Nature BGI Stereo-seq helps reveal the mechanisms of melanoma cell growth and metastasis

Because of its high heterogeneity and variability, melanoma has long been an important challenge for cancer research and public health worldwide
.

The research results were published on Nature on
the evening of September 21, Beijing time.

Article: A cellular hierarchy in melanoma uncouples growth and metastasis
Published: 2022-09-21Published
Journal: Nature Main Research Team: VIB Center for Cancer Biology
, Leuven, Belgium, etc.

Impact Factor: 69.
504
DOI：10.
1038/s41586-022-05242-7

Research background

Melanoma, also known as malignant melanoma, is mainly a highly malignant tumor derived from melanocytes with high heterogeneity and variability
.
There is substantial evidence that melanoma cells can present as a cell transcription state
such as melanocytes (proliferative) or mesenchymal-like (aggressive).
The scRNA-seq analysis confirmed that these two cellular states can coexist
.
However, the latest research from in vivo, in vitro models suggests that melanoma cells may be present in other states, meaning that the current understanding of the heterogeneity within melanoma may be incomplete
.
The evidence from these observations also raises the possibility that these different cellular states may contribute differently
to tumor growth and metastasis.
Therefore, it is necessary to further study the phenotypic heterogeneity
of melanoma.

Research strategies

Using an allogeneic melanoma transplant model, potential subsets of melanoma cells are mined and injected into mice
.
Then, combined with scRNA-seq, lineage tracking and spatial transcriptome techniques such as Stereo-seq, the growth structure of melanoma cell subsets and the spatial environment of their locations are studied to identify key factors
that promote tumor growth and metastasis.

Research Results

1.
Re-explore the diversity of melanoma cell states

Based on mouse models carrying the Tyr::Nras^Q61K/o allele and lacking Cdkn2a (also known as Ink4a or p16^INK4a), the researchers found that these mice developed cutaneous monoclonal melanoma lesions and subsequent metastases
to lymph nodes, liver, and lungs.
To increase the versatility of this model, the researchers derived homogeneous cell lines in C57BL/6 mice and evaluated transcriptome maps
of such melanoma lesions with scRNA-seq 。 Using published cell type-specific gene sets, researchers identified T cells, B cells, cancer-associated fibroblasts, dendritic cells, monocytes/macrophages, endothelial cells (EC), pericytes, and malignant cell populations, indicating the complexity of the tumor microenvironment
。 Further analysis of the malignant cell population identified 7 different melanoma cell states by unsupervised clustering and UMAP analysis (Figure 1a
).

MITF is a key regulator of the differentiated state of melanoma
.
The researchers detected extensive Mitf transcriptional activity in all malignant cells, but relatively speaking, proliferative cells had the highest MITF activity and the lowest
infiltrating cells.
The researchers observed that cells from cluster 4 exhibited the lowest MITF activity, indicating that this cluster cell was the highest degree of dedifferentiation of all malignant cells (Figure 1b
).

The researchers functionally annotated each cluster using the gene set enrichment tool and established the gene expression characteristics of each of these clusters (Figures 1c, d), used AUCell to calculate the predicted values of different cells and mapped them to UMAP space (Figure 1e), and the results showed that these cells showed traces of pre-epithelial–mesenchymal transition (pre-EMT).
These cells are the first neural crest (NC) directed cells to appear in embryonic form and constitute a self-renewing pool of NC stem cells (Figure 1f
).
In addition, scRNA-seq analysis showed that in these lesions, melanoma cells with melanin-like, mesenchymal-like, NC-like, and pre-EMT NC-stem-like features can be identified
.

2.

To confirm whether these primary melanomas have a hierarchical structure, the researchers used an induceable genetic lineage tracing method of the multicolor Rosa26-LSL-Confetti allele and found that Tyr::Nras^Q61K/°; Ink4a−^/−; Tyr::cre^ERT2/°; Rosa26-LSL-Confetti^+/+ composite mice develop primary melanoma lesions at approximately 8±2 months of age (Figure 2a
).

Next, the researchers used density-based assays to continuously reconstruct and assess the total clone size at single-cell resolution (Figure 2d), where they found significant differences in clone size across all tumor samples (Figure 2e), with further evidence demonstrating the bimodal behavior of the clone size accumulation distribution (Figure 2f).

In summary, primary melanoma has a hierarchy, including at least two cell populations exhibiting different growth kinetics and fate behaviors – stem cell-like and progenitor cell populations, and the properties of stem cells are related to
the environment of spatial localization.

3.

To find evidence to support the bionic pattern tissue of the pre-EMT NC stem group, the researchers analyzed the mouse tissue slices of the model using Visium technology (Figure 3a) and integrated the spatial transcriptome data with the scRNA-seq data to find an expected positive correlation between the stress sample (hypoxia) prediction score and the distance to the nearest blood vessel, while the pre-EMT NC stem sample group prediction score is inversely proportional to the distance to the nearest blood vessel (Figure 3b).

PRRX1 labels cells that promote melanoma metastasis

PRRX1 promotes epithelial-to-mesenchymal transition (EMT) in epithelial carcinoma and is a putative driver of mesenchymoid cells in mice and humans
.

To study the subsequent development of these cells, the researchers treated mice with TAM and collected lesions after 30 days and found that the proportion of tdTomato⁺ cells decreased from about 35% in the early days to about 4% (Figure 4f
).

conclusion

By integrating mouse genetics, single-cell and spatial transcriptomes, lineage tracking, and quantitative modeling, the researchers found that melanoma growth is supported
by a relatively limited population of tumor-causing cells.

This study provides raw data for scRNA-seq, Visium, and stereo-seq:
_msthash="320101" _msttexthash="275639">About Stereo-seq

Stereo-seq, a spatio-temporal omics technology independently developed by BGI, is a panoramic spatiotemporal transcriptome technology
that achieves ultra-high throughput and ultra-high precision.

Stereo-seq has been successfully applied to animal embryonic development, plant development, brain science, tumors and many other different fields, proving its universality and can be applied to various fields of life science and clinical medical research in
the future.
Especially in the direction of disease treatment, spatiotemporal omics is ultimately expected to bring great improvements
to human health.

At present, 11 articles including CNS have been successfully published using Stereo-seq, as follows: