absence; number of olfactory

(purple), RMS, and OB in the postnatal brain. oINs are generated in the V-SVZ,

; G, glomerular layer; M, mitral cell layer; GCL, granule cell layer; E, ependymal

Neuron 94, 517–533, May 3, 2017 525

GFP

/ /

tdT

DA

PI

P0P7

P21

6Mo

12M

oLgl1-MADM / -/- Lgl1- wtcontrol-MADM / wt- wt

A B

C D

E F

HG

I J

K

0

100

1 2-3 4-5 6-10 11-15

50

25

75

>15

# cells / cluster

% n

orm

. fre

quen

cy +/+Lgl1

+/+Lgl1

ns

ns

L

# cells / cluster

0

100

1 2-3 4-5 6-10 11-15

50

25

75

>15

+/+Lgl1

-/-Lgl1

% n

orm

. fre

quen

cy

**

***

*** *** *

*

*

*

*

*

* *

**

**

**

Qua

ntifi

catio

n

ratio / cells-/-Lgl1 wt

control-MADM

Lgl1-MADM

0

15

10

P0 P7 P21 6Mo 12Mo

5

**

*****

*** ***

20

Quantification

M

Lgl1

-MA

DM

/ -

-/- L

gl1

wt

cont

rol-M

AD

M /

w

t-

wt

P0

Lgl1

-MA

DM

/ -/-

L

gl1

-w

t

cont

rol-M

AD

M

/ -

wt

wt

P21P7

N

O

P

Q

R S

Figure 4. Lgl1 Is Cell-Autonomously Required for V-SVZ NSC Lineage Progression

(A–J) Time course analysis of MADM-labeling pattern in V-SVZ in dorsal wall (DW) of the lateral ventricle (LV) in control-MADM (A, C, E, G, and I) and Lgl1-MADM

(B, D, F, H, and J) at P0 (A and B), P7 (C and D), P21 (E and F), 6 months (6Mo; G and H), and 1 year (12Mo; I and J). White asterisks (*) mark aberrant large clusters

of green Lgl1�/� cells in V-SVZ. Nuclei are stained using DAPI. Scale bars, 75 (A and B), 60 (C and D), 30 (E and F), and 50 mm (G–J).

(legend continued on next page)

526 Neuron 94, 517–533, May 3, 2017

in cortical development, which disrupts the architecture of the

entire hemisphere (Baek et al., 2015). These findings suggest

that alteration of the properties of individual neurons collectively

may affect the entire community. Such a phenomenon is also

observed in distinct cellular contexts including collective cell

migration and tissue morphogenesis (Heisenberg and Bellaıche,

2013). The cellular and molecular mechanisms orchestrating

community effects during brain development are mostly un-

known due to the lack of experimental assays enabling the

visualization and quantitative assessment of the non-cell-auton-

omous elements in full or whole tissue conditional loss-of-func-

tion phenotypic analysis (Greenman et al., 2015). To this end,

we have established an unprecedented genetic strategy to visu-

alize and dissect the interplay of relative cell-autonomous gene

function and the contribution of non-cell-autonomous commu-

nity effects to the overall phenotype presentation. Our assay

relies uponMADM-based single-cell phenotypic analysis of indi-

vidual mutant cells in (1) normal environment, and (2) homozy-

gous mutant environment. The approach includes the sparse

mosaic versus global/whole tissue wide ablation of a candidate

gene, resulting in distinct cellular environments permitting the

direct assessment of the non-cell-autonomous influence on the

single-cell phenotype. This assay can be applied in principle to

any candidate gene of interest, provided that knockout or condi-

tional alleles are available, and MADM cassettes have been in-

serted on the particular chromosome where the gene is located.

Here, we utilized such a genetic paradigm to genetically dissect

the function of Lgl1 in controlling NSC proliferation behavior in

neurogenesis and glia production.

Role of Lgl1 in Embryonic Cortical NeurogenesisAblation of Lgl1 from all NESCs results in a severe non-cell-

autonomous community effect. Albeit, the formation of ectopic

clusters of exuberantly proliferating Lgl1�/� progenitors in

cKO-Lgl1-MADM (where all cells lack Lgl1) occurred in a non-

stereotypic fashion but was correlated with absence or misex-

pression of apical and basolateral components. Strikingly, the

phenotype of cKO-Lgl1-MADM is almost congruent to the

phenotype ofNumb/Numbl doublemutants (Li et al., 2003; Rasin

et al., 2007). NUMB has been proposed to control the trafficking

of adherens junction components (Rasin et al., 2007). Because

LGL1 has also been suggested to regulate polarized secretion

and exocytosis (M€usch et al., 2002), it will be interesting in the

future to determine a potential functional relationship of LGL1

and NUMB in the regulation of adherens junctions and possibly

the control of progenitor proliferation. The SBH in cKO-Lgl1-

MADM was composed of neurons with upper and lower cortical

layer identity as well as glia cells, indicating that the generation of

faithful cell fates by embryonic Lgl1�/� mutant RGPs is not

(K and L) Cluster size (number of cells/cluster) distribution of V-SVZ cells at P21 in

green bars in control-MADM and red bar in Lgl1-MADM) are not forming cluster

(M) Quantification of the green/red ratio of MADM-labeled cells in DW V-SVZ at

MADM (Lgl1�/�/WT, gray bars). Scale bar, 7.5 mm.

(N–S) Time course analysis of cellular morphology in V-SVZ in control-MADM (M

P21 (Q and R). Scale bars, 7.5 (N–P) and 9 mm (Q and R).

Values represent mean ± SEM. ns, nonsignificant; *p < 0.05, **p < 0.01, ***p < 0.

See also Figures S4 and S8.

severely disturbed. However, it will be intriguing to determine

the global tissue-wide homeostatic properties and possible

changes associated with the SBH formation and maintenance

in the adult cKO-Lgl1-MADM. In Lgl1-MADM with only sparse

Lgl1 knockout, the individual Lgl1�/� mutant progenitors appear

to proliferate normally, presumably because their integration

within the neuroepithelium is rescued by the surrounding

‘‘normal’’ progenitors maintaining the cell adhesion of mutant

cells in a non-cell-autonomous manner. Future efforts aiming

to determine the genetic fingerprints of Lgl1�/� mutant progeni-

tors in Lgl1-MADM and cKO-Lgl1-MADM by single cell RNA

sequencing (RNA-seq) could help (provided that a large enough

set of progenitors can be faithfully isolated) to identify the

intrinsic signaling pathways associated with either ‘‘rescue’’ of

mutant progenitors in Lgl1-MADM or aberrant proliferation and

SBH formation in cKO-Lgl1-MADM.

