Temporal requirements for ISL1 in sympathetic neuron proliferation, differentiation, and diversification

Malformations of the sympathetic nervous system have been associated with cardiovascular instability, gastrointestinal dysfunction, and neuroblastoma. A better understanding of the factors regulating sympathetic nervous system development is critical to the development of potential therapies. Here, we have uncovered a temporal requirement for the LIM homeodomain transcription factor ISL1 during sympathetic nervous system development by the analysis of two mutant mouse lines: an Isl1 hypomorphic line and mice with Isl1 ablated in neural crest lineages. During early development, ISL1 is required for sympathetic neuronal fate determination, differentiation, and repression of glial differentiation, although it is dispensable for initial noradrenergic differentiation. ISL1 also plays an essential role in sympathetic neuron proliferation by controlling cell cycle gene expression. During later development, ISL1 is required for axon growth and sympathetic neuron diversification by maintaining noradrenergic differentiation, but repressing cholinergic differentiation. RNA-seq analyses of sympathetic ganglia from Isl1 mutant and control embryos, together with ISL1 ChIP-seq analysis on sympathetic ganglia, demonstrated that ISL1 regulates directly or indirectly several distinct signaling pathways that orchestrate sympathetic neurogenesis. A number of genes implicated in neuroblastoma pathogenesis are direct downstream targets of ISL1. Our study revealed a temporal requirement for ISL1 in multiple aspects of sympathetic neuron development, and suggested Isl1 as a candidate gene for neuroblastoma.


Introduction
Neurons of the sympathetic nervous system (SNS) innervate visceral organs, glands, and blood vessels, and play important roles in maintaining homeostasis. Aberrant development of the SNS causes neurocristopathies associated with cardiovascular instability, gastrointestinal dysfunction, as well as neuroblastoma that arises from developing sympathetic neurons 1 . The SNS is derived from trunk neural crest (NC) cells. As NC cells commit to the sympathetic lineage, they downregulate the expression of genes characteristic of NC cells such as Sox10, and express several sympathetic neuron-specific genes such as Phox2b and markers of catecholamine biosynthesis (TH, DBH). The newly generated sympathetic neurons transiently exit the cell cycle, and then reenter the cell cycle and are proliferating. Thereafter, an increasing number of sympathetic neurons exit the cell cycle, which peaks at embryonic day 14.5 (E14.5), and undergo post-mitotic differentiation [2][3][4][5] .
Segregation of initially bimodal sympathetic neurons into distinct noradrenergic and cholinergic neurons occurs relatively later during development, and is controlled by intrinsic as well as target derived signals 11,12 . Recent studies have uncovered a cross-repressive and maintenance mechanism for sympathetic neuron diversification, which involves antagonistic actions between pro-cholinergic factors Ntrk3 (TrkC)/Ret/Tlx3 and pro-noradrenergic factors Hmx1/Ntrk1 (TrkA) 13,14 . TrkC and Ret repress Hmx1 expression and induce cholinergic differentiation. Conversely, Hmx1 represses Tlx3 and Ret, induces TrkC, and maintains noradrenergic phenotypes.
ISL1 is an LIM-homeodomain transcription factor expressed in all peripheral neurons 15 . ISL1 plays an essential role in sensory neuron differentiation by activating genes critical for sensory neuron phenotype, and repressing genes expressed earlier during sensory neuron development 15 . ISL1 is required for early sympathetic neuron development 16 ; however, genes and signaling pathways regulated by ISL1 and direct targets of ISL1 in sympathetic neurons remain unexplored. In this study, we uncovered a temporal requirement for ISL1 in various aspects of SNS development. ISL1 regulates directly or indirectly a number of genes essential for sympathetic neuron proliferation and differentiation, many of which have been implicated in neuroblastoma pathogenesis, suggesting Isl1 as a candidate gene for neuroblastoma.
