A Unidirectional Transition from Migratory to Perivascular Macrophage Is Required for Tumor Cell Intravasation

Abstract

Tumor-associated macrophages (TAMs) are critical for tumor metastasis. Two TAM subsets support cancer cell intravasation: migratory macrophages guide cancer cells toward blood vessels, where sessile perivascular macrophages assist their entry into the blood. However, little is known about the inter-relationship between these functionally distinct TAMs or their possible inter-conversion. We show that motile, streaming TAMs are newly arrived monocytes, recruited via CCR2 signaling, that then differentiate into the sessile perivascular macrophages. This unidirectional process is regulated by CXCL12 and CXCR4. Cancer cells induce TGF-β-dependent upregulation of CXCR4 in monocytes, while CXCL12 expressed by perivascular fibroblasts attracts these motile TAMs toward the blood vessels, bringing motile cancer cells with them. Once on the blood vessel, the migratory TAMs differentiate into perivascular macrophages, promoting vascular leakiness and intravasation.

Keywords: CCR2; CXCR4; Mena; TAMs; TGF beta; TMEM; breast cancer; metastasis; tumor associated macrophages.

Graphical abstract

Figure 1

Newly Arriving Monocytes Become Perivascular Macrophages (A) Flow cytometry quantification of the proportion of DiI+ CD11b+ cells in the tumor, measured different days after DiI liposome delivery. (B) Immunofluorescence (IF) of a PyMT tumor at different days after DiI liposome injection, showing cells that ingested DiI (red), endothelial cells (green), and the nuclear counter stain DAPI (blue). Scale bar is 10 μm. Inserts show magnification of one of the cells from the image. (C) Quantification of IF staining showing the proportion of perivascular DiI+ cells. (D) Still frames from Video S1 showing a DiI-labeled macrophage (red) among other macrophages (green) in a tumor (cyan) with collagen fibers (dark blue). The arrowhead indicates a labeled monocyte in the blood stream. (E) Quantification of EdU+ cells in blood smears at different times after EdU injection (n = 3). (F) Quantification of the location of the absolute number of CD68+/EdU+ TAMs, based on IF images as seen in (F) at different times after EdU administration. Per time point, seven PyMT mice were analyzed and the number of CD68+/EdU+ cells analyzed were as follows: days 3–5, 84 cells; days 7–8, 63 cells; and days 10–14, 72 cells. (G) IF imaging of PyMT tumor sections at different days after EdU injection. TAMs are stained for CD68 (green), blood vessels are stained for CD31 (cyan), with DAPI (blue), and EdU+ nuclei indicate cells that were in S-phage at the time of EdU administration (red). Scale bar is 30 μm. Chi-square analysis comparing perivascular versus non-perivascular macrophages. p < 0.0001. Data show mean ± SEM, and each data point represents an individual animal (in A and C).

Figure 2

TAM Numbers Quickly Recover after Depletion but Are Initially Non-perivascular and Differentiate into Functional Perivascular TAMs over Time (A) Schematic overview of experiments. (B) IF imaging of PyMT tumor sections at different days after final B/B treatment. Orange arrows show non-perivascular TAMs, while white arrows show perivascular TAMs. TAMs are stained for CD68 (green), blood vessels are stained for CD31 (red), and nuclei are stained with DAPI (blue). Scale bar is 20 μm. (C and D) Quantification of the number of CD68-positive cells found in the tumor tissue (C) and their location in relation to the vasculature (D) at different time points after final B/B treatment. (E) IF imaging of orthotopic PyMT tumor sections at different days after final B/B treatment. Blood vessels are stained with streptavidin against CD31-biotin injected 5 min before sacrifice (green) and dextran (red), and nuclei are stained with DAPI (blue). Scale bar is 20 μm. (F and G) Quantification of extravascular 155 kDa dextran TMR as a measurement of vascular leakiness (F) and number of CTCs found per milliliter of blood (G) different days after the last B/B injection. Data show mean ± SEM, and each data point represents an individual animal (in C, D, F, and G).

Figure 3

Recruitment of TAMs into the Tumor Depends on CCR2, while Recruitment of Perivascular TAMs to the Blood Vessels Depends on CXCR4 Signaling (A and B) FACS analysis of the ratio of CCR2 KO/WT monocytes found in bone marrow or tumor 2 days (A) or 6 days (B) after adoptive transfer (n = 4 mice per time point). The purple line indicates the ratio of monocytes upon injection. (C) IF imaging of PyMT tumor sections at different days after final B/B treatment. TAMs are visualized with CD68 (green) and CXCR4 (red), and nuclei are stained with DAPI (blue). Scale bar is 10 μm. (D) Quantification of the number of CXCR4+ TAMS found at different times after final B/B injection. (E) Schematic overview of experiments with AMD3100/CXCR4 inhibitor. (F) IF imaging of PyMT tumor sections after 5 days of B/B treatment, followed by 7 days of AMD3100 compared to controls. TAMs are visualized with CD68 (green), vasculature is visualized with endomucin (red), and nuclei are stained with DAPI (blue). Scale bar is 20 μm. (G–I) Quantification of the area blood vessels per field of view (FoV) (G), macrophages per FoV (H), and perivascular macrophages per FoV (I). (J) IF imaging of PyMT tumor after 5 days of B/B treatment, followed by 7 days of AMD3100. Blood vessels are stained with streptavidin against CD31-biotin injected 5 min before sacrifice (green) or extravascular dextran (red), and nuclei are stained with DAPI (blue). Scale bar is 20 μm. (K) Quantification of extravascular 155 kDa dextran TMR as a measurement of vascular leakiness. Data show mean ± SEM, and each data point represents an individual animal (in A, B, D, G–I, and K).

Figure 4

Cancer Cell-Derived TGF-β Upregulates CXCR4 on BMMs, and CXCL12 Is Produced by Fibroblasts (A) IF of BMMs co-cultured with PyMT cancer cells. Cells are stained for F4/80 (green), CXCR4 (red), and DAPI (blue). cc, cancer cells. Scale bar is 10 μm. (B) IF of BMMs cultured in BMM media with or