Figure 6. Specific oxidative stress in patient-derived lung metastasis tissues.

(A–D) Expression of nuclear Complex I and antioxidant genes in a gene expression data set of matched primary tumors and lung metastases from patients with breast cancer (Siegel et al., 2018). (A) GSEA analysis of the expression of Complex I genes shows higher expression of these genes in lung metastases (LM) compared to primary tumors (Primary); (B) Complex I genes, lung antioxidant genes (from Figure 5D), and NRF2 response signature genes are upregulated in lung metastases (LM) compared to matched primary tumor (Pr). Gene set variation analysis (GSVA) analysis for transcriptomic data from primary tumors and matched lung metastases of individual patients (letter and color coded); (C) Association of anti-antioxidant gene expression with mitochondrial Complex I expression in lung metastasis. Patients were divided into two groups based on the upregulation of Complex I genes in the lung metastases relative to their corresponding primary tumor. The Complex I-High group consisted of five patients with more than 25 out of 43 Complex I genes upregulated by more than twofold in lung metastases relative to the corresponding primary tumor. The Complex I-Low group consisted of six patients with less than 25 Complex I genes upregulated by twofold in the lung metastases compared to the corresponding primary tumor. GSVA signature analysis of Complex I genes, lung antioxidant genes, and NRF2 signature genes was performed in the Complex I-High and -Low groups. p-Values were calculated by unpaired two-tailed student’s t test; (D) Heatmap of the relative expression of individual mitochondrial Complex I genes and antioxidant genes in lung metastases relative to the corresponding primary tumor. Complex I-High and –Low patient samples are shown as separate groups, in order to highlight the association of antioxidant gene expression with Complex I gene expression. Red dots, antioxidant genes that were also identified to be upregulated in mouse lung micrometastases by Flura-seq (shown in Figure 5D); (E, F) IHC analysis of oxidative stress marker 4-HNE (E) and NRF2 (F) in tissue microarrays of brain metastases (BrM) and lung metastases (LM) from breast cancer patients. Shown are representative images and the quantifications based on the degree of staining (0, no signal: 3, highest signal). (n = 55 samples for BrM and n = 45 for LM for 4-HNE; n = 48 for BrM and n = 41 for LM for NRF2). Scale bar, 20 μm; (G) Association between oxidative stress (4-HNE) and NRF2 scores in lung metastases and brain metastases of breast cancer patients. Heatmap of the IHC staining of 4-HNE (E) and NRF2 (F) was plotted for each patient sample in the TMAs; (H) Hazard Ratio plots of the predictive ability of NRF2 signatures in brain (BrM), lung (LM) and bone (BM) metastasis-free survival outcomes in EMC-MSK dataset (GSE2603, GSE5327, GSE2034 and GSE12276). p-Values were calculated using Log-rank test.