Multimodal Analysis of Composition and Spatial Architecture in Human Squamous Cell Carcinoma.
Ji AL, Rubin AJ, Thrane K, Jiang S, Reynolds DL, Meyers RM, Guo MG, George BM, Mollbrink A, Bergenstråhle J, Larsson L, Bai Y, Zhu B, Bhaduri A, Meyers JM, Rovira-Clavé X, Hollmig ST, Aasi SZ, Nolan GP, Lundeberg J, Khavari PA. · 2020
Read Multimodal Analysis of Composition and Spatial Architecture in Human Squamous Cell Carcinoma. by Ji AL, Rubin AJ, Thrane K, Jiang S, Reynolds DL, Meyers RM, Guo MG, George BM, Mollbrink A, Bergenstråhle J, Larsson L, Bai Y, Zhu B, Bhaduri A, Meyers JM, Rovira-Clavé X, Hollmig ST, Aasi SZ, Nolan GP, Lundeberg J, Khavari PA. (2020) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc7391009.txt.
Introduction
Ji and co-authors build a multi-layer atlas of skin squamous cell carcinoma in Cell. They combined single-cell gene sequencing, spatial gene mapping, and ion beam imaging on human tumors and matched normal skin. The aim was to define which cells the cancer holds and where each type sits.
Tumors contained four cancer subpopulations. Three mimicked normal skin states, but a fourth, tumor-specific keratinocyte group appeared only in cancer and settled in a niche rich in fibroblasts and blood vessels. Those cells acted as a communication hub. Suppressive features showed too, with regulatory T cells parked beside killer T cells in walled-off stroma. Screens in tumor grafts found gene networks each subpopulation needs to grow.
Catalog record
| CREATOR | Ji AL, Rubin AJ, Thrane K, Jiang S, Reynolds DL, Meyers RM, Guo MG, George BM, Mollbrink A, Bergenstråhle J, Larsson L, Bai Y, Zhu B, Bhaduri A, Meyers JM, Rovira-Clavé X, Hollmig ST, Aasi SZ, Nolan GP, Lundeberg J, Khavari PA. |
|---|---|
| YEAR | 2020 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 128 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Multimodal Analysis of Composition and Spatial Architecture in Human Squamous Cell Carcinoma By Andrew L Ji, Adam J Rubin, Kim Thrane, Sizun Jiang, David L Reynolds, Robin M Meyers, Margaret G Guo, Benson M George, Annelie Mollbrink, Joseph Bergenstråhle, Ludvig Larsson, Yunhao Bai Published in: Cell DOI: 10.1016/j.cell.2020.05.039 ABSTRACT Summary To define the cellular composition and architecture of cutaneous squamous cell carcinoma (cSCC), we combined single-cell RNA sequencing with spatial transcriptomics and multiplexed ion beam imaging from a series of human cSCCs and matched normal skin. cSCC exhibited four tumor subpopulations, three recapitulating normal epidermal states, and a tumor-specific keratinocyte (TSK) population unique to cancer, which localized to a fibrovascular niche. Integration of single-cell and spatial data mapped ligand-receptor networks to specific cell types, revealing TSK cells as a hub for intercellular communication. Multiple features of potential immunosuppression were observed, including T regulatory cell (Treg) co-localization with CD8 T cells in compartmentalized tumor stroma. Finally, single-cell characterization of human tumor xenografts and in vivo CRISPR screens identified essential roles for specific tumor subpopulation-enriched gene networks in tumorigenesis. These data define cSCC tumor and stromal cell subpopulations, the spatial niches where they interact, and the communicating gene networks that they engage in cancer. Keywords: squamous cell carcinoma, intra-tumoral heterogeneity, scRNA-seq, spatial transcriptomics, MIBI, tumor microenvironment, tumor immunology, skin cancer, multi-omics, CRISPR screen INTRODUCTION Epithelial cancers comprise ~90% of human malignancies. Among these, cutaneous squamous cell carcinoma (cSCC) serves as a prototype, given its cardinal features such as disruption of tissue polarity and basement membrane invasion. cSCC is the second most common malignancy in the U.S., with an annual incidence over 1 million (Rogers et al., 2015). Although excision can be curative, up to 4% of cSCCs develop nodal metastases and approximately 1.5% of patients succumb to the disease (Karia et al., 2013). The recent approval of cemiplimab, an inhibitor of programmed death 1 (PD-1), for advanced and metastatic disease is promising (Migden et al., 2018), but resistance and inadequacy of biomarkers to predict response (Keenan et al., 2019) emphasize the need for better characterization of immunosuppression in the tumor microenvironment (TME). Single-cell transcriptomics has revealed intra-tumoral heterogeneity (ITH) within many cancer types, identifying cell populations that drive drug resistance, predict metastatic risk, and mediate plasticity (Neftel et al., 2019, Puram et al., 2017, Tirosh et al., 2016a). However, studies often suffer from lack of normal tissue comparisons needed to identify tumor-specific biology and loss of spatial information during tissue dissociation. Recent advances, however, enable simultaneous capture of the locations of dozens of cell types within the TME, which is critical for understanding tumor-stroma crosstalk (Angelo et al., 2014, Keren et al., 2018, Salmén et al., 2018a, Thrane et al., 2018). Orthogonal integration of single-cell