Identification of Embryonic Neural Plate Border Stem Cells and Their Generation by Direct Reprogramming from Adult Human Blood Cells

Article

Identification of Embryonic Neural Plate Border Stem Cells and Their Generation by Direct Reprogramming from Adult Human Blood Cells Graphical Abstract Reprogramming into “induced Neural Plate Border Stem Cells“

Authors Multipotent self-renewing (i)NBSCs in mouse and human

ar

ve

on

ys

tc

ou

ec im

CRISPR Cas9 Genome editing

Chir Alk.5-Inh. Purmorphamine Feeder

Chir ALK5-Inh. Purmorphamine Tranylcypromine

Differentiation

(i)NBSCs low density Chir BMP4 Chir FGF8 IGF1 DAPT

Development

Chir SB Purmorphamine

[email protected] (M.C.T.), [email protected] (A.T.)

E8.5 mouse embryo

Pr

ion

rs

ADFs

Neural Crest

bFGF EGF LIF

RG-like stem cells

+ TetOP

Correspondence

rce

Dir

PBMCs

FPFs

Marc Christian Thier, Oliver Hommerding, Jasper Panten, ..., Hannah Monyer, Frank Edenhofer, Andreas Trumpp

BRN2 2A KLF4 IRES SOX2 2A

ZIC3

LoxP

Reprogramming

Chir ALK5-Inh. Purmorphamine Feeder induced Neural Plate Border Stem Cells (iNBSCs)

Mature neurons and glia

Sensory neurons and mesenchymal crest cells

Highlights d

Four factors directly reprogram human somatic cells into selfrenewing iNBSCs

d

Multipotent iNBSCs can generate CNS and neural crest progeny

d

NBSCs can also be obtained by iPSC differentiation or from E8.5 mouse neural folds

d

Expandable iNBSCs can be easily modified via CRISPR/Cas9 to model neural diseases

Thier et al., 2019, Cell Stem Cell 24, 166–182 January 3, 2019 ª 2018 Elsevier Inc. https://doi.org/10.1016/j.stem.2018.11.015

In Brief Thier, Trumpp, et al. demonstrate how to reprogram human somatic cells into induced neural plate border stem cells (iNBSCs). Self-renewing iNBSCs show multilineage differentiation toward CNS and neural crest and are modifiable by CRISPR/Cas9 technology. NBSCs can also be established from mouse embryos and might have a direct counterpart during development.

Cell Stem Cell

Article Identification of Embryonic Neural Plate Border Stem Cells and Their Generation by Direct Reprogramming from Adult Human Blood Cells Marc Christian Thier,1,2,* Oliver Hommerding,5,15 Jasper Panten,1,2,15 Roberta Pinna,4 Diego Garcı´a-Gonza´lez,4 Thomas Berger,1,2 Philipp Wo¨rsdo¨rfer,6 Yassen Assenov,7 Roberta Scognamiglio,1,2 Adriana Przybylla,1,2 Paul Kaschutnig,1,8 Lisa Becker,1,2,3 Michael D. Milsom,1,8 Anna Jauch,9 Jochen Utikal,10,11 Carl Herrmann,12,13 Hannah Monyer,4 Frank Edenhofer,5,14,16 and Andreas Trumpp1,2,3,16,17,* 1Heidelberg Institute for Stem Cell Technology and Experimental Medicine (HI-STEM gGmbH), Im Neuenheimer Feld 280, 69120 Heidelberg, Germany 2Division of Stem Cells and Cancer, German Cancer Research Center (DKFZ), and DKFZ-ZMBH Alliance, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany 3German Cancer Consortium (DKTK), 69120 Heidelberg, Germany 4Department of Clinical Neurobiology, Medical Faculty of Heidelberg University and German Cancer Research Center (DKFZ), 69120 Heidelberg, Germany 5Stem Cell Engineering Group, Institute of Reconstructive Neurobiology, Universita €t Bonn Life and Brain Center and Hertie Foundation, Sigmund-Freud Strasse 25, 53105 Bonn, Germany 6Stem Cell and Regenerative Medicine Group, Institute of Anatomy and Cell Biology, Julius-Maximilians-Universita €t Wu €rzburg, €rzburg, Germany Koellikerstrasse 6, 97070 Wu 7Division of Epigenomics and Cancer Risk Factors, German Cancer Research Center (DKFZ), Heidelberg, Germany 8Division of Experimental Hematology, German Cancer Research Center (DKFZ) and DKFZ-ZMBH Alliance, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany 9Institute of Human Genetics, University Hospital of Heidelberg, Heidelberg, Germany 10Skin Cancer Unit, German Cancer Research Center (DKFZ), Heidelberg, Germany 11Department of Dermatology, Venereology, and Allergology, University Medical Center Mannheim, Ruprecht-Karl University of Heidelberg, Mannheim, Germany 12Division of Theoretical Bioinformatics, German Cancer Research Center (DKFZ), 69120 Heidelberg, Germany 13Health Data Science Unit and Bioquant Center, Medical Faculty of Heidelberg University, 69120 Heidelberg, Germany 14Leopold-Franzens-University Innsbruck, Institute of Molecular Biology & CMBI, Department Genomics, Stem Cell Biology & Regenerative Medicine, 6020 Innsbruck, Austria 15These authors contributed equally 16Senior author 17Lead Contact *Correspondence: [email protected] (M.C.T.), [email protected] (A.T.) https://doi.org/10.1016/j.stem.2018.11.015

SUMMARY

We report the direct reprogramming of both adult human fibroblasts and blood cells into induced neural plate border stem cells (iNBSCs) by ectopic expression of four neural transcription factors. Self-renewing, clonal iNBSCs can be robustly expanded in defined media while retaining multilineage differentiation potential. They generate functional cell types of neural crest and CNS lineages and could be used to model a human pain syndrome via gene editing of SCN9A in iNBSCs. NBSCs can also be derived from human pluripotent stem cells and share functional and molecular features with NBSCs isolated from embryonic day 8.5 (E8.5) mouse neural folds. Single-cell RNA sequencing identified the anterior hindbrain as the origin of mouse NBSCs, with human iNBSCs sharing a similar regional identity. In summary, we identify embryonic NBSCs and report their generation by direct reprogramming in human, which 166 Cell Stem Cell 24, 166–182, January 3, 2019 ª 2018 Elsevier Inc.

may facilitate insights into neural development and provide a neural stem cell source for applications in regenerative medicine.

INTRODUCTION Various types of neural stem and progenitor cells (NSPCs) can be derived by directed differentiation from mouse and human pluripotent stem cells (PSCs) as well as fetal and adult brain tissue (Conti and Cattaneo, 2010). However, the generation of induced PSC (iPSC)-derived NSPCs suffers from lengthy reprogramming protocols and difficult validation of individual NSPC lines and the concurrent risk of propagating tumor-prone pluripotent remnants. Direct conversion of somatic cells into specific neural cells, especially mature neurons, has become feasible during recent years (Pfisterer et al., 2011; Vierbuchen et al., 2010). Nevertheless, direct reprogramming into expandable neural progenitors remains a major challenge. Although we and others have previously described the direct conversion from mouse embryonic

A TetOP

BRN2 2A KLF4 IRES SOX2 2A

ZIC3

OR

TetOP

BRN2 2A KLF4 IRES SOX2 2A

ZIC3

LoxP

PAX6

B

BKSZ

PBMCs

SOX1

ADFs

Doxycycline + CAPT

CAP +/- CreRecombinase

days 0

ADF C5

24

Conversion

FPFs

C

12

D

P17

DAPI

SOX1

PAX6

MSX1

DAPI

NESTIN

SOX2

merge

DAPI

ZIC1

PAX6

merge

merge

CAP

E

****

Clonogenicity of parent [%]

40 30 20 10 0

F

CD133- CD133+ CD133- CD133+ CXCR4- CXCR4- CXCR4+ CXCR4+

G

40

Biological Process (GO) pathway ID GO:0010721 GO:0097150 GO:0021915 GO:2000179

Stem Cell Identity

20

PC2

0 ADF-derived iNBSCs Blood-derived iNBSCs iPSC-derived NBSCs hESCs (GSE34904) iPSCs ADFs

-20 -40

CNS related pathways

GO:0007399 GO:0007420 GO:0007417

nervous system development brain development central nervous system development

53 22 24

NES, FGFR3, ZIC3 PAX6, ZIC2, FABP7 POU3F2, NHLH2, PTPRZ1

7.04E-11 2.47E-05 1.01E-04

Crest related pathways

GO:0007423 GO:0060322 GO:0043583

sensory organ development head development ear development

20 22 10

ASCL1, LEF1, CHD7 HOXB2, FZD3, SOX3 VANGL2, FGFR3, FRZB

1.13E-05 6.25E-05 1.78E-03

KEGG Pathways 4110 4330 4310

pathway description Cell cycle Notch signaling pathway Wnt signaling pathway

-60 Signaling

-80 -100

-50

0

50

100

pathway description count in gene set e.g. false discovery rate negative regulation of cell development 14 LIN28A, SOX11, DLL1 3.19E-05 neuronal stem cell population maintenance 4 HES5, SOX2, NOTCH1 4.05E-03 neural tube development 7 SALL4, ZIC2, PAX6 3.41E-02 positive regulation of neural precursor cell proliferation 7 INSM1, NOTCH1, FZD3 6.21E-05

150

count in gene set e.g. false discovery rate 10 CDC7, E2F2, MCM2 4.37E-05 5 HES5, NOTCH1,DLL3 9.27E-03 7 CCND2, LEF1, FZD9 2.14E-02

PC1

H

2

4

6

GO_REGIONALIZATION

75

GO_EMBRYONIC_ORGAN_DEVELOPMENT GO_EMBRYONIC_SKELETAL_SYSTEM_DEVELOPMENT

50

GO_CELL_FATE_COMMITMENT

ADF-derived iNBSCs iPSC-derived NBSCs ESCs (GSE61461) ADFs ADFs (GSE52025)

25

X2

THY1

0 GO_PATTERN_SPECIFICATION_PROCESS

PAX6 CLDN5 ASCL1 POU3F2 HES5 MSX1 HOXB2 IRX3 COL6A3 GREM1 TGFBI

NES

J

I

0

GO_EMBRYONIC_MORPHOGENESIS GO_EMBRYONIC_SKELETAL_SYSTEM_MORPHOGENESIS GO_NEURON_FATE_SPECIFICATION GO_ANTERIOR_POSTERIOR_PATTERN_SPECIFICATION GO_SKELETAL_SYSTEM_DEVELOPMENT

−25

GO_EMBRYONIC_ORGAN_MORPHOGENESIS

POU5F1

iPSC C2

iPSC C4

iPSC C11

ADF2 C6 iNBSC

GSE34904_ESC_H1_2

ADF2 C2 iNBSC

GSE34904_ESC_H1_1

ADF2 C5 iNBSC

PBMC1 C4 iNBSC

iPSC C2 NBSC

PBMC1 C3 iNBSC

PBMC1 C17 iNBSC

iPSC C4 NBSC

GSE69486_FIB

iPSC C11 NBSC

ADF2

ADF1

GSE51980_FIB_y

ZFP42 CDH1 EPCAM

GO_CELL_FATE_SPECIFICATION

−50

GO_LIMB_DEVELOPMENT

−50

2 1 0 -1 -2

0

X1

50 GO_NEURON_FATE_COMMITMENT

(legend on next page)

Cell Stem Cell 24, 166–182, January 3, 2019 167

fibroblasts into neural progenitor cells (Han et al., 2012; Lujan et al., 2012; Thier et al., 2012), direct reprogramming toward defined and expandable neural progenitors in humans has remained difficult. Typically attempted by reprogramming using pluripotency factors, it has mostly resulted in heterogeneous cell types with an unclear developmental stage and unknown equivalent in vivo (Cairns et al., 2016; Hou et al., 2017; Lee et al., 2015). In this study, we present a robust strategy to reprogram human adult somatic cells into early, defined, and long-term selfrenewing induced neural plate border stem cells (iNBSCs). They can be expanded in culture and show broad neural crest and neural lineage differentiation potential and resemble cells present in the anterior hindbrain region of neurulation-stage embryos. RESULTS The formation of the nervous system initiates with the neural plate stage shortly after gastrulation. Signaling pathways, such as WNTs, bone morphogenetic proteins (BMPs), and sonic hedgehog (SHH), orchestrate the diversification of neural committed cells, which underlie the development of the various brain regions, spinal cord, as well as the neural crest (Garnett et al., 2012; Le Dre´au and Martı´, 2012). We hypothesized that overexpression of stage-specific transcription factors, in combination with signaling cues provided by the growth medium, might allow for the direct reprogramming of adult somatic cells to early embryonic neural progenitors with stem cell features, including self-renewal and multipotency. We therefore treated human adult dermal fibroblasts (ADFs) with various combinations of transcription factors (BRN2, SOX2, KLF4, MYC, TLX, and ZIC3) and small molecules that we hypothesized would allow the generation of early neural stages (Table S6). This approach led to the discovery that neural reprogramming could be facilitated by the combination of the four transcription factors BRN2, KLF4, SOX2, and ZIC3 (BKSZ), with the four small-molecule inhibitors Chir99021 (GSK-3 inhibitor), Alk5 inhibitor II, purmorphamine (hedgehogsmoothened agonist), and tranylcypromine (inhibitor of monoamine oxidase [MAO] and CYP2 enzymes: A6; C19; and D6) (CAPT). ADFs were transduced with a polycistronic, doxycycline (DOX)-inducible lentiviral vector containing BKSZ and subse-

quently cultured in the presence of CAPT. We also derived a modified lentiviral vector containing a loxP-flanked expression cassette (Figure 1A) to allow subsequent Cre-mediated transgene removal. At day 14 post-transduction, colony formation was observed. Upon DOX withdrawal, the colonies continued to grow and expressed the early neural markers PAX6 and SOX1 (Figure 1B). Individual colonies were clonally expanded, resulting in stable lines 19–24 days after transduction (Figure 1C). The minimum time period of BKSZ expression required for reprogramming in the presence of CAPT was 12 days, and the efficiency of reprogramming further increased with prolonged DOX application (Figure S1A). Importantly, overexpression of BKSZ and growth without small molecules or treatment with CAPT alone did not result in colony formation (Figure S1B). BKSZmediated reprogramming did not seem to pass through a functional pluripotent intermediate state because induction of the transgenes did not result in a significant increase of OCT4 expression (Figure S1C), and iPSCs could not be derived by BKSZ reprogramming of ADFs, even after switching to pluripotent stem cell medium (data not shown). We were also able to successfully convert human fetal pancreas fibroblasts (FPFs) and peripheral blood mononuclear cells (PBMCs) and established more than 30 stable neural progenitor lines (Figure S1D). In concordance with earlier reports, the conversion efficiency varied with the degree of maturity of the cell of origin, being highest for the FPFs (0.166%) and lowest for PBMCs (0.015%; Eminli et al., 2009; Figure S1E). Importantly, once stable lines were established, the cell type of origin had no apparent impact on proliferation, neural marker expression, or differentiation capacity (Figures S1D, S1F, and S1I). Stable neural lines with sustained >1,000-fold downregulation of the polycistronic transgene cassette after DOX withdrawal could be established (Figure S1H). To rule out potential effects of Tet-on-system-intrinsic leakiness, we also derived cell lines containing the LoxP-modified version of the BKSZ vector (Figure 1A). Next, we excised the vector using Cre-recombinase (Scognamiglio et al., 2016) and established subclones that showed neither transgene expression (Figure S1G) nor detectable vector integration on the genomic level (Figure S1H). Taken together, these results show that the derived neural clones are transgene independent.

Figure 1. Direct Conversion of Somatic Cells into Neural Plate Border Stem Cells (A) Reprogramming vectors used for neural conversion of somatic cells. (B) A schematic overview of the conversion of somatic cells into neural progenitors. Scale bars, 50 mm. (C) Bright field (BF) image of a representative neural progenitor line derived from ADFs; scale bar, 100 mm. (D) Representative confocal images of an ADF-derived neural progenitor line staining positive for neural plate border and neural stem cell markers; arrowheads mark double- and triple-positive cells, respectively; scale bars, 50 mm. (E) Clonogenicity of iNBSCs in respect to surface marker expression of CD133/2 and CXCR4 in the moment of single-cell sorting. 6 independent iNBSC cell lines were sorted in two technical replicates each. Numbers indicate relative values of parent gate (t test; ****p < 0.0001; mean with SEM). (F) Microarray-based transcriptional profiling of ADF-, blood-, and iPSC-derived (i)NBSCs, PSCs, and ADFs. Principal-component analysis shows three clearly distinct clusters, representing (i)NBSCs (ADF-, blood-, and iPSC-derived) and its somatic cell (ADF) and PSC (hESCs or iPSCs) precursor, respectively. (G) Gene Ontology (GO) processes based on 200 top upregulated genes in iNBSC lines compared to ADFs. (H) Expression heatmap of (i)NBSCs, ADFs, and PSCs based on the 200 probes showing the highest contribution for the segregation of samples in the principalcomponent analysis. Specific markers for iNBSCs, ADFs, and PSCs are highlighted. (I) Array-based methylation profiling of ADF-derived iNBSCs, PSCs, and ADFs. Multidimensional scaling plot reveals distinct clusters for each population. (J) Gene set enrichment analysis of hypomethylated promoter sites of iNBSCs compared to ADFs. Terms are ranked by normalized enrichment score (NES). See also Figure S1 and Tables S1 and S2.