RNA-seq analysis reveals downregulation of a large number of genes essential for cell cycle progression at early developmental stages To better understand the genetic program regulated by ISL1, we performed RNA-seq analysis on sympathetic ganglia of Isl1 CKO and control embryos at E11.5-12, a stage when minimal cell loss occurred in Isl1 CKO sympathetic ganglia. We identified 358 downregulated and 555 upregulated transcripts in Isl1 CKO mutants (|fold change Isl1 CKO vs. ctrl | ≥ 1.5, p ≤ 0.05) ( Fig. 2a and Supplementary Table 1). Gene Set Enrichment Analysis (GSEA) revealed significant downregulation of gene sets representing the cell cycle, response to DNA damage stimulus, and sympathetic neuron development, but upregulation of gene sets representing NC development and glial differentiation ( Fig. 2b-f). Selected cell cycle genes were verified by qPCR on FACS sorted sympathetic neurons from E11.5 Isl1 CKO and control embryos (Rosa26-tdTomato+). These included, e.g., Mki67 (all phases), Ccnd1 (G1-S phase), Smc2/4, Bub1 (M phase), Arhgap11a, Pbk (proliferative signaling) (Fig. 2g). Reduced expression of Ccnd1 and Mki67 was further confirmed by immunostaining ( Fig. 2h-m).

ISL1 is required for effective downregulation of NC and glial progenitor factors during initial sympathetic neuronal differentiation
Our RNA-seq and qPCR analyses revealed significant upregulation of genes associated with NC (Sox10, Foxd3, Snai1, Sox2, and Crabp2) and glial (Fabp7, Metrn, Mpz, and Phgdh) progenitors ( Fig. 3a and Supplementary Table 1), suggesting a failure in their developmental downregulation in sympathetic neuroblasts, and an increase in gliogenesis. To investigate the role of ISL1 in early sympathetic neurogenesis, we further analyzed coexpression of Sox10 (NC and glial progenitors, and sympathetic neuronal progenitors), Phox2b (sympathetic progenitors and neurons), and ISL1 during early sympathetic development. At E10.5 and 11.5, a proportion of sympathetic progenitors co-expressed Phox2b and Sox10; however, few, if any, ISL1-expressing or Isl1-Cre lineagelabeled sympathetic neurons co-expressed Sox10 (0-1 per section) ( Fig. 3b-d, arrow). At E11.5, all ISL1expressing sympathetic neurons co-expressed Phox2b; however, a subset of Phox2b+ cells did not express ISL1 and were presumably progenitors for NC, sympathetic neurons, or glia (Fig. 3e, arrow). In Isl1 CKO sympathetic ganglia, the number of Phox2b+ sympathetic neurons was decreased; however, the number of Sox10+ cells, and Sox10+/Phox2b+ progenitors was significantly increased (Fig. 3f-h). GSEA showing downregulated gene sets representing cell cycle, response to DNA damage stimulus, and sympathetic neuron development, but upregulated gene sets representing neural crest cell development and neuroglial differentiation. g qPCR validation of selected genes involved in the cell cycle and proliferation. Results are shown as relative expression CKO vs. ctrl. Error bars represent the s. d. n = 3, p < 0.05*, or p < 0.01**, two-tailed t-test. h-m Immunostaining and quantitative analysis showing a marked reduction in expression of Mki67 and Ccnd1 in Isl1 CKO sympathetic ganglia marked by Rosa26-Tomato compared with controls at E11.5. Error bars represent the s.d., n = 3, p = 0.0008 (j), p = 0.0001 (m), two-tailed t-test. Scale bar, 50 µm. * p < 0.05, or ** p < 0.01 BMP and Notch signals are required for coordinating sympathetic differentiation and progenitor maintenance [6][7][8][9] . We found that ablation of Isl1 during early sympathetic development resulted in reduced BMP signaling (Bmpr1b), but upregulated BMP antagonist signaling (Ctgf and Fst), and a significant increase in expression of Notch ligands (Dll1, Dll3) and downstream target genes (Hes1, Heyl, and Sox2) 17 ( Fig. 3i and Supplementary Table 1). Together, these data suggest that ISL1 plays an important role in initial sympathetic neurogenesis, likely by modulating BMP and Notch signaling pathways.
Role of Isl1 in the gene regulatory network essential for sympathetic neuron differentiation at early developmental stages Our RNA-seq data revealed significant downregulation of genes involved in neuronal development. We further examined significantly downregulated genes for those that might contribute to the observed phenotypes of Isl1 CKO SNS, confirming downregulation in their expression by qPCR on E11.5 sympathetic neurons (Fig. 4a). These included genes involved in transcriptional regulation (e.g., Gata2/3, Insm2, Tlx3), chromatin modification (e.g., Dnmt1, Kdm6b), and RNA binding (Elavl2/3, Lin28a/b), many of which are known to play critical roles in sympathetic neuron development [2][3][4] (Fig. 4a). Downregulation of selected genes was further confirmed by in situ hybridization or by immunostaining ( Fig. 4b-m).
