Article
topic: Single-cell profiling of tumor heterogeneity and the microenvironment in advanced non-small cell lung cancer Journal: Nature Communication (IF:12) Date: May 2021 DOI: https://doi.org/10.1038/s41467-021-22801-0
_blockquote img18 strong
img1
0p8 overview
Lung cancer is a highly heterogeneous disease. Cancer cells and cells in the tumor microenvironment jointly determine the progression of the disease and the response or escape of treatment. In order to map the cell type-specific transcriptome of cancer cells in advanced non-small cell lung cancer (NSCLC) and their tumor microenvironment, the authorThrough single-cell RNA sequencing, 42 tissue biopsy samples from patients with stage III/IV NSCLC were analyzed, showing the condition of advanced non-small cell lung cancer at single-cell resolution. In addition to the cell types described in previous early-stage lung cancer single-cell studies, the authors also identified rare cell types in tumors, such as follicular dendritic cells and T helper 17 cells. Tumors from different patients show great heterogeneity in cell composition, chromosome structure, developmental trajectory, intercellular signal network and phenotypic advantage. The author's study also revealed the correlation between tumor heterogeneity and tumor-related neutrophil , which helps to clarify their function in NSCLC.
Sample information
The primary tumors of 42 advanced (stage III/IV) NSCLCs were collected. The cancer type and smoking status are shown in the figure.
single cell sequencing technology
microfluidic chip (see the original text for details)
_span11span
_span11span single cell sequencing technology
microfluidic chip (see the original text for details) , A total of 90,406 cell transcriptomes were analyzed.
main single-cell analysis results
Late NSCLC cell profile
p0
Including cancer cells, myeloid cells, fibroblasts, T cells, B cells, neutrophils, alveolar cells, epithelial cells, endothelial cells, mast cells and follicular trees Dendrite cells. The lower left image shows that non-tumor cells from different patients tend to cluster together, while cancer cells have patient heterogeneity. The two graphs on the right show the heat map of marker genes and the percentage of cell subpopulations.
lung squamous cell carcinoma has higher inter-tumor and intra-tumor heterogeneity than lung adenocarcinoma
img4
_p0 p2 p2 p img
img 42 The CNA spectrum shows heterogeneity between and within patients (a). For patients with lung adenocarcinoma (LUAD), significant arm-level insertions were found on chromosomes 7 and 8q, and deletions were found on chromosome 10. It is worth noting that lung adenocarcinomas with known driver mutations have additional amplification in the 1q and 5p arms. In contrast, most patients with lung squamous cell carcinoma (LUSC) have 3q insertions and 5q deletions. Interestingly, some patients with lung adenocarcinoma without driver mutations have a CNA profile similar to that of lung squamous cell carcinoma. Although the expression profile and composition of the cancer cell transcriptome are largely patient-specific, some patients have cancer cells that are more similar than others (b). In most cases, cancer cells from patients with lung adenocarcinoma and lung squamous cell carcinoma are divided into different clusters. More than half of lung adenocarcinoma patients clustered into one group, and most lung squamous carcinoma tumors formed patient-specific clusters, indicating that the difference between tumors in LUSC was higher than that in lung adenocarcinoma. Most patients, especially lung adenocarcinoma patients, such as P16, P20 and P32,There are dominant clones, while in a few lung squamous cell carcinomas (such as P27 and P37), the malignant cells spread to multiple clusters.
In order to quantify intra-tumor heterogeneity, the authors defined CNA-based and expression-based intra-tumor heterogeneity scores, expressed as ITHCNA and ITHGEX. The authors observed varying degrees of heterogeneity within the tumor. ITHCNA and ITHGEX showed a moderate correlation (Figure d), which may be due to non-driver genome changes or tumor phenotypes formed by the microenvironment. The authors further divided the patients into three groups based on cancer type and mutation: LUAD patients with driver mutations (n = 12), denoted as LUADm, LUAD patients without driver mutations (n = 6), denoted as LUADn, and no driver mutations Of LUSC patients with mutations (n = 16), denoted as LUSCn. Interestingly, compared with LUADm patients, LUSCn patients have significantly higher ITHCNA, which is not statistically significant in terms of ITHGEX (e). This finding also suggests that patients with driver mutations may be affected by phenotypes other than genomic changes.
