Dissecting transcriptional heterogeneity in primary gastric adenocarcinoma by single cell RNA sequencing

Abstract

Objective: Tumour heterogeneity represents a major obstacle to accurate diagnosis and treatment in gastric adenocarcinoma (GA). Here, we report a systematic transcriptional atlas to delineate molecular and cellular heterogeneity in GA using single-cell RNA sequencing (scRNA-seq).

Design: We performed unbiased transcriptome-wide scRNA-seq analysis on 27 677 cells from 9 tumour and 3 non-tumour samples. Analysis results were validated using large-scale histological assays and bulk transcriptomic datasets.

Results: Our integrative analysis of tumour cells identified five cell subgroups with distinct expression profiles. A panel of differentiation-related genes reveals a high diversity of differentiation degrees within and between tumours. Low differentiation degrees can predict poor prognosis in GA. Among them, three subgroups exhibited different differentiation grade which corresponded well to histopathological features of Lauren's subtypes. Interestingly, the other two subgroups displayed unique transcriptome features. One subgroup expressing chief-cell markers (eg, LIPF and PGC) and RNF43 with Wnt/β-catenin signalling pathway activated is consistent with the previously described entity fundic gland-type GA (chief cell-predominant, GA-FG-CCP). We further confirmed the presence of GA-FG-CCP in two public bulk datasets using transcriptomic profiles and histological images. The other subgroup specifically expressed immune-related signature genes (eg, LY6K and major histocompatibility complex class II) with the infection of Epstein-Barr virus. In addition, we also analysed non-malignant epithelium and provided molecular evidences for potential transition from gastric chief cells into MUC6⁺TFF2⁺ spasmolytic polypeptide expressing metaplasia.

Conclusion: Altogether, our study offers valuable resource for deciphering gastric tumour heterogeneity, which will provide assistance for precision diagnosis and prognosis.

Keywords: endoscopic gastrostomy; epithelial cells; gastric cancer; molecular pathology.

Figure 1

Cellular atlas of gastric tumours and non-tumour gastric tissues. (A) Schematic diagram for the generation of single-cell RNA sequencing (scRNA-seq) data. Nine gastric tumours (seven endoscopic samples and two surgical samples) and three non-tumour samples (one chronic gastritis sample and two normal samples) were collected. (B) t Stochastic neighbour Embedding (tSNE) plots for the 27 677 high-quality cells showing sample origin and Lauren’s classification. (C) tSNE plots showing cell types for the 27 677 cells. (D) Violin plots showing the smoothed expression distribution of marker genes in nine cell types. (E) tSNE plots showing the expression levels of canonical marker genes for nine cell types. (F) The proportion of each cell type in 12 samples.

Figure 2

Classification of 10 411 epithelial cells as malignant or non-malignant. (A) tSNE of 10 411 epithelial cells, colour-coded according to malignant score minus non-malignant score. (B) Scatter plot displaying the distribution of malignant scores (x-axis) and non-malignant scores (y-axis). Each point corresponds to a cell and is colour-coded to reflect density. (C) tSNE plot of the classification of malignant and non-malignant cells. (D) Violin plots and corresponding box plots showing the expression of eight representative genes with differential expression between malignant and non-malignant cells. (E) Expression of eight representative genes with differential expression, shown using tSNE plots. (F) Bar plot showing fold changes of signature genes in malignant cells and non-malignant cells. (G) Gene set enrichment analysis (GSEA) results showing the enrichment of six gastric tumour-associated gene sets in malignant epithelial cells. EMT, epithelial-mesenchymal transition; IL6, interleukin-6; JAK, Janus kinase; NF-κB, nuclear factor-kappa B; TNF, tumour necrosis factor-α; STAT, signal transducer and activator of transcription.

