Metabolic shifts in lipid utilization and reciprocal interactions within the lung metastatic niche of triple-negative breast cancer revealed by spatial multi-omics

Abstract

The Triple-Negative Breast Cancer (TNBC) subtype constitutes 15-20% of breast cancer cases and is associated with the poorest clinical outcomes. Distant metastasis, particularly to the lungs, is a major contributor to the high mortality rates in breast cancer patients. Despite this, there has been a lack of comprehensive insights into the heterogeneity of metastatic tumors and their surrounding ecosystem in the lungs. In this study, we utilized spatial RNA sequencing technology to investigate the heterogeneity of lung metastatic tumors and their microenvironment in two spontaneous lung metastatic mouse models. Our findings revealed an increase in metabolic-related genes within the cancer cells, with the hub gene Dlat (Dihydrolipoamide S-Acetyltransferase) showing a significant association with the development of lung metastatic tumors. Upregulation of Dlat led to the reprogramming of fatty acid utilization, markedly enhancing the bioenergetic capacity of cancer cells. This finding was corroborated by the increased dependence on fatty acid utilization in lung metastatic cancer cells, and inhibition of Dlat in breast cancer cells exhibited a reduced oxygen consumption rate. Consequently, inhibition of Dlat resulted in decreased survival capacity of breast cancer by reducing cancer stem cell properties and cell adhesion in the lung in vivo. The three cell components within the lung metastatic niche, including CD163⁺ macrophages, neutrophils, and endothelial cells, expressed elevated levels of ApoE, leading to the secretion of various protumorigenic molecules that promote cancer cell growth in the lung. These molecules include galectin-1, S100A10, S100A4, and S100A6. Collectively, our findings highlight the lipid metabolism reprogramming of cancer and components of the tumor microenvironment that support lung metastasis of TNBC breast cancer.

Fig. 1. Spatial transcriptomics (ST) of metastatic lung nodules in the 4T1 mouse model.

A Experimental workflow. Lung sections from two mouse models were utilized to generate spatial data. H&E stain, t-SNE representation, and spatial distribution were identified by clustering the integrated ST dataset across sections from the lungs of the 4T1 model (n = 6 for spontaneous metastasis, n = 2 for ST) (B) and the MMTV-PyVT (C) models (n = 2 for ST). Malignant status of different tumor nodules in the lung of 4T1 model (D) and MMTV-PyVT (E) model. CytoTRACE analysis of varying tumor nodules in the lung of the 4T1 model (F) and the MMTV-PyVT model (G). H The survival rate of breast cancer expressing the top 100 genes of aggressive cancer nodules (Cancer-B1). OS, overall survival; PSF, progression-free survival; DSS, disease-specific survival; DFI, disease-free interval.

Fig. 2. The hub networks contributed to cancers and TMEs.

A The trajectory analysis of lung metastatic cancer nodules in the 4T1 mouse model. B The soluble factors expressed in different tumor nodules of the lung in the 4T1 model. C The KEGG pathway of genes contributing to the transition of different tumor nodules in the lung in the 4T1 model. D The change of Dlat expression in trajectory analysis. E The expression of Dlat in different tumor nodules in the lung of the 4T1 model. F The trajectory analysis of primary and lung metastatic tumor nodules in the MMTV-PyVT mouse model. G The change of Dlat expression in trajectory analysis and (H) The expression of Dlat in different tumor nodules in the primary sites and lung of mice with the MMTV-PyVT mouse model.

Fig. 3. Elevated Dlat associated with lung metastasis of TNBC.

The expression of Dlat in the tumor of the primary site and lung nodules in the 4T1 model (n = 6) (A) and the MMTV-PyVT model (n = 6) (B). C The expression of Dlat in 4T1 cells and MMTV cells. D Inhibition of Dlat decreased cell adhesion (24 h) and tumor growth (7 days) in lung of mice. Knockout of Dlat reduced cell proliferation of 4T1-L (E) and MMTV-L cells (F). Inhibition of Dlat reduced CSC property in 4T1-L (G) and MMTV-L cells (H) and, as determined by tumor spheroid formation and ALDH activity (I, J). All results were representative of at least three independent experiments. Graphs shown as mean ± S.D. **p-value < 0.01.

