Human metastatic melanoma cells express a dedifferentiated, plastic phenotype, which may serve as a selective advantage, because melanoma cells invade various microenvironments. Over the last three decades, there has been an increased focus on the role of the tumor microenvironment in cancer progression, with the goal of reversing the metastatic phenotype. Here, using an embryonic chick model, we explore the possibility of reverting the metastatic melanoma phenotype to its cell type of origin, the neural-crest-derived melanocyte. GFP-labeled adult human metastatic melanoma cells were transplanted in ovo adjacent to host chick premigratory neural crest cells and analyzed 48 and 96 h after egg reincubation. Interestingly, the transplanted melanoma cells do not form tumors. Instead, we find that transplanted melanoma cells invade surrounding chick tissues in a programmed manner, distributing along host neural-crest-cell migratory pathways. The invading melanoma cells display neural-crest-cell-like morphologies and populate host peripheral structures, including the branchial arches, dorsal root and sympathetic ganglia. Analysis of a melanocyte-specific phenotype marker (MART-1) and a neuronal marker (Tuj1) revealed a subpopulation of melanoma cells that invade the chick periphery and express MART-1 and Tuj1. Our results demonstrate the ability of adult human metastatic melanoma cells to respond to chick embryonic environmental cues, a subset of which may undergo a reprogramming of their metastatic phenotype. This model has the potential to provide insights into the regulation of tumor cell plasticity by an embryonic milieu, which may hold significant therapeutic promise.plasticity ͉ chick ͉ MART-1 ͉ epigenetic

The proper assembly of craniofacial structures and the peripheral nervous system requires neural crest cells to emerge from the neural tube and navigate over long distances to the branchial arches. Cell and molecular studies have shed light on potential intrinsic and extrinsic cues, which, in combination,are thought to ensure the induction and specification of cranial neural crest cells. However, much less is known about how migrating neural crest cells interpret and integrate signals from the microenvironment and other neural crest cells to sort into and maintain the stereotypical pattern of three spatially segregated streams. Here, we explore the extent to which cranial neural crest cells use cell-to-cell and cell-environment interactions to pathfind. The cell membrane and cytoskeletal elements in chick premigratory neural crest cells were labeled in vivo. Three-dimensional reconstructions of migrating neural crest cells were then obtained using confocal static and time-lapse imaging. It was found that neural crest cells maintained nearly constant contact with other migrating neural crest cells, in addition to the microenvironment. Cells used lamellipodia or short, thin filopodia (1-2 μm wide) for local contacts (<20 μm). Non-local, long distance contact (up to 100 μm) was initiated by filopodia that extended and retracted, extended and tracked, or tethered two non-neighboring cells. Intriguingly, the cell-to-cell contacts often stimulated a cell to change direction in favor of a neighboring cell's trajectory. In summary, our results present in vivo evidence for local and long-range neural crest cell interactions, suggesting a possible role for these contacts in directional guidance.

Neural crest (NC) cell migration is crucial to the formation of peripheral tissues during vertebrate development. However, how NC cells respond to different microenvironments to maintain persistence of direction and cohesion in multicellular streams remains unclear. To address this, we profiled eight subregions of a typical cranial NC cell migratory stream. Hierarchical clustering showed significant differences in the expression profiles of the lead three subregions compared with newly emerged cells. Multiplexed imaging of mRNA expression using fluorescent hybridization chain reaction (HCR) quantitatively confirmed the expression profiles of lead cells. Computational modeling predicted that a small fraction of lead cells that detect directional information is optimal for successful stream migration. Single-cell profiling then revealed a unique molecular signature that is consistent and stable over time in a subset of lead cells within the most advanced portion of the migratory front, which we term trailblazers. Model simulations that forced a lead cell behavior in the trailing subpopulation predicted cell bunching near the migratory domain entrance. Misexpression of the trailblazer molecular signature by perturbation of two upstream transcription factors agreed with the in silico prediction and showed alterations to NC cell migration distance and stream shape. These data are the first to characterize the molecular diversity within an NC cell migratory stream and offer insights into how molecular patterns are transduced into cell behaviors.

The neural crest is an excellent model to study embryonic cell migration, since cell behaviors can be studied in vivo with advanced optical imaging and molecular intervention. What is unclear is how molecular signals direct neural crest cell (NCC) migration through multiple microenvironments and into specific targets. Here, we tested the hypothesis that the invasion of cranial NCCs, specifically the rhombomere 4 (r4) migratory stream into branchial arch 2 (ba2), is due to chemoattraction through neuropilin-1-vascular endothelial growth factor (VEGF) interactions. We found that the spatio-temporal expression pattern of VEGF in the ectoderm correlated with the NCC migratory front. RT-PCR analysis of the r4 migratory stream showed that ba2 tissue expressed VEGF and r4 NCCs expressed VEGF receptor 2. When soluble VEGF receptor 1 (sVEGFR1) was injected distal to the r4 migratory front, to bind up endogenous VEGF, NCCs failed to completely invade ba2. Time-lapse imaging revealed that cranial NCCs were attracted to ba2 tissue or VEGF sources in vitro. VEGF-soaked beads or VEGF-expressing cells placed adjacent to the r4 migratory stream caused NCCs to divert from stereotypical pathways and move towards an ectopic VEGF source. Our results suggest a model in which NCC entry and invasion of ba2 is dependent on chemoattractive signaling through neuropilin-1-VEGF interactions.

Embryonic neural crest cells travel in discrete streams to precise locations throughout the head and body. We previously showed that cranial neural crest cells respond chemotactically to vascular endothelial growth factor (VEGF) and that cells within the migratory front have distinct behaviors and gene expression. We proposed a cell-induced gradient model in which lead neural crest cells read out directional information from a chemoattractant profile and instruct trailers to follow. In this study, we show that migrating chick neural crest cells do not display distinct lead and trailer gene expression profiles in culture. However, exposure to VEGF in vitro results in the upregulation of a small subset of genes associated with an in vivo lead cell signature. Timed addition and removal of VEGF in culture reveals the changes in neural crest cell gene expression are rapid. A computational model incorporating an integrate-and-switch mechanism between cellular phenotypes predicts migration efficiency is influenced by the timescale of cell behavior switching. To test the model hypothesis that neural crest cellular phenotypes respond to changes in the VEGF chemoattractant profile, we presented ectopic sources of VEGF to the trailer neural crest cell subpopulation and show diverted cell trajectories and stream alterations consistent with model predictions. Gene profiling of trailer cells that diverted and encountered VEGF revealed upregulation of a subset of 'lead' genes. Injection of neuropilin1 (Np1)-Fc into the trailer subpopulation or electroporation of VEGF morpholino to reduce VEGF signaling failed to alter trailer neural crest cell trajectories, suggesting trailers do not require VEGF to maintain coordinated migration. These results indicate that VEGF is one of the signals that establishes lead cell identity and its chemoattractant profile is critical to neural crest cell migration.