Insights into the Steps of Breast Cancer-Brain Metastases Development: Tumor Cell Interactions with the Blood-Brain Barrier

Abstract

Brain metastases (BM) represent a growing problem for breast cancer (BC) patients. Recent studies have demonstrated a strong impact of the BC molecular subtype on the incidence of BM development. This study explores the interaction between BC cells of different molecular subtypes and the blood-brain barrier (BBB). We compared the ability of BC cells of different molecular subtypes to overcome several steps (adhesion to the brain endothelium, disruption of the BBB, and invasion through the endothelial layer) during cerebral metastases formation, in vitro as well as in vivo. Further, the impact of these cells on the BBB was deciphered at the molecular level by transcriptome analysis of the triple-negative (TNBC) cells themselves as well as of hBMECs after cocultivation with BC cell secretomes. Compared to luminal BC cells, TNBC cells have a greater ability to influence the BBB in vitro and consequently develop BM in vivo. The brain-seeking subline and parental TNBC cells behaved similarly in terms of adhesion, whereas the first showed a stronger impact on the brain endothelium integrity and increased invasive ability. The comparative transcriptome revealed potential brain-metastatic-specific key regulators involved in the aforementioned processes, e.g., the angiogenesis-related factors TNXIP and CXCL1. In addition, the transcriptomes of the two TNBC cell lines strongly differed in certain angiogenesis-associated factors and in several genes related to cell migration and invasion. Based on the present study, we hypothesize that the tumor cell's ability to disrupt the BBB via angiogenesis activation, together with increased cellular motility, is required for BC cells to overcome the BBB and develop brain metastases.

Keywords: adhesion; angiogenesis; blood–brain barrier; breast cancer molecular subtypes; breast cancer–brain metastasis; gap junction assembly; invasion.

Figure 1

Properties of BC cell lines during brain metastatic cascade. (A) BC cell (MCF-7, MDA-MB-231, and MDA-MB-231-BR) adhesion to activated (+TNF-α, 10 ng/mL for 4 h) and not activated (-TNF-α, untreated) hBMECs was analyzed under static conditions. Relative amount (to MCF-7 untreated situation = 1) of adhesive cells is shown (representative experiment; n = 5); (B) immunofluorescence staining of ZO-1 (green) and nuclei (DAPI, blue) in hBMECs, magnification 60×; (C) the effect of different BC cells on BBB integrity are shown as normalized resistance values at 4 kHz (values were set at 1 before treatment (=0 min)) measured with the ECIS system over 20 min (representative experiment; n = 3); (D) bar graphs displaying relative resistance values under the influence of different tumor cells (n = 3); (E) invasion potential of MCF-7, MDA-MB-231, and MDA-MB-231-BR through hBMECs measured in a transwell assay (n = 3). Values are means ± s.d. * p < 0.05, ** p < 0.005.

Figure 2

BM development in an intracardiac mouse model depending on different BC cell lines. (A) Kaplan–Meier plot of mouse survival, intracardiac injected with 1 × 10⁶ cells of each BC cell line (MCF-7: n = 5, MDA-MB-231: n = 2, MDA-MB-231-BR: n = 12); (B) representative BLI pictures of whole mice of each group 21 days after injection; (C) ex vivo BLI signal quantification from brains of all three test groups (MCF-7, MDA-MB-231, and MDA-MB-231-BR); (D) representative pictures of BLI-measured brains at the final time point of each group; (E) sagittal sections of mouse brains immunohistochemically stained on luciferase. From each test group, one representative picture of the whole brain and detailed metastases staining are shown. Corresponding scales are indicated above the pictures. Values are means ± s.d. ** p < 0.005.

Figure 3

Transcriptome analysis of hBMECs after treatment with BC secretomes. (A) Schematic representation of the experimental design: hBMECs were treated with CM of BC cell lines MCF-7, MDA-MB-231, and their corresponding brain metastatic subline (MDA-MB-231-BR), and RNA sequencing was subsequently performed; (B) Venn diagrams showing RNA sequencing results (n = 4). A total of 58.611 genes were analyzed from each data set. The number of up- and downregulated and overlapping genes compared between different groups of treatment (hBMEC^Ctrl, hBMEC^MCF-7, hBMEC^MDA-MB-231, and hBMEC^{MDA-MB-231-BR}) are displayed; (C) reactome pathway analysis represents pathways enriched due to different treatments; (D) heatmap displaying log2FC values of the strongest differential gene expression by the influence of TNBC secretomes relative to control (untreated).

Figure 4

Brain metastatic-specific effects of BC cells on the brain endothelium. (A) Heatmap displaying log2FC values corresponding to all genes included in signaling by NTRK1, interleukin-7 signaling, interleukin-6 family signaling, and gap junction assembly pathways for hBMEC^{MDA-MB-231-BR} as well as hBMEC^MDA-MB-231, both vs. hBMEC^Ctrl; (B) list of endothelial genes significantly deregulated by MDA-MB-231-BR in comparison to the parental cell line MDA-MB-231; (C) relative brain endothelial expression of CXCL1 and TXNIP after 4 h of treatment with CM of MDA-MB-231 (hBMEC^MDA-MB-231) and CM of MDA-MB-231-BR (hBMEC^{MDA-MB-231-BR}); (D) gene concept network displaying significantly differentially expressed genes of MDA-MB-231-BR cells compared with MDA-MB-231 in enriched GO-terms cell adhesion and locomotion. Genes of interest are marked up. Values are normalized to corresponding GAPDH expression (n = 3). Values are means ± s.d. ** p < 0.005.