Lgl1-Dependent Lineage Progression in PostnatalV-SVZ NSCsThe V-SVZ in the DW of the LV was not properly established in

cKO-Lgl1-MADM, and postnatal OB interneuron (oIN) generation

was virtually absent. It is likely that the strong community effects

observed in the embryonic VZ resulted in a dispersal of the DW

ependymal cells in the SBH. This phenotype reflects the impor-

tance of an intact E1 cell layer that derives from a discrete pool

of RGPs (Jacquet et al., 2011; Paez-Gonzalez et al., 2011) and

that is severely affected in the cKO-Lgl1-MADM. In contrast,

Lgl1-MADM mice show a normal E1 layer. Nevertheless, post-

natal neurogenesis is still compromised in Lgl1-MADM, although

not to the extent like in cKO-Lgl1-MADM, suggesting a distinct

underlying basis and critical cell-autonomous role for Lgl1 in

postnatal neurogenesis. In sparse Lgl1-MADM, mutant Lgl1�/�

type B1 cells display aberrant cellularmorphology and frequently

appear in clusters or syncytia of nuclei with long cytoplasmic

bridges that could represent incomplete cell division. It is

tempting to speculate that Lgl1 could regulate very specific intra-

cellular trafficking events required for correct cytokinesis (Schiel

et al., 2013). In any case, our results indicate a delicate Lgl1-

dependent step in the proliferation pattern of type B1 cells. Lgl1

function could dictate the balance of symmetric versus asym-

metric V-SVZ progenitor division pattern and thereby indirectly

regulate the level of progenitor expansion versus differentiation

and thus quantitative output during postnatal neurogenesis.

Because Lgl1 has been associated with polarity-driven pro-

cesses and is involved in controlling the switch from symmetric

to asymmetric RGP divisions, it will be intriguing to assess in

future studies whether the loss of Lgl1 in asymmetrically dividing

type B1 cells results in ‘‘more symmetric’’ division that may be

not fully completed due to possible cell-cycle checkpoints.

control-MADM (K) and Lgl1-MADM (L). Note that�90% of Lgl+/+ cells (red and

s and occur as single cells but <40% Lgl�/� appear as single cells.

P0, P7, P21, 6Mo, and 12Mo in control-MADM (WT/WT, white bars) and Lgl1-

, O, and Q) and Lgl1-MADM (N, P, and R) at P0 (M and N), P7 (O, and P), and

001.

Neuron 94, 517–533, May 3, 2017 527

Clonal Analysis

GFP / / tdT DAPI

TM /

E11

- A

naly

sis

/ P21

Cre

ER/+

Lgl1

-MA

DM

; Em

x1 /

-/-

Lgl

1-

wt

Cre

/+Lg

l1-M

AD

M ;

Emx1

/ -/-

L

gl1

-w

t

Cre

ER/+

cont

rol-M

AD

M ;

Emx1

/

wt

-w

t

Cre

/+co

ntro

l-MA

DM

; Em

x1 /

w

t-

wt

Clonal AnalysisPopulation Analysis

I

II/III

IV

V

VI

I

II/III

IV

V

VI

I

II/III

IV

V

VI

I

II/III

IV

V

VI

NCX

HC

HC

NCX

A C

F

G

DB

E

Population Analysis

H

ratio

/

ast

rocy

tes

gr

0

wt/wt

-/-

Lgl1/w

t

20

10

-/-

Lgl1/w

t-/-

Lgl1/w

twt/w

twt/w

t

P21 3Mo 12Mo

******

# astrocytes /red subclone

0

contro

l-MADM - w

t

Lgl1-MADM - w

t

10075

50

25ns

# astrocytes /green subclone

0

contro

l-MADM - w

t

Lgl1-MADM -

-/-

Lgl1

10075

50

25

***

Figure 5. Lgl1 Function Is Cell-Autono-

mously Required for Cortical Astrocyte

Generation

(A–D) Population analysis of MADM-labeled

cortical astrocytes in control-MADM (A and C) and

Lgl1-MADM (B and D) at P21. White arrows in (D)

mark groups with increased numbers of Lgl1�/�

astrocytes. For high-resolution analysis of cell size

and arborization, refer to Figure S5. Scale bars,

500 (A and B) and 60 mm (C and D).

(E and F) Clonal analysis ofMADM-labeled cortical

astrocytes in control-MADM (E) and Lgl1-MADM

(F). Tamoxifen (TM) was applied at E11 and brain

samples analyzed at P21. G2-X clones are illus-

trated. Cortical layers are indicated (roman nu-

merals). Scale bars, 70 mm.

(G) Quantification of the green/red ratio of cortical

astrocytes in population analysis at P21, 3 months

(3Mo), and 1 year (12Mo) in control-MADM (WT/

WT) and Lgl1-MADM (Lgl1�/�/WT).

(H) Quantification of the number of astrocytes in

red (left) and green (right) MADM subclones. Note

that red cells within red subclones in both control-

MADM and Lgl1-MADM are WT, and green cells

within green subclones are WT in control-MADM

but Lgl1�/� in Lgl1-MADM.

Values represent mean ± SEM. ns, nonsignificant,

***p < 0.001.

See also Figures S5–S7.

Alternatively, Lgl1 could act as an intracellular sensor and/or

mediator of extracellular signals to regulate the transition from

quiescent to active status in type B1 cells. Importantly, however,

typeB1 cells produce typeCcells that in turn can give rise to neu-

roblasts migrating toward the RMS. Because there is a lineage

relationship from type B1 to type C to neuroblast, the loss of

Lgl1 function in type B1 cells results in a block of lineage progres-

sion and thus affects all downstream progenitor cells and their

output. Type C cells not only produce neuroblasts destined for

528 Neuron 94, 517–533, May 3, 2017

the olfactory granule cell layer but also

give rise to oligodendrocyte lineage

(Kriegstein and Alvarez-Buylla, 2009). It

will thus be important in future lineage

tracing experiments to evaluate whether,

and to what extent, a block in lineage

progression from type B1 to type C cells

may affect the oligodendrocyte lineage

emerging from type C cells. Perhaps

even more important will be the determi-

nation of progenitor cell-type diversity

among type B and type C cells and to

evaluate whether the loss of Lgl1 may

result in changes of progenitor cell fates

and how it affects neuron and glia output.