To understand the role of ISL1 in sympathetic axon development and neurotransmission, we examined sympathetic innervation by TH immunostaining. We observed a marked reduction or nearly absence of sympathetic innervations in the heart at E14. 5   Error bars represent the s.d., n = 3, two-tailed t-test. b-e Co-immunostaining of ISL1/Sox10, Tom/Sox10, and Phox2b/ISL1 at E10.5 and E11.5. Scale bar, 50 µm. f-h Co-immunostaining of Sox10 and Phox2b in Isl1 CKO and control sympathetic ganglia showing a significant decrease in the number of Phox2b+ neurons (h2), but a significant increase in the number of Sox10+ (h1) and Phox2b+/Sox10+ (h3) progenitors. Error bars represent the s.d., n = 4, p = 0.0135 (h1), p = 0.0109 (h2), p = 0.0134 (h3), two-tailed ttest. Scale bar, 50 µm. i qPCR verification of upregulation of a number of genes of Notch signaling pathway and BMP antagonists at E11.5-12. Error bars represent the s.d., n = 3, two-tailed t-test. * p < 0.05, or ** p < 0.01 However, defects in sympathetic innervation may be secondary to neuronal loss in Isl1 CKO sympathetic ganglia. Therefore, we cultured sympathetic ganglia from E12 Isl1 CKO and control mice, and axon growth was assessed after 24-48 h. We observed a significant 5-fold reduction in neurite outgrowth in Isl1 CKO mutants compared to controls (Fig. 4q, r). Furthermore, RNA-seq analysis revealed significantly altered expression of genes involved in neuronal projection morphogenesis and function. qPCR confirmed downregulation of genes involved in axon growth (e.g., Bmpr1b, Plxna4), neurotransmission (e.g., Cplx1, Scg3, Snap25), and altered expression of genes of neurotransmitters or related enzymes ( Fig. 4s; Supplementary Figure 1e and f). Reduced Isl1 expression leads to impaired sympathetic innervation, and delayed cell cycle withdrawal during later development Marked cell loss in Isl1 CKO sympathetic ganglia occurred well before robust sympathetic neuronal differentiation, precluding the study of the role of ISL1 during later sympathetic neuronal development. Therefore, we analyzed a mouse line with reduced Isl1 expression (hypomorphic) 18 (Supplementary Figure 2a-e). TH immunostaining revealed that the size of sympathetic ganglia and the number of neurons in Isl1 hypomorphic embryos were not significantly altered prior to E17.5,  (h-m). n-p Increased proliferation in E14.5 Isl1 hypomorphic sympathetic ganglia compared with controls revealed by co-immunostaining for BrdU and Phox2b. Error bars represent the s.d., n = 4, p = 0.011, two-tailed t-test. Scale bar, 100 µm. q-s Increased neuronal death revealed by co-immunostaining for activated caspase-3 and neurofilament (NF) in E14.5 Isl1 hypomorphic sympathetic ganglia compared with controls. Error bars represent the s.d., n = 4, p = 0.024, two-tailed t-test. Scale bar, 100 µm. * p < 0.05, or ** p < 0.01 although a reduction in sympathetic neuron number was observed in Isl1 hypomorphic mice at E17.5 (Fig. 5a-e). This allowed us to examine the role of ISL1 during later developmental stages.
In control embryos, neurites of sympathetic ganglia began to extend branches to innervate the dorsa aorta (da, dashed line) at E14.5, and formed a complex nerve plexus around the dorsal aorta at E17.5 (Fig. 5a, c, arrow). In Isl1 hypomorphic embryos, neurites of sympathetic ganglia appeared to be shorter with fewer branches at E14.5, and failed to form the aortic plexus at E17.5 (Fig. 5b, d, arrow). Similarly, sympathetic innervation to other target organs such as the salivary gland, heart, and stomach were markedly reduced in Isl1 hypomorphic embryos at E14.5 and E17.5 ( Fig. 5f-m).