168 Cell Stem Cell 24, 166–182, January 3, 2019

Converted clonal neural cell lines derived from ADFs, FPFs, and PBMCs could be cultured on a layer of supportive feeders and in medium containing Chir99021, Alk5 inhibitor II, and purmorphamine (CAP) for more than 40 passages (>7 months), without loss of proliferative potential. Analysis of SNPs confirmed that clonal lines were indeed derived from PBMCs and ADFs of the respective donor (data not shown). Expanded iNBSCs maintained high expression of early neural markers, such as PAX6 and SOX1 (Figures 1D and S1I). Moreover, the cultures were homogenously positive for the stem cell markers NESTIN and SOX2, showed high levels of CD133/2 and CXCR4, and also expressed MSX1, ZIC1, and PAX3, suggesting a neural plate border-like identity of the converted cells (Figures 1D, S1F, S1I, and S1J). In concordance with their sustained selfrenewal capacity and expression of neural plate border markers, we named these cells iNBSCs. To formally prove that the obtained neural cell lines were indeed multipotent at the clonal level, we performed index single-cell sorting of six independent iNBSC lines (951 cells in total). Indeed, the iNBSC lines were highly clonogenic (mean 20.64% ± SD = 5.66%), and we could establish more than 100 single-cellderived subclones that showed similar proliferation, stem cell marker expression, and differentiation capacity as compared to their parental cell lines (Figures S1J and S3G). Interestingly, retrospective analysis of the surface marker expression for CD133/2 and CXCR4 revealed that subclones predominantly originated from CD133/2+CXCR4+ cells, and single-positive and double-negative iNBSCs gave rise to only very few clones (Figure 1E). This result is in line with their known role as neural progenitor markers (Yuan et al., 2011) and confirms the high frequency of self-renewing neural progenitors within the iNBSC lines before and after single-cell sorting (Figure S1J). Next, we performed bead-chip-based comparative global gene expression analysis of iNBSCs, iPSCs, and the cell type of origin. Principal-component analysis revealed that each cell type clustered distinctly in separate expression clusters (Figures 1F and S1K). All iNBSC clones clustered closely, indicating that the cell of origin had no major impact on the iNBSC identity. Next, we explored whether NBSCs can also be obtained by directed differentiation from iPSCs, as this would exclude an artificial state that occurs only as an outcome of the reprogramming process. To this end, we analyzed neural plate border marker expression in clonal lines from differentiated iPSCs (see STAR Methods for details) and found that it was indeed similar to that of ADF- and PBMC-derived iNBSCs (Figures S1F and S1L). Moreover, as the expression profile of iPSCderived clones also clustered with that of iNBSCs, we conclude that the NBSC state can be established also by directed in vitro differentiation from pluripotent stem cells (Figures 1F and S1K). Comparison of the expression profiles of iNBSCs with that of ADFs revealed 10,917 differentially expressed genes (DEGs) (p < 0.05), reflecting the robust change of cellular identity (Table S1). Gene ontology (GO) analysis of the top 200 upregulated genes in iNBSCs unveiled an enrichment for stem cell-related processes, such as ‘‘neuronal stem cell population maintenance’’ and ‘‘positive regulation of neural precursor cell proliferation,’’ as well as processes related to CNS and neural crest (NC) identity, such as ‘‘brain development’’ and ‘‘head development,’’ respectively (Figure 1G; Table S1). In agreement with

their proliferative nature, the ‘‘cell cycle’’ and ‘‘Notch signaling pathway’’ were also enriched in iNBSCs compared to ADFs (Figure 1G; Table S1). Numerous neural genes were among the genes that contributed strongest to the segregation of iNBSCs from ADFs and pluripotent stem cells (Figure 1H). These include the neural stem cell markers NESTIN, PAX6, and HES5, as well as neural plate border and neural crest markers, including MSX1, TFAP2a, SOX3, and HOXB2 (Figures 1H and S1M). In contrast, pluripotency markers (OCT4/POU5F1 and NANOG) were not upregulated (Figures 1H and S1M). Interestingly, although most mesoderm and fibroblast markers, such as BRACHYURY/ T and THY1, were strongly downregulated, some residual expression of COL3A1 was detected in ADFderived, but not PBMC-derived, iNBSCs, possibly indicative of residual epigenetic memory (Figure S1M). To further investigate the regional identity of iNBSCs, the expression of region-specific transcription factors was analyzed. qPCRs demonstrated expression of anterior hindbrain markers GBX2, IRX3, HOXA2, and HOXB2, and there was either no or low expression of forebrain markers (EMX1 and FOXG1) and the anterior midbrain marker OTX2 (Figure S1O). Along the dorsoventral axis, iNBSCs expressed dorsal markers, such as MSX1, PAX3, and PAX7 (Figures S1F and S1O). Therefore, the regional identity of iNBSCs is most compatible with a dorsal anterior hindbrain fate. In contrast to the transcriptome, the DNA methylation pattern of cells is rather stably associated with the identity and fate of the cell and independent of the cellular state. Comparison of the DNA methylation profiles of ADF-converted iNBSCs and iPSC-derived NBSCs with ADFs and human embryonic stem cells (hESCs) by 450K bead chip arrays revealed three distinct clusters (Figure 1I). ADF-converted iNBSCs and iPSC-differentiated NBSCs (thereafter referred to as (i)NBSCs) clustered closely, confirming high similarity also at the level of their methylome. Although global methylation of CpG islands remained at a similar range between all groups, the direct comparison of ADFs with ADF-derived and iPSC-differentiated (i)NBSCs resulted in 9,067 differentially methylated promoter regions (false discovery rate [FDR]-adjusted p value < 0.05). Gene set enrichment analysis (GSEA) of hypomethylated promoter sites of iNBSCs compared to ADFs revealed enrichments for GO terms, such as ‘‘pattern specification,’’ ‘‘skeletal system development,’’ and ‘‘neuron fate commitment’’ (Figure 1J; Table S2). These data are consistent with the reprogramming into an NBSC identity also at the epigenetic level. The analysis of hypermethylated promoter sites revealed processes related to a fibroblast identity, such as ‘‘cell-cell adhesion via plasma membrane adhesion molecules,’’ ‘‘calcium-dependent cell-cell adhesion,’’ ‘‘innate immune response,’’ and ‘‘collagen fibril organization’’ (Figure S1N; Table S2), which apparently became silenced. Together, these results demonstrate robust changes in the methylation and transcriptional landscape after reprogramming of ADFs into iNBSCs. In summary, we present a robust strategy to directly convert various adult human cell types into iNBSCs that are characterized by high expression of neural plate border markers, high clonogenicity, and sustained self-renewal capacity and the acquisition of an expression and epigenomic landscape consistent with a neuralplate-border-like phenotype. Cell Stem Cell 24, 166–182, January 3, 2019 169

A

B

D

C

E

F

H

G

I

J

K

Figure 2. iNBSCs Recapitulate Early Neural Development and Give Rise to CNS and Neural Crest Progenitors (A) Directed differentiation of iNBSCs toward neural crest (NC) fate. The plot shows representative cytometry data after differentiation of iNBSCs toward NCSClike cells. (B) Confocal images of stainings for neural crest markers on SSEA-1neg/CD133neg/P75+/HNK1+-sorted cells. (C) Top 20 upregulated DEGs between SSEA-1neg/CD133neg/P75+/HNK1+-sorted NCSC-like cells and iNBSCs. (D) Derivation of RG-like stem cells from iNBSCs. FACS plot shows a representative differentiation prior to FACS sorting. (E) Immunofluorescence stainings for radial glia markers on iNBSC-derived RG-like SCs; scale bars, 50 mm. (F) mRNA expression of radial-glia-associated markers in independent RG-like SCs (n = 3) and iNBSCs (n = 3; t test; *p < 0.05; mean with SEM). (G) Temporal identity of RG-like SCs and iNBSCs. Expression data of iNBSCs and RG-like stem cells (CD133+/SSEA1+/GLAST+) were analyzed by machine learning framework (CoNTExT) to match data with the transcriptome atlas of the developing human brain. (H) The transcriptional landscape of iNBSCs and their progeny. Principal-component analysis (PCA) of iNBSCs, NCSC-like cells (SSEA-1neg/CD133neg/P75+/ HNK1+), and RG-like SCs (CD133+/SSEA1+/GLAST+) reveals three distinct clusters with increasing distance to iNBSCs in the process of differentiation. (legend continued on next page)

170 Cell Stem Cell 24, 166–182, January 3, 2019

iNBSCs Give Rise to CNS and Neural Crest Progenitors To examine whether iNBSCs represent a developmental stage prior to the separation into the CNS and neural crest lineage, we aimed to derive lineage-committed CNS and neural crest progenitors by specific modulation of signaling cues. To first derive neural crest progeny, we sought to mimic the conditions prevailing at the neural plate border during neural crest specification, where WNT and BMP signaling have been shown to play pivotal roles (Garnett et al., 2012; Villanueva et al., 2002). Therefore, we cultured iNBSCs for three days in neural crestpriming conditions, comprising Chir99021, Alk-5 inhibitor, and BMP4 (CAB). Thereafter, cultures were switched to neural crest stem cell (NCSC) medium, containing Chir99021, fibroblast growth factor 8 (FGF8), insulin growth factor 1 (IGF1), and N-[N-(3,5-Difluorophenacetyl)-L-alanyl]-S-phenylglycine t-butyl ester (DAPT) known to provide cues for induction and maintenance of neural crest progenitors (Noisa et al., 2014; Shao et al., 2015). Robust induction of migratory crest markers P75 and HNK1 were observed to parallel a decrease in CD133 and SSEA1 expression levels (Figures 2A and S2A). Immunofluorescence and qPCR analysis of SSEA-1neg/CD133neg/P75+/HNK1+ sorted cells revealed expression of SOX10, P75, and HNK1 triple-positive cells, but not of PAX6, which is indicative of an NCSC-like identity (Figures 2B and S2B). Furthermore, comparison of expression profiles of iNBSCs and SSEA-1neg/CD133neg/ P75+/HNK1+ NCSC-like cells revealed migratory crest and neural crest stem cell markers, such as KANK4, BGN, TFAP2A, and SOX10, to be among the top regulated genes, and neural progenitor markers, such as HES5 and PAX6, were robustly downregulated (Figures 2C and S2C; Table S1). GO analysis of the top 100 upregulated differentially expressed genes in NCSC-like cells versus iNBSCs confirmed this finding and pointed to biological processes, such as ‘‘neural crest differentiation,’’ further indicating a robust development of iNBSC identity toward neural crest lineage (Figure S2D). Neural crest-primed cells could also be differentiated to cells with a CD105+/CD44+/CD13+/CD90+/ CD146+ phenotype resembling mesenchymal stem cells (MSCs) (Figure S2E; Nery et al., 2013). The analysis of gene expression profiles of MSC-like cells compared to iNBSCs demonstrated an upregulation of genes involved in maintenance and differentiation of MSCs, such as COL1A1, GREMLIN1, and TAGLN, and was further supported by GO analysis resulting in enrichment for biological processes, such as ‘‘extracellular matrix organization’’ and ‘‘skeletal system development’’ (Figures S2F and S2G; Table S1). Taken together, these data demonstrate that neural crest-primed differentiation of iNBSCs recapitulates physiological stages of neural crest development and that corresponding progenitors can be enriched when applying specific culture conditions. Second, to explore whether iNBSCs would also give rise to a defined CNS progenitor, we aimed to derive radial glia-like stem cells (Gorris et al., 2015). To this end, we primed iNBSCs toward the CNS by culturing the cells in the presence of

Chir99021, SB431542 (ALK 4, 5, and 7 inhibitor) and purmorphamine (CSP) for 5 days, followed by addition of basic fibroblast growth factor (bFGF) for one week. Subsequently, cultures were allowed to mature for 7 weeks in absence of small molecules before changing to bFGF-, epidermal growth factor (EGF)-, and Leukemia inhibiting factor (LIF)-supplemented expansion medium. In doing so, we could obtain a sub-population of SSEA1+, CD133+, and glutamate-aspartate-transporter (GLAST; SLC1A3)-positive cells, i.e., a phenotype associated with radial glia-like stem cell (RG-like SC) identity (Gorris et al., 2015; Figures 2D and S2H). Cells, triple-positive for SSEA1+, CD133+, and GLAST+, strongly expressed both the glial markers VIMENTIN, GFAP, and GLAST and the stem cell markers PAX6, NESTIN, SOX1, and BLBP (Figures 2E, 2F, and S2I). To study the developmental stage of RG-like stem cells compared to their parental iNBSCs, we made use of a recently described machine learning framework (CoNTExT), which was developed to match in vitro derived neural cultures with the spatiotemporal transcriptome atlas of the human brain (Stein et al., 2014). This analysis revealed a clear maturation from iNBSCs to RG-like SCs. Although iNBSCs mapped to embryonic and early fetal stages, our RG-like SCs mapped more toward the later developmental stages of early to late fetal and childhood (Figure 2G). This is also reflected in the principal-component analysis when combining expression data of iNBSCs and their progeny, showing three clearly distinct clusters for iNBSCs, RG-like SCs, and NCSCs (Figure 2H). Taken together, these data indicate that the iNBSCs display multipotent neural progenitor activity, prior to the CNS versus neural crest lineage commitment branch point, and allows the derivation and enrichment of restricted progenitors that correspond to previously defined developmental stages in vitro. To confirm that iNBSCs are a defined multipotent population, rather than a mixture of different types of committed progenitors, we examined iNBSC heterogeneity by performing single-cell RNA sequencing analysis. We single-cell-sorted and sequenced 310 cells derived from six iNBSC lines that were reprogrammed from ADFs or PBMCs (3 clones each). Principal-component analysis (PCA) revealed that most cells clustered closely together, independent of their clonal origin, and that only a small fraction of cells, derived from all clones, showed an increasingly larger distance to the main cluster. GO analysis of the genes that showed the strongest contribution to PC1 revealed that positively contributing genes (directing toward the right side of the plot) were enriched for terms such as ‘‘mitotic cell cycle,’’ ‘‘cell cycle,’’ and ‘‘cell cycle process,’’ and genes contributing negatively (directing toward the left side of the plot) were enriched for terms such as ‘‘nervous system development,’’ ‘‘generation of neurons,’’ and ‘‘neurogenesis’’ (Figures 2I and S2J). This observation indicated that, although most cells appeared to be in a proliferative stem cell state, some cells undergo spontaneous differentiation. To further investigate the nature of these cells, we analyzed the genes contributing to this differentiated

(I) Single-cell RNA sequencing of iNBSCs. PCA of 310 single iNBSCs, derived from 6 clonal lines is shown: three blood- (red) and ADF-derived (blue), respectively. (J) Overlay of pseudo-time prediction on PCA of single-cell RNA sequencing data of 6 iNBSC lines. (K) Plotting of specific stem-cell- and differentiation-related genes over the course of pseudo-time. See also Figure S2 and Table S1.

Cell Stem Cell 24, 166–182, January 3, 2019 171

Regionalization of progenitors (1 week)

B

Maturation (5-8 weeks)

*

*

40

Motoneuron Differentiation Purmorphamin + RA

iNBSCs

20

undirected CNS Differentiation Purmorphamin

TUJ1+ CHAT+

FOXA2+ IAP+ NURR1+ TH+ Frequence of singlets [%]

Dopaminergic Differentiation Purmorphamin + FGF8

Frequence of singlets [%]

A

20 15 10

0

TUJ1/vGLUT2

CHAT/ HB9

FOXA2/TH/TUJ1

iNBSCs

GAD67/GABA

5 0

Dopa Diff

iNBSCs

Moto Diff

Frequence of singletss [%]

TUJ1+ GATA3+ TPH2+ 10

GATA3/TPH2/TUJ1

TUJ1/SYN1

GFAP/S100ß

MBP/Olig2

5

0

iNBSCs

Sero Diff

C

D 40 mV 200 ms

20 pA 500 ms

E

F

GFP/GFAP

G

SOX1/NESTIN

CNS priming

40 mV

8 days

Chir; SB; Purmorphamine; bFGF (4 days)

200 ms

iNBSCs Tx into striatum

SB; Purmorphamine (9 days)

of NSG mice

GFP/NeuN

SOX10/AP2a/PAX6

GFP/MBP

single cell suspension of iNBSCs in MG

H

Chir; Alk5-Inh.; BMP4; bFGF

neural crest priming

100 μm

J

BRN3a/Peripherin

Maturation (5-8 weeks)

L

Differentiation into sensory neurons

CRISPR Cas9 Genome editing

Fluo-3 AM + α,β-me-ATP

Calcium flux measurements

Activity?

O DAPI

BRN3A

Peripherin

MERGE

Relative Fluorescence (F/F0)

N

WT WT + A-317491 SCN9A-/-

1.2 1.0 120

240

360

Time [s]

iNBSCs

SCN9A-/-

WT

P

α,β-me-ATP (30 μM)

1.6 1.4

Oil Red

iNBSC-derived sensory neurons

M SCN9A-/Sensory neurons

SCN9A-/iNBSCs

iNBSCs

Alcian Blue

SMA

K

Mesenchymal Crest Differentiation 10% FCS or MSC Medium

Nav1.7

Chir iNBSCs BMP4 (1 week)

ACTIN

low density

Sensory Neuron Differentiation Chir + DAPT + Su

480

Max. fluoreescence intesity relative to baseline

I

*

2.0

*

1.5

1.0 WT

WT + A-317491

SCN9A-/-

(legend on next page)

172 Cell Stem Cell 24, 166–182, January 3, 2019

state and found that sensory neuron markers, such as POU4F1 (BRN3a), P2RX3, and NEFM, were strongly represented among those. We next performed GSEA, making use of genes shown to be upregulated in sensory neurons compared to iPSC-derived neural progenitor cultures (Chambers et al., 2012). Cells were ordered by the gradual transition of their transcriptomes (pseudo-time), and GSEA was performed on genes ranked by their correlation with the progression of pseudo-time (Trapnell et al., 2014; Figures 2J and S2K). Indeed, GSEA confirmed a strong enrichment (normalized enrichment score [NES] = 3.34) of sensory neuron genes in the differentiated cells (Figure S2K). Further, plotting of specific genes for undifferentiated NBSCs, such as PAX6, SOX2, and PAX3, and sensory neuron markers, such as P2RX3, DCX, and POU4F1, showed the dynamic progression of differentiation by downregulation of stem-cellrelated genes and upregulation of sensory neuron markers over the course of pseudo-time (Figure 2K). Moreover, analyses of the index information for CD133 and CXCR4, recorded during single cell sorting, revealed that the CD133+CXCR4+ cells were assigned to the undifferentiated fraction of the culture by the pseudo-time algorithm. In contrast, double-negative or CXCR4 single-positive cells were mainly assigned to differentiated cells (Figure S2L). These data support and complement at the molecular and single-cell level the clonogenicity data shown above, demonstrating the functional stem cell characteristics of CD133+ CXCR4+ cells. Collectively, we show that cultured iNBSCs are mostly comprised of cells with a molecular and functional neural-plate-

border-like identity, although some cells show signs of spontaneous differentiation toward neural crest sensory neurons. Differentiation of iNBSCs into Mature CNS and Neural Crest Cells For the derivation of neurons with CNS identity, cells were first cultured in presence of purmorphamine for 7 days, supplemented with either FGF8 (dopaminergic neurons), all-trans retinoic acid (ATRA) (motoneurons) or without additional cues (GABAergic and glutamatergic differentiation). Subsequently, cultures were switched to maturation medium and analyzed five weeks later (Figure 3A). iNBSCs displayed a high neurogenic potential and differentiated rapidly upon removal of CAP medium. Immunofluorescence staining and qPCR analysis revealed the presence of glutamatergic (vGLUT2), dopaminergic (FOXA2, TH, and EN1), motoneurons (CHAT and HB9), GABAergic neurons (GABA and GAD67), and serotonergic neurons (THP2 and GATA3; Figures 3A, S3A, and S3B). Cytometric quantification further confirmed a robust induction of aforementioned subtypes ranging from 18% (dopaminergic neurons) to 5% (serotonergic neurons) of total cells (Figures 3B and S3C). Positivity for SYN1 indicated the formation of synapses, which is associated with mature neurons (Figure 3A). In addition to neuronal subtypes, we detected astrocytes and oligodendrocytes after 10 weeks of differentiation, as shown by GFAP/S100b and Olig2/MBP expression, respectively (Figure 3A). To examine the functional activity of in vitro differentiated iNBSC-derived neurons, we performed whole-cell patch-clamp