Fig. 4. The alternation of lipid metabolism.

GSEA analysis of Dlat-positive and negative spots in the lung metastatic nodules in the 4T1 (A) and MMTV-PyVT (B) mouse model. The KEGG pathway of DEG between Dlat-positive and negative spots in the 4T1 (C) and MMTV-PyVT (D) mouse models. The oxygen consumption rate (OCR) (E), various parameters of respiration (F) in 4T1 cells. The OCR (G) and various respiration parameters (H) in MMTV. Percentage of dependency of mitochondrial respiration to oxidize three main energetic fuels: glucose, glutamine, and fatty acids in 4T1 (I) and MMTV (J). The OCR of Dlat-knockdown (K) and knockout (L) 4T1 cells. M Knockdown of Dlat reduced lung metastasis in 4T1 cells in vivo (n = 6). Graphs shown as means ± SD of experimental triplicates. *p-value < 0.05; **p-value < 0.01; ***p-value < 0.001; ****p-value < 0.0001. DEGs, differentially expressed genes.

Fig. 5. The infiltration of Cd68⁺Cd74⁺ApoE⁺ macrophage in TMEs.

The score of macrophages (A), dendritic cells (B), and Neutrophils (C) in TMEs. KEGG pathway of top100 genes of Cancer-B1/2/3-TME (D) and Cancer-B4-TME (E). F M2 macrophage score of Cd68⁺Cd74⁺ApoE⁺ and Cd68⁺Cd74⁺ApoE^- spots. G The cell number of Cd68⁺Cd74⁺ApoE⁺ macrophages in the lungs of mice (n = 10 for control, n = 6 for 4T1 model). H The expression of M2 markers (Cd163 and Arginase I) in ApoE⁺ and ApoE^-Cd68⁺Cd74+ cells isolated from the lung of mice with 4T1 (n = 6 for 4T1 model). H The expression of AopE in THP-1 after Hs578T stimulation. The expression of ApoE (I), Cd206 (J), Cd163 (K), Arginase I (L), and IL-10 (M) in macrophage co-cultured with Hs578T cells. All results were representative of at least three independent experiments in vitro. Graphs shown as mean ± S.D. *p-value < 0.05; **p-value < 0.01; ***p-value < 0.001; ****p-value < 0.0001.

Fig. 6. The factors contributed to the remodeling of TMEs.

A Trajectory analysis of all TMEs. The genes associated with Cancer-B1/2/3-TMEs (B), Cancer-B4-TME (C), and Diffusive cancer-TMEs (D). E The correlation of various soluble factors with macrophage markers. F The presence of CD68⁺ApoE⁺S100A4⁺S100A6⁺Galectin-1⁺ macrophages in lung TME in vivo (n = 6). G, H The interaction of Cancer-B1/2/3 with their TMEs. I, J The interactions between Cancer-B1/2/3-TMEs with Cancer-B1/2. K The interactions within Cancer-B1/2/3-TMEs.

Fig. 7. Spatial proteomics of lung metastatic niche.

A The cell clusters of lung TME were conducted using CODEX staining. B The spatial analysis of cell clusters in lung TME. C The odd ratio of cell communication between different cell clusters. D CD163⁺ApoE⁺ macrophages expressed S100A10 and S100A4 protein. E ApoE⁺Ki67⁺ cancer cells expressed higher levels of vimentin. F CD31⁺ApoE⁺ endothelial cells expressed S100A10 and Thbs1 protein. G The expression of S100A4 and Thbs1 protein in Ly6g⁺ neutrophils. H The levels of ApoE, S100A4, and S100A10 in the macrophages isolated from the lung with or without 4T1 tumor nodules (n = 6). I The levels of ApoE and S100A4 in the neutrophils isolated from the lung with or without 4T1 tumor nodules. Graphs are shown as means ± SD of experimental triplicates. *p-value < 0.05; **p-value < 0.01; ***p-value < 0.001; ****p-value < 0.0001.