RNA-seq analysis revealed dysregulation of genes important for axon growth and neurotransmission in Isl1 hypomorphic sympathetic neurons during later development We performed RNA-seq analysis on Isl1 hypomorphic and wildtype sympathetic ganglia at E14.5 and identified 641 upregulated and 526 downregulated genes (p < 0.05) ( Fig. 6a and Supplementary Table 2). GO analysis revealed that downregulated genes were over-represented in categories such as neuron development and neurotransmission, whereas upregulated genes were mainly cell cycle related (Fig. 6b, c; Supplementary Table 2). Similarly, GSEA analysis revealed downregulation of gene sets representing neuron differentiation and axonogenesis, but upregulation of a gene set representing the cell cycle in Isl1 hypomorphic embryos (Fig. 6d-f). Selected genes involved in axon growth, neurotransmission, and the cell cycle were verified by qPCR (Fig. 6g, h).
Intersection of the two RNA-seq data sets from Isl1 CKO and hypomorphic mutants identified only 84 genes commonly regulated by ISL1 at early (9.2% overlap) and later (7.1%) developmental stages ( Supplementary Figure 4). The commonly downregulated genes (46) were over-represented in categories such as protein secretion and synaptic vesicle transport, and upregulated genes (38) were involved in regulation of apoptosis (Supplementary Figure 4). Consistent with observed phenotypes, a subset of cell cycle-related genes were activated by ISL1 at early developmental stages but failed to downregulate in Isl1 hypomorphic sympathetic neurons at later developmental stages. Another subset of genes involved in protein secretion and nervous system development were repressed early but activated later by ISL1 during sympathetic neuron development (Supplementary Figure 4). These results suggested a distinct temporal requirement for ISL1 during sympathetic neuron development.

ISL1 is required for diversification of noradrenergic and cholinergic sympathetic neurons
RNA-seq and GSEA analyses revealed downregulation of a gene set representing noradrenergic differentiation, but upregulation of a gene set representing cholinergic differentiation in Isl1 hypomorphic sympathetic neurons (Fig. 7a, b; Supplementary Table 3). These included cholinergic genes (Ret, Tlx3, Sst, and Vip) and noradrenergic genes (Hmx1, Lmo1, Ntrk1, and TH) 19 . Alterations in their expression were confirmed by qPCR, in situ and immunostaining ( Fig. 7c-r). At E14.5 and E17.5, Tlx3, Vip, and Sst were upregulated, but Hmx1, Lmo1, and Ntrk1 were downregulated in Isl1 hypomorphic sympathetic ganglia. Ret expression was transiently upregulated at E14.5, but back to normal level at E17.5, whereas expression of Ntrk3, Chat, and Vacht was relatively normal at E14.5 but became upregulated at E17.5. Conversely, TH expression was unchanged at E14.5, but downregulated at E17.5. These data suggested that reduced ISL1 led to a failure both in repression of a cholinergic gene program and in maintenance of a noradrenergic program.

ISL1 directly regulates a number of genes required for sympathetic development
To gain an insight into the direct downstream targets of ISL1 in sympathetic neurons, we performed ISL1 ChIPseq analyses on E12.5-14.5 sympathetic ganglia. Analyses of these data revealed 3031 ChIP-seq peaks for ISL1, which also contained other motifs for Phox2a, Nkx6.1, Lhx3, Nanog, and Gata3 (Fig. 8a, b; Supplementary Table 4 and 5). Spatial annotation of ISL1 binding sites revealed 2046 potential ISL1 targets, including those involved in neuron differentiation and axonogenesis (Fig. 8c, Supplementary Figure 5a and Table 4). Intersection of ChIP-seq data with RNA-seq data from Isl1 CKO and hypomorphic mutants revealed 130 (71 downregulated, 59 upregulated) and 186 (86 downregulated, 100 upregulated) genes that were directly regulated by ISL1 during early and later sympathetic development, respectively (Fig. 8d, g; Supplementary Table 6 and 7). For direct targets of ISL1 downregulated in E12 Isl1 CKO mutants, over-represented categories included neurotransmission and cell-cell signaling, whereas upregulated targets of ISL1 were involved in, e.g., cell adhesion and signal transduction (Fig. 8e, f; Supplementary Table 6). Similarly, for direct targets of ISL1 downregulated in E14.5 Isl1 hypomorphic mutants, over-represented categories included neuron differentiation and axonogenesis. For upregulated targets of ISL1, categories such as cell morphogenesis, neuron differentiation, and cell proliferation were prominent (Fig. 8h, i; Supplementary  Table 7).