Figure 3. Differentiation of iNBSCs into Mature CNS and Neural Crest Progeny (A) Immunofluorescence pictures of iNBSCs differentiated into various neuronal and glial subtypes. Arrowheads indicate glutamatergic neurons, dopaminergic neurons, motoneurons, and GABAergic neurons (left to right, upper panel), as well as serotonin neurons, synapse formation, astrocytes, and oligodendrocytes (left to right, lower panel), respectively. Scale bars, 50 mm. (B) Cytometric quantification of dopaminergic neurons (green; n = 6), motoneurons (blue; n = 6), and serotonergic neurons (mint green; n = 6) relative to undifferentiated iNBSCs (gray; n R 3; t test; *p < 0.05; mean with SEM). (C) Electrophysiological properties of a >6 weeks in vitro differentiated neuron. (D) Spontaneous post-synaptic currents of a >6 weeks in vitro differentiated neuron. (E) iNBSCs form neural-tube-like structures and migratory-crest-like cells upon culture in 3D. iNBSCs were embedded as a single-cell suspension in MG and directed toward CNS or neural crest fate. Upon CNS priming, neural-tube-like structures could be observed (upper BF image, scale bar, 100 mm); neural crest priming resulted in cells migrating throughout the matrix (lower BF image, scale bar, 100 mm). Arrowheads indicate representative SOX1/NESTIN double-positive neuroepithelial structures and SOX10+/AP2a+/PAX6neg cells after CNS and neural crest priming, respectively; scale bars, 50 mm. (F) iNBSCs give rise to all three major neural lineages in vivo. Arrowheads in confocal images mark co-staining of iNBSC progeny (GFP) and markers for neurons (NeuN), astrocytes (GFAP), and oligodendrocytes (MBP). (G) Electrophysiological properties of >6 weeks in vivo differentiated neurons. (H) Morphological reconstruction of transplanted neuronal progeny by filling of GFP-positive neurons with biocytin and subsequent staining; scale bar, 100 mm. (I) Schematic overview of the differentiation of iNBSCs toward neuronal and mesenchymal neural crest progeny. (J) Immunofluorescence images of iNBSCs differentiated into peripheral neurons. Arrowheads indicate sensory neurons (BRN3a/Peripherin double positive); scale bar, 50 mm. (K) Mesenchymal crest differentiation of neural crest-primed iNBSC. Stainings of mesenchymal crest differentiations reveal smooth muscle cells (SMA), chondrocytes (Alcian blue), and adipocytes (oil red); scale bars, 100 mm (SMA) and 250 mm (Alcian blue/oil red). (L) Schematic outline of CRISPR/Cas9-mediated knockout of SCN9A and functional analysis by calcium flux measurements. (M) Western blot analysis of undifferentiated iNBSCs and sensory neurons derived from wild-type (WT) and SCN9A
/
iNBSCs. Arrowhead marks molecular weight (MW) of 200 kDa. (N) Representative confocal image of a sensory neuron differentiation derived from an SCN9A
/
iNBSC clone. Arrowhead marks BRN3a/Peripherin doublepositive cells; scale bar, 50 mm. (O) Calcium flux measurements of sensory neurons derived from WT and SCN9A
/
iNBSCs before and after stimulation with 30 mM a,b-me-ATP. Fluorescence intensity (Fluo-3 AM) of the whole picture is shown as fold change relative to the baseline measurement. Plot shows mean (dark-colored lines) and SD (light color) of measurements for four independent WT sensory neuron differentiations, three independent WT differentiations with additional treatment of the P2X2/3 antagonist A-317491, and three independent SCN9A
/
sensory neuron differentiations. (P) Quantification of calcium flux. The graph shows the comparison of maximal fluorescence intensities (Fluo-3 AM) after treatment with a,b-me-ATP relative to the baseline. Four independent WT sensory neuron differentiations, three independent WT differentiations with additional treatment of the P2X2/3 antagonist A-317491, and four independent SCN9A
/
sensory neuron differentiations were analyzed (t test; *p < 0.05; mean with SEM). See also Figure S3.

Cell Stem Cell 24, 166–182, January 3, 2019 173

recordings of 10-week-old neural cultures. This revealed repetitive trains of action potentials in response to depolarizing voltage steps (12 cells in 3 cultures; Figure 3C). Moreover, patched neurons also exhibited spontaneous post-synaptic currents, in line with a mature and functional neuronal phenotype (7 cells in 3 cultures; Figure 3D). Three-dimensional and organoid cultures have been recently appreciated as a valuable system to spatially model neural development and recapitulate the underlying differentiation processes in vitro (Di Lullo and Kriegstein, 2017). We embedded iNBSCs as a single-cell suspension in Matrigel and expanded them in CNS- or neural crest-priming culture conditions, respectively (Figure 3E). iNBSCs were first cultured in SB431542 and purmorphamine (SP) for nine days and subsequently switched to CSP supplemented with bFGF. Of note, single iNBSCs efficiently gave rise to three-dimensional spheres, some of which developed a central lumen during the differentiation process, resembling the morphology of a neural tube (Figure 3E; see below). Immunofluorescence analysis revealed that 3D cultures stained positive for the neural progenitor markers SOX1 and NESTIN (Figure 3E). The expression of PAX6 and SOX2 further confirmed a CNS progenitor identity (Figure S3D). In contrast, when MG-embedded iNBSCs were grown in neural crest-priming conditions and bFGF (CABF), no spheres were obtained. Instead, iNBSCs differentiated in loosely attached clusters of mesenchymal-type cells that migrated throughout the matrix (Figure 3E). Stainings of 3D cultures revealed cells double-positive for AP2a and SOX10 or SOX10 only, which indicates different stages of neural crest development (Figure 3E). Expression analysis of neural crest-primed 3D cultures confirmed downregulation of PAX6 and SOX1 and concomitant induction of SOX10 and SOX9. Furthermore, there was an upregulation of the crest-associated epithelial-mesenchymal transition (EMT) marker SNAIL (Figure S3D). Thus, there is a clear divergence in phenotype and gene expression of iNBSC-derived 3D cultures grown in CNS- or neural crest-priming conditions. To assess the in vivo differentiation potential of iNBSCs, we pre-differentiated GFP-labeled iNBSCs for 8 days and transplanted them into the striatum of 6-week-old non-obese diabetic (NOD).Prkdcscid.Il2rgnull (NSG) mice. Eight weeks post-transplantation, mice were sacrificed, and the engrafted cells were detected by immunofluorescence. GFP-positive cells included NeuN-positive neurons, GFAP-positive astrocytes, and myelin basic protein (MBP)-positive oligodendrocytes, demonstrating the differentiation potential of iNBSCs to all three major neural lineages and survival of the progeny in vivo (Figures 3F and S3E). To confirm that iNBSC-derived neurons exhibit mature functional properties in vivo, whole-cell patch-clamp analysis of transplanted neurons in acute brain slices was performed. Indeed, GFP-positive neurons exhibited repetitive action potentials upon depolarizing current injection (6 cells from 3 mice; Figure 3G). The neuronal identity was further confirmed by morphological reconstructions of biocytin-filled patched neurons (Figure 3H). To investigate the maturation potential of iNBSCs toward differentiated neural crest cell types, iNBSCs were first cultured in neural crest induction medium and then switched to neural (Chambers et al., 2012) or mesenchymal maturation media (Figure 3I). Four weeks after differentiation was initiated, BRN3a/ 174 Cell Stem Cell 24, 166–182, January 3, 2019

Peripherin/NAV1.7 triple-positive neurons, characteristic for peripheral sensory neurons, could be detected (Figures 3J and S3F). In contrast, differentiation of iNBSCs in mesenchymal differentiation media produced smooth muscle actin (SMA)positive cells, Alcian-blue-positive cartilage, and oil-red-positive fat cells (Figure 3K). Finally, we also derived iNBSC subclones by single-cell sorting and successfully differentiated these toward CNS or neural crest progeny (Figure S3G). Collectively, our data demonstrate that iNBSCs can differentiate into functional neurons both in vitro and in vivo, show broad neuronal differentiation capacity, and can give rise to glial and mesenchymal cells of the CNS and neural crest. The self-renewal and extensive differentiation potential of iNBSCs make them a powerful tool to model genetically based human neuronal syndromes. To directly demonstrate this as a proof-of-concept, we used CRISPR/Cas9-mediated gene editing to mutate the voltage-gated sodium-channel Nav1.7, a nociceptor encoded by the SCN9A gene. This channel is expressed in the peripheral nervous system, and although gain-of-function mutations in SCN9A lead to primary erythromelalgia and paroxysmal pain disorder, loss-of-function mutations result in congenital insensitivity to pain (Dib-Hajj et al., 2007). To model the loss-of-function situation in human iNBSC-derived sensory neurons, we targeted SCN9A with specific guide RNAs that resulted in mutations in exon 22/27 leading to frameshifts of the coding region and nonsense mutations (Figures 3L and S3H). SCN9A
/
iNBSC subclones were expanded and subsequently differentiated into sensory neurons. Western blot analysis confirmed the absence of SCN9A in neurons derived from SCN9A
/
iNBSCs, but not controls (Figures 3M and S3I). Staining of >3-week-old neuronal cultures for the sensory neuron markers BRN3a and Peripherin ruled out that the mutations in SCN9A interfered with sensory neuron differentiation (Figures 3N and S3J). Next, we assessed neuronal activity by calcium flux measurements upon stimulation with a,b-methylene-ATP, a selective agonist of P2RX3 receptors (Chambers et al., 2012), which are expressed in sensory neurons and whose activation mediates inflammatory pain. As expected, the addition of a,b-methyleneATP resulted in a robust increase of activity associated with a synchronous calcium flux in most cells (Figure 3O; Video S1). In contrast, SCN9A
/
cultures exhibited less activity and a paucity of synchronous calcium flux (Figures 3O and 3P; Video S1). To confirm that a,b-methylene-ATP was mediated by P2RX3 receptors, we performed a,b-methylene-ATP-induced calcium flux measurements in the presence of the selective P2RX3 antagonist A-317491, and indeed, neuronal activity was significantly reduced compared to controls (Figures 3O and 3P). In sum, the data show the requirement of SCN9A for painmediated signaling and provides proof of principle of CRISPR/ Cas9-mediated genome editing in iNBSCs. Derivation of Primary NBSCs from Mouse Embryos The existence of iNBSC and the ability to generate and stabilize them from pluripotent cells in vitro raises the question of whether ‘‘primary NBSCs’’ (pNBSCs) also emerge during normal embryogenesis. To this end, we isolated embryonic day 7.5 (E7.5)–E9.5 mouse embryos. After mechanic removal of optic and nonneural tissue and enzymatic digestion, the resulting single-cell suspension was seeded onto a layer of supportive mouse

embryonic fibroblasts and cultured in the presence of CAP (Figures 4A and S4A). Two to three days after plating, clonal lines were established by picking single colonies. Notably, although all embryonic stages gave rise to a variety of neural precursors, stably proliferating lines could only be established from E8.5, but not from E9.5 and later stage embryos, which indicates that their counterpart in vivo is only present during a tight developmental time window during embryonic development (Figure S4B). Expression of neural markers was similar to that of human iNBSCs described above. Thus, we detected expression of Sox1, Sox2, Pax6, and Zfp521 as well as the neural plate border markers Msx1 and Pax3 in pNBSCs (Figures 4B, 4C, S4C, and S4D). Furthermore, pNBSCs showed sustained self-renewal and could be kept in culture for more than 40 passages (>4 months) without loss of early marker expression (Figure S4E). Moreover, upon single-cell sorting, pNBSCs showed up to 10 times higher clonogenic capacity compared to cultured primary radial glia stem cells isolated from E13.5 stage embryos (Figure S4F; Conti et al., 2005). Interestingly, the culture of pNBSC was also feeder dependent, although they differentiated even more rapidly in the absence of feeders compared to human iNBSCs. The rapid differentiation in the absence of feeders could be partially halted when pNBSCs were cultured on Matrigel (MG) in the presence of the Notch ligands Dll4 and Jagged1 (Figure S4G). This indicates that the feeder-mediated maintenance of the NBSC niche is at least in part mediated by Notch signaling. To investigate the regional identity of pNBSCs, the expression of region-specific transcription factors was analyzed. This analysis revealed that they expressed the anterior hindbrain markers Gbx2, Irx3, HoxA2, and HoxB2 (Figure S4H). Taken together, mouse pNBSCs show a similar regionalization as human iNBSCs, a scenario that is compatible with the notion that they share an anterior hindbrain identity. We next addressed the differentiation capability of pNBSCs. When pNBSCs were induced to differentiate in the presence of purmorphamine, PLZF/ZO1 double-positive rosettes formed within two days, suggesting CNS progenitor activity (Figure 4D). The switch to Fgf2 and Egf at this point led to the stable expansion of OLIG2-, SOX2-, and NESTIN-positive RG-like cells (Figure 4F). In contrast, treatment of pNBSCs with Chir99021 and Bmp4 led to the induction of P75 expression and gave rise to AP2a/SOX10 double-positive neural crest cells (Figures 4E and S4I). These data suggest that primary mouse NBSCs have the potential to give rise to CNS, RG-like stem cells, and neural crest progeny and thus constitute self-renewing, multipotent neural progenitors, akin to human iNBSCs obtained by direct reprogramming. When pNBSCs were allowed to further differentiate, subsequent to CNS priming by purmorphamine, robust neuronal differentiation (TUJ1) prevailed, but glial progeny, such as oligodendrocytes (O4) and astrocytes (GFAP), were also detected. Within the neural cultures, we identified GABA-, TH-, and serotonin-positive neurons (Figures 4G and S4J), indicating a broad neuronal differentiation potential. Electrophysiological recordings from neuronal cultures that had matured for three weeks provided functional evidence for the neuronal identity. Patchclamped cells held in the whole-cell mode exhibited repetitive trains of action potentials in response to depolarizing voltage steps, clearly showing that pNBSCs differentiated into mature neurons (Figure 4H). Upon neural crest-primed differentiation,

peripheral neurons, smooth muscle cells, oil-red-positive fat, and Alcian-blue-positive cartilage could be observed (Figure 4I). Finally, we embedded pNBSCs as a single-cell suspension in Matrigel and switched them to CNS- or neural crest-priming culture conditions, respectively (Figure S4K). When pNBSCs were cultured in CSP, small epithelial clusters became apparent after 2 days, some of which formed neural-tube-like structures within two additional days (Figure S4K; Video S2). In contrast, as reported above for iNBSCs, continuous culturing in CABF did not result in neural-tube-like structures but instead gave rise to loosely connected mesenchymal-like cells that started migrating throughout the matrix (Figure S4K). Next, the molecular expression landscape of pNBSCs and their progeny was analyzed. pNBSCs were differentiated and subsequently sorted by fluorescence-activated cell sorting (FACS) for RG-like SCs (Ssea1+ and Glast+) or neural crests (P75+, Glast
, and Ssea1
) and analyzed together with pNBSCs by Illumina gene expression arrays. PCA revealed that pNBSCs and their RG-like and neural crest progeny clustered separately (Figure 5A). Notably, RG-like SC isolated from E13.5-stage embryos showed great overlap with RG-like SCs derived from pNBSC, confirming their identity (Figure 5A). In line with the results from human iNBSCs, GO analysis of pNBSCs compared to mouse embryonic fibroblasts (MEFs) showed that processes such as ‘‘tube development,’’ ‘‘regulation of neural precursor cell proliferation,’’ ‘‘brain development,’’ and ‘‘head development’’ were significantly enriched (Figure S4L; Table S3). Comparison of RG-like SCs or neural crest with pNBSCs showed an upregulation of processes such as ‘‘gliogenesis,’’ ‘‘oligodendrocyte differentiation,’’ ‘‘nervous system development,’’ and ‘‘CNS development’’ for RG-like SCs and processes related to ‘‘embryonic cranial skeleton morphogenesis,’’ ‘‘regulation of cell migration,’’ and ‘‘tissue development’’ for the neural crest derivates (Figure S4L; Table S3). The analysis of genes identified by GO analysis in pNBSCs and their progeny also validated their respective identity. Although pNBSCs showed a specific expression of progenitor markers, such as Hes5, Sox1, and Lin28, RGlike cells expressed glia-related markers, such as Olig2, Omg, and S100b (Figure 5B). In contrast, the neural crest progenitors showed expression of neural crest-associated genes, such as Dlx2 and Pdgf receptors, and absence of neural progenitor and glia markers, respectively (Figure 5B). Of note, neither pNBSCs nor their progeny showed expression of pluripotencyassociated markers, such as Nanog, Oct4, or Utf1 (Figure 5B). To investigate the cellular heterogeneity within pNBSC cultures, single-cell RNA sequencing of two pNBSC clonal lines (73 cells) was performed and data were analyzed analogously to the single-cell iNBSC data above (Figures 2I–2K). Highly analogous to iNBSCs, PCA revealed that the majority of cells clustered independently of their clonal origin (Figure S5A). GO analysis of the genes showing the strongest contribution to PC1 revealed that negatively contributing genes were attributed to terms related to cell cycle, and positively contributing cells were enriched for terms such as ‘‘neurogenesis’’ (Figure S5B). GSEA on pseudo-time-ranked genes showed a strong enrichment for sensory neuron genes (NES = 3.05; Chambers et al., 2012), which was also recapitulated by progression from stem cell (Pax3 and Sox2) to sensory neuron markers (Pou4f1 and Dcx) over the course of pseudo-time (Figures S5C and S5D). Cell Stem Cell 24, 166–182, January 3, 2019 175

A

pNBSCs

B

pNBSCs

SOX1

P26

MSX1

Merge

CAP

C

mRNA expression relative to whole fetal brain

E8.5 embryo

D

1500

1.5

Sox1

60

Pax6

3

Msx1

1000

1.0

40

2

500

0.5

20

1

0

0.0 fetal brain

ESCs

pNBSCs

fetal brain

ESCs

pNBSCs

0

fetal brain

ESCs

pNBSCs

E

PLZF/ZO1

Purmorphamine

0

Pax3

fetal brain

ESCs

AP2a

pNBSCs

SOX10

Chir + BMP4

pNBSCs

OLIG2/NESTIN

F

NESTIN/SOX2

+ bFGF + EGF

TUJ1

O4

GABA/TUJ1

TH

Differentation

G

GFAP

H

I

Alcian Blue

Oil Red

Peripherin

SMA

40 mV 200 ms

Figure 4. Derivation of Primary Neural Plate Border Stem Cells from Mouse Embryos (A) Scheme for the derivation of pNBSCs. (B) BF image and immunofluorescence stainings of a representative neural progenitor line derived from E8.5-stage embryos. BF picture shows representative morphology after 9 weeks of culture (P26). Immunofluorescence stainings demonstrate expression of neural progenitor (SOX1) and neural plate border (MSX1) markers. Scale bar, 50 mm. (C) mRNA expression of neural progenitor and neural plate border markers in pNBSCs and controls. qRT-PCRs were performed on three independent pNBSCs lines and normalized to whole E13.5-stage fetal brains; mouse ESCs served as negative control. (D) Immunofluorescence stainings of CNS-primed pNBSCs. pNBSCs were differentiated in presence of purmorphamine for 3 days and stained for rosette stage markers PLZF and ZO1; scale bar, 100 mm. (E) Immunofluorescence stainings for neural crest stem cell markers AP2a and SOX10 on neural crest primed pNBSCs. Scale bars, 50 mm. (F) Stabilization of radial-glia-like stem cells from pNBSCs. Arrowheads indicate co-expression of radial glia and stem cell markers OLIG2/NESTIN and SOX2/ NESTIN; scale bars, 50 mm. (G) Immunofluorescence pictures of pNBSCs differentiated into various neuronal and glial subtypes. Stainings were performed after >2 weeks of differentiation. pNBSCs showed high neurogenic potential (TUJ1) and also gave rise to oligodendrocytes (O4) and astrocytes (GFAP). Stainings for GABA and TH reveal differentiation into specific subtypes; scale bars, 100 mm. (H) Electrophysiological properties of >3 weeks in vitro differentiated neurons. (I) Differentiation of pNBSCs into neural crest derivates. Stainings of neural crest differentiations reveal chondrocytes (Alcian blue), adipocytes (oil red), peripheral neurons (Peripherin), and smooth muscle cells (SMA); scale bars, 100 mm. See also Figure S4.

