Single-cell atlas of lung adenocarcinoma (LUAD) tissues from the control, and NCT group. (a) Workflow depicting collection and processing of LUAD samples for scRNA-seq analysis. (b) Consensus clustering based on the correlations among the 20 clusters identified through the tSNE algorithm. (c) TSNE of the 83622 cells enrolled here, with each cell color indicating: its sample type of origin, the corresponding patient, predicted cell type, and the transcript counts. (d) Expression of marker genes for the cell types defined above each panel. (e) The proportion of each cell type in different groups and samples. (f) For each of the eight epithelial subclusters and 43 non-epithelial clusters (left to right): the fraction of cells originating from the three groups, the fraction of cells originating from each of the nine patients, the number of cells and box plots of the number of transcripts (with plot center, box, and whiskers corresponding to the median, IQR and 1.5 × IQR, respectively). NCT: Neoadjuvant chemotherapy.

Metabolic reprogramming in lung adenocarcinoma driven by neoadjuvant chemotherapy. (a) The metabolic pathway activities of different cells from the Control group and NCT group showed significant differences. (b) The metabolic pathway activity of macrophages and malignant cells increased significantly after chemotherapy.

Tumor cells and epithelial cells had significant phenotypic changes before and after chemotherapy. (a) The tSNE plots and overview of the tumor cells and epithelial cells. (b) The proportion of malignant cells and nonmalignant cells in the Control group and NCT group. (c) Flow cytometry showed that the proportion of malignant cells was significantly reduced after chemotherapy and immunotherapy. (d) Marker genes of MalignantSA cells, MalignantNCT cells, and Nonmalignant cells. (e) Pseudotime analysis showed that nonmalignant cells evolved in two directions. (f) The heat map showed that a series of genes play an important role in transforming epithelial cells into tumor cells. (g) GSVA analysis was performed for malignant and nonmalignant cells. (h) SCENIC analysis revealed the hub genes in the malignant transformation of epithelial cells. (i) The tSNE plots for re-clustered malignant cells. (j) Marker genes of 13 sub-clusters from malignant cells. (k) Metabolic characteristics in different malignant cell sub-clusters. (l) GSVA analysis reveals the characteristics of pathway activity in different malignant cell sub-clusters.

The scRNA profile of stromal cells derived from LUAD samples in control and NCT groups. (a) The tSNE plots an overview of the 6 clusters of stromal cells. (b) Proportions of the six predicted clusters of stromal cells in different groups and samples. (c) Heatmap exhibiting the expression level of marker genes in each stromal cell cluster. (d) GSVA analysis estimated the pathway activation levels of different stromal cell subtypes. The scores have been normalized. (e) GSVA analysis revealed the activation level of hallmark pathways in stromal cells (control vs. NCT groups) (f) Heatmap exhibiting the expression level of marker genes in each fibroblast cluster. (g-h) The tSNE plots revealed the group origins (g) and predicted subclusters (h) of fibroblast. (i-j) GSVA analysis estivated the pathway activation levels of different fibroblast subtypes.

Three newly identified subtypes of tumor-associated macrophages (TAMs) displayed distinct genetic and metabolic features. (a) The tSNE plots showed the group origins, sample origins, and clusters of TAMs. (b) Proportions of the 10 clusters of TAMs in different groups and samples. (c) Marker genes of the 10 clusters of TAMs. (d) GSVA analysis was performed for the 10 clusters of TAMs. (e) The activity of various metabolic processes in the 10 clusters of TAMs. (f) GSVA analysis was performed for TAMs from the Control group and NCT group. (g) SCENIC analysis was performed for the 10 clusters of TAMs. (h) Consensus clustering based on the correlations among the 10 clusters of TAMs identified through the tSNE algorithm. (i) Polarization score (left) and inflammatory score (right) for 10 clusters of TAMs based on the expression of polarization marker genes and inflammatory genes. (j) The tSNE plots for three types of TAMs. (k) Proportions of the three subtypes of TAMs in different groups and samples. (l) Development trajectory analysis for the three subtypes of TAMs. (m) Pseudotime analysis revealed a series of genes that affect the differentiation and development of macrophages.