A close examination of direct targets of ISL1 downregulated in Isl1 CKO revealed a number of genes essential for sympathetic neuron development (e.g., Gata3, Insm1, Tlx3, Lin28b), axon growth, and neurotransmission (e.g., Bmpr1b, Plxna4, Snap25). Similarly, many direct targets of ISL1 downregulated in Isl1 hypomorphic sympathetic ganglia were involved in neuronal differentiation (e.g., Tbx20, Etv5, Phox2a) and neurotransmission (e.g., Ryr2, Kcnd2, Kcnmb2). Intersection of direct targets of ISL1 found in Isl1 CKO and hypomorphic mutants revealed 41 common targets of ISL1, many of which, e.g., Lmo1, Egln3, Tbx20, were positively regulated by ISL1 at both early and late stages (Supplementary  Table 8). However, several direct targets of ISL1, such as Prox1, Tlx3, and several cell cycle genes, were downregulated in Isl1 CKO during early development, but upregulated in Isl1 hypomorphic mutants during later development (Supplementary Table 8). Selected targets of ISL1 were verified by ChIP-qPCR, and genome browser views of representative common targets of ISL1 were shown ( Fig. 8j and Supplementary Figure 5b). Error bars represent the s.d., n = 3, two-tailed t-test. e-r Verification of selected genes associated with cholinergic neurons (Ret, Vip, Sst, Tlx3) and noradrenergic neurons (Hmx1, Ntrk1, TH) by in situ hybridization (e-j, m, n) and immunostaining (k, l, o-r) in E14.5 control and Isl1 hypomorphic sympathetic ganglia. Scale bar, 100 µm, ns = not significant, * p < 0.05, or ** p < 0.01 Fig. 8 ISL1 directly regulates a number of genes required for sympathetic neuron development. a ChIP-seq ISL1-binding regions were mapped relative to their nearest transcription start site (TSS). Annotation includes whether a peak is in the TSS, transcription termination site, exon, 5ʹUTR, 3ʹUTR, intronic, or intergenic. b Top motifs enriched in the vicinity of ISL1-binding sites. c GO functional clustering of genes associated with ISL1 ChIP-seq peaks. d Overlay of ChIP-seq data and RNA-seq results from E11.5 Isl1 CKO sympathetic ganglia revealed 71 downregulated and 59 upregulated genes as potential direct targets of ISL1 in sympathetic neurons; p = 1.56 × 10 −7 , Fisher's exact test. e, f GO functional clustering of these genes bound by ISL1 and downregulated (e) or upregulated (f) in E11.5 Isl1 CKO sympathetic neurons. g Intersection of ChIP-seq data and RNA-seq results from E14.5 Isl1 hypomorphic sympathetic ganglia revealed 86 downregulated and 100 upregulated genes as potential direct targets of ISL1 in sympathetic neurons; p = 4.6 × 10 −13 , Fisher's exact test. h, i GO functional clustering of these genes bound by ISL1 and dysregulated in E14.5 Isl1 hypomorphic sympathetic neurons. j ChIP-qPCR validation of ISL1 binding at selected targets in E12.5-E14.5 sympathetic neurons, including those genes involved in the cell cycle, transcription regulation, neurotransmission, and ion channels. Error bars represent the s.d., n = 3, two-tailed t-test. *p < 0.05, or ** p < 0.01 Notably, a number of targets of ISL1 downregulated in Isl1 CKO mutants were implicated in neuroblastoma pathogenesis [20][21][22][23] , including Lmo1, Gata3, Prox1, Lin28b, and Alk. We found that ISL1 binds to an evolutionary conserved region within the first intron of Lmo1 (mm9: chr7:116289070-116289656) in mouse sympathetic neurons that, by whole-genome alignment, overlaps with the super-enhancer of LMO1 identified in human neuroblastoma 20 .

Discussion
Distinct temporal requirements for ISL1 in sympathetic neuron proliferation and differentiation during sympathetic neuron development Sympathetic neurons continue to proliferate after the acquisition of pan-neuronal and noradrenergic properties 5 . From E14.5 onward, a majority of sympathetic neurons undergo cell cycle withdrawal and terminal differentiation and maturation; however, the mechanisms underlying these temporal events remain largely unknown. At early developmental stages, ablation of Isl1 resulted in reduced sympathetic neuron proliferation and reduced expression of cell cycle genes, including Ccnd1 and Foxm1. Expression of a number of transcription factors critical for sympathetic neuron proliferation was also downregulated in Isl1 CKO mutants [24][25][26][27] , including Gata3, Insm1, and Prox1. Although mechanisms by which these transcription factors control sympathetic neuron proliferation are unclear, these data suggest that ISL1 may regulate sympathetic neuron proliferation by controlling cell cycle gene expression.