176 Cell Stem Cell 24, 166–182, January 3, 2019

A

pNBSCs

B

40

RG-like SCs

NC

ESCs

pNBSCs

20

PC2

0

2 1 0 -1 -2

pNBSCs pNBSC-derived NC pNBSC-derived RG E13.5 stage-derived RG

RG-like SCs

-20

-40

Neural Crest -60

-80

-60

-40

-20

PC1

0

20

40

60

E13.5 pNBSC derived derived

D

C

OTX2

GBX2

IRX3

8

8

10

pNBSC Clone 6 pNBSC Clone 10

6

4

Gene Expression

E8.5 Embryo 2

Gene Expression

E8.5 Embryo 1

Gene Expression

7.5

Samples

5.0

6

Forebrain-assigned E8.5 cells Mid-hindbrain-assigned E8.5 cells

4

Neural Tube-assigned E8.5 cells

2.5

0

Prediction

2

pNBSCs

2

midHindbrain

0.0

0

neuralCrest

HoxA2

8

neuralTube −10

−20

−20

−10

0

PC1

10

20

30

Gene Expression

7.5

non−neural cell−types

5.0

Gene Expression

placodes

0

PAX6

PAX3 7.5

6

4

Gene Expression

PC2

forebrain

5.0

2.5

2

2.5

0.0

0

0.0

Forebrain-assigned E8.5 cells Mid-hindbrain-assigned E8.5 cells Neural Tube-assigned E8.5 cells pNBSCs

Figure 5. Expression Profiling and Direct Comparison of Primary Neural Plate Border Stem Cells to E8.5-Stage Mouse Embryos (A) Microarray-based transcriptional profiling of pNBSCs, pNBSC-derived neural crest, pNBSC-derived radial glia-like stem cells (RG-like SCs), and E13.5-stagederived RG-like SCs. PCA shows three distinct clusters, representing pNBSCs and its neural crest and RG-like SC progeny. Note clustering of E13.5-stagederived RG-like SCs together with pNBSCs-derived RG-like SCs. (B) Gene expression heatmap of genes identified by GO analysis. The heatmap shows a selection of genes related to GO terms enriched in pNBSCs, its neural crest, and RG-like SC progeny as well as a selection of pluripotency markers (see also Figure S4L). E13.5-stage-derived RG SCs and ESCs serve as controls. (C) PCA of pNBSCs (73 cells from two independent clones) and E8.5 primary cells (173 cells from two embryos). Color scheme indicates the prediction of the classification algorithm trained on published whole E8.25 mouse single-cell RNA sequencing dataset (Ibarra-Soria et al., 2018). For the classification of ‘‘nonneural cells,’’ see Figure S5E. (D) Violin plot of regional and pNBSCs-specific marker gene expression at the single-cell level. The plot shows gene expression for pNBSCs and forebrain-, mid-hindbrain-, or neural-tube-assigned E8.5 primary cells at the single-cell level. See also Figures S4 and S5 and Table S3.

Collectively, mouse pNBSC and human iNBSC cultures appear to be characterized by a similar expression landscape containing a major cluster of proliferative stem cells, with some cells displaying signs of spontaneous neuronal differentiation. In summary, these results suggest that mouse-embryo-derived pNBSCs are highly similar to directly reprogrammed human iNBSCs. Notably, pNBSCs represent a so far unidentified, pre-rosette stage neural progenitor with sustained expansion potential. Because we were only able to isolate pNBSCs from embryonic stages prior to E9.5, we hypothesized that cultured pNBSCs might have an in vivo equivalent only present during this time of brain development. We thus performed single-cell transcriptomic analysis comparing pNBSCs to uncultured, manually isolated, and enzymatically separated cells from E8.5 mid-hindbrain regions (173 cells total from 2 embryos). PCA of the transcriptomic data of pNBSCs and E8.5 primary cells revealed five separate cell clusters (Figures 5C and S5E). In order to characterize the nature of those, we trained a classification algorithm

(Ben-hur, 2008) on the recently described dataset of whole E8.25 mouse embryos, defining numerous cell types by single-cell RNA sequencing analysis (Ibarra-Soria et al., 2018), and finally applied this model onto our dataset (Figure 5C). As expected, the manual isolation of embryonic tissue resulted in neural and non-neural cell types, which was recapitulated by terms such as ‘‘neural tube,’’ ‘‘forebrain,’’ and ‘‘midbrain’’ or ‘‘cardiac cells,’’ ‘‘pharyngeal mesoderm,’’ and ‘‘somatic mesoderm’’ (Figure S5E). Interestingly, pNBSCs clustered closest to neural cells and were predominantly assigned to ‘‘neural tube cells’’ (undifferentiated pNBSCs) or ‘‘placode cells’’ (differentiated pNBSCs; Figure 5C). To further investigate the similarities and differences of pNBSCs and E8.5 cells, we normalized the gene expression of cells from both sources relative to non-neural ‘‘pharyngeal mesoderm cells’’ and determined their correlation in relation to the assigned groups. Analyses of individual E8.5 cells confirmed that those assigned to ‘‘neural tube’’ showed the strongest correlation with pNBSCs, and non-neural cells showed much lower similarities (Figure S5F). Moreover, the direct comparison of Cell Stem Cell 24, 166–182, January 3, 2019 177

differentially expressed genes from all ‘‘neural-tube’’-assigned cells from the E8.5 embryos and pNBSCs showed a positive correlation of 0.65 with 74 shared upregulated and 139 shared downregulated genes, as well as 114 and 165 uni-directionally upregulated genes for E8.5 tissue and pNBSCs, respectively (Figure S5G). GO analysis of the shared genes confirmed the neural tube signature and further showed significant enrichment for terms such as ‘‘positive regulation of Notch signaling,’’ ‘‘positive regulation of proliferation,’’ and ‘‘tube development’’ (Figure S5G). Unique genes of pNBSCs related to terms such as ‘‘metabolic processes,’’ ‘‘L-serine biosynthetic process,’’ and ‘‘cellular amino acid biosynthetic process,’’ which are consistent with metabolic adaptations to the in vitro culture conditions. As a negative control, correlation of pNBSCs with ‘‘cardiac’’assigned cells resulted in a very low correlation of 0.25 without significant enrichment of GO terms for shared upregulated genes (Figure S5H). Finally, we investigated the likely regional identity of pNBSCs and E8.5 neural cells by gene expression analysis at the single-cell level. Again, ‘‘neural-tube’’-assigned cells showed the greatest overlap to pNBSCs, showing expression of Hoxa2, Gbx2, Irx3, and Fgf15 and absence of the more anterior marker Otx2 (Figures 5D and S5I). Moreover, some neural tube cells also showed expression of neural-plate-border-associated markers Msx1 and Pax3, as well as stem-cell-related markers Pax6 and Sox21 (Figures 5D and S5I). In summary, our single-cell data suggest that pNBSCs and ex vivo obtained E8.5 neural tube cells share a similar regional identity and expression signature, suggesting that pNBSCs are likely to be derived from the neural plate border contained in this area of the developing embryonic brain, which is also in accordance to their functional characteristics. Comparison of iNBSCs and Mouse pNBSCs Human iNBSCs and mouse pNBSCs share many characteristics, including long-term self-renewal capacity, expression of specific markers, and developmental potential. To further extend the comparison to the molecular expression landscape, we compared the global gene expression profile of mouse pNBSCs with human iNBSCs. To this end, differential gene expression of iNBSCs versus ADFs and pNBSCs versus MEFs was evaluated for similarity by applying the agreement of differential expression procedure (AGDEX), developed to enable cross-species comparisons (Pounds et al., 2011). This analysis revealed a positive correlation of 0.457 for differential gene expression of iNBSCs and pNBSCs (Figure 6A). GO analysis of the top shared upregulated genes (log 2 fold change > 1; 248 genes) identified processes such as ‘‘neural precursor cell proliferation,’’ ‘‘nervous system development,’’ and ‘‘head development’’ to be upregulated, and downregulated DEGs were related to ‘‘response to wound healing,’’ ‘‘tissue development,’’ and ‘‘extracellular matrix organization’’ (Figures 6B and 6C; Table S4). Network analysis of the shared top upregulated genes (log 2 fold change > 1.75; 74 genes) using the Search Tool for the Retrieval of Interaction Genes/Proteins (STRING) revealed the core network of genes expressed in pNBSCs and iNBSCs (Figure 6D; Szklarczyk et al., 2015). This comprises members of the Notch signaling pathway, such as HES5, DLL1, and NOTCH1, as well as prominent neural transcription factors, such as SOX2, ASCL1, and PAX6. Notably, the network comprised two of the four factors used for iNBSC 178 Cell Stem Cell 24, 166–182, January 3, 2019

reprogramming, such as POU3f2 (BRN2), SOX2, as well as the ZIC family member ZIC2. To further corroborate iNBSC and pNBSC culture dynamics, we additionally compared the expression signatures derived from single-cell RNA sequencing data from both species. We then performed GSEA on genes ranked by their fold-change expression comparing the top and bottom 10% of pseudotime-ranked cells, reflecting the least and most differentiated cells within the cultures, respectively, and correlated the resulting enrichment scores for iNBSCs and pNBSCs (Figure 6E). Intriguingly, pNBSCs and iNBSCs shared most gene sets and showed a strong positive correlation (0.542) of their respective enrichment scores. The gene sets showing the highest enrichment included the sensory-neuron-specific gene set (Chambers et al., 2012) and terms such as ‘‘cell morphogenesis involved in differentiation,’’ ‘‘MAPKKK cascade,’’ and ‘‘axon guidance’’ or ‘‘DNA replication,’’ ‘‘DNA metabolic process,’’ and ‘‘M phase’’ for the differentiation- and stem-cell-related portion of the cultures, respectively (Figure 6E). Taken together, mouse pNBSCs and human iNBSCs show a high functional overlap and also robust similarities in their global expression signatures. They share expression of an embryonic neural progenitor network and display analogous spontaneous differentiation in culture, further supporting the notion that NBSCs in mouse and human appear to be physiological, embryonic cellular counterparts, irrespective of whether they were obtained in vivo, ex vivo, by directed differentiation, or by direct reprogramming. DISCUSSION In recent years, various neural progenitor populations obtained by directed differentiation of mouse and human PSCs have been described to model specific stages of physiological neural development in vitro (for review, see Conti and Cattaneo, 2010). For example, PSC-derived rosette-type NSC populations (R-NSCs) have been reported to be progenitors capable of differentiating toward CNS and neural crest progeny. However, the R-NSC state is transient and long-term cultures result in rather heterogeneous populations (Elkabetz et al., 2008). More recently, the use of small molecules has opened up new possibilities to stabilize expandable NPCs. The combination of Chir99021, SB431542, and LIF has been reported to maintain an early mid-hindbrain precursor that retains high neurogenic potential and shows sustained proliferation in culture (Li et al., 2011). These precursors show multi-CNS lineage contribution, but their potential to give rise to cells of the neural crest lineage is very low. Reinhardt and colleagues described neural progenitors with CNS- and neural crest-differentiation potential that can be maintained in the presence of Chir99021 and purmorphamine on Matrigel-coated plates (Reinhardt et al., 2013). These progenitors show some overlapping functional properties to iNBSCs but display a distinct molecular signature (data not shown) and culture requirements. Particularly, (i)NBSCs grow in the presence of Chir99021, Alk5-inhibitor, and purmorphamine and are dependent on a supportive layer of fibroblasts. The latter retains their neuroepithelial and neural plate border phenotype likely by activating Notch signaling. Our data suggest that clonally derived (i)NBSCs

Figure 6. Comparison of iNBSCs and Primary Mouse NBSCs (A) Comparison of human iNBSC and mouse pNBSC expression data. Differential gene expression of iNBSCs versus ADFs and pNBSCs versus MEFs was evaluated for its similarity by applying the agreement of differential expression procedure (AGDEX). The analysis shows a positive correlation of 0.475 (permutated p value of 0.0256). (B) Shared biological processes enriched in iNBSCs and pNBSCs. GO processes are based on upregulated genes with log 2 fold change > 1 (258 genes) in iNBSC and pNBSCs lines. (C) Biological processes downregulated in iNBSCs and pNBSCs compared to ADFs and MEFs. GO processes are based on shared 200 most downregulated genes in iNBSC and pNBSCs lines. (D) Shared core network of iNBSCs and pNBSCs. Network analysis was performed on shared upregulated genes of iNBSCs and pNBSCs with a log 2 fold change > 1.75 (74 genes) using the STRING v10 web interface; only connected nodes of the network are displayed. (E) Comparison of human and mouse differentiation dynamics. GSEA was performed on gene rankings based on fold changes comparing top 10% to bottom 10% of pseudo-time-ranked cells of pNBSC and iNBSC cultures; the resulting enrichment scores were plotted and correlated. See also Table S4.

represent a thus far unidentified embryonic neural progenitor with broad CNS and neural crest differentiation capacity. Because development at the neural plate border progresses rapidly in the early embryo, the properties of the stem and progenitors dynamically change and are transient in nature. In agreement with this phenomenon, large numbers of human progenitors of such defined stages are difficult to obtain (Tchieu et al., 2017). However, applying our direct conversion protocol in conjunction with specifically defined culture conditions, we have been able to stabilize such a human progenitor and developed conditions for its robust expansion while maintaining its multipotency.

Importantly, the NBSC state is independent of transgene activity and can also be reached by differentiation from human iPSCs, which excludes an artificial reprogramming-induced state. Applying single-cell RNA sequencing of clonal iNBSC lines, we further demonstrate that CD133+CXCR4+ iNBSCs represent a rather homogeneous population, with only marginal spontaneous differentiation toward sensory neurons. Moreover, iNBSCs share a similar expression profile as well as functional features with pNBSCs isolated and expanded from E8.5 mouse neural folds. Therefore, NBSCs represent a bona fide embryonic, multipotent, and self-renewing neural progenitor population that can be derived either by direct reprogramming of human adult Cell Stem Cell 24, 166–182, January 3, 2019 179

rs ion

Pr im

ar

ve

on

ys

tc

ou

rce

c re Di

E8.5 mouse embryo

CRISPR Cas9 Genome editing

Chir Alk.5-Inh. Purmorphamine Feeder

Differentiation

bFGF EGF LIF

(i)NBSCs low density Chir BMP4 Chir FGF8 IGF1 DAPT

Development

Chir SB Purmorphamine

Neural Crest

RG-like stem cells

Mature neurons and glia

Sensory neurons and mesenchymal crest cells

Figure 7. Summary of iNBSCs’ and Primary Mouse NBSCs’ Features A schematic overview of the direct conversion and isolation of (i)NBSCs, its manipulation via CRISPR-Cas9, and differentiation into neural progeny from the CNS and neural crest lineages.

somatic cells or by isolation from primary mouse embryonic brain tissue (Figure 7). Starting from NBSCs, we then characterized a broad variety of downstream CNS- and neural crest-lineage progenitors, such as RG-like SCs, NCSCs, and MSC-like cells (Conti and Cattaneo, 2010; Gorris et al., 2015; Liu and Cheung, 2016; Nery et al., 2013). Moreover, iNBSCs can differentiate into mature progeny of the CNS and neural crest and also recapitulate early neural development when embedded into a 3D matrix. Expandable iNBSCs can also be genetically engineered as demonstrated by generating an SCN9a-mediated pain model. The pathophysiological response of SCN9a mutant cells in vitro accurately models the functional deficit of the SCN9A gene in patients associated to congenital insensitivity to pain and who do not experience any modality of pain except for neuropathic pain (Cox et al., 2006). To date, the developmental stage and in vivo counterpart of directly converted neural progenitors often remains elusive, limiting their use to model physiological development (Cairns et al., 2016; Hou et al., 2017; Lee et al., 2015). In fact, the identification of a direct in vivo counterpart of in vitro expanded stem cells is challenging. Even mouse embryonic stem cells that without doubt are the functional in vitro counterpart of inner cell mass cells in the blastocyst show significant transcriptomic differences (Boroviak et al., 2014). Along these lines, our data suggest that an NBSC state occurs naturally during normal neu180 Cell Stem Cell 24, 166–182, January 3, 2019

ral development. First, mouse NBSCs can only be isolated from E8.5 embryos, but not later stages of development. Second, the comparison of single-cell transcriptomes from pNBSCs and embryonic brain tissue hints toward an endogenous counterpart. Although direct comparison of E8.5 neural tissue and pNBSCs did not result in perfectly matching expression profiles at the single-cell level, we nevertheless identified endogenous neural tube cells with an anterior-hindbrain signature as the most likely candidates from which cultured pNBSCs may originate. Our data suggest the presence of a transient, NBSC-like progenitor during the neurulation stage embryo of mouse and likely also human. As assumed, the frequency of embryonic NBSC-like cells appears to be low. Sequencing of a larger number of embryonic neural cells, including additional developmental intermediates, might further refine the exact nature of endogenous NBSCs in the future. Collectively, we show that (i)NBSCs and progenitors derived thereof can be generated from various somatic cells, thus offering alternative approaches in obtaining defined and scalable neural progenitors. This strategy circumvents the reprogramming into pluripotent and also teratogenic iPSCs and therefore shortcuts the derivation of defined NBSCs and progeny by omitting the consecutive and labor-intensive validation of iPSC and subsequent neural progenitor cell lines. (i)NBSCs might prove valuable to model and repair genetic defects underlying neural diseases, such as, for example, Parkinson’s disease, ataxia, or neuropathic diseases, by gene editing and will eventually open new options for regenerative medicine. STAR+METHODS Detailed methods are provided in the online version of this paper and include the following: d d d

d

KEY RESOURCES TABLE CONTACT FOR REAGENT AND RESOURCE SHARING EXPERIMENTAL MODEL AND SUBJECT DETAILS B Derivation of human PBMCs and ADFs B Culture of human fetal pancreas fibroblasts B Culture of iNBSCs and pNBSCs B Culture of human iPSCs B Mice METHOD DETAILS B Identification of optimal growth and reprogramming conditions for induced NBSCs B Reprogramming into induced NBSCs B Cre-mediated deletion of BKSZ flox B Differentiation of human iPSCs into NBSCs B Differentiation of iNBSCs toward RG-like cells B Differentiation of iNBSCs toward NCSC-like cells B Differentiation of iNBSCs toward MSCs and neural crest progeny B Differentiation of iNBSCs toward mature CNS progeny B 3D Differentiation of iNBSCs B Transplantation of iNBSCs B Derivation of pNBSCs B Differentiation of pNBSCs B Derivation of primary RG from mouse embryos