Metabolic switching in tumor-associated macrophages (TAMs) contributed to diametrical effects on tumor cells. (a) The heatmap showed the essential marker genes for three subtypes of TAMs. (b) Based on Pro-mac and Anti-mac marker genes, these two types of cells were sorted by flow cytometry from lung adenocarcinoma tissue. (c) Flow cytometry verified the sorted cells. (d) The proportion changes of Pro-mac and Anti-mac cells in lung adenocarcinoma tissues before and after chemotherapy. (e) Immunofluorescence showed the changes in the proportion of TAMs with high CD206 and CD86 after neoadjuvant therapy. (f) The heatmap showed the differences in the cytokines secreted by the three subtypes of macrophages. (g) ELISA detected the secretion of VEGF, EGF, IL10, and MMP9. (h) The intensity of fluorescence changes in Luciferase-labeled A549 cells mixed with different TAMs. (i) The histogram showed the average fluorescence intensity emitted by the subcutaneous tumor. (j) GSVA analysis performed for Pro-mac, Anti-mac, and Mix. (k) Glucose uptake and lactate production in the TAMs cell subtypes. (l) Seahorse XFe96 cell outflow analyzer detected the glycolysis level of TAMs cell subtypes (Extracellular acidification rate: ECAR). (m) Transwell experiment detected the influence of TAMs subtypes on the invasion ability of A549 cells. (n) The 3D cell culture experiment detected the effect of 2-DG on the spheroidization ability of A549 cells when cultured with subtypes of TAMs. (o) In vivo experiments verified the effect of 2-DG on the tumorigenesis ability of A549 after inhibiting glycolysis of TAMs. All error bars are mean ± SD. NS, not significant. ***P < 0.001, **P < 0.01, *P < 0.05; determined by two-tailed Student’s t-test (95% confidence interval).

Figure related to Figure 1. (a) The tSNE plots of all clusters in this research. (b) The heatmap showed the marker genes in the different cell types.

Figure related to Figure 3. (a) Copy number variations (CNVs) of malignant and nonmalignant epithelial cells. (b) The proportion of malignant cells and nonmalignant cells from different patients. (c) The heatmap showed differentially expressed genes of epithelial cells in the Control group and NCT group. (d) Immunofluorescence showed LPCAT1, FOXL2, and RAC1 were highly expressed in normal lung tissues, the Control group, and the NCT group, respectively. (e) Some genes played an important role in the transformation of epithelial cells into tumor cells. (f) Changes in the activity of several important pathways during the transformation of epithelial cells to tumor cells. (g) The tSNE plots and overview of the non-malignant cells. (h) Heat map of marker genes for nonmalignant cells sub-clusters. (i) Metabolic characteristics in different nonmalignant cell sub-clusters. (j) GSVA analysis revealed the characteristics of pathway activity in different nonmalignant cell sub-clusters.

Figure related to Figure 4. (a) The activity of various metabolic processes in fibroblasts from the Control group and NCT group. (b) GSVA analysis was performed for fibroblasts from the Control group and NCT group. (c) SCENIC analysis revealed the hub genes in fibroblast.

The scRNA profile of B cells derived from LUAD samples in the control, neoadjuvant chemotherapy, and immunotherapy group. (a) The tSNE plots revealed the sample origins, group origins, and predicted clusters of B cells. (b) The two predicted clusters of B cells (plasma cells, class-switched memory B-cells) were reported in different groups and samples. (c) Heatmap exhibiting the expression level of marker genes in each B cell cluster. (d) GSVA analysis estivated the pathway activation levels of different B cell subtypes.

Crosstalk between cancer and immune cells. (a) Each cell type and the other cell types expressed some of the ligands. (b) Bubble plot revealing the specific ligand-receptor interactions between cancer cells and immune cells in the control group. The circle size indicates P values, with the scale to the right (permutation test), and color indicates communication probability.

The scRNA profile of T cells derived from LUAD samples in the control, neoadjuvant chemotherapy, and immunotherapy group. (a-c) The tSNE plots revealed the sample origins (a), group origins (b), and predicted clusters (c) of T cells. (d) The three indicated clusters of T cells (CD4+ T, CD8+ T, and Tregs) were reported in different groups and samples. (e) Bubble plot exhibiting the expression level of marker genes in each T cell cluster. (f-h) The tSNE plots revealed the sample origins (f), group origins (g), and predicted subclusters (h) of CD8+ T cells. (i) Heatmap exhibiting the expression level of marker genes corresponding to naïve, Treg/exhausted, and cytotoxic phenotypes in each CD8+ T cell subcluster. (j-k) Proportions of the five predicted clusters of CD8+ T cells in different samples (j) and groups (k). (l) Dynamic changes in gene expression of CD8+ T cells during the transition (divided into three phases). (m-p) Pseudotime-ordered analysis of CD8+ T cells (m-n) revealing the dynamics of their cytotoxic (o) and exhausted levels (p). (l) SCENIC analysis of CD8+ T cells.

Crosstalk between cancer and immune cells. (a) Overview of selected ligand-receptor interactions of cancer cells and immune cells in control and NCT groups. The line thickness indicates the number of ligands when cognate receptors are present in the recipient cell type. The loops indicate autocrine circuits. (b) Detailed view of the ligands expressed by each cell type and the other cell types. (c) Bubble plot revealing the specific ligand-receptor interactions between cancer cells and immune cells in the NCT group. The circle size indicates P values, with the scale to the right (permutation test), and color indicates communication probability.