Differentiation of noradrenergic and pan-neuronal phenotypes is controlled by an early gene regulatory network 3 , including Mash1, Phox2b/2a, Hand2, Gata2/3, and Insm1. We found that ISL1 is downstream of Phox2b, but upstream of Gata2/3, Insm1/2, and Hand1. ISL1 acts parallel to Hand2, since Hand2 expression is not changed in Isl1 CKO mutants and vice versa 28 . ISL1 is dispensable for initial noradrenergic differentiation, since expression of TH and DBH is not altered in Isl1 CKO mutants. But instead, ISL1 directly or indirectly regulates sets of genes essential for sympathetic neuron development, axon growth, and neurotransmission, suggesting that ISL1 is essential for generic sympathetic neuronal differentiation during early sympathetic neuron development.
Reduced Isl1 expression in Isl1 hypomorphic mice caused impaired sympathetic innervations, delayed cell cycle withdrawal, and reduced expression of genes involved in axon growth and neurotransmission, demonstrating a continued requirement for ISL1 in sympathetic neuron differentiation and maturation at later development stages. Analyses of ChIP-seq and RNAseq data sets from both Isl1 CKO and hypomorphic mutants revealed ISL1 regulates common, yet distinct gene programs during early and later sympathetic neuronogenesis. Tlx3, Prox1, and several cell cycle-related genes were among the direct targets of ISL1 that were downregulated in Isl1 CKO but upregulated in Isl1 hypomorphic mutants, suggesting that ISL1 may act both as an activator and a repressor in a context-dependent manner during sympathetic neuron development. Interestingly, a comparison of ISL1 targets in different cell types (sympathetic neurons, cardiac progenitors 29 , and sinus node pacemakers 30 ) revealed few, if any, common targets that could account for the observed phenotypes in each of these mutants, suggesting that the action of ISL1 is cellular context-dependent.
Molecular mechanisms of ISL1 in sympathetic neurons remain unknown. Interestingly, ablation of Isl1 causes downregulation of a number of epigenetic regulators 29,31,32 (Dnmt1, Kdm6a, and Satb1) critical for neuronal development. In the heart, ISL1 physically interacts with Kdm6a to promote H3K27me3 demethylation at the enhancers of target genes encoding key cardiac transcription factors 29 . It is tempting to speculate that ISL1 might regulate the expression of sympathetic developmental genes via epigenetic mechanisms.
Essential role of ISL1 in coordination of initial neurogenesis and gliogenesis by modulating BMP and Notch signaling pathways BMP and Notch pathways are essential for coordinating progenitor maintenance and neuronal differentiation during sympathetic nneurogenesis 6 . Ablation of Isl1 resulted in increased Notch signaling but decreased BMP signaling. Accordingly, genes associated with NC and glial progenitors were significantly upregulated, suggesting that during initial sympathetic neurogenesis, ISL1 is required for coordination of lineage differentiation of sympathetic neurons and glia by modulating BMP and Notch pathways. Interestingly, previous studies have revealed a role of ISL1 in cell fate specification and determination in sensory neurons of the dorsal root ganglion and spinal motor neurons. Ablation of Isl1 in NC cells leads to reduced expression of genes essential for nociceptive neuron differentiation, but a failure in developmental downregulation of genes characteristic of sensory progenitors 15 . In spinal motor neurons, reduced Isl1 expression leads to conversion of motor neurons to V2 interneurons, although underlying mechanism remains unknown 18 .

ISL1 is required for diversification of noradrenergic and cholinergic sympathetic neurons
Segregation of initially bimodal sympathetic neurons into noradrenergic and cholinergic neurons occurs later during development and involves antagonistic actions between pro-cholinergic (Ntrk3/Ret/Tlx3) and pro-noradrenergic (Hmx1/Ntrk1) factors 13,14 . In Isl1 hypomorphic sympathetic ganglia, we observed significantly reduced expression of genes essential for noradrenergic differentiation (e.g., Hmx1, Lmo1, Ntrk1, TH), but increased expression of genes essential for cholinergic differentiation (Ret, Tlx3, Sst, Vip, Ntrk3, Chat, Vacht), demonstrating that ISL1 is required for repression of cholinergic characteristics and maintenance of noradrenergic characteristics during later developmental stages. Previous studies have shown that, in forebrain and spinal cholinergic motor neurons, ISL1, complexed with LHX3 or LHX8, directly binds to and activates the expression of genes of cholinergic traits 33 . However, we did not find direct binding of ISL1 to the enhancers/ promoters of Chat, Vacht, or other cholinergic genes.