B

3D Differentiation of pNBSCs Slice preparation for electrophysiology B Whole-cell recordings B Cell identification and reconstruction B Calcium Imaging B Flow Cytometry B Immunofluorescence B Immunohistochemical analysis B Oil Red o staining B Alcian blue staining B Western Blot B CRISPR/Cas9-mediated knockout B Gene Expression Analysis by Quantitative PCR B Microarray analysis B Methylation analysis B Single-cell RNA sequencing protocol B Raw data processing and quality control B Single-cell RNA sequencing data analysis QUANTIFICATION AND STATISTICAL ANALYSIS DATA AND SOFTWARE AVAILABILITY B

d d

SUPPLEMENTAL INFORMATION Supplemental Information includes five figures, six tables, and two video files and can be found with this article online at https://doi.org/10.1016/j.stem. 2018.11.015. ACKNOWLEDGMENTS We thank Corinna Klein and Markus Sohn for technical assistance. We would like to thank Dr. Steffen Schmitt, Dr. Marcus Eich, and the DKFZ Flow Cytometry Core facility for their assistance and Dr. Kurt Reifenberg, Dr. Michaela Socher, and all members of the DKFZ Laboratory Animal Core Facility for excellent animal welfare and husbandry. Support by Damir Krunic and the DKFZ Light Microscopy facility is gratefully acknowledged. We thank the microarray unit of the DKFZ Genomics and Proteomics Core Facility for providing the whole-genome expression and methylation analysis and related services. We thank Dr. Simon Haas and all other colleagues of the Trumpp, Edenhofer, and HI-STEM laboratories for helpful discussions. The authors would like to €stle, thank Dr. Johannes Jungverdorben, Dr. Michael Karus, Dr. Oliver Bru and all members of the Institute of Reconstructive Neurobiology (Bonn, Germany) for helpful discussions during the initial phase of the project. This work was supported by the Stiftung Sybille Assmus (M.C.T.), the SFB873 funded by the Deutsche Forschungsgemeinschaft (DFG) and the Dietmar Hopp Foundation (all to A.T.), and funds from the ERA-NET E-RARE research program as well as the Austrian Science Fund (FWF) (grant number I 3029-B30; to F.E.). AUTHOR CONTRIBUTIONS Conceptualization, M.C.T., F.E., and A.T.; Methodology, M.C.T., O.H., J.P., R.P., D.G.-G., T.B., P.W., Y.A., R.S., A.P., P.K., L.B., M.D.M., A.J., J.U., C.H., and H.M.; Investigation, M.C.T., O.H., J.P., R.S., D.G.-G., P.W., and Y.A.; Writing – Original Draft, M.C.T., H.M., and A.T.; Writing – Reviewing & Editing, M.C.T., O.H., J.P., D.G.-G., P.W., Y.A., M.D.M., H.M., F.E., and A.T.; Supervision, M.C.T., F.E., and A.T. DECLARATION OF INTERESTS M.C.T., O.H., F.E., and A.T. filed a patent application related to this work. Received: October 9, 2017 Revised: May 30, 2018 Accepted: November 9, 2018 Published: December 20, 2018

REFERENCES €ller, F., Lutsik, P., Walter, J., Lengauer, T., and Bock, C. (2014). Assenov, Y., Mu Comprehensive analysis of DNA methylation data with RnBeads. Nat. Methods 11, 1138–1140. Ben-Hur, A. (2008). Support vector clustering. Scholarpedia 3, 5187. Boroviak, T., Loos, R., Bertone, P., Smith, A., and Nichols, J. (2014). The ability of inner-cell-mass cells to self-renew as embryonic stem cells is acquired following epiblast specification. Nat. Cell Biol. 16, 516–528. Brambrink, T., Foreman, R., Welstead, G.G., Lengner, C.J., Wernig, M., Suh, H., and Jaenisch, R. (2008). Sequential expression of pluripotency markers during direct reprogramming of mouse somatic cells. Cell Stem Cell 2, 151–159. Bray, N.L., Pimentel, H., Melsted, P., and Pachter, L. (2016). Near-optimal probabilistic RNA-seq quantification. Nat. Biotechnol. 34, 525–527. Cairns, D.M., Chwalek, K., Moore, Y.E., Kelley, M.R., Abbott, R.D., Moss, S., and Kaplan, D.L. (2016). Expandable and rapidly differentiating human induced neural stem cell lines for multiple tissue engineering applications. Stem Cell Reports 7, 557–570. Chambers, S.M., Qi, Y., Mica, Y., Lee, G., Zhang, X.-J., Niu, L., Bilsland, J., Cao, L., Stevens, E., Whiting, P., et al. (2012). Combined small-molecule inhibition accelerates developmental timing and converts human pluripotent stem cells into nociceptors. Nat. Biotechnol. 30, 715–720. Chen, E.Y., Tan, C.M., Kou, Y., Duan, Q., Wang, Z., Meirelles, G.V., Clark, N.R., and Ma’ayan, A. (2013). Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinformatics 14, 128. Conti, L., and Cattaneo, E. (2010). Neural stem cell systems: physiological players or in vitro entities? Nat. Rev. Neurosci. 11, 176–187. Conti, L., Pollard, S.M., Gorba, T., Reitano, E., Toselli, M., Biella, G., Sun, Y., Sanzone, S., Ying, Q.-L., Cattaneo, E., and Smith, A. (2005). Niche-independent symmetrical self-renewal of a mammalian tissue stem cell. PLoS Biol. 3, e283. Cox, J.J., Reimann, F., Nicholas, A.K., Thornton, G., Roberts, E., Springell, K., Karbani, G., Jafri, H., Mannan, J., Raashid, Y., et al. (2006). An SCN9A channelopathy causes congenital inability to experience pain. Nature 444, 894–898. Dehairs, J., Talebi, A., Cherifi, Y., and Swinnen, J.V. (2016). CRISP-ID: decoding CRISPR mediated indels by Sanger sequencing. Sci. Rep. 6, 28973. Di Lullo, E., and Kriegstein, A.R. (2017). The use of brain organoids to investigate neural development and disease. Nat. Rev. Neurosci. 18, 573–584. Dib-Hajj, S.D., Cummins, T.R., Black, J.A., and Waxman, S.G. (2007). From genes to pain: Na v 1.7 and human pain disorders. Trends Neurosci. 30, 555–563. Elkabetz, Y., Panagiotakos, G., Al Shamy, G., Socci, N.D., Tabar, V., and Studer, L. (2008). Human ES cell-derived neural rosettes reveal a functionally distinct early neural stem cell stage. Genes Dev. 22, 152–165. Eminli, S., Foudi, A., Stadtfeld, M., Maherali, N., Ahfeldt, T., Mostoslavsky, G., Hock, H., and Hochedlinger, K. (2009). Differentiation stage determines potential of hematopoietic cells for reprogramming into induced pluripotent stem cells. Nat. Genet. 41, 968–976. Garnett, A.T.A., Square, T.A.T., and Medeiros, D.M.D. (2012). BMP, Wnt and FGF signals are integrated through evolutionarily conserved enhancers to achieve robust expression of Pax3 and Zic genes at the zebrafish neural plate border. Development 139, 4220–4231. Gorris, R., Fischer, J., Erwes, K.L., Kesavan, J., Peterson, D.A., Alexander, M., €stle, O. (2015). No¨then, M.M., Peitz, M., Quandel, T., Karus, M., and Bru Pluripotent stem cell-derived radial glia-like cells as stable intermediate for efficient generation of human oligodendrocytes. Glia 63, 2152–2167. Han, D.W., Tapia, N., Hermann, A., Hemmer, K., Ho¨ing, S., Arau´zo-Bravo, M.J., Zaehres, H., Wu, G., Frank, S., Moritz, S., et al. (2012). Direct reprogramming of fibroblasts into neural stem cells by defined factors. Cell Stem Cell 10, 465–472. Heigwer, F., Kerr, G., and Boutros, M. (2014). E-CRISP: fast CRISPR target site identification. Nat. Methods 11, 122–123.

Cell Stem Cell 24, 166–182, January 3, 2019 181

Hockemeyer, D., Soldner, F., Cook, E.G., Gao, Q., Mitalipova, M., and Jaenisch, R. (2008). A drug-inducible system for direct reprogramming of human somatic cells to pluripotency. Cell Stem Cell 3, 346–353. Hou, P.-S., Chuang, C.-Y., Yeh, C.-H., Chiang, W., Liu, H.-J., Lin, T.-N., and Kuo, H.-C. (2017). Direct conversion of human fibroblasts into neural progenitors using transcription factors enriched in human ESC-derived neural progenitors. Stem Cell Reports 8, 54–68. Ibarra-Soria, X., Jawaid, W., Pijuan-Sala, B., Ladopoulos, V., Scialdone, A., Jo¨rg, D.J., Tyser, R.C.V., Calero-Nieto, F.J., Mulas, C., Nichols, J., et al. (2018). Defining murine organogenesis at single-cell resolution reveals a role for the leukotriene pathway in regulating blood progenitor formation. Nat. Cell Biol. 20, 127–134. Le Dre´au, G., and Martı´, E. (2012). Dorsal-ventral patterning of the neural tube: a tale of three signals. Dev. Neurobiol. 72, 1471–1481. Lee, J.H., Mitchell, R.R., McNicol, J.D., Shapovalova, Z., Laronde, S., Tanasijevic, B., Milsom, C., Casado, F., Fiebig-Comyn, A., Collins, T.J., et al. (2015). Single transcription factor conversion of human blood fate to NPCs with CNS and PNS developmental capacity. Cell Rep. 11, 1367–1376. Li, W., Sun, W., Zhang, Y., Wei, W., Ambasudhan, R., Xia, P., Talantova, M., Lin, T., Kim, J., Wang, X., et al. (2011). Rapid induction and long-term selfrenewal of primitive neural precursors from human embryonic stem cells by small molecule inhibitors. Proc. Natl. Acad. Sci. USA 108, 8299–8304.

Reinhardt, P., Glatza, M., Hemmer, K., Tsytsyura, Y., Thiel, C.S., Ho¨ing, S., Moritz, S., Parga, J.A., Wagner, L., Bruder, J.M., et al. (2013). Derivation and expansion using only small molecules of human neural progenitors for neurodegenerative disease modeling. PLoS ONE 8, e59252. Scognamiglio, R., Cabezas-Wallscheid, N., Thier, M.-C., Altamura, S., Reyes, €rtner, D., Carnevalli, L.S., Atzberger, A., Haas, A., Prendergast, A´.M., Baumga S., et al. (2016). Myc depletion induces a pluripotent dormant state mimicking diapause. Cell 164, 668–680. Shao, M., Liu, C., Song, Y., Ye, W., He, W., Yuan, G., Gu, S., Lin, C., Ma, L., Zhang, Y., et al. (2015). FGF8 signaling sustains progenitor status and multipotency of cranial neural crest-derived mesenchymal cells in vivo and in vitro. J. Mol. Cell Biol. 7, 441–454. Sommer, C.A., Stadtfeld, M., Murphy, G.J., Hochedlinger, K., Kotton, D.N., and Mostoslavsky, G. (2009). Induced pluripotent stem cell generation using a single lentiviral stem cell cassette. Stem Cells 27, 543–549. Sommer, A.G., Rozelle, S.S., Sullivan, S., Mills, J.A., Park, S.-M., Smith, B.W., Iyer, A.M., French, D.L., Kotton, D.N., Gadue, P., et al. (2012). Generation of human induced pluripotent stem cells from peripheral blood using the STEMCCA lentiviral vector. J. Vis. Exp. 4327.

Liu, J.A., and Cheung, M. (2016). Neural crest stem cells and their potential therapeutic applications. Dev. Biol. 419, 199–216.

Stein, J.L., de la Torre-Ubieta, L., Tian, Y., Parikshak, N.N., Herna´ndez, I.A., Marchetto, M.C., Baker, D.K., Lu, D., Hinman, C.R., Lowe, J.K., et al. (2014). A quantitative framework to evaluate modeling of cortical development by neural stem cells. Neuron 83, 69–86.

€dhof, T.C., and Wernig, M. (2012). Direct Lujan, E., Chanda, S., Ahlenius, H., Su conversion of mouse fibroblasts to self-renewing, tripotent neural precursor cells. Proc. Natl. Acad. Sci. USA 109, 2527–2532.

Subramanian, A., Kuehn, H., Gould, J., Tamayo, P., and Mesirov, J.P. (2007). GSEA-P: a desktop application for gene set enrichment analysis. Bioinformatics 23, 3251–3253.

Lun, A.T.L., Bach, K., and Marioni, J.C. (2016). Pooling across cells to normalize single-cell RNA sequencing data with many zero counts. Genome Biol. 17, 75. €nther, K., Thier, M., and Edenhofer, F. (2015). Meyer, S., Wo¨rsdo¨rfer, P., Gu Derivation of adult human fibroblasts and their direct conversion into expandable neural progenitor cells. J. Vis. Exp. e52831. Nery, A.A., Nascimento, I.C., Glaser, T., Bassaneze, V., Krieger, J.E., and Ulrich, H. (2013). Human mesenchymal stem cells: from immunophenotyping by flow cytometry to clinical applications. Cytometry A 83, 48–61. €hdesma €ki, H., Lahesmaa, R., Noisa, P., Lund, C., Kanduri, K., Lund, R., La Lundin, K., Chokechuwattanalert, H., Otonkoski, T., Tuuri, T., and Raivio, T. (2014). Notch signaling regulates the differentiation of neural crest from human pluripotent stem cells. J. Cell Sci. 127, 2083–2094. Pfisterer, U., Kirkeby, A., Torper, O., Wood, J., Nelander, J., Dufour, A., Bjo¨rklund, A., Lindvall, O., Jakobsson, J., and Parmar, M. (2011). Direct conversion of human fibroblasts to dopaminergic neurons. Proc. Natl. Acad. Sci. USA 108, 10343–10348. Picelli, S., Faridani, O.R., Bjo¨rklund, A˚.K., Winberg, G., Sagasser, S., and Sandberg, R. (2014). Full-length RNA-seq from single cells using Smartseq2. Nat. Protoc. 9, 171–181. Pounds, S., Gao, C.L., Johnson, R.A., Wright, K.D., Poppleton, H., Finkelstein, D., Leary, S.E.S., and Gilbertson, R.J. (2011). A procedure to statistically evaluate agreement of differential expression for cross-species genomics. Bioinformatics 27, 2098–2103. Ran, F.A., Hsu, P.D., Wright, J., Agarwala, V., Scott, D.A., and Zhang, F. (2013). Genome engineering using the CRISPR-Cas9 system. Nat. Protoc. 8, 2281–2308.

182 Cell Stem Cell 24, 166–182, January 3, 2019

Szklarczyk, D., Franceschini, A., Wyder, S., Forslund, K., Heller, D., HuertaCepas, J., Simonovic, M., Roth, A., Santos, A., Tsafou, K.P., et al. (2015). STRING v10: protein-protein interaction networks, integrated over the tree of life. Nucleic Acids Res. 43, D447–D452. Tchieu, J., Zimmer, B., Fattahi, F., Amin, S., Zeltner, N., Chen, S., and Studer, L. (2017). A modular platform for differentiation of human PSCs into all major ectodermal lineages. Cell Stem Cell 21, 399–410.e7. Thier, M., Wo¨rsdo¨rfer, P., Lakes, Y.B., Gorris, R., Herms, S., Opitz, T., Seiferling, D., Quandel, T., Hoffmann, P., No¨then, M.M., et al. (2012). Direct conversion of fibroblasts into stably expandable neural stem cells. Cell Stem Cell 10, 473–479. Trapnell, C., Cacchiarelli, D., Grimsby, J., Pokharel, P., Li, S., Morse, M., Lennon, N.J., Livak, K.J., Mikkelsen, T.S., and Rinn, J.L. (2014). The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. 32, 381–386. €dhof, T.C., and Vierbuchen, T., Ostermeier, A., Pang, Z.P., Kokubu, Y., Su Wernig, M. (2010). Direct conversion of fibroblasts to functional neurons by defined factors. Nature 463, 1035–1041. Villanueva, S., Glavic, A., Ruiz, P., and Mayor, R. (2002). Posteriorization by FGF, Wnt, and retinoic acid is required for neural crest induction. Dev. Biol. 241, 289–301. Yuan, S.H., Martin, J., Elia, J., Flippin, J., Paramban, R.I., Hefferan, M.P., Vidal, J.G., Mu, Y., Killian, R.L., Israel, M.A., et al. (2011). Cell-surface marker signatures for the isolation of neural stem cells, glia and neurons derived from human pluripotent stem cells. PLoS ONE 6, e17540.