Cell-Autonomous Lgl1 Function inPostnatal Cortical AstrocyteProductionMADM is a powerful approach to study

cell-autonomous gene function at high

spatiotemporal resolution (Hippenmeyer et al., 2010; Joo et al.,

2014). Here, the analysis of Lgl1-MADM allowed us to identify

as-yet-unknown cell-autonomous Lgl1 functions in NSC prolifer-

ation behavior (Figure 8). Once RGPs cease the production of

projection neurons, they adopt gliogenic potential (Bayraktar

et al., 2014; Kriegstein and Alvarez-Buylla, 2009). However, the

cellular and molecular mechanisms controlling glia production

and quantitative output are not well understood. In our analysis

of Lgl1-MADM, we discovered that the production of astrocytes

BLBP / / / GFP tdT DAPI

control-MADMLgl1-MADMcontrol-MADM Lgl1-MADM

E16 E17

A B C D

F

E

H

I

J

G

+ + +% BLBP ,MADM / total MADM

0

6

4

E16 E17 E18 P0

2

+/+Lgl1

-/-Lgl1***

8

**

***

+ +% MADM astrocytes / total MADM cells

0

50

P7 P14 P21

25

P210

50

25

control-MADM +/+(R+G Lgl1 cells)

Lgl1-MADM -/-(G Lgl1 mutant cells)

cKO-Lgl1-MADM -/-(R+G Lgl1 mutant cells)

M

****** ***

+ + +% Ki67 ,MADM / total MADM

0

6

4

E17 E18 P0 P7

2

+/+Lgl1

-/-Lgl1

8

***

******

***

E16

LK

Figure 6. Increased Numbers of Astrocyte Progenitors Precede Overproduction of Postnatal Cortical Astrocytes

(A–J) Analysis of BLBP (white) expression pattern in CP in control-MADM (A and C–F) and Lgl1-MADM (B and G–J) at E16 (A and B) and E17 (C–J). Higher

magnification of yellow boxed areas in (C and G) illustrate MADM-labeled cells expressing BLBP (marked by cyan arrows) in control-MADM (D–F) and Lgl1-

MADM (H–J). Scale bars, 50 (A–C and G) and 60 mm (D–F and H–J).

(K) Quantification of fraction (%) BLBP+/MADM+ double-positive cells of total number of MADM-labeled cells (red tdT+ or green GFP+) in Lgl1-MADM at E16, E17,

E18, and P0. Red bars represent tdT+/BLBP+ double-positive Lgl1+/+ cells, and green bars represent GFP+/BLBP+ Lgl1�/� cells.

(L) Quantification of fraction (%) Ki67+/MADM+ double-positive cells of total number of MADM-labeled cells (red tdT+ or green GFP+) in Lgl1-MADM at E16, E17,

E18, P0, and P7. Red bars represent tdT+/Ki67+ double-positive Lgl1+/+ cells, and green bars represent GFP+/Ki67+ Lgl1�/� cells.

(M) Quantification of fraction (%) of MADM-labeled astrocytes of total number of MADM-labeled cells (neurons and astrocytes). Population analysis in control-

MADM, Lgl1-MADM, and cKO-Lgl1-MADM at P7 and P14 (control-MADM, Lgl1-MADM) and P21 (control-MADM, Lgl1-MADM, and cKO-Lgl1-MADM). Note that

the fraction of mutant Lgl1�/� astrocytes in cKO-Lgl1-MADM is elevated to similar levels like the mutant Lgl1�/� astrocytes in Lgl1-MADM at P21.

Values represent mean ± SEM. **p < 0.01, ***p < 0.001.

See also Figure S7.

Neuron 94, 517–533, May 3, 2017 529

++

% M

AD

M a

stro

cyte

s / t

otal

MA

DM

cel

ls

0

80

40

20

60

ns++

% M

AD

M a

stro

cyte

s / t

otal

MA

DM

cel

ls

0

80

40

20

60

***

ns

++

% M

AD

M a

stro

cyte

s / t

otal

MA

DM

cel

ls

0

80

40

20

60

++

% M

AD

M a

stro

cyte

s / t

otal

MA

DM

cel

ls

0

80

40

20

60 ***

++

% M

AD

M a

stro

cyte

s / t

otal

MA

DM

cel

ls

0

80

40

20

60 ***

++

% M

AD

M a

stro

cyte

s / t

otal

MA

DM

cel

ls

0

80

40

20

60 ***

GT/TG Cre/+MADM-11 ; Emx1 GT/TG,Lgl1 Cre/+MADM-11 ; Emx1

GT/TG,Egfr,Lgl1 Cre/+MADM-11 ; Emx1GT/TG,Egfr Cre/+MADM-11 ; Emx1

+/+Lgl1 +/- Lgl1-/-Lgl1wt wtwt

GT,Egfr/TG,Egfr,Lgl1 Cre/+MADM-11 ; Emx1GT,Lgl1/TG,Egfr Cre/+MADM-11 ; Emx1

+/+ +/+Lgl1 ; Egfr +/- +/- Lgl1 ; Egfr-/- -/-Lgl1 ; Egfr+/+Egfr +/- Egfr-/-Egfr

+/+ -/-Lgl1 ; Egfr +/- -/- Lgl1 ; Egfr-/- -/-Lgl1 ; Egfr-/- +/+Lgl1 ; Egfr +/- +/- Lgl1 ; Egfr+/+ -/-Lgl1 ; Egfr

GFP / / tdT DAPI

I

II/III

IV

V

VI

I

II/III

IV

V

VI

I

II/III

IV

V

VI

I

II/III

IV

V

VI

I

II/III

IV

V

VI

I

II/III

IV

V

VI

HC

NCX

HC

NCX

HCNCX

HC

NCX

HCNCX

HCNCX

A B C D FE

G H I J LK

M N O P RQ

Figure 7. Genetic Interaction of Lgl1 with Egfr Reveals Functional Relationship in Cortical Astrocyte Generation

Analysis of MADM-labeled cortical astrocytes in MADM-11GT/TG;Emx1Cre/+ (A–C; control-MADM; all cells WT), MADM-11GT/TG,Lgl1-D;Emx1Cre/+ (D–F, Lgl1-

MADM; red cells: Lgl1+/+; green cells: Lgl1�/�; background: Lgl1+/�), MADM-11GT/TG,Egfr-flox;Emx1Cre/+ (G–I, Egfr-MADM; red cells: Egfr+/+, green cells: Egfr�/�,background: Egfr+/�), MADM-11GT/TG,Egfr-flox,Lgl1-D;Emx1Cre/+ (J–L; Egfr-Lgl1-MADM; red cells: Egfr+/+,Lgl1+/+; green cells: Egfr�/�,Lgl1�/�; background:

Egfr+/�,Lgl1+/�), MADM-11GT,Lgl1-D/TG,Egfr-flox;Emx1Cre/+ (M–O; Egfr-Lgl1-MADM; red cells: Egfr+/+,Lgl1�/�; green cells: Egfr�/�,Lgl1+/+; background:

Egfr+/�,Lgl1+/�), and MADM-11GT,Egfr-flox/TG,Egfr-flox,Lgl1-D;Emx1Cre/+ (P–R; cKO-Egfr-Lgl1-MADM; red cells: Egfr�/�,Lgl1+/+; green cells: Egfr�/�,Lgl1�/�; back-ground: Egfr�/�,Lgl1+/�) at P21. Quantification (C, F, I, L, O, and R) of fraction (%) of MADM-labeled cortical astrocytes of total number of MADM-labeled cells

(neurons and astrocytes) is indicated. Note that in Egfr-MADM (G–I) the relative fraction of red Egfr+/+ astrocytes is increased up to �40% (presumably due to

dosage sensitivity and growth advantage in an Egfr+/� background) whereas green Egfr�/� astrocytes are almost completely absent. Introduction of Lgl1�/� into

Egfr�/� green cells (J–L) does not rescue the number of green Egfr�/�,Lgl1�/� double mutant astrocytes. In contrast, the relative fraction of red Egfr+/+ astrocytes

that also lack Lgl1 (i.e., Egfr+/+,Lgl1�/�) is increased up to�70% (M–O). Global ablation of Egfr (P–R) results in near complete loss of cortical astrocytes regardless

of Lgl1 genotype. Cortical layers are indicated (roman numerals). Values represent mean ± SEM. ns, nonsignificant, ***p < 0.001. Scale bars, 500 (A, D, G, J,

N, and P) and 60 mm (B, E, H, K, N, and Q).