ISL1 and neuroblastoma
Neuroblastoma is an aggressive pediatric tumor arising from developing sympathetic neurons. ISL1 is overexpressed in most types of neuroblastoma, particularly in poorly differentiated neuroblastomas that indicates a poor prognosis 34,35 . ISL1 is also a minimal residual disease marker of neuroblastoma 34 . However, the role of ISL1 in neuroblastoma tumorigenesis is unknown. We found that a number of genes implicated in neuroblastoma pathogenesis are downstream of ISL1 (e.g., Lmo1, Ccnd1, Tfap2b, Prox1, Casz1, Foxm1, Lin28b, Ccnd1, and Alk) 20-23,34,36-38 , many of which are direct targets of ISL1. Notably, LMO1 is a neuroblastoma oncogene associated with more proliferative and aggressive phenotype 39 . A polymorphism within a super-enhancer in the first intron of LMO1 that preserves an evolutionarily conserved GATA factor binding motif predisposes to neuroblastoma. Knockdown of GATA3 results in decreased LMO1 expression and suppression of neuroblastoma cell growth 20 . We found that ISL1 binds to an evolutionary conserved region within the first intron of Lmo1 (chr7:116289070-116289656) in mouse sympathetic neurons that overlaps with the super-enhancer of LMO1 identified in human neuroblastoma, suggesting coregulation of LMO1 expression by ISL1 and GATA3 and a potential role of ISL1 in controlling neuroblastoma cell proliferation.
Taken together, our study uncovered a temporal requirement for ISL1 during sympathetic neurogenesis (Supplementary Figure 6), and implicated Isl1 as a candidate gene for neuroblastoma.

Transgenic mice
Generation of Wnt1-Cre;Isl1 f/f and Isl1 hypomorphic mice(Isl1 f-neo/f-neo ) have been reported previously 18 . To facilitate the visualization and lineage tracing of sympathetic neurons, Isl1 f/f mice were crossed onto a Rosa26-LacZ or -Tomato reporter background 40 . All the experiments involving mice were carried out in accordance with protocols approved by the Animal Committee of Tongji University School of Medicine (TJmed-010-10).
In situ hybridization, immunostaining, and X-gal staining In situ hybridization, immunostaining, and X-gal staining were performed as described 30 . Fragments amplified from mouse cDNA were used to generate the RNA probes for Insm2, Tlx3, Tfap2b, Hand1, and Lmo1. Riboprobe for mouse Vip, Sst, and Ret were kindly provided by Dr. Leping Cheng (Chinese Academy of Sciences). For BrdU staining, pregnant mice were injected with 300 μl of BrdU (100 μg/g body weight, Sigma). For cell counting, sections of sympathetic ganglia of the appropriate developmental stages were cut at 10 μm, every fourth section was stained, and positive cells were counted. For proliferation, the total number of BrdU+ sympathetic neurons (Phox2b+) were counted and expressed as percentage of total Phox2b+ neurons. For cell death assay, the total number of Casp-3+ neurons (Tomato+) were counted and expressed as percentage of total Tomato+ neurons. At least 5-8 matched sections were analyzed and three samples per genotype per time point were analyzed. Detailed information of antibodies is available in the Supplementary material.

RNA-seq and ChIP-seq data analysis
RNA-seq analyses with two biological repeats were performed as described 30 . ChIP-seq was performed as previously described 41 . The complete RNA-seq and ChIPseq data sets are available from the Gene Expressing Omnibus database (http://www.ncbi.nlm.nih.gov/geo/) under the accession number GSE93308. For experimental and data analysis details, see the Online Data Supplement.

Statistical analysis
All experiments were performed for at least three independent times and respective data were used for statistical analyses. Data were presented as mean ± s.d., and a two-tailed t-test was used for two-group comparisons. Differences were considered statistically significant at a value of p ≤ 0.05*, p ≤ 0.01**. For experimental details, see the online Supplementary material.