STAR+METHODS KEY RESOURCES TABLE

REAGENT or RESOURCE

SOURCE

IDENTIFIER

Antibodies Rabbit anti-BRN3A

LifeSpan BioSciences

Cat# LS-C12182

Goat anti-CHAT

Merck Millipore

Cat# AB144P

Goat anti-EN1

Santa Cruz

Cat# D-20

Mouse anti-FOXA2

Abcam

Cat# ab60721

Goat anti-FOXA2

R&D Systems

Cat# AF2400

Rabbit anti-GABA

Sigma-Aldrich

Cat# A-2052

Rabbit anti-GABA

Sigma-Aldrich

Cat# A-2052

Mouse anti-GAD67

Millipore

Cat# MAB5406

Mouse anti-GATA3

R&D Systems

Cat# MAB6330

Rabbit anti-GFAP/ Gfap

Dako

Cat# Z0334

Goat anti-GFP

Abcam

Cat# ab5450

Rabbit anti-GLAST

Abcam

Cat# ab416

Mouse anti-HB9

Santa Cruz

Cat# sc-515769

Mouse anti-HNK1

Sigma-Aldrich

Cat#C6680-50TST

Rat anti-MBP

Abcam

Cat# ab7349

Mouse anti-MSX1/ Msx1

DSHB

4G1

Mouse anti-NAV1.7

Abcam

Cat# ab85015

Mouse anti-NESTIN

Abcam

Cat# ab22035

Mouse anti-Nestin

Millipore

Cat# MAB353

Biotin anti-NeuN

Millipore

Cat# MAB377B

Mouse anti-NURR1

Santa Cruz

Cat# sc-376984

Mouse anti-O4

R&D Systems

Cat# MAB1326

Goat anti-OLIG2

R&D Systems

Cat# AF2418

Rabbit anti-Olig2

Millipore

Cat# AB9610

Mouse anti-Pax3

DSHB

Pax3 (clone C2)

Rabbit anti-PAX6/ Pax6

Life Technologies

Cat# 42-6600

Chicken anti-PERIPHERIN

Abcam

Cat# ab39374

Rabbit anti-PERIPHERIN/ Peripherin

Abcam

Cat# ab4666

Mouse anti-Plzf

Calbiochem

Cat# OP128

Biotin anti-PROMININ1

Miltenyi Biotec

Cat# 130-101-851

Mouse anti-S100B

Sigma-Aldrich

Cat# S2532

Mouse anti-SCN9A

Abcam

Cat# ab85015

Rabbit anti-Serotonin

Sigma-Aldrich

Cat# S5545

Mouse anti-SMA/ Sma

DAKO

Cat# M0851

Goat anti-SOX1/ Sox1

R&D Systems

Cat# AF3369

Goat anti-Sox10

Santa Cruz

Cat# sc-17342

Goat anti-SOX10

R&D Systems

Cat# AF2864-SP

Rabbit anti-SOX2

Millipore

Cat# AB5603

Mouse anti-Sox2

R&D Systems

Cat# MAB2018

Mouse anti-SYN1

Synaptic Systems

Cat# 106011

Mouse anti-TFAP2A/ Tfap2a

DSHB

3B5

Rabbit anti-TH/ Th

Abcam

Cat# ab112

Rabbit anti-TPH2

Novus Biologicals

Cat# NB100-74555

Chicken anti-TUBB3

Abcam

Cat# ab41489 (Continued on next page)

Cell Stem Cell 24, 166–182.e1–e13, January 3, 2019 e1

Continued REAGENT or RESOURCE

SOURCE

IDENTIFIER

Rabbit anti-TUBB3/ Tubb3

Covance

Cat# PRB-435P

Mouse anti-TUBB3/ Tubb3

Covance

Cat# MMS-435P

Rabbit anti-VGLUT2

Abcam

Cat# ab18105

Mouse anti-VIMENTIN

Sigma-Aldrich

Cat# V2258

Mouse anti-ZIC1

DSHB

PCRP-ZIC1-1E3

Rabbit anti-Zo1

Thermo Fisher Scientific

Cat# 61-7300

APCCy7 anti-human CD13 (WM15)

BioLegend

Cat# 301710

PE anti-human CD44 (G44-26)

BD Bioscience

Cat# 555479

PerCP/Cy5.5 anti-human CD47 (CC2C6)

BioLegend

Cat# 323110

PE-Cy7 anti-human CD57 (TBO1)

eBioscience

Cat# 25-0577-42

BV421 anti-human CD90 (5E10)

BioLegend

Cat# 328122

FITC anti-human CD105 (43A3)

BioLegend

Cat# 323204

FITC anti-human CD133/2 (293C3)

Milteny Biotech

Cat# 130-104-273

Alexa647 anti-human CD146 (P1H12)

BioLegend

Cat# 361014

PE/Cy5 anti-human CXCR4 (12G5)

Biozol Diagnostica

Cat# BLD-306508

APC Anti-GLAST (ACSA-1)

Milteny Biotech

Cat# 130-095-814

FITC anti-mouse/ human NGFR / p75 NGF Receptor

ThermoFisher LifeScience

Cat# MA1-18421

PE anti-human P75 (ME20.4-1.H4)

Milteny Biotech

Cat# 130-091-885

V450 anti-mouse/ human SSEA-1 V450 (MC480)

Becton Dickinson

Cat# 561561

Goat anti-chicken 488

Abcam

Cat# ab150169

Donkey anti-Goat-488

Abcam

Cat# ab150129

Goat anti-mouse-488

Abcam

Cat# ab150117

Donkey anti-mouse-488

Abcam

Cat# 150105

Donkey anti-mouse-555

Abcam

Cat# ab150106

Goat anti-mouse-647

Abcam

Cat# ab150115

Goat anti-rabbit-488

Cell Signaling

Cat# 4412S

Donkey anti-rabbit-555

Abcam

Cat# ab150074

Donkey anti-rabbit-647

Abcam

Cat# ab150075

Donkey anti-rat-647

Abcam

Cat# ab150155

SAV-488

eBioscience

Cat# 11431787

SAV-555

eBioscience

Cat# 12431787

Goat anti-Actin

Santa Cruz, USA

Cat# sc-1616

Goat anti-HRP

Pierce, Thermo Scientific

Cat# 31402

Mouse anti-HRP

DAKO

Cat# P0447

Life Technologies

Cat# 11635018

human dermal fibroblasts

skin biopsy, this study

n/a

Human Primary Pancreatic Fibroblast

VitroBiopharma

Cat# SC00A5

peripheral blood mononuclear cells

peripheral blood, this study

n/a

Bacterial and Virus Strains ElectroMAX Stbl4 Competent Cells Biological Samples

Chemicals, Peptides, and Recombinant Proteins 3,30 ,5-Triiodo-L-Thyronine sodium

Sigma-Aldrich

Cat# T6397-100MG

A317491

Sigma-Aldrich

Cat# A2979

A-803467

Sigma-Aldrich

Cat# A3109-10MG

Albumax I

GIBCO

Cat# 11020039

Alcian Blue 8GX

Sigma-Aldrich

Cat# A5268

Alk5 Inhibitor II

enzolifesciences

Cat# ALX-270-445-M005

BD Matrigel Growth Factor reduced

BD

Cat# 354230

BDNF

Peprotech

Cat# 450-02-10 (Continued on next page)

e2 Cell Stem Cell 24, 166–182.e1–e13, January 3, 2019

Continued REAGENT or RESOURCE

SOURCE

IDENTIFIER

Bmp-4, rec. Murine

Preprotech

Cat# 315-27-10

bNGF

Peprotech

Cat# 450-01-250

Cell Tracker Red CMTPX

Thermo Scientific

Cat# C34552

Chir 99021

Sigma-Aldrich

Cat# SML1046-25MG

DAPT

Sigma-Aldrich

Cat# D5942-5MG

Dapi

Sigma-Aldrich

Cat# D9542

db camp

Sigma-Aldrich

Cat# D0260-25MG

Dexamethason

Sigma-Aldrich

Cat# D4902

Doxycyline

Sigma-Aldrich

Cat# D9891

EPO

R&D Systems

Cat# CAA26094

FGF4

Peprotech

Cat# AF-100-31-25

FGF8

Preprotech

Catt# 100-25-25

Fibronectin

Calbiochem

Cat# 341631-1MG

Fluo-3 AM

Thermo Scientific

Cat# F1242

Formaldehyde

Thermo Scientific

Cat# 28908

Forskolin

Sigma-Aldrich

Cat# F6886-10MG

GDNF

Peprotech

Cat# 450-10-10

HALT Protease Inhibitor

Thermo Scientific

Cat# 78440

human bFGF

Peprotech

Cat# 100-18B-100

human EGF

Peprotech

Cat# AF-100-15-1000

human LIF

Peprotech

Cat# 300-05-25

human rec. BMP4

Peprotech

Cat# 120-05-5

IL-3

R&D Systems

Cat# AAC08706

IGF-I

Peprotech

Cat# 100-11-500

L-ascorbic acid 2-phosphate magnesium

Sigma-Aldrich

Cat# A8960-5G

Matrigel, growth factor reduced

BD

Cat# 354230

Midi Protein Gel, 4–15% Criterion TGX Precast

BioRad

Cat# 5671084

NT-3

Peprotech

Cat# 450-03-100

Oil Red O

Sigma-Aldrich

Cat# 00625

RIPA Buffer

NEB

Cat# 9806S

Polybrene

Sigma-Aldrich

Cat# TR-1003-G

Purmorphamine

Merck Millipore

Cat# 540223-5MG

PVDF membran

BioRad

Cat# 1704157

Retinoic acid

Sigma-Aldrich

Cat# R2625

SB43152

Sigma-Aldrich

Cat# S4317

SCF

R&D Systems

Cat# P21583.1

SU-5402

Sigma-Aldrich

Cat# SML0443-5MG

TGF-b3

Peprotech

Cat# 100-36E-10

TGFß1

PeproTech

Cat# AF-100-21C-10

Tranylcypromine

Sigma-Aldrich

Cat# P8511-250MG

Triton X-100

Sigma-Aldrich

Cat# T8787

a,b-Methyleneadenosine 50 -triphosphate lithium salt

Sigma-Aldrich

Cat# M6517-5MG

Agencourt AMPure XP

Beckman Coulter

Cat# A63881

BD Cytofix/Cytoperm

BD Biosciences

Cat# 554714

Dneasy Blood and Tissue Kit

QIAGEN

Cat# 69506

SYBR Green Master Mix

Applied Biosystems

Cat# 4368708

GeneChip Mouse Genome 430 2.0 Array

Affimetrix

Cat# 900495

high capacity cDNA reverse transcription kit

Applied biosystems

Cat#4374966

Critical Commercial Assays

(Continued on next page)

Cell Stem Cell 24, 166–182.e1–e13, January 3, 2019 e3

Continued REAGENT or RESOURCE

SOURCE

IDENTIFIER

high Sensitivity DNA KIT

Agilent

Cat# 5067-4626

HumanHT-12 v4.0 Expression BeadChip

Illumina

Cat# BD-103-0604

Infinium HumanMethylation450K BeadChip

Illumina

discontinued

KAPA HiFi HotStart Ready Mix

Peqlab/VWR

Cat#KAPBKK2602 Cat# VPG-1004

Mouse Neural Stem Cell (NSC) Nucleofector Kit

Lonza Verviers

Nextera XT DNA Sample Preparation Kit

Illumina

Cat# FC-131-1096

Nextera XT Index Kit

Illumina

Cat# FC-131-1002

Nextera XT Index Kit v2 Set D

Illumina

Cat# FC-131-2004

PicoPure RNA Isolation Kit

Applied Biosystems

Cat# KIT0202

RNasin Plus RNase Inhibitor, 10,000u

Promega

Cat# N2615

Rneasy Mini Kit

QIAGEN

Cat# 74106

SMARTScribe Reverse Transcriptase

Takara Bio Clonetech

Cat# 639538

ERCC RNA Spike-In Mix

ThermoFisher

Cat# 4456740

Expression profiling of induced Neural Plate Border Stem Cells, Neural Border Stem Cell progeny and biological controls

This study, ArrayExpress

E-MTAB-5804

DNA Methylation profiling of iNBSCs, iPSC-derived NBSCs and adult dermal fibroblasts

This study, ArrayExpress

E-MTAB-5808

Expression profiling of primary murine Neural Plate Border Stem Cells, differentiated progeny and biological controls

This study, ArrayExpress

E-MTAB-5805

Single cell RNA-sequencing of primary Neural Plate Border Stem Cells

This study, ArrayExpress

E-MTAB-6925

Single cell RNA-sequencing of induced Neural Plate Border Stem Cells

This study, ArrayExpress

E-MTAB-6911

Single cell RNA-sequencing of E8.5 stage, embryonic neural tissue from mouse

This study, ArrayExpress

E-MTAB-5728

ThermoFisher Scientific

Cat# R70007

Deposited Data

Experimental Models: Cell Lines 293FT cell line Experimental Models: Organisms/Strains Mouse: C57BL/6J

The Jackson Labratory

Cat# 000664

Mouse: Gt(ROSA)26Sortm4(ACTB-tdTomato, -EGFP)Luo

The Jackson Labratory

Cat# 7576

Mouse: NOD.Prkdcscid.Il2rgnull

The Jackson Labratory

Cat# 5557

Oligonucleotides See Table S5 for list of primers SCN9A targeting sequence

N/A This study, GCAAAACAT TCGGAACGATTTGG

N/A

Recombinant DNA m2RTTa

Addgene

Cat# 20342

Cherry-Cre

(Scognamiglio et al., 2016)

N/A

phage-tetO-STEMCCA

(Sommer et al., 2009)

N/A

phage2-tetO-BKZS

this study

N/A

phage2-tetO-BKZS-flox

this study

N/A

pMD2.G

Addgene

Cat# 12259

psPAX2

Addgene

Cat# 12260

pSpCas9(BB)-2A-GFP vector (PX458)

Addgene

Cat# 48138 (Continued on next page)

e4 Cell Stem Cell 24, 166–182.e1–e13, January 3, 2019

Continued REAGENT or RESOURCE

SOURCE

IDENTIFIER

AGDEX package

Bioconductor, (Pounds et al., 2011)

https://bioconductor.org/packages/release/ bioc/html/AGDEX.html

biomaRt

Bioconductor

https://bioconductor.org/packages/release/ bioc/html/biomaRt.html

Broad Institute GSEA software

(Subramanian et al., 2007)

http://software.broadinstitute.org/ gsea/index.jsp

Clampfit

Molecular Devices

N/A

CoNTExT

(Stein et al., 2014)

http://context.semel.ucla.edu

CRISP-ID

(Dehairs et al., 2016)

http://crispid.gbiomed.kuleuven.be/

e1071

CRAN

https://cran.r-project.org/web/packages/ e1071/index.html

E-CRISP

(Heigwer et al., 2014)

http://www.e-crisp.org/E-CRISP/

Enrichr

(Chen et al., 2013)

http://amp.pharm.mssm.edu/Enrichr/

FlowCore

Bioconducor

https://bioconductor.org/packages/release/ bioc/html/flowCore.html

Fiji

ImageJ

https://imagej.net/Fiji/Downloads

Fitmaster

HEKA

N/A

GAGE

Bioconductor

https://bioconductor.org/packages/release/ bioc/html/gage.html

ggplot2

CRAN

https://ggplot2.tidyverse.org

Kallisto

(Bray et al., 2016)

https://github.com/pachterlab/kallisto

Leica Application Suite AF software

Leica

https://www.leica-microsystems.com/de/ produkte/mikroskop-software/details/ product/leica-application-suite/

M3Drop

Bioconductor

https://bioconductor.org/packages/ release/bioc/html/M3Drop.html

Monocle

Bioconductor, (Trapnell et al., 2014)

https://bioconductor.org/packages/ release/bioc/html/monocle.html

Python

Python Software Foundation

https://docs.python.org/2.7/

RnBeads

Bioconductor, (Assenov et al., 2014)

https://bioconductor.org/packages/ release/bioc/html/RnBeads.html

Rstudio

Rstudio

https://www.rstudio.com/products/ rstudio/download/

Scran

Bioconductor, (Lun et al., 2016)

http://bioconductor.org/packages/ release/bioc/html/scran.html

Stats

CRAN

https://stat.ethz.ch/R-manual/R-devel/ library/stats/html/stats-package.html

Software and Algorithms

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, Andreas Trumpp ([email protected]). EXPERIMENTAL MODEL AND SUBJECT DETAILS Derivation of human PBMCs and ADFs PBMCs were derived from healthy male donors (Age 24-30) and isolated using the Ficoll gradient procedure. PBMCs were used either directly or frozen in 90% Serum Replacement/ 10% DMSO as previously described (for details see (Sommer et al., 2012)). Adult dermal fibroblasts were derived from skin biopsies of healthy male donors (Age 24-30) as described in Meyer et al., 2015 (Meyer et al., 2015). ADFs were expanded until passage 4 and frozen in 90% Serum/ 10% DMSO prior to use. All cultures were grown at 37 C in 5% CO2 and 20% O2. Cells were derived with informed consent from all donors and handled in accordance with Ethics Committee II of Heidelberg University approval no. 2009-350N-MA. Cell Stem Cell 24, 166–182.e1–e13, January 3, 2019 e5

Culture of human fetal pancreas fibroblasts Human Primary Pancreatic Fibroblasts were obtained from Vitro Biopharma and cultured in MSC-Gro medium (Vitro Biopharma) supplemented with 10% FCS. Cells were expanded for 4 passages prior to use. Cells were grown at 37 C in 5% CO2 and 20% O2. Culture of iNBSCs and pNBSCs NBSC maintenance medium is composed of DMEM/ F-12 Glutamax (for iNBSCs) or Adv. DMEM/ F12 Glutamax (for pNBSCs) and Neurobasal Medium (1:1) containing 64 mg/ml LAAP, N2 Supplement (1:100, Life Technologies), B27 without RA (1:50, Life Technologies), 18 mg/ml Albumax I (Life Technologies) and Glutamax (1:200, Life Technologies), referred to as ‘basal medium’, and 4 mM Chir99021, 5mM Alk5 Inhibitor II and 0.5 mM Purmorphamine. iNBSCs and pNBSCs were cultured in maintenance medium on a layer of inactivated mouse embryonic fibroblasts (feeder) at 37 C in 5% O2 and 5% CO2. Cell lines were routinely split after 2-3 days (pNBSCs) or 4-5 days (iNBSCs) by treatment with Accutase (Life Technologies) and transferred onto fresh feeders. Cell lines were tested for mycoplasma, Squirrel monkey retrovirus, and Epstein-Barr virus contamination prior to analysis. Culture of human iPSCs Human iPSCs were derived from adult dermal fibroblasts obtained from a healthy male donor (Age 26) and routinely grown on feeder cells in DMEM/ F12, 64 mg/mL LAAP, 1x NEAA, 15% Serum Replacement and 20 ng/ml bFGF at 37 C in 20% O2 and 5% CO2. Mice Six- to twelve-week-old mice were used throughout the study. For isolation of pNBSCs embryos at different gestation stages were derived from Tomato-mice (STOCK Gt(ROSA)26Sortm4(ACTB-tdTomato,-EGFP)Luo/J; Jackson Laboratory, Stock No: 007576). Single-cell RNA sequencing data were obtained from E8.5 stage C57BL6/J mice after manual isolation of the mid-hindbrain area. Transplantations were performed into female NOD.Prkdcscid.Il2rgnull (NSG) mice. All mice were maintained at the DKFZ under specific pathogenfree (SPF) conditions in individually ventilated cages (IVCs). Animal procedures were performed according to protocols approved by €sidium Karlsruhe (Nr.Z110/02, DKFZ 299 and G184-13). the German authorities, Regierungspra METHOD DETAILS Identification of optimal growth and reprogramming conditions for induced NBSCs In order to identify the optimal conditions for iNBSC induction and maintenance, an iterative approach was chosen that comprised testing of different culture conditions on neural tissue from E8.5 mouse embryos and its subsequent test during neural reprogramming. Table S6 describes a selection of relevant molecules and coatings used during the exploratory screen. Table S6 contains a list of transcription factors tested for neural reprogramming of ADFs in combination with 4 mM Chir99021, 5mM Alk5-Inhibitor II, 0.5 mM Purmorphamine, and 5mM Tranylcypromine. Combinations that did not result in the expected reprogramming phenotype in the first trial have not been analyzed further. Reprogramming was performed as described below. Reprogramming into induced NBSCs For the production of lentiviral particles plasmids encoding for pHAGE2-TetO-BKSZ, pHAGE2-TetO-BKSZ-flox or m2RTTA ((Hockemeyer et al., 2008), Addgene plasmid #20342) were transfected together with helper plasmids psPAX2 (Addgene plasmid #12260) and pMD2.G (Addgene plasmid #12259) into 293FT cell lines (Life Technologies) as described elsewhere (Brambrink et al., 2008). Lentiviral supernatants containing BKSZ or BKSZ-flox and m2RTTA were mixed in a ratio of 2:1, supplemented with 5 mg/ml polybrene (Sigma) and used freshly for transduction of 8x105/ 6 Well primary ADF and pancreas fibroblasts. For reprogramming of PBMCs the lentiviral supernatant was concentrated via ultracentrifugation and PBMCs were transduced in QBSF-60 Stem Cell Medium (Quality Biological) containing 50 mg/ml Ascorbic Acid (Sigma), 50 ng/ml SCF (R&D Systems), 10 ng/ml IL-3 (R&D Systems), 2 U/ml EPO (R&D Systems), 40 ng/ml IGF-1 (Peprotech), 1 mM Dexamethasone (Sigma) and 5 mg/ml polybrene (for details see (Sommer et al., 2012)). The other day 2x105 transduced PBMCs were transferred onto one well of a 6 Well plate, coated with inactivated MEFs and reprogramming was initiated one day thereafter. For reprogramming transduced cells were cultured in DMEM/F-12 Glutamax (Life Technologies) containing 64 mg/ml L-Ascorbic acid 2-phosphate (LAAP, Sigma), ITS-X (1:100, Life Technologies), NEAA (1:100, Life Technologies), 2% FCS, 8% Serum Replacement (Life Technologies) supplemented with 4 mM Chir99021 (Sigma), 5 mM Alk5 Inhibitor II (Enzo Life Sciences), 0.5 mM Purmorphamine (Sigma), 5mM Tranylcypromine (Sigma) and 1mg/ml Doxycycline (Sigma) and incubated at 37 C in 5% O2 and 5% CO2. During reprogramming, the medium was changed every other day. When first colonies became visible or latest after 19 days, the medium was changed to NBSC maintenance medium. NBSC maintenance medium is composed of DMEM/ F-12 Glutamax (for iNBSCs) or Adv. DMEM/ F12 Glutamax (for pNBSCs) and Neurobasal Medium (1:1) containing 64 mg/ml LAAP, N2 Supplement (1:100, Life Technologies), B27 without RA (1:50, Life Technologies), 18 mg/ml Albumax I (Life Technologies) and Glutamax (1:200, Life Technologies), referred to as ‘basal medium’, and 4 mM Chir99021, 5mM Alk5 Inhibitor II and 0.5 mM Purmorphamine.