See also Figure S8.

530 Neuron 94, 517–533, May 3, 2017

E14E12E8 - E10

Neuroepithelialcell

Neurogenic radial gliaprogenitor cell

Postnatal / Adult

V-SVZ type B1 cell

IV

I

II

III

VI

V

V-SVZ

Neuroblasts

WM

VentricularZone(VZ)

CorticalPlate

SubventricularZone (SVZ)

VZ

E18 / Birth

Gliogenic radial gliaprogenitor cell

II-IV

VI

V

SVZ

VZ

Astrocytes

Lgl1 function required for VZ integrity(control of embryonic neurogenesis)

Lgl1 regulates corticalastrocyte production

Lgl1 controls postnatalneurogenesis in V-SVZ

LGL1

LGL1 LGL1

Figure 8. Lgl1 Controls NSC Lineage Progression in Embryonic and Postnatal Stem Cell Niches

Schematic model of NSC lineage progression and the role of Lgl1 non-cell-autonomous (indicated in green) and cell-autonomous (indicated in blue) functions at

discrete sequential steps.

is strongly increased upon cell-autonomous loss of Lgl1. Inter-

estingly, the relative number of astrocytes was also significantly

higher in cKO-Lgl1-MADM indicating a dominant cell-autono-

mous component, not strongly influenced by non-cell-autono-

mous community effects. What can we learn from the Lgl1

loss-of-function astrocyte phenotype with regard to the general

principles of RGP proliferation behavior? It has been suggested

that RGPs give rise to aIPCs that locally amplify astrocyte pro-

duction in a tightly controlled manner (Ge et al., 2012). It is

currently not clear whether astrocyte production follows a strictly

deterministic program similar to neurogenic RGPs (Gao et al.,

2014) or whether a certain degree of stochasticity contributes

to the proliferation dynamics of aIPCs. Regardless of the precise

mechanism, clonally related astrocytes do not disperse broadly

(Gao et al., 2014; Molofsky et al., 2014), and because astrocytes

exhibit precise tiling (i.e., do not overlap their fine projections), it

has been suggested that astrocyte production is controlled by

homeostatic cues to ensure complete coverage of the local neu-

ropil (Molofsky and Deneen, 2015). In any case, loss of Lgl1 func-

tion results in a scalable overproduction of Lgl1�/� astrocytes.

Given that astrocytes tile the neuropil, do Lgl1�/� aIPCs have a

competitive advantage over WT progenitors? And if yes, which

Lgl1-dependent signaling cascades are misregulated in aIPCs?

The astrocyte overproduction in Lgl1-MADM could reflect the

loss of a specific Lgl1-dependent function in polarized secretion

and/or exocytosis in order to regulate cell-surface abundance of

astrocyte production stimulating and/or inhibiting factors. It is

intriguing to note that the control of polarized secretion, exocy-

tosis (M€usch et al., 2002), and possibly further intracellular traf-

ficking events, could actually represent one unifying function of

Lgl1 in the control of proliferating NSCs. In such a mechanistic

framework, Lgl1 could regulate the cell-surface abundance of

junctional complex components in embryonic RGPs, control traf-

ficking events critical for cytokinesis in type B1 NSC, and tune

growth factor receptor levels at the plasmamembrane in prolifer-

ating aIPCs. In this regard, we could observe genetic interaction

between Lgl1 and Egfr suggesting, indeed, a functional relation-

ship. Although the precise nature of Lgl1/Egfr interaction remains

to be determined, it seems highly specific for cortical astrocyte

generation but not V-SVZ NSC behavior (Figure S8). It will be

revealing toprobewhether LGL1andEGFR interact at the protein

level and to assay EGFR cell surface levels and/or turnover in

Lgl1�/� context during astrocyte generation. We cannot exclude

that Egfr and Lgl1 also play independent functions in astrocyte

generation, and it will be interesting to dissect those putative

functions. Alternatively, but not mutually exclusive, Lgl1 could

regulate the number of symmetric amplification versus asym-

metric differentiation divisions by regulating intracellular polarity

Neuron 94, 517–533, May 3, 2017 531

and/or the symmetry of the division plane in aIPCs. It will be inter-

esting in the future to assess the mechanisms and dependence

on Lgl1 function dictating the total astrocyte unit production in

distinct functional areas in the cortex and beyond in other brain

areas. Lastly, Lgl1 is highly expressed in mature astrocytes

(Zhang et al., 2016), and it will be intriguing to assess the expres-

sion status of Lgl1 in reactive astrocytes during injury and

whether the local response in quiescent astrocyte progenitors

require the downregulation and/or inhibition of Lgl1 function in

order to initiate the astrocyte production at injury sites.

Collectively, by using sparse and whole tissue genetic MADM

approaches to ablate Lgl1 gene function in NSCs, we define

distinct sequential Lgl1 functions in neurogenesis, astrocyte pro-

duction, and postnatal stem cell behavior. Our study emphasizes

the importance of the local stem cell niche environment and thus

non-cell-autonomous contributions in concert with cell-autono-

mous gene function in the control of NSC proliferation behavior.

More generally, single-cell phenotypes in conditional or full

knockout reflect a combination of both cell-autonomous gene

function andenvironment-derivedcues thatmay remedyor exac-

erbate any observed phenotype. It will thus be important in future

genetic loss-of-function paradigms to qualitatively and quantita-

tively determine the relative contributions of the intrinsic and

extrinsic components to the overall loss-of-function phenotype.