Figure 3

Cell clusters of non-malignant epithelial cells and potential transition between cell types. (A) tSNE plot of 4776 non-malignant epithelial cells colour-coded for sample origin. (B) tSNE plot of non-malignant epithelial cells colour-coded to reflect cell cluster. (C) Violin plots showing the smoothed expression distribution for marker genes for each cell type. (D) tSNE plot of non-malignant epithelial cells, colour-coded for the expression of marker genes. (E) Immunostaining showing the spatial distribution of chief cell, neck cell and spasmolytic polypeptide expressing metaplasia (SPEM). (F) Monocle 2 pseudotime analysis for chief cells, neck cells and SPEM. (G) Heatmap showing scaled expression of dynamic genes along the pseudotime. Rows of the heatmap represent genes that show dynamic changes along the pseudotime, and these genes were clustered into three groups according to their expression pattern along the pseudotime. (H) A simple model for the origin and progression of gastric metaplasia cells in human stomach.

Figure 4

The landscape of intratumour and intertumour heterogeneity in gastric adenocarcinoma (GA). (A) tSNE plot of 5635 tumour cells, colour-coded for pathology and five molecular clusters. (B) Sankey diagram showing the distribution of Lauren’s histological characteristics and Epstein-Barr virus (EBV) infection in five clusters. (C) Fraction of five cell subgroups in nine patients. (D) Relative expression of signature genes with adjusted p value <0.01 and fold change >1. (E) Average expression of seven epithelial differentiation-related genes in five subgroups. Right, Pearson’s correlation coefficient between seven genes and KRT20. (F) tSNE plot showing differentiation scores of 5635 tumour cells. (G) Histogram of differentiation scores in nine patients. The red dash lines partition two peaks of differentiation score distribution. (H) Trajectory of malignant cells in IGC1 and IGC4 constructed by Monocle 2. Each point corresponds to a single cell and is colour-coded by cell subgroup (top) and differentiation score (bottom). (I) The overall survival of Diff-high and Diff-low groups of patients. P=0.0032, calculated using log rank test.

Figure 5

Molecular characteristics of subgroup C1, C2 and C3. (A) tSNE plot of tumour cells, colour-coded for C1, C2 and C3. (B and C) Violin plots showing the expression of CLDN18, TFF2, APOA1 and FABP2. (D) Violin plots showing differentiation scores of five clusters. (E) Violin plots showing the expression of KRT20, PHGR1 and MDK in five cell subgroups.

Figure 6

Molecular characteristics of subgroup C4 and definition of GA-FG-CCP. (A) tSNE plot of tumour cells, coloured for C4. (B) Violin plots showing the expression of chief cell markers, differentiation-related marker and C4-specific genes in five subgroups. (C) The enriched gene ontology terms for top 50 signature genes in C4 using Metascape (www.metascape.org). (D) Immunofluorescence staining indicates the co-expression of MUC6, PGA3 and DAPI (nuclei) on malignant cells of GA-FG-CCP. Scale bars, 100 µm. (E) Heatmap showing the clustering result of 86 intestinal-type patients in TCGA STAD dataset. The colours in heatmap correspond to the concordance between samples (see ‘Methods’ section in online supplementary materials). Upper panel showing the relative expression of four specific genes in C4 and two differentiation-related markers. (F) Box plot showing the average expression of C4-specific genes selected for clustering of TCGA samples in four clusters. (G) Box plot showing the expression CTNNB1 in four clusters of TCGA samples.

Figure 7

Molecular characteristics of subgroup C5. (A) tSNE plot of tumour cells, coloured for C5. (B) tSNE plots showing the expression of HLA-DPA1, HLA-DPB1 and LY6K in tumour cells. (C) Violin plots showing the expression of HLA-DPA1, HLA-DPB1 and LY6K in five subgroups. (D) The enriched gene ontology terms for top 50 signature genes in C5 using Metascape (www.metascape.org). (E) Immunofluorescence staining indicates the co-expression of HLA-DP, KRT18 and DAPI (nuclei) on malignant cells of Epstein-Barr virus (EBV)+ gastric adenocarcinoma (GA). Scale bars, 100 µm. (F) Immunohistochemistry staining indicates the expression of LY6K on EBV+ and EBV− tumour samples, Scale bars, 100 µm. (G) Box plots showing the expression HLA-DPA1 and HLA-DPB1 in EBV+ and EBV− patients of TCGA STAD dataset, p value calculated by Student’s t-test. (H) Fraction of G1, S and G2M B cells in 12 samples.