Transcriptome analysis reveals a role for the endothelial ANP-GC-A signaling in interfering with pre-metastatic niche formation by solid cancers

Abstract

Cancer establishes a microenvironment called the pre-metastatic niche in distant organs where disseminated cancer cells can efficiently metastasize. Pre-metastatic niche formation requires various genetic factors. Previous studies suggest that inhibiting a single niche-factor is insufficient to completely block pre-metastatic niche formation especially in human patients. Here we show that the atrial natriuretic peptide (ANP), an endogenous hormone produced by the heart, inhibits pre-metastatic niche formation and metastasis of murine solid cancer models when pharmacologically supplied in vivo. On the basis of a wealth of comprehensive RNA-seq data, we demonstrated that ANP globally suppressed expression of cancer-induced genes including known niche-factors in the lung. The lungs of mice overexpressing GC-A, a receptor for ANP in endothelial cells, were conferred resistance against pre-metastatic niche formation. Importantly, neither ANP administration nor GC-A overexpression had a detrimental effect on lung gene expression in a cancer-free condition. The current study establishes endothelial ANP-GC-A signaling as a therapeutic target to control the pre-metastatic niche.

Keywords: RNA-seq analyses; atrial natriuretic peptide; cancer metastasis; pre-metastatic niche; vascular endothelial cells.

Figure 1. ANP suppresses lung metastasis of 4T1 breast cancer

(A) Representative images of lung metastasis in vehicle- or ANP-treated 4T1-bearing mice. Mice were sacrificed four weeks after cancer cell transplantation. Scale bars represent 10 mm. (B) Dot plot showing the number of nodules representing lung metastasis of 4T1-EGFP cells in mice grouped as in (A) (10 mice per a group). (C) Primary cancer volume in vehicle- or ANP-treated 4T1-bearing mice on day 10, 17, and 24 after cancer cell transplantation. Data are means ± s.e.m. (10 mice per a group). (D) Primary cancer weight in vehicle- or ANP-treated 4T1-bearing mice on 28 dpt (8 mice for vehicle and 10 for ANP).

Figure 2. ANP represses 4T1-induced gene expression changes representative of pre-metastatic niche formation in the lung

(A) Scatter plot showing log2 fold changes between the lungs of 4T1-bearing or sham-operated mice. Genes exhibiting more than 2-fold changes are highlighted. Data from two biological replicates are shown (rep1 and rep2). (B) Gene ontology analysis. Top 100 up-regulated genes were subjected to GO analysis and signatures highly significantly enriched in the group are shown. (C) Scatter plot comparing log2 fold changes between 4T1-bearing or sham-operated mice with or without ANP treatment. Genes exhibiting ≥ 3-fold or ≤ 0.3-fold changes in vehicle are highlighted. Data from two biological replicates are averaged. (D) Heatmap of genes exhibiting more than 3-fold increases in the lung of 4T1-bearing mice is shown. Gene expression changes of the indicated genes in ANP-treated sham group are also shown. Validation for (A)-(D) by qPCR and immunohistochemistry is presented in Supplementary Figure 4. (E) Representative images of the lungs stained with hematoxylin-eosin (HE) and anti-Mac3 antibody obtained 7 days after 4T1 cancer-transplantation in each group. Scale bar represents 100 µm. Higher resolution pictures are shown in Supplementary Figure 6. (F) Dot plot showing the number of Mac3–positive cells per a filed in mice grouped as in (E) (9 mice per a group except for 4T1-ANP (n = 8)). p-values were calculated using one-way ANOVA. Representative low-magnification images are shown in Supplementary Figure 5.

Figure 3. ANP suppresses LLC-induced gene expression changes representative of pre-metastatic niche formation in the lung

(A) Gene set enrichment analysis comparing 4T1- and LLC-induced gene expression changes in the lung. (B) Scatter plot comparing log2 fold changes between LLC-bearing or sham-operated mice with or without ANP treatment. Genes exhibiting ≥ 3-fold or ≤ 0.3-fold changes in vehicle are highlighted. Data from two biological replicates are averaged. (C) Heatmap of genes exhibiting more than 3-fold increases in the lung of LLC-bearing mice is shown. Gene expression changes of the indicated genes in the ANP-treated sham group are also shown. Validation for (A)-(C) by qPCR and immunohistochemistry is presented in Supplementary Figure 9. (D) Representative images of the lungs stained with HE or anti-Mac3 antibody obtained 10 days after LLC cancer-transplantation in each group. Scale bars represent 100 µm. Higher resolution pictures are shown in Supplementary Figure 11. (E) Dot plot showing the number of Mac3–positive cells per a filed in mice grouped as in (D) (9 mice per a group except for LLC-ANP (n = 8)). p-values were calculated using one-way ANOVA. Representative low-magnification images are shown in Supplementary Figure 10.

Figure 4. The endothelial GC-A overexpression attenuates pre-metastatic niche formation in the lung

(A) Scatter plot comparing log2 fold changes between LLC-bearing or sham-operated mice in WT and EC GC-A-Tg. Genes exhibiting ≥ 3-fold or ≤ 0.3-fold changes are highlighted. Data from two biological replicates are averaged. (B) Heatmap of genes exhibiting more than 3-fold increases in the lung of LLC-bearing WT mice is shown. Gene expression changes of the indicated genes between WT and EC GC-A-Tg (sham-operated) are also shown. (C) Venn diagram demonstrates extensive overlap between ANP- or GC-A-regulated genes. Validation for (A)-(C) by qPCR and immunohistochemistry is presented in Supplementary Figure 13. (D) Representative images of the lungs stained with HE or anti-Mac3 antibody obtained 10 days after LLC tissue-transplantation in each group. Scale bars represent 100 µm. Higher resolution pictures are shown in Supplementary Figure 16. (E) Dot plot showing the number of Mac3–positive cells per a filed in mice grouped as in (D) (n = 9 for the WT group and n = 8 for the EC GC-A Tg group). p-values were calculated using one-way ANOVA. Representative low-magnification images are shown in Supplementary Figure 15.

Figure 5. Summary of this study

The ANP-GC-A signaling inhibits adhesion of cancer cells in blood vessels [22] and the pre-metastatic niche (this study) to protect different murine cancer models from metastasis. Although the detailed mechanism remains unknown, the ANP-GC-A signaling appears to simultaneously repress multiple cancer-induced signals (TGFβ, inflammation, and so on).