STAR+METHODS

Detailed methods are provided in the online version of this paper

and include the following:

d KEY RESOURCES TABLE

d CONTACT FOR REAGENT AND RESOURCE SHARING

d EXPERIMENTAL MODEL AND SUBJECT DETAILS

B Mouse Lines and Maintenance

d METHOD DETAILS

B Preparation of MADM-Labeled Tissue

B Immunostaining of MADM-Labeled Brains

B Imaging and Analysis of Marker Expression in MADM-

Labeled Brains

B Generation of MADM Clones in the Neocortex

B Astrocyte Morphology Filament Tracing

d QUANTIFICATION AND STATISTICAL ANALYSIS

SUPPLEMENTAL INFORMATION

Supplemental Information includes eight figures and two tables and can be

found with this article online at http://dx.doi.org/10.1016/j.neuron.2017.

04.012.

AUTHOR CONTRIBUTIONS

Conceptualization, S.H.; Methodology, R.B., C.S., and S.H.; Investigation,

R.B., M.P.P., L.E.B., C.S., S.L., F.M.P., and S.H.; Resources, G.X., T.H.G.,

O.K., and V.V.; Writing – Original Draft, S.H.; Writing – Review & Editing, all

authors; Funding Acquisition and Supervision, S.H.

ACKNOWLEDGMENTS

We thank Drs. N. Sans andM.Montcouquiol for transferring Lgl1-floxmice and

Dr. Threadgill for providing Egfr-flox mice; A. Heger (Preclinical Facility) and

532 Neuron 94, 517–533, May 3, 2017

E. Papusheva (Bioimaging Facility) for technical support; C. Schwayer,

E. Fisher, P. Hirschfeld, M. Frank, and J. Rodarte for initial experiments and/or

assistance; J. Knoblich, M. Loose, C.P. Heisenberg, and members of the

Hippenmeyer lab for discussion; and A. Kicheva, C. Mieck, N. Amberg, and

C. Duellberg for comments on the manuscript. This work was supported by

IST Austria institutional funds, NO Forschung & Bildung n[f+b] (C13-002), the

European Union (FP7-CIG618444), and a program grant from the Human Fron-

tiers Science Program (RGP0053/2014) to S.H.

Received: October 3, 2016

Revised: March 2, 2017

Accepted: April 5, 2017

Published: May 3, 2017

REFERENCES

Baek, S.T., Copeland, B., Yun, E.J., Kwon, S.K., Guemez-Gamboa, A.,

Schaffer, A.E., Kim, S., Kang, H.C., Song, S., Mathern, G.W., and Gleeson,

J.G. (2015). An AKT3-FOXG1-reelin network underlies defective migration in

human focal malformations of cortical development. Nat. Med. 21, 1445–1454.

Bayraktar, O.A., Fuentealba, L.C., Alvarez-Buylla, A., and Rowitch, D.H.

(2014). Astrocyte development and heterogeneity. Cold Spring Harb.

Perspect. Biol. 7, a020362.

Betschinger, J., Mechtler, K., and Knoblich, J.A. (2003). The Par complex di-

rects asymmetric cell division by phosphorylating the cytoskeletal protein

Lgl. Nature 422, 326–330.

Burrows, R.C., Wancio, D., Levitt, P., and Lillien, L. (1997). Response diversity

and the timing of progenitor cell maturation are regulated by developmental

changes in EGFR expression in the cortex. Neuron 19, 251–267.

Chung,W.S., Allen, N.J., and Eroglu, C. (2015). Astrocytes control synapse for-

mation, function, and elimination. Cold Spring Harb. Perspect. Biol. 7,

a020370.

Costa, M.R., Bucholz, O., Schroeder, T., and Gotz, M. (2009). Late origin of

glia-restricted progenitors in the developing mouse cerebral cortex. Cereb.

Cortex 19 (Suppl 1 ), i135–i143.

Doetsch, F., Caille, I., Lim, D.A., Garcıa-Verdugo, J.M., and Alvarez-Buylla, A.

(1999). Subventricular zone astrocytes are neural stem cells in the adult

mammalian brain. Cell 97, 703–716.

Feng, L., Hatten, M.E., and Heintz, N. (1994). Brain lipid-binding protein

(BLBP): a novel signaling system in the developing mammalian CNS. Neuron

12, 895–908.

Fietz, S.A., Kelava, I., Vogt, J., Wilsch-Br€auninger, M., Stenzel, D., Fish, J.L.,

Corbeil, D., Riehn, A., Distler, W., Nitsch, R., and Huttner, W.B. (2010).

OSVZ progenitors of human and ferret neocortex are epithelial-like and

expand by integrin signaling. Nat. Neurosci. 13, 690–699.

Freeman, M.R., and Rowitch, D.H. (2013). Evolving concepts of gliogenesis:

a look way back and ahead to the next 25 years. Neuron 80, 613–623.

Gao, P., Postiglione, M.P., Krieger, T.G., Hernandez, L., Wang, C., Han, Z.,

Streicher, C., Papusheva, E., Insolera, R., Chugh, K., et al. (2014).

Deterministic progenitor behavior and unitary production of neurons in the

neocortex. Cell 159, 775–788.

Ge, W.P., Miyawaki, A., Gage, F.H., Jan, Y.N., and Jan, L.Y. (2012). Local gen-

eration of glia is a major astrocyte source in postnatal cortex. Nature 484,

376–380.

Gleeson, J.G., Allen, K.M., Fox, J.W., Lamperti, E.D., Berkovic, S., Scheffer, I.,

Cooper, E.C., Dobyns, W.B., Minnerath, S.R., Ross, M.E., and Walsh, C.A.

(1998). Doublecortin, a brain-specific gene mutated in human X-linked lissen-

cephaly and double cortex syndrome, encodes a putative signaling protein.

Cell 92, 63–72.

Gorski, J.A., Talley, T., Qiu, M., Puelles, L., Rubenstein, J.L., and Jones, K.R.

(2002). Cortical excitatory neurons and glia, but not GABAergic neurons, are

produced in the Emx1-expressing lineage. J. Neurosci. 22, 6309–6314.

Greenman, R., Gorelik, A., Sapir, T., Baumgart, J., Zamor, V., Segal-Salto, M.,

Levin-Zaidman, S., Aidinis, V., Aoki, J., Nitsch, R., et al. (2015). Non-cell

autonomous and non-catalytic activities of ATX in the developing brain. Front.

Neurosci. 9, 53.

Hansen, D.V., Lui, J.H., Parker, P.R., and Kriegstein, A.R. (2010). Neurogenic

radial glia in the outer subventricular zone of human neocortex. Nature 464,

554–561.

Heisenberg, C.P., and Bellaıche, Y. (2013). Forces in tissue morphogenesis

and patterning. Cell 153, 948–962.

Hippenmeyer, S. (2013). Dissection of gene function at clonal level using

mosaic analysis with double markers. Front. Biol. 8, 557–568.

Hippenmeyer, S. (2014). Molecular pathways controlling the sequential steps

of cortical projection neuron migration. Adv. Exp. Med. Biol. 800, 1–24.