e6 Cell Stem Cell 24, 166–182.e1–e13, January 3, 2019

Once visible, distinct colonies were manually picked and cultured in NBSC maintenance medium on a layer of inactivated mouse embryonic fibroblasts (feeder) at 37 C in 5% O2 and 5% CO2. iNBSC lines were routinely split after 4-5 days by treatment with Accutase (Life Technologies) and transferred onto fresh feeders. Cre-mediated deletion of BKSZ flox For the excision of the transgene cassette, 5x105 iNBSCs were seeded onto fresh feeders and transfected with a plasmid encoding for a Cherry-Cre as previously described (Scognamiglio et al., 2016). 48 hours later cells were harvested and sorted for Cherry fluorescence. 5000 cherry positive cells were seeded onto a 10 cm dish coated with feeders and incubated until colonies became visible. Single colonies were picked and checked for transgene removal by transgene-specific PCRs on genomic DNA. Differentiation of human iPSCs into NBSCs Human iPSCs were routinely grown on feeder cells in DMEM/ F12, 64 mg/mL LAAP, 1x NEAA, 15% Serum Replacement and 20 ng/ml bFGF. Prior to the differentiation into NBSCs, human iPSCs were cultured for one passage on Matrigel-coated plates in basal medium supplemented with 20 ng/ml bFGF and 1 ng/ml TGFß1. When cells reached confluency, iPSCs were harvested by treatment with 1 mg/ml Collagenase II (Life Technologies) and cell clumps were transferred into uncoated Petri dishes containing NBSC maintenance medium (CAP). At this point, the culture was switched from 20% O2 to 5% O2 and subsequently grown under this condition. After five days spheres were plated onto feeder cells and grown in NBSC maintenance medium (CAP). The other day, attached spheres were split by treatment with Accutase and 3x 104 cells were seeded on feeder containing 10 cm dishes. Thereafter, single colonies were manually picked and clonal lines established. Differentiation of iNBSCs toward RG-like cells NBSCs were seeded on Matrigel-coated plates and cultured in basal medium supplemented with 1 mM Purmorphamine and 10 ng/ml FGF8 for one week, followed by culture in basal medium with 1 mM Purmorphamine for one additional day. Thereafter, cultures were grown in basal medium containing 10 ng/ml BDNF and 10 ng/ml GDNF for 7 more weeks. Subsequently, cells were cultured in radial glia medium, comprised of basal medium, 20 ng/ml bFGF, 20 ng/ml EGF and 10 ng/ml LIF. When proliferative, RG-like cells became apparent, cultures were treated with Accutase and expanded on Matrigel-coated plates in radial glia medium. Finally, cultures were enriched for RG-like cells by cell sorting for CD133/2, SSEA1 and GLAST. All cultures were grown at 37 C, 5% CO2 and 20% O2. Differentiation of iNBSCs toward NCSC-like cells To initiate neural crest differentiation 1x105/ 6 Well iNBSCs were seeded on Matrigel-coated plates and grown in basal medium supplemented with 4 mM Chir99021, 5 mM Alk5 Inhibitor II and 10 ng/ml BMP4 (Peprotech) (CAB) for three days. Thereafter, the medium was switched to basal medium supplemented with 4 mM Chir99021, 10 ng/ml FGF8 (Peprotech), 10 ng/ml IGF1 (Peprotech) and 10 mM DAPT (Sigma) (adapted from (Noisa et al., 2014)). Five days later NCSC-like cells were purified by cell sorting for SSEA-1negCD133negP75+HNK1+. All cultures were grown at 37 C, 5% CO2 and 20% O2. Differentiation of iNBSCs toward MSCs and neural crest progeny To derive mesenchymal crest cells, iNBSCs were first seeded on Matrigel-coated plates and cultured in 4 mM Chir99021, 10 ng/ml BMP4 and 10 mM DAPT for 7 days. Thereafter, cells were cultured in basal medium containing 10 ng/ml bFGF and 10 ng/ml IGF-1 for > 5 passages. Subsequently, MSC-like cells were stabilized by switching cultures to mesenchymal stem cell medium (GIBCO). MSC-like cells were cultured for > 2 passages prior to analysis. In order to derive mature mesenchymal crest progeny, neural crest-primed cultures were treated for 5 days in mesenchymal induction medium comprising DMEM/ F12, 64 mg/ml LAAP and 10% FCS. Thereafter, cells were cultured in StemPro Adipogenesis Differentiation Kit (Life Technologies), StemPro Chondrocyte Differentiation Kit (Life Technologies) or kept in mesenchymal induction medium to generate adipocytes, chondrocytes or smooth muscle, respectively. To induce sensory neurons, iNBSCs were grown on MG-coated plates in basal medium in presence of 3 mM Chir99021, 10 mM DAPT (Sigma) and 10 mM SU5402 (Sigma) for 10 days. Subsequently, cultures were grown in maturation medium comprising basal medium, 10 ng/ml BDNF, 10 ng/ml GDNF, 10 ng/ml NT-3 (Peprotech) and 25 ng/ml NGF (Peprotech) for at least another three weeks (Chambers et al., 2012). Differentiation of iNBSCs toward mature CNS progeny To induce neuronal differentiation iNBSCs were seeded on MG-coated plates and iNBSCs maintenance medium was switched to neural induction medium containing 1 mM Purmorphamine (undirected differentiation), 1 mM Purmorphamine and 10 ng/ml FGF8 (dopaminergic differentiation) or 1 mM Purmorhamine and 1 mM all-trans retinoic acid (Sigma) (motoneuron differentiation) for one week (adapted from ((Reinhardt et al., 2013)). Serotonergic differentiation was initiated by culture of iNBSCs in basal medium, 3 mM Chir99021, 3 mM SB431542 and 1 mM Purmorphamine for one week, followed by culture in 3 mM Chir99021, 3 mM SB431542, 1 mM Purmorphamine and 10 ng/ml FGF4 for another week and finally switching to 1 mM Purmorphamine for two days. Subsequent to neural induction, cultures were grown in neuronal maturation medium comprising basal medium, 500 mM dbcAMP (Sigma), 1 ng/ml TGFß3, 10 ng/ml BDNF and 10 ng/ml GDNF for at least 5 more weeks. Astrocytes and oligodendrocytes could be found within neuronal cultures, starting after 5 weeks of differentiation. Cell Stem Cell 24, 166–182.e1–e13, January 3, 2019 e7

3D Differentiation of iNBSCs In order to differentiate iNBSCs as three-dimensional cultures, a single cell suspension of 2x104 cells was resuspended in 10 ml of basal medium and mixed with 150 ml of Matrigel on ice. Next, the mix was distributed on 5 coverslips (1x 30 ml drop per coverslip) and incubated at 37 C for 6 minutes. After gelling the differentiation medium was applied onto the embedded cells. For CNS-primed differentiation, embedded iNBSCs cultures were first treated with basal medium supplemented with 3 mM SB431542 and 1 mM Purmorphamine for nine days. Thereafter cultures were switched to basal medium supplemented with 3 mM Chir99021, 3 mM SB431542, 1 mM Purmorphamine and 20 ng/ml bFGF and cultured for 4-6 days. For neural crest-primed differentiation, embedded iNBSC cultures were cultured in basal medium supplemented with 4 mM Chir99021, 5 mM Alk5-Inh., 10 ng/ml BMP4 and 20 ng/ml bFGF for at least 12 days. Cultures were either fixed with 4% paraformaldehyde and analyzed by confocal microscopy or total RNA was extracted using the ARCTURUS PicoPure RNA Isolation Kit. Transplantation of iNBSCs Prior to transplantation iNBSCs were allowed to initiate differentiation by cultivation in basal medium supplemented with 1 mM Purmorphamine and 10 ng/mL FGF8 on MG coated plates for 8 days. On the day of transplantation primed cultures were dissociated to a single cell suspension and resuspended in medium at a concentration of 5x104 cells per ml. 6 to 8 weeks old female NSG mice were anesthetized with isoflurane, mounted in a stereotactic apparatus and kept under isoflurane anesthesia during surgery. 3 ml of neural cell suspension was bilaterally transfused into the striatum using a glass micropipette. The following coordinates were used for transplantation: from bregma and the brain surface, anterior/posterior: 0 mm; medial/lateral ± 2.5; dorsal/ventral
2.5 mm. The scalp incision was sutured, and post-surgery analgesics were given to aid recovery (0.03 mg/kg KG Metamizole). Electrophysiological experiments were performed 8-12 weeks after the treatments. No statistical methods were used to estimate sample size and mouse experiments were not randomized nor blinded. Derivation of pNBSCs Male and female tomato mice were paired and checked for vaginal plug the following morning. Positive plug test was considered 0.5 days post coitum. Embryos were collected at day 8.5 post coitum and optic and non-neural tissues were removed mechanically. Neural tissue was digested with Accutase and the resulting single cell suspension was seeded onto a layer of mouse embryonic fibroblasts and cultured in NBSC maintenance medium in 5% CO2, 5% O2 at 37 C. Two to three days after seeding, single colonies were mechanically picked and clonal pNBSC lines established. pNBSC lines were routinely split every three days by treatment with Accutase and seeded onto fresh feeder cells. pNBSCs could be maintained in culture for > 40 passages (> 4 months). Differentiation of pNBSCs In order to differentiate pNBSCs toward the CNS, pNBSCs were seeded on Matrigel-coated plates and cultured in basal medium supplemented with 1 mM Purmorphamine for 3 days. Rosette-like structures could be observed 2-4 days after differentiation was initiated. RG-like SCs were stabilized by switching culture medium of rosette-like cells to basal medium, supplemented with 20 ng/ml bFgf and 20 ng/ml Egf. Cells were expanded for > 3 passages before RG-like SCs were further enriched by FACS sorting for Ssea1+Glast+. Mature progeny such as neurons, astrocytes, and oligodendrocytes were derived by culturing pNBSCs in basal medium supplemented with1 mM Purmorphamine for 3 days and subsequent switch to basal medium containing 10 ng/ml Bdnf and 10 ng/ml Gdnf for > 3 weeks. To derive NCSC-like cells, 1x104 pNBSCs were seeded onto a Matrigel-coated 6 well plate and cultured in basal medium supplemented with 4 mM Chir99021 and 10 ng/ml Bmp4 for 3 days. Thereafter, cells were cultured in basal medium supplemented with 10ng/ml bFgf and 10 ng/ml Egf for 4 days and enriched for NCSC-like cells by FASC sorting for P75+Glast-Ssea1-. Oil red positive adipocytes were obtained by culture of pNBSCs in 4 mM Chir99021 and 10 ng/ml Bmp4 for 3 days, followed by addition of 10 ng/ml bFGF to the medium for 2 days and switch to basal medium supplemented with 10 ng/ml bFGF and 10 ng/ml Egf for another 3 weeks. Chondrocytes and smooth muscle cells were derived by culturing pNBSCs in 4 mM Chir99021 and 10 ng/ml Bmp4 for 3 days, followed by addition of 10 ng/ml bFgf to the medium for 2 days and subsequent switch to basal medium supplemented with 10% FCS. Peripheral neurons were differentiated from pNBSCs by culture in Chir99021 and 10 ng/ml Bmp4 for 3 days, followed by switching to basal medium supplemented with 10 ng/ml Bdnf, 10 ng/ml Gdnf, and 10 ng/ml Ngf for two weeks. Derivation of primary RG from mouse embryos Male and female tomato mice were paired and checked for vaginal plug the following morning. Positive plug test was considered 0.5 days post coitum. Embryos were collected at day 13.5 post coitum and medial and lateral ganglionic eminences were mechanically isolated. Neural tissue from individual embryos was digested by treatment with Accutase. The resulting single cell suspension was transferred to Petri dishes and cultured in basal medium supplemented with 20 ng/ ml bFgf and 20 ng/ml Egf for 5 days. The resulting spheres were plated onto fibronectin-coated cell culture dishes and expanded for at least 3 passages before FACS sorting for Glast+Ssea1+ cells was performed. Glast+Ssea1+ RG-type stem cells were either used directly for expression profiling or plated onto fibronectin for additional downstream analysis.

e8 Cell Stem Cell 24, 166–182.e1–e13, January 3, 2019

3D Differentiation of pNBSCs A single cell suspension of 2x104 pNBSCs was resuspended in 10 ml of basal medium and mixed with 150 ml of Matrigel on ice. Next, the mix was distributed on 5 coverslips (1x 30 ml drop per coverslip) and incubated at 37 C for 6 minutes. After gelling, the differentiation medium was applied onto the embedded cells. For CNS-primed differentiation, embedded pNBSCs cultures were treated with basal medium supplemented with 4 mM Chir99021, 3 mM SB431542 and 1 mM Purmorphamine for six days until neural tube-like structures formed. For neural crest-primed differentiation, embedded pNBSC cultures were cultured in basal medium supplemented with 4 mM Chir99021, 5 mM Alk5-Inh., 10 ng/ml BMP4 and 20 ng/ml bFGF for at least 10 days. Slice preparation for electrophysiology Electrophysiological recordings were performed from 6 to 12 weeks old female mice. We recorded from acute coronal slices (300 mm) containing the striatum. Mice were deeply anaesthetized with inhaled isoflurane, followed by transcardially perfusion with 30 mL ice-cold sucrose solution containing (in mM) 212 sucrose, 26 NaHCO3, 1.25 NaH2PO4, 3 KCl, 7 MgCl2, 10 glucose and 0.2 CaCl2, oxygenated with carbogen gas (95% O2/ 5% CO2, pH 7.4). Sections were cut in ice-cold oxygenated sucrose solution, followed by incubation in oxygenated extracellular solution containing (in mM) 12.5 NaCl, 2.5 NaHCO3, 0.125 NaH2PO4, 0.25 KCl, 2 CaCl2, 1 MgCl2 and 25 glucose. Cells in the striatum were visualized with DIC optics and epifluorescence was used to detect GFP fluorescence. Whole-cell recordings Whole-cell patch-clamp recordings were performed at 30 to 32 C bath temperature. Individual slices or coverslips were placed in a submerged recording chamber mounted on an upright microscope (Olympus BW-X51) and continuously perfused with an oxygenated extracellular solution. Recording pipettes were pulled from borosilicate glass capillaries and had tip resistances of 5-8 MU. Liquid junction potentials were not corrected. Series resistance was maximally compensated and continuously monitored during the recordings. Cells were discarded if no ‘‘Giga seal’’ was initially obtained or series resistance changed more than 20% or was higher than 40 MU. The following intracellular solutions were used: low Cl- potassium-based solution containing (in mM) 130 K-gluconate, 10 Nagluconate, 10 HEPES, 10 phosphocreatine, 4 NaCl, 4 Mg-ATP and 0.3 GTP, pH adjusted to 7.2 with KOH for firing patterns. High Clsolution containing (in mM) 127.5 KCl, 11 EGTA, 10 HEPES, 1 CaCl2, 2 MgCl2 and 2 Mg-ATP (pH 7.3) for spontaneous activity in cell culture. Cs+-based solution containing (in mM) 120 Cs+-gluconate, 10 CsCl, 10 HEPES, 0.2 EGTA, 8 NaCl, 10 phosphocreatine, 2 Mg-ATP and 0.3 GTP, pH 7.3 adjusted with CsOH for spontaneous activity in transplanted neurons. For subsequent morphological and immunocytochemical characterization of patched cells, biocytin (circa 10 mg/ ml; Sigma) was added to the respective intracellular solution. Cells were initially kept in cell-attached mode. After achieving a ‘‘Giga seal,’’ whole-cell configuration was established and firing patterns were analyzed in current-clamp mode at resting membrane potential by applying 1 s current pulses, starting from
30 pA in 5 pA steps until maximal firing frequency was reached for cell culture and from
200 pA in 20 pA steps for acute slices. Individual traces upon
30 pA/-200 pA current injection, at action potential threshold (for intermediated excitatory cell types), were selected for illustration of firing pattern. Spontaneous activity of the neurons was recorded at a holding potential of
70 mV. All recordings were made using HEKA PatchMaster EPC 10 amplifier and signals were filtered at 3 kHz, sampled at 10 or 20 kHz. Data analysis was done offline using HEKA software FitMaster and Clampfit (Molecular Devices, USA). Cell identification and reconstruction Acute slices with biocytin-filled cells in the MEC were fixed overnight in 4% paraformaldehyde, followed by extensive washing with PBS. For morphological reconstructions, biocytin-filled MEC cells were identified via 3,30 -diaminobenzidine (DAB) staining. Sections quenched in 1% H2O2 for 5 minutes. After renewed washing, sections were permeabilized in PBS with 1% Triton X-100 for 1 hr. Subsequently, sections were incubated with avidin-biotin-horseradish-peroxidase complex in PBS for 2 hours at room temperature. Following washing in PBS, sections were developed in DAB and mounted in Mowiol. Labeled cells were reconstructed using the Neurolucida (MBF bioscience, Willston, VT, USA) tracing program. Calcium Imaging Cell cultures were incubated in a solution containing the cell-permeable fluorescent Ca2+ indicator Fluo-3 AM (2 mM; F1242, Thermo ScientificTM) and Cell Tracker Red CMTPX (1 mM; C34552, Thermo ScientificTM) to delineate cell membranes. Dye loading was carried out in an incubator at 37 C, 5% CO2 for 30 min. Imaging of fluorescently labeled cells was performed on a TCS SP5 microscope (Leica) equipped with a 20x (1 numerical aperture) water-immersion objective. Images (512x512 pixels) were acquired at 1000 Hz speed every 0.8-1.6 s with 0.5 mm per pixel resolution in the xy dimension, and 3-4 mm steps in the z dimension. Argon and HeNe-543 lasers were used to excite Fluo-3 AM and Cell Tracker Red CMTPX dyes, respectively. Artificial cerebrospinal fluid (ACSF) containing (in mM): 120 NaCl, 3.5 KCl, 2.5 CaCl2, 1.3 MgSO4, 1.25 NaH2PO4, 25 NaHCO3, 25 glucose (pH 7.2) was applied by a pump perfusion system with a constant flux (1.5 ml/min) that Cell Stem Cell 24, 166–182.e1–e13, January 3, 2019 e9

continuously renewed the buffer in the recording chamber. After recording the baseline fluorescence for 2 minutes, a,b Methylene ATP was applied (30 mM in ACSF; Sigma) for 2 minutes. Recordings continued for up to 10 minutes in total. We added a non-nucleotide P2X3 and P2X2/3 receptor antagonist (1 mM; A-317491, Sigma) together with the dye loading and were also present in ACSF during the whole recording. Leica Application Suite AF software and FIJI software was used to record and measure fluorescent activity, respectively, in 3-5 independent experiments per group. We obtained relative fluorescence changes (Fi/F0), where F0 is the fluorescence image formed by averaging the first 50 frames of the sequence, and F(i) represents each (i) frame of the recording. Curves were normalized by subtracting a linear regression line fitted through the first and last 50 values and maximum peak intensity was aligned at respective time points were applicable (WT measurements). We then studied global calcium activity by averaging fluorescence intensity in the whole image before and in response to stimulation with a,b Methlynene ATP in WT- (n = 3) and SCN9A
/
(n = 4) neuronal cultures. Statistical analyses were performed using Prism 6. Differences between groups were examined using Student’s t test. Values of p < 0.05 were considered statistically significant for the rest of the analyses. Flow Cytometry Cells were harvested by treatment with Accutase and washed twice with unsupplemented DMEM/ F12. Thereafter cells were resuspended in medium and stained with for 30 minutes at 4 C using the following antibodies: anti-SSEA1 (MC480)-V450 (Becton Dickinson), anti-CD133/2 (293C3)-FITC (Miltenyi Biotec), anti-CD271 (ME20.4-1.H4)-PE (Miltenyi Biotec), anti-GLAST (ACSA-1)-APC (Miltenyi Biotec), anti-HNK1 (TB01)-PeCy7 (eBioscience), ms anti-HNK1 (clone VC1.1, Sigma), Donkey anti-Mouse Alexa Fluor 488 (Abcam, ab150105), anti-CXCR4 (12G5)-PeCy5 (BioLegend), anti-CD44(G44-26)-PE (BD Bioscience), anti-CD105 (43A3)FITC (BioLegend), anti-CD90 (5E10)-BV421 (BioLegend), anti-CD146 (P1H12)-Alexa647 (Biolegend), anti-CD13 (WM15)-APCCy7 (BioLegend). FACS analysis was performed on LSRII or LSR Fortessa flow cytometers (Becton Dickinson, San Jose, CA). Data were analyzed using the FlowJo software (Tree Star, Ashland, OR). For intracellular FACS staining cells were detached using Accumax (Sigma). After centrifugation cells were resuspended in Cytofix/ Cytoperm solution (BD Biosciences) and incubated for 10 min on ice. Subsequently, cells were washed three times in Perm/Wash buffer (BD Biosciences) followed by staining with primary antibodies in the same buffer. Antibodies used in this study comprised: anti-CHAT (AB144P; Merck Millipore), anti-FOXA2 (AF2400, R&D Systems), anti-GATA3 (MAB6330, R&D Systems), anti-NURR1 (sc-376984, Santa Cruz), anti-TH (ab112, Abcam), anti-TPH2 (NB100-74555, Novus Biologicals), anti-TUJ1 (PRB-435P, Covance), anti-TUJ1 (ab41489, Abcam), anti-chicken-488 (ab150169, Abcam), anti-mouse-488 (ab150117, Abcam), anti-goat-488 (ab150129, Abcam), anti-rabbit-555 (ab150074, abcam), and anti-mouse-647 (ab150115, Abcam). Primary antibodies were applied for 1 h at 4 C and secondary antibodies were applied for 45 min at 4 C. When stainings comprised the directly coupled antibody against IAP (CC2C6, BioLegends), the latter was applied separately after staining with secondary antibody had been completed. Finally, cells were stained with DAPI (ThermoFisher) to exclude fragmented cells and duplets. Cell sorting experiments were carried out on a BD FACSAria III sorter (Becton Dickinson, San Jose, CA). The following sort parameters were used: 100 mm nozzle; up to 2000 events/second. Single cell sorting of iNBSCs and pNBSCs was performed in ‘‘single cell’’ mode in conjunction with the recording of the index information for each well. Cells were sorted onto 96 Well plates coated with feeder cells and colony formation was monitored from 10 days onward after sorting. Clonogenicity was determined for 6 independent iNBSC and 3 independent pNBSC lines, in which R 2x 96 Well plates were analyzed for each cell line. Immunofluorescence Cells were fixed in PBS with 4% paraformaldehyde (Electron Microscopy Sciences, 19208) for 15 minutes. Fixed cells were then washed three times with PBS and blocked for one hour in PBS containing 0.1% Triton X-100 and 1% BSA. Primary antibodies were applied in 0.1% Triton X-100 and 1% BSA at 4 C overnight. For the list of primary antibodies used in this study see Key Resources Table. Subsequent to three times washing with PBS, cells were incubated with secondary antibodies for two hours at room temperature. Following DAPI (Sigma, D9542) staining, cells were mounted (DAKO, S3035) and analyzed on an LSM 710 ConfoCor 3 confocal microscope (Zeiss). Immunohistochemical analysis For IHC analysis, transplanted mouse brains were fixed with 4% paraformaldehyde overnight and consecutively dehydrated using 20% sucrose in PBS. Thereafter, brains were embedded in 3% agarose and 100 mm coronal sections were prepared using a Leica VT1000S sliding microtome. Brain slices were blocked in 3% goat serum, 0.25% Triton-X in TBS for one hour at 4 C. Primary antibodies (see Key Resources Table) were applied in 3% goat serum, 0.25% Triton-X in TBS for 72 hours at 4 C. After washing slices in TBS three times, secondary antibodies were incubated in 3% goat serum, 0.25% Triton-X in TBS for two hours at room temperature. Following DAPI staining, brain slices were mounted and analyzed on a LSM 710 ConfoCor 3 confocal microscope (Zeiss).

e10 Cell Stem Cell 24, 166–182.e1–e13, January 3, 2019

Oil Red o staining Adipocyte differentiations were fixed in 4% paraformaldehyde for 15 minutes at room temperature. After washing twice with ddH20, cells were incubated with 60% isopropanol for 5 minutes and completely dried afterward. Fresh Oil Red staining solution, consisting of 3.5 mg/ml Red O (Sigma) in 60% isopropanol, was applied on cells for 10 minutes at room temperature. After washing cells 4 times with ddH2O microscope images (Nikon Eclipse Ti-E) were acquired. Alcian blue staining Chondrocyte differentiations were fixed in 4% paraformaldehyde for 15 minutes and rinsed three times with PBS. Alcian blue solution, consisting of 10 mg/ml Alcian Blue GX (Sigma) in 3% acetic acid, was applied for 1 hour and rinsed twice with 0.1 M HCl. After washing cells twice with PBS microscope images (Nikon Eclipse Ti-E) were acquired. Western Blot WT and SCN9A
/
iNBSCs were differentiated into sensory neurons for at least three weeks before whole cell lysates were prepared using RIPA buffer (Cell Signaling Technology), 1 mM PMSF (Sigma), 1 mM EDTA and Halt Protease-Phosphates Inhibitor Cocktail (Pierce). After denaturing lysates in 5% SDS at 95 C, protein samples were resolved on 4%–12% TGX gels (Criterion, Bio-Rad) with TGS (Tris-Glycine-SDS) running buffer (Bio-Rad) and blotted onto PVDF membranes (Trans-Blot TURBO, Bio-Rad). Membranes were blocked with TBS containing 0.3% (vol/vol) Tween-20 and 5% (wt/vol) BSA powder for 1 hour. Primary antibodies (see Key Resources Table) were incubated overnight at 4 C on a shaker. After thorough washing, Secondary HRP-coupled antibodies were incubated in TBS containing 0.3% Tween-20 for 1 hour at RT. Membranes were washed and immunocomplexes were visualized using the ECL kit (Amersham International). CRISPR/Cas9-mediated knockout Guide RNAs were ordered as DNA oligos (Sigma; for sequence see Key Resources Table) and cloned into the pSpCas9(BB)-2A-GFP vector (PX458) (A gift from Feng Zhang, via Addgene) as described elsewhere (Ran et al., 2013). 0.5-2x106 iNBSCs were nucleofected with 0.5-2mg of plasmid DNA using Nucleofector Kits for Mouse Neural Stem Cells (Lonza) with program A-033 and seeded on fresh feeder cells. Transfected cultures were harvested after 48 hours, sorted for GFP fluorescence and plated onto feeders. Individual colonies were manually isolated 5-7 days later. Genomic DNA was extracted using the Dneasy Blood and Tissue Kit (QIAGEN) and guide RNA target sites were amplified and analyzed by Sanger sequencing (GATC). Biallelic sequences were deconvoluted using CRISP-ID (Dehairs et al., 2016). Gene Expression Analysis by Quantitative PCR Total RNA was isolated using the ARCTURUS PicoPure RNA Isolation Kit (Life Technologies, Invitrogen) including on-column DNA digestion (QIAGEN, 79254). Reverse transcription was performed using the high capacity cDNA synthesis kit (Applied Biosystems). Real-time quantitative PCRs were run in ABI Power SYBR Green Mastermix (Applied Biosystems, 4368708) on a ViiA7 machine. Results were analyzed using the ViiA7-software V1.2.4. Expression was normalized to the housekeeping gene GAPDH (human) or Oaz1 (mouse). All PCR reactions were carried out as technical triplicates. Samples showing low RNA quality and/ or detection of the housekeeping gene for amplification cycles > 25 were excluded from the study. For primers used in this study see Table S5. Microarray analysis Total RNA was isolated using the ARCTURUS PicoPure RNA Isolation Kit including on-column DNA digestion. RNA expression profiling using the human HT-12 v4 Expression BeadChip (Illumina) or Mouse 430 V2.0 GeneChips (Affymetrix) was performed according to the manufacturer’s instructions at the DKFZ Genomics and Proteomics Core Facility (Heidelberg, Germany). Raw data were obtained from DKFZ Genomics and Proteomics Core Facility and merged with publically available expression data from the Gene Expression Omnibus database. Datasets used in this study: GSE40838 GSM1003015 GSM1003016 GSM1003021 GSM1003022 GSM1003027 GSE51980 GSM1256637 GSE69486 GSM1701950

Cell Stem Cell 24, 166–182.e1–e13, January 3, 2019 e11

GSE34904 GSM857336 GSM857337 The whole dataset was quantile normalized (self-written function) and log2-transformed. Principal component analyses were computed using the PCA function from sklearn.decomposition and visualized (seaborn). Gene expression heatmaps were visualized as clustered heatmap (seaborn, clustermap) and show z-transformed expression values unless indicated otherwise. In order to generate PCA-based expression heatmaps, the 200 probes showing highest contribution were extracted applying the formula sqrt(PC1+PC2). Differentially expressed genes between two samples were identified using ttest_ind from the scipy.stats package, p values were corrected using the Benjamini-Hochberg false discovery control (stats, R-package). GO analysis and gene set enrichment analysis were performed using STRING Version 10.0 (Szklarczyk et al., 2015), EnrichR (Chen et al., 2013), and the Broad Institute GSEA software (Subramanian et al., 2007). In order to map neural populations with the spatiotemporal atlas of the human brain, iNBSCs-derived RG-like SCs and iNBSCs were submitted to the online tool of the machine-learning framework CoNTExT (https://context.semel.ucla.edu/instructions, (Stein et al., 2014)). Data were uploaded as suggested by the authors, using an identifier file and the respective subset of identifiers from the expression data. To compare human and mouse data, iNBSCs/ADFs and pNBSCs/Mefs were analyzed using the AGDEX package (R, (Pounds et al., 2011)) using homology information from BioMart (R). Log2 expression fold changes were correlated and genes that were differentially expressed in both comparisons (p < 0.05) to an absolute fold change of > 1 were analyzed for GO enrichment. Methylation analysis Genomic DNA was extracted using the DNeasy Blood and Tissue Kit (QIAGEN) according to the manufacturer’s instructions. DNA methylation profiling using the Illumina Infinium HumanMethylation450 BeadChip array was performed according to the manufacturer’s instructions at the DKFZ Genomics and Proteomics Core Facility (Heidelberg, Germany). Reference data for hESCs and human ADFs were obtained from the Gene Expression Ominbus database: GSE52025 GSM1257669: GM02704 GSM1257670: GM02706 GSM1257671: GM01650 GSM1257672: GM01653 GSE61461 GSM1505345 B105-ES GSM1505346 B152-ES GSM1505347 B160-ES GSM1505348 B209-ES GSM1505349 B220-ES GSM1505350 B312-ES All methylation data was processed using the minfi package from the Bioconductor suite. Probes with a detection p value < 0.01 or coinciding with known SNPs and all probes on X- and Y chromosomes were excluded. The dataset was normalized using the preprocessIllumina function and visualized using classical multidimensional scaling. For differential methylation analysis, promoter methylation was analyzed using the RnBeads (Assenov et al., 2014) package (Bioconductor). Promoters were ranked based on their combined rank statistic (RnBeads) and ordered from most significantly hyper- to most significantly hypomethylated and used as input for GSEA with a gene set file reduced to genes covered by the 450k array. GSEA was run with the ‘‘classical’’ enrichment statistic (Subramanian et al., 2007). Single-cell RNA sequencing protocol Single-cell RNA sequencing was performed following the Smart-seq2 protocol (Picelli et al., 2014), with adaptation as outlined below. First, cells were harvested by treatment with Accutase and stained as aforementioned (iNBSCs) or used directly (E8.5 embryos, pNBSCs). Cell sorting was carried out on a BD FACSAria III sorter (Becton Dickinson, San Jose, CA). The following sort parameters were used: 100mm nozzle; up to 800 events/second. Single cells were sorted into lysis buffer (2.86ml H2O, 1ml dT30VN (10mM), 1ml dNTPs (10mM), 0.1ml RNase inhibitor, 0.04ml Triton X-100, 0.081ml ERCC spike-ins) in separate 96 wells and flash-frozen in liquid N2 until further processing. Plates were thawed on ice, incubated at 72 C for 3 minutes and reverse transcribed using the SmartScribe RT Kit (Takara Bio Clontech): 2ml Smart FS buffer, 0.5ml DTT, 0.1ml (100mM TSO), 1ml SMARTScribe reverse transcriptase, 1.4ml H2O per cell. PCR preamplification and cDNA purification was performed as described by Picelli and colleagues (Picelli et al., 2014), applying 19 amplification cycles.

e12 Cell Stem Cell 24, 166–182.e1–e13, January 3, 2019

For tagmentation 0.15-0.2 ng cDNA was used per reaction: 1.25ml diluted sample, 2.5ml Tagment DNA buffer, and 1.25ml Amplification Tagment Mix. The reaction was neutralized with 1.25mL NT buffer. The adaptor ligation was performed with 5ml of the neutralized tagmentation reaction, 3.75ml of Nextera PCR master mix and 1.25ml of Index-Primer 1 and Index-Primer 2 and amplified as described in the Smart-seq2 protocol. Finally, 192 reactions were pooled (1ml per reaction) and purified according to the Smartseq2 protocol. Quality of pre-amplified cDNA and final libraries was controlled by analysis of fragment size distribution using the Bioanalyzer highsensitivity DNA chip (Agilent). DNA sequencing was performed on a HiSeq2500 device (Illumina) with 200 cells per lane. Raw data processing and quality control Reads were demultiplexed and gene expression was quantified using pseudo-alignment by kallisto (v0.43, bioconductor), using the reference files from http://www.ensembl.org//useast.ensembl.org/info/data/ftp/index.html?redirectsrc=//www.ensembl.org% 2Finfo%2Fdata%2Fftp%2Findex.html (Homo_sapiens.GRCh38.cdna.all.fa and Mus_musculus.GRCm38.cdna.all.fa) and adding ERCC spike-in sequences. The FACS-index intensity information was recovered from fcs-files using flowCore (bioconductor) and was further compensated and logicle-transformed. ScRNaseq data was filtered according to the following criteria: Cells were required to have > 100000 reads, > 10000 detected transcripts and < 10% reads mapping to ERCC spike-ins. Transcripts were kept if they were detected in at least 10 cells. For all analyses, transcript-level expression was collapsed to gene-level expression and library-size normalization was performed according to Lun et al. (Lun et al., 2016). After quality control, 310 out of 336 cells remained for the iNBSC-, 73 out of 96 for the pNBSC- and 173 out of 191 for the embryo dataset. Single-cell RNA sequencing data analysis Principal component analysis was performed using the prcomp function (R, stats) after detection of highly variable genes (M3Drop, bioconductor). Monocle (bioconductor) was used to infer single-cell differentiation trajectories. Gene expression along the pseudotemporal ordering was visualized by scaling all expression values to the genes maximal expression across all cells and calculating a loess-fit (ggplot2, bioconductor). Gene set enrichment analysis was done on gene lists ranked by their correlation to pseudo-time ranking using the Broad Institute GSEA software. Single cell trajectories were compared between human and mouse by calculating gene expression fold changes between the top and bottom 10% of cells according to pseudo-temporal ordering and performing gene set enrichment analysis on the resulting fold-change lists using GAGE (bioconductor) on databases containing GO terms for biological processes and kegg-pathways. The genesets for nociceptive neurogenesis was derived by comparing iPSC-derived nociceptive neurons to iPSC-derived anterior neuroectoderm and using differentially expressed genes with the highest positive or negative regulation (GSE26867, (Chambers et al., 2012)). For supervised annotation of embryonic cell types, single-cell RNA sequencing data were obtained from ArrayExpress ((Ibarra-Soria et al., 2018) E-MTAB-5728) with annotated cluster terms. We summarized sub-stratified clusters and sampled 1000 cells as training data. Training data were genes variable in both the published dataset and the Smart-seq2 dataset of this study and both datasets were z-transformed. A multi-class support vector machine classifier with a radial-basis kernel was trained on the published dataset to predict cluster labels (R, e1071). The trained model robustly predicted cell identity as evidenced by > 80% prediction accuracy during 10-fold cross-validation and was then used to annotate the presented Smart-seq2 dataset. To identify shared and divergent transcriptomic signatures in embryos and pNBSCs, differential expression was calculated relative to annotated pharyngeal mesoderm cells and genes were classified according to their differential expression state between two cell populations. QUANTIFICATION AND STATISTICAL ANALYSIS In all experiments, at least three biological replicates were used. Quantitative results were analyzed by two-sided unpaired Student’s t test using GraphPad Prism 6. Estimation of variation within each group was determined by s.e.m. unless otherwise indicated. Levels of significance were determined as follows: * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. DATA AND SOFTWARE AVAILABILITY Gene expression and DNA methylation data that support the findings of this study have been deposited at the Gene Expression Omnibus database and is available at ArrayExpress (accession number E-MTAB-5804, E-MTAB-5805, E-MTAB-5808, E-MTAB6911, E-MTAB-6925, E-MTAB-6912). A detailed description of data analyses and software used for each section can be found in METHOD DETAILS.

Cell Stem Cell 24, 166–182.e1–e13, January 3, 2019 e13