and high-dimensional spatial data from both normal and diseased tissue should therefore facilitate the dissection of TME cellular communication. Here, we combine single-cell and spatial transcriptomics (ST) with single-cell resolution multiplexed protein imaging of a series of primary human cSCCs, along with matched normal skin. 48,164 single-cell transcriptomes were characterized from 10 tumor-normal skin pairs using single-cell RNA sequencing (scRNA-seq). These data were integrated with ST profiling of 17,064 spots and with multiplexed ion beam imaging (MIBI) of 55,832 cells. Tumors included basal, cycling, and differentiating keratinocyte (KC) populations similar to normal skin, along with a tumor-specific keratinocyte (TSK) population. TSK cells localized to leading edges, as did a population of basal tumor cells. TSK, basal, and adjacent stromal and immune cell types exhibited invasive and immunosuppressive features associated with physical proximity and distinct sets of ligands and receptors. Xenograft tumors and TME recapitulated spontaneous human cSCC, facilitating in vivo CRISPR screens that identified an essential tumorigenic function for TSK-enriched integrin signaling genes ITGB1, FERMT1, and CD151. Together, these results uncover cSCC spatial heterogeneity, highlight potential intercellular signals controlling the positioning and states of associated cell types, and serve as a resource for further investigation of tumor and immune dynamics. RESULTS Multimodal Profiling of cSCC We performed scRNA-seq on tumors and patient- and site-matched normal skin collected from 10 individuals presenting for surgical resection using the 10X Genomics Chromium platform (Figure S1A; STAR Methods). Prior to dissociation, portions of tumors were fresh frozen and formalin-fixed for ST (Salmén et al., 2018b, Ståhl et al., 2016) and MIBI (Angelo et al., 2014), respectively (Figure 1A). 50,009 cells were sequenced at a median depth of 115,216 reads/cell, with 48,164 cells passing filtering utilizing Seurat (Butler et al., 2018) (Table S1; STAR Methods). Unsupervised clustering identified tumor and normal epithelial cells as well as stromal subtypes (Figure 1B; STAR Methods). In addition to fibroblasts, endothelial cells, and melanocytes, the latter included CLEC9A, CD1C, plasmacytoid, and AXL+SIGLEC6+ (AS) dendritic cells (DCs) (Villani et al., 2017), Langerhans cells, macrophages, and myeloid-derived suppressor cells (MDSCs) (Figures 1C and S1C). Natural killer (NK) and T cell clusters included CD4 and CD8 conventional and exhausted T cells, regulatory T cells (Treg), and naive-like T cells (Figure 1D), all bearing resemblance to subsets from prior solid tumor scRNA-seq studies (Guo et al., 2018, Puram et al., 2017, Savas et al., 2018, Tirosh et al., 2016a, Yost et al., 2019, Zhang et al., 2018, Zheng et al., 2017). Non-tumor clustering was driven predominantly by cell type rather than batch effects (Figures S1B and S1C). Cell-type proportion analysis reflected consistent sample processing (Figures 1E and 1F). Whole-exome sequencing found a mutational landscape consistent with prior cSCC studies (Lee et al., 2014, Li et al., 2015, Pickering et al., 2014, South et al., 2014) (Figures S1D–S1F). Taken together, these data provide a representative cellular atlas for cSCC. [Figure: scRNA-seq and Whole-Exome Sequencing of Normal Skin and cSCC, Related to Figure 1(A) Patient cohort characteristic table. (B) UMAP of all cells (k = 48,164) recovered labeled by patient on the top and tissue type (normal skin or tumor) on the bottom. (C) UMAP feature plots showing expression of marker genes by annotated cell types. (D) Types of mutations in select genes in patient cSCC samples. (E) Proportion of transitions and transversions mutations across patients. (F) Tumor mutational burden across patients. MB = megabase.] [Figure: A Single-Cell Transcriptomic Atlas of Normal Skin and cSCC(A) Workflow of patient sample processing for scRNA-seq, MIBI, and ST with integration of a xenograft CRISPR screen.(B) Uniform manifold approximation and projection (UMAP) of scRNA-seq cells recovered from both normal skin and cSCC labeled by cell type.(C) UMAP of myeloid subsets, including dendritic cells (DCs), macrophages, and monocytic myeloid-derived suppressor cells (MDSCs) (marker genes in parentheses).(D) UMAP of subsets of natural killer (NK) and T cell subsets, including CD4+ and CD8+ conventional T cells and associated subsets, and regulatory T cells (Treg). Pre-Exh, pre-exhausted; TEM, effector memory T cell; TEMRA, recently activated effector memory T cell.(E) Bar plots of proportion of cell type by patient, tumor or normal origin, and total cell number.(F) Similar to (E), for NK cell and T cell subsets.See also Figure S1 and Table S1.] A TSK Subpopulation After re-clustering normal epithelial cells, three major KC subpopulations were observed in normal skin, each expressing known representative genes, including (1) basal (COL17A1), (2) cycling (MKI67), and (3) differentiating (KRT1) cells (Figures 2A and S2A–S2D; STAR Methods). These correspond to known functions of cells within