Hippenmeyer, S., Youn, Y.H., Moon, H.M., Miyamichi, K., Zong, H., Wynshaw-

Boris, A., and Luo, L. (2010). Genetic mosaic dissection of Lis1 and Ndel1 in

neuronal migration. Neuron 68, 695–709.

Hippenmeyer, S., Johnson, R.L., and Luo, L. (2013). Mosaic analysis with dou-

blemarkers reveals cell-type-specific paternal growth dominance. Cell Rep. 3,

960–967.

Homem, C.C., Repic, M., and Knoblich, J.A. (2015). Proliferation control in

neural stem and progenitor cells. Nat. Rev. Neurosci. 16, 647–659.

Jacquet, B.V., Muthusamy, N., Sommerville, L.J., Xiao, G., Liang, H., Zhang,

Y., Holtzman, M.J., and Ghashghaei, H.T. (2011). Specification of a Foxj1-

dependent lineage in the forebrain is required for embryonic-to-postnatal tran-

sition of neurogenesis in the olfactory bulb. J. Neurosci. 31, 9368–9382.

Joo, W., Hippenmeyer, S., and Luo, L. (2014). Neurodevelopment. Dendrite

morphogenesis depends on relative levels of NT-3/TrkC signaling. Science

346, 626–629.

Kessaris, N., Fogarty, M., Iannarelli, P., Grist, M., Wegner, M., and Richardson,

W.D. (2006). Competing waves of oligodendrocytes in the forebrain and post-

natal elimination of an embryonic lineage. Nat. Neurosci. 9, 173–179.

Klezovitch, O., Fernandez, T.E., Tapscott, S.J., and Vasioukhin, V. (2004). Loss

of cell polarity causes severe brain dysplasia in Lgl1 knockout mice. Genes

Dev. 18, 559–571.

Kriegstein, A., and Alvarez-Buylla, A. (2009). The glial nature of embryonic and

adult neural stem cells. Annu. Rev. Neurosci. 32, 149–184.

Lee, T.C., and Threadgill, D.W. (2009). Generation and validation of mice car-

rying a conditional allele of the epidermal growth factor receptor. Genesis

47, 85–92.

Li, H.S., Wang, D., Shen, Q., Schonemann, M.D., Gorski, J.A., Jones, K.R.,

Temple, S., Jan, L.Y., and Jan, Y.N. (2003). Inactivation of Numb and

Numblike in embryonic dorsal forebrain impairs neurogenesis and disrupts

cortical morphogenesis. Neuron 40, 1105–1118.

Lim, D.A., and Alvarez-Buylla, A. (2016). The adult ventricular-subventricular

zone (V-SVZ) and olfactory bulb (OB) neurogenesis. Cold Spring Harb.

Perspect. Biol. 8, http://dx.doi.org/10.1101/cshperspect.a018820.

Lui, J.H., Hansen, D.V., and Kriegstein, A.R. (2011). Development and evolu-

tion of the human neocortex. Cell 146, 18–36.

Magavi, S., Friedmann, D., Banks, G., Stolfi, A., and Lois, C. (2012). Coincident

generation of pyramidal neurons and protoplasmic astrocytes in neocortical

columns. J. Neurosci. 32, 4762–4772.

Merkle, F.T., Tramontin, A.D., Garcıa-Verdugo, J.M., and Alvarez-Buylla, A.

(2004). Radial glia give rise to adult neural stem cells in the subventricular

zone. Proc. Natl. Acad. Sci. USA 101, 17528–17532.

Molofsky, A.V., and Deneen, B. (2015). Astrocyte development: a guide for the

perplexed. Glia 63, 1320–1329.

Molofsky, A.V., Kelley, K.W., Tsai, H.H., Redmond, S.A., Chang, S.M.,

Madireddy, L., Chan, J.R., Baranzini, S.E., Ullian, E.M., and Rowitch, D.H.

(2014). Astrocyte-encoded positional cues maintain sensorimotor circuit

integrity. Nature 509, 189–194.

M€usch, A., Cohen, D., Yeaman, C., Nelson, W.J., Rodriguez-Boulan, E., and

Brennwald, P.J. (2002). Mammalian homolog of Drosophila tumor suppressor

lethal (2) giant larvae interacts with basolateral exocytic machinery in Madin-

Darby canine kidney cells. Mol. Biol. Cell 13, 158–168.

Noctor, S.C., Martınez-Cerdeno, V., Ivic, L., and Kriegstein, A.R. (2004).

Cortical neurons arise in symmetric and asymmetric division zones and

migrate through specific phases. Nat. Neurosci. 7, 136–144.

Paez-Gonzalez, P., Abdi, K., Luciano, D., Liu, Y., Soriano-Navarro, M.,

Rawlins, E., Bennett, V., Garcia-Verdugo, J.M., and Kuo, C.T. (2011). Ank3-

dependent SVZ niche assembly is required for the continued production of

new neurons. Neuron 71, 61–75.

Postiglione, M.P., and Hippenmeyer, S. (2014). Monitoring neurogenesis in the

cerebral cortex: an update. Future Neurol. 9, 323–340.

Rasin, M.R., Gazula, V.R., Breunig, J.J., Kwan, K.Y., Johnson, M.B., Liu-Chen,

S., Li, H.S., Jan, L.Y., Jan, Y.N., Rakic, P., and Sestan, N. (2007). Numb and

Numbl are required for maintenance of cadherin-based adhesion and polarity

of neural progenitors. Nat. Neurosci. 10, 819–827.

Schiel, J.A., Childs, C., and Prekeris, R. (2013). Endocytic transport and cyto-

kinesis: from regulation of the cytoskeleton to midbody inheritance. Trends

Cell Biol. 23, 319–327.

Schmechel, D.E., and Rakic, P. (1979). A Golgi study of radial glial cells in

developing monkey telencephalon: morphogenesis and transformation into

astrocytes. Anat. Embryol. (Berl.) 156, 115–152.

Sholl, D.A. (1953). Dendritic organization in the neurons of the visual andmotor

cortices of the cat. J. Anat. 87, 387–406.

Sibilia, M., Steinbach, J.P., Stingl, L., Aguzzi, A., and Wagner, E.F. (1998).

A strain-independent postnatal neurodegeneration in mice lacking the EGF re-

ceptor. EMBO J. 17, 719–731.

Spassky, N., Merkle, F.T., Flames, N., Tramontin, A.D., Garcıa-Verdugo, J.M.,

and Alvarez-Buylla, A. (2005). Adult ependymal cells are postmitotic and are

derived from radial glial cells during embryogenesis. J. Neurosci. 25, 10–18.

Stancik, E.K., Navarro-Quiroga, I., Sellke, R., and Haydar, T.F. (2010).

Heterogeneity in ventricular zone neural precursors contributes to neuronal

fate diversity in the postnatal neocortex. J. Neurosci. 30, 7028–7036.