The plasticity of lung epithelial cells and their developmental trajectory to malignant tumor cells
This part is a pseudo-chronological analysis. All identified alveolar cells express the typical markers of type 2 alveolar cells (AT2) (CLDN18, SFTPA1, SFTPC), but do not express the type 1 alveolar cell markers (CAV1, AGER). Further cluster analysis of revealed two different AT2 cell clusters, denoted AT2-1 and AT2-2 (a). AT2-2 is similar to the normal AT2 phenotype, with the common AT2 marker SFTPA and the transporter ABCA3 up-regulated (b). In contrast, AT2-1 expressed cell proliferation and cell migration related genes,Such as CEACAM6, KITLG and FOXC1, suggesting phenotypic changes to malignant tumors. Epithelial cells can be further divided into ciliated epithelial cells, rod-shaped cells and basal cells (c, d).
Previous studies have shown that both AT2 cells and club cells can develop into LUAD cells, and basal cells and club cells are potential progenitor cells of LUSC. Therefore, the author organized AT2 cells, club cells and LUAD cancer cells according to their developmental trajectories (e). The inferred pseudo-time path shows that AT2 cells and club cells independently transform into LUAD tumors. In contrast, basal cells appear to act as a transitional state between rod-shaped cells and LUSC tumor cells (f). In addition to these different characteristics, we found that some patients' tumor cells clustered tightly at the ends of the branches, which means homogenous and terminal phenotypes, while other patients have more diverse and heterogeneous characteristics, which develop along the cancer. The trajectory spreads.
Th17-like cells and their potential interconversion with Tregs
Among tumor-invasive T cells, CD4+ TCD4+ TCD4+ naïve Ts help cells, CD4+ TCD4+ naïve Ts help 17 -like T cells (Th17-like), CD8+ effector T cells, CD8+ exhausted T cells, and Natural Killer (NK) cells (a), T cell subtypes are annotated by supervised cell types based on the previously studied T subtype expression profile Confirm (a). In order to further characterize the two NK clusters (CD3D-, KLRD1+, NKG7+), the author refers to the CD16+ (FCGR3A) cluster as NK-1 and the CD16- cluster as NK-2 (Figure 4b). NK-1 contains the up-regulated transcript fractalkine receptor (CX3CR1) and fibroblast growth factor binding protein 2 (FGFBP2),Both are involved in the cytotoxic function of lymphocytes. NK-2 has higher expression of tissue resident markers, such as CD49a (ITGA1), CD103 (ITGAE) and ZNF683. Co-suppressive immune checkpoints including CTLA4 and TIGIT are rich in CD4+ Treg and CD8+ depleted T cells (c). However, LAG3 is mainly expressed in CD8+ depleted T cells, which is consistent with the results of previous studies.
Then, the author used Slingshot18 and Monocle19 to analyze the trajectory of CD4+ T cells to determine their developmental pathways in the TME. Slingshot revealed the transitional relationship between Tregs and Th17-like cells, which originated from naive cells (d). Monocle found that immature cells differentiate into two main branches, Tregs and proliferating populations (e, f). Interestingly, Th17-like cells confirmed by the expression of its master transcription factor RORC show a transitional phenotype that spreads along the developmental pathway from naive cells to Tregs (f). According to the author, the CD4+ Th17-like cell population marked by the high expression of the KLRB120 gene is the first report of Th17-like cells identified by scRNA-seq in the NSCLC tumor environment. It has been reported in the literature that natural Tregs (nTregs), a subset of Tregs, are considered to transform into Th17-like cells. This result reveals the complex and subtle interactions between Tregs and Th17-like cells, and emphasizes the importance of their balance in the adaptive immune response to tumor antigens.
.
