A human neural crest model reveals the developmental impact of neuroblastoma-associated chromosomal aberrations

nature communications

A human neural crest model reveals the developmental impact of neuroblastoma-associated chromosomal aberrations

Early childhood tumours arise from transformed embryonic cells, which often carry large copy number alterations. However, it remains unclear how copy number alterations contribute to embryonic tumourigenesis due to a lack of suitable models. Here we employ female human embryonic stem cell differentiation and single-cell transcriptome and epigenome analysis to assess the effects of chromosome seventeen q/one q gains, which are prevalent in the embryonal tumour neuroblastoma. We show that copy number alterations impair the specification of trunk neural crest cells and their sympathoadrenal derivatives, the putative cells-of-origin of neuroblastoma. This effect is exacerbated upon overexpression of MYCN, whose amplification co-occurs with copy number alterations in neuroblastoma. Moreover, copy number alterations potentiate the pro-tumourigenic effects of MYCN and mutant neural crest cells resemble neuroblastoma cells in tumours. These changes correlate with a stepwise aberration of developmental transcription factor networks. Together, our results sketch a mechanistic framework for the copy number alteration-driven initiation of embryonal tumours.

Cancers in early childhood are driven by sparse genetic aberrations arising in utero, which are thought to lead to defective differentiation and uncontrolled proliferation. Most tumours harbour large genomic rearrangements and chromosomal copy number alterations, which co-occur with mutations in tumour suppressors or tumourigenic transcription factors. The mechanistic interactions between different mutations and early developmental processes are likely foundational drivers of tumour heterogeneity. However,

since visible tumours are only detected long after their initiation, early mutation-driven interactions leading to the healthy-to-tumour transition have remained largely intractable.

Neuroblastoma is the most common extra-cranial solid tumour in infants and an archetypal "developmental cancer". Neuroblastoma tumours are usually found in the adrenal gland or sympathetic ganglia, tissues derived from the trunk neural crest lineage during embryonic development, and studies using transgenic animal models and transcriptome analysis have anchored neuroblastoma tumourigenesis in impaired sympathoadrenal differentiation of trunk neural crest cells. Copy number alterations, such as gains of the long arms of chromosomes seventeen and one have been identified in the majority (up to sixty-five percent) of neuroblastoma tumours and their emergence is considered an early tumourigenesis "priming" event. Chromosome seventeen q/one q gains often co-occur with amplification of the MYCN oncogene (at least one copy number alteration in greater than ninety-five percent of MYCN-amplified tumours), suggesting they may jointly contribute to tumourigenesis. However, despite our advanced understanding of the genetic and developmental origin of neuroblastoma, it remains unclear to date how copy number alterations disrupt embryonic cell differentiation and lead to neuroblastoma initiation.

Here, we used a human embryonic stem cell-based model to experimentally dissect the links between neuroblastoma-associated copy number alterations, MYCN amplification, and tumour initiation. We interrogated the stepwise specification of trunk neural crest and sympathoadrenal lineages using directed differentiation of isogenic human embryonic stem cell lines with chromosome seventeen q/one q gains and inducible MYCN overexpression. We found that (i) Copy number alterations derail differentiation by potentiating immature neural crest progenitor phenotypes. Combining copy number alterations with MYCN overexpression completely disrupted normal neural crest differentiation; (ii) Mutant neural crest cells acquired tumourigenic hallmarks in vitro, the capacity to form tumours in xenografts, and resemble distinct subpopulations of heterogeneous neuroblastoma tumours; (iii) An extensive re-wiring of chromatin connects the observed transcriptional and functional aberrations with a dysregulated network of developmental transcription factors. Collectively, our data put forward a copy number alteration-driven distortion of trunk neural crest and sympathoadrenal differentiation as a priming mechanism for subsequent MYCN-induced tumour initiation.

Results

Differentiation of human embryonic stem cells recapitulates key stages of trunk neural crest and sympathoadrenal development

To model the initiation stage and cell types relevant to neuroblastoma tumourigenesis, we turned to an in-vitro modelling approach. We have previously described an efficient strategy to produce human trunk neural crest, sympathoadrenal progenitors, and sympathetic neurons from human embryonic stem cells. Our protocol involves treatment with defined cocktails of signalling pathway agonists/antagonists that induce neuromesodermal-potent axial progenitors at day three of differentiation and subsequently steer axial progenitors toward trunk neural crest cells and their sympathoadrenal derivatives. At day nineteen, the protocol yields catecholamine-producing sympathetic neurons marked by peripherin-expressing axons.

As a prerequisite for studying the effects of copy number alterations on trunk neural crest differentiation, we first needed to define a molecular roadmap of normal human embryonic stem cell differentiation as a control. Therefore, we employed our protocol for the differentiation of karyotypically normal human embryonic stem cells and performed droplet-based single-cell RNA sequencing at key differentiation stages to examine the resulting cell populations. We obtained twenty-nine thousand eight hundred fifty-seven cells that passed quality control, which we allocated to fourteen distinct clusters. We bioinformatically annotated these cell clusters using two complementary approaches: (i) by identifying characteristic marker genes, and (ii) by mapping our data to single-cell transcriptomes of trunk neural crest derivatives in human embryos. This strategy identified cells at different stages of trunk neural crest development, including axial progenitor-like cells marked by CDX1/2, NKX1-2, and FGF signalling-associated transcripts; cluster two in supplementary figure two, and later cell populations of a predominantly trunk axial identity exhibiting characteristics of Schwann cell precursors, sympathoblasts, as well as mesenchymal features. For example, day nine cells split into subpopulations expressing markers of trunk NC/early SCPs (C three; e.g., SOX ten sixteen, Figure one c; weak SCP-like signature, Figure one e) and sensory neurons (C five; ONECUT thirteen nine, Figure one c; weak SYM-like signature, Figure one f). At D fourteen, cells started to assume a sympathoadrenal/autonomic progenitor (C eight; ASCL one) or mesenchymal (C eleven; FN one) identity, and by day D nineteen, we observed three distinct fractions: mature SCP-like cells (C nine; POSTN forty; strong SCP signature), autonomic sympathoblasts (C twelve through C fourteen; PHOX two A/B, ELAVL forty-one six, forty-one; strong SYM signature), and MES-like cells (C eleven; COL one A one, FN one). This is in line with findings showing that trunk NC and SCPs are competent to generate mesenchyme. Interestingly, we also found cells at the intersection of MES and SYM identity, as observed in mice and NB cell lines (Supplementary Figure three; Supplementary Data three). After four weeks (D twenty-eight), we also observed some cells with a partial chromaffin-like cell identity (part of C fourteen; CHGA positive, PMNT) (Figure one d).

Together, these data confirm that our hESC-based model successfully captures trunk NC and sympathoadrenal cells as found in embryos during the onset of NB tumourigenesis. Moreover, they reveal two major developmental branching events: (i) an early commitment of trunk NC toward a sensory neuron fate; (ii) the late generation of multipotent SCP/sympathoadrenal progenitors, which subsequently give rise to three distinct cell types: mature SCPs, MES, and SYM.