Tang, S.H., Silva, F.J., Tsark, W.M., and Mann, J.R. (2002). A Cre/loxP-deleter

transgenic line in mouse strain 129S1/SvImJ. Genesis 32, 199–202.

Taverna, E., Gotz, M., and Huttner, W.B. (2014). The cell biology of neurogen-

esis: toward an understanding of the development and evolution of the

neocortex. Annu. Rev. Cell Dev. Biol. 30, 465–502.

Voigt, T. (1989). Development of glial cells in the cerebral wall of ferrets: direct

tracing of their transformation from radial glia into astrocytes. J. Comp. Neurol.

289, 74–88.

Wang, X., Tsai, J.W., LaMonica, B., and Kriegstein, A.R. (2011). A new subtype

of progenitor cell in the mouse embryonic neocortex. Nat. Neurosci. 14,

555–561.

Yamanaka, T., Horikoshi, Y., Sugiyama, Y., Ishiyama, C., Suzuki, A., Hirose, T.,

Iwamatsu, A., Shinohara, A., and Ohno, S. (2003). Mammalian Lgl forms a pro-

tein complex with PAR-6 and aPKC independently of PAR-3 to regulate epithe-

lial cell polarity. Curr. Biol. 13, 734–743.

Zhang, Y., Sloan, S.A., Clarke, L.E., Caneda, C., Plaza, C.A., Blumenthal, P.D.,

Vogel, H., Steinberg, G.K., Edwards, M.S., Li, G., et al. (2016). Purification and

characterization of progenitor andmature human astrocytes reveals transcrip-

tional and functional differences with mouse. Neuron 89, 37–53.

Zong, H., Espinosa, J.S., Su, H.H., Muzumdar, M.D., and Luo, L. (2005).

Mosaic analysis with double markers in mice. Cell 121, 479–492.

Neuron 94, 517–533, May 3, 2017 533

STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER

Antibodies

GFP – Chick Aves Labs Cat#GFP-1020; RRID: AB_10000240

RFP – Rabbit MBL Cat#PM005; RRID: AB_591279

tdTomato – Goat Sicgen Antibodies Cat#AB8181-200

GFAP – Rabbit Dako Cat#Z0334; RRID: AB_10013382

Ki67 – Rabbit Leica Microsystems Cat#NCL-Ki67p; RRID: AB_442102

BLBP – Rabbit Millipore Cat#AB9558; RRID: AB_2314014

S100b – Mouse Sigma-Aldrich Cat#S2532; RRID: AB_477499

Prominin1 – Rat Affymetrix eBioscience Cat#14-1331; RRID: AB_2314136

Satb2 – Mouse Abcam Cat#AB51502; RRID: AB_882455

Ctip2 – Rat Abcam Cat#AB18465; RRID: AB_2064130

Pax6 – Rabbit BioLegend Cat#901301; RRID: AB_2565003

Tbr1 – Rabbit Abcam Cat#AB31940; RRID: AB_2200219

N-cadherin (CDH2) – Mouse Thermo Fisher Scientific Cat#333900; RRID: AB_2313779

BrdU – Mouse BD Bioscience Cat#347580; RRID: AB_400326

Caspase-3 pAb – Rabbit Promega Cat#G7481; RRID: AB_430875

Glast (EAAT1) – Rabbit Abcam Cat#AB416; RRID: AB_304334

g-tubulin – Mouse Sigma-Aldrich Cat#T3559; RRID: AB_477575

b-catenin – Mouse BD Transduction labs Cat#610153; RRID: AB_397554

Pals1 – Rabbit Proteintech Cat#17710-1-AP; RRID: AB_2282012

DyLight 405 Anti-Mouse IgG Jackson ImmunoResearch Labs Cat#715-475-150; RRID: AB_2340839

Alexa Fluor 488 Anti-Chicken IgG Jackson ImmunoResearch Labs Cat#703-545-155; RRID: AB_2340375

Cy3 Anti-Goat IgG Jackson ImmunoResearch Labs Cat#705-165-147; RRID: AB_2307351

Alexa Fluor 647 Anti-Goat IgG Jackson ImmunoResearch Labs Cat#705-605-147; RRID: AB_2340437

Alexa Fluor 647 Anti-Mouse IgG Jackson ImmunoResearch Labs Cat#715-605-151; RRID: AB_2340863

Cy3 Anti-Rabbit IgG Jackson ImmunoResearch Labs Cat#711-165-152; RRID: AB_2307443

Alexa Fluor 647 Anti-Rabbit IgG Jackson ImmunoResearch Labs Cat#711-605-152; RRID: AB_2340624

Alexa Fluor 647 Anti-Rat IgG Jackson ImmunoResearch Labs Cat#712-605-153; RRID: AB_2340694

Experimental Models: Organisms/Strains

Mouse: MADM-11-GT The Jackson Laboratory RRID: IMSR_JAX:013749

Mouse: MADM-11-TG The Jackson Laboratory RRID: IMSR_JAX:013751

Mouse: Lgl1-Flox Klezovitch et al., 2004 MGI: 102682

Mouse: Egfr-Flox Lee and Threadgill, 2009 MGI: 3513096

Mouse: Emx1-Cre The Jackson Laboratory RRID: IMSR_JAX:005628

Mouse: Emx1-CreER The Jackson Laboratory RRID: IMSR_JAX:027784

Mouse: Hprt-Cre The Jackson Laboratory RRID: IMSR_JAX:004302

Software and Algorithms

IMARIS 7.7.1 Bitplane http://www.bitplane.com/imaris/imaris

MATLAB Mathworks https://www.mathworks.com/products/matlab/index.

html?s_tid=gn_loc_drop

ZEN Digital Imaging for Light Microscopy Zeiss https://www.zeiss.com/microscopy/en_us/products/

microscope-software/zen.html#introduction

Other

LSM 800 Confocal Zeiss N/A

LSM 880 with Airy Scan Confocal Zeiss N/A

Cryostat Cryostar NX70 Thermo Fisher Scientific N/A

e1 Neuron 94, 517–533.e1–e3, May 3, 2017

CONTACT FOR REAGENT AND RESOURCE SHARING

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Simon

Hippenmeyer (simon.hippenmeyer@ist.ac.at).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

See Table S1 for complete information regarding genotypes used in this study.

Mouse Lines and MaintenanceMouse protocols were reviewed by institutional preclinical core facility (PCF) at IST Austria and all breeding and experimentation was

performed under a license approved by the Austrian Federal Ministry of Science and Research in accordance with the Austrian and

EU animal laws.Mice weremaintained and housed in animal facilities with a 12 hr day/night cycle and adequate food andwater under

conditions according to IST Austria institutional regulations. Mouse lines with chromosome 11MADM cassettes (Hippenmeyer et al.,

2010), Lgl1-flox allele (Klezovitch et al., 2004), Egfr-flox (Lee and Threadgill, 2009), Emx1-Cre (Gorski et al., 2002), and Emx1-CreER

(Kessaris et al., 2006) have been described previously. For the generation of Lgl1-D allele, Lgl1-flox was crossed toHprt-Cre germline

delete (Tang et al., 2002) and double heterozygous Lgl1-flox; Hprt-Cre backcrossed to wild-type. All analyses were carried out in a

mixed C57/Bl6, CD1 genetic background. Littermates were randomly assigned to experimental groups based on genotype. No sex

specific differences were observed under any experimental conditions or in any genotype.

METHOD DETAILS

Preparation of MADM-Labeled TissueMice were deeply anesthetized through injection of a ketamine/xylazine/acepromazine solution (65 mg, 13 mg and 2 mg/kg body

weight, respectively), and confirmed to be unresponsive through pinching the paw. Perfusion was performed with ice-cold phos-

phate-buffered saline (PBS) followed immediately by 4% PFA prepared in PBS. Tissue was further fixed in 4% PFA over night to

ensure complete fixation. Brains were washed with PBS 4-5 times, and cryopreserved with 30% sucrose solution in PBS for approx-

imately 48 hr. Brains were then embedded in Tissue-Tek O.C.T. (Sakura) and for postnatal and adult time points, 30mm coronal

sections were collected in 24 multi-well dishes (Greiner Bio-one) and stored at�20�C in antifreeze solution (30% v/v ethyleneglycol,

30% v/v glycerol, 10% v/v 0.244M PO4 buffer) until use. For embryonic time points, 20mmcoronal sections were collected directly on

Superfrost glass slides (Thermo Fisher Scientific). For clonal analysis with P21 time points, coronal 45mm sections were collected

sequentially in individual wells of 24 multi-well dishes, and mounted onto glass slides while maintaining the order of sectioning.

Immunostaining of MADM-Labeled BrainsFor sections fixed to glass slides, tissue was first rehydrated by incubating for 15 min in PBS at room temperature. All following steps

apply to both tissue sections fixed to glass slides and floating sections in multi-well plates. Tissue sections were blocked for 45min in

a buffer solution containing 2% normal donkey serum (Life Technologies), 1% Triton X-100 in PBS. Primary antibodies were diluted

(see table) in blocking buffer and incubated overnight at 4�Con either a rocking table (floating sections) or a stationary platform (tissue

on slides). Sections were washed 4 times for 10 min each with PBS and incubated with corresponding secondary antibody diluted in

blocking buffer solution for 1 hr. Sections were washed 4 times with PBS, with the penultimate wash step containing the nuclear stain

DAPI (Invitrogen). For primary antibodies requiring antigen retrieval, tissue was first processed by incubating in sodium citrate buffer

(10mM Sodium Citrate, 0.05% Tween 20, pH 6.0) at 80�C for 10 min. For primary antibodies requiring DNA denaturation, sections

were first treated with 2N HCl at 37�C for 30 min. Once cooled to room temperature, sections were washed with PBS and standard

immunostaining protocol as described above was carried out. Floating sections were mounted on Superfrost glass slides (Thermo

Fisher Scientific), dehydrated and embedded in mounting medium containing 1,4-diazabicyclooctane (DABCO; Roth) and Mowiol

4-88 (Roth).

Imaging and Analysis of Marker Expression in MADM-Labeled BrainsSections were imaged using either an inverted LSM800 or LSM880 with airy scan confocal microscope (Zeiss) and processed using

Zeiss Zen Blue software. Tiled images, encompassing the entire region of interest, were taken for a minimum of three brain sections

per animal. Confocal images were imported into Photoshop software (Adobe) and the boundaries for the region of interest were

traced. MADM-labeled cells were manually counted based on respective marker expression. Quantification of cell clusters lining

the V-SVZ was performed by first measuring the average cell area of control-MADM cells in the V-SVZ. Next, the number of cells

in Lgl1�/� mutant cell clusters was estimated by measuring their total area and dividing by the average cell area. These estimates

were confirmed by manually counting DAPI labeled nuclei. Quantification of BrdU and Caspase-3 labeling was performed using a

40x oil objective.

Neuron 94, 517–533.e1–e3, May 3, 2017 e2

Generation of MADM Clones in the NeocortexMADM clone induction was performed as described previously (Gao et al., 2014; Hippenmeyer et al., 2010). In brief, timed pregnant

females injected intraperitoneally with tamoxifen (TM) (Sigma) dissolved in corn oil (Sigma) at E11 or E12 at a dose of 2-3mg/pregnant

dam. Live embryos were recovered at E18–E19 through caesarean section, fostered, and raised for further analysis at P21. For

embryonic time point analysis, caesarean section and analysis was performed at either E13 or E16. Brains containing MADM clones

were isolated, and tissue sections processed as described above. 3D reconstruction was performed by using a custom MATLAB

script and IMARIS-based imaging analysis platform. Cortical areas were identified by using the Allen Brain Atlas (http://mouse.

brain-map.org/static/atlas).

Astrocyte Morphology Filament TracingThe Sholl method of concentric spheres was used tomeasure astrocyte branching complexity (Sholl, 1953). First, brain sectionswere

stained for GFAP, labeling the main processes of the majority of cortical astrocytes. Then individual GFP+ layer II/III astrocytes from

control-MADM and Lgl1-MADM mice were imaged using a 63x oil objective. 3D reconstruction and analysis was performed using

IMARIS software. Briefly, using the Filament Tracer algorithm, the GFAP labeled processes were 3D reconstructed, concentric

sphere Sholl analysis was performed, and total cell volume and cell body volume was calculated from the 3D structure.

QUANTIFICATION AND STATISTICAL ANALYSIS

See Table S2 for complete information regarding quantifications and statistics used in this study. This table includes all graphed

values, including SEMs, p values, and exact values of n. For all graphs except Figures S5L, S5N, and S8M, statistical analysis

was performed in Excel. Values represent mean ± SEM and significance was determined using the two-tailed unpaired Student’s

t test. n was defined as an individual section green/red ratio quantification. For Figures S5L, S5N, and S8M, statistical analysis

was performed in Graphpad Prism 7.0. For Figures S5L and S5N, multiple t tests were performed and discovery determined using

the two-stage linear step-up procedure of Benjamini, Krieger and Yekutieli, with Q = 1%. Each rowwas analyzed individually, without

assuming a consistent SD n was defined as an individual reconstructed astrocyte. For Figure S8M, significance was determined

using One-way Anova and corrected with Tukey’s multiple comparisons test, with n being defined as one single section green/

red ratio quantification. By plotting the raw data we were able to determine that our data met the criteria of normal distribution. Sig-

nificance was established at *p < 0.05, **p < 0.01, ***p < 0.001.

e3 Neuron 94, 517–533.e1–e3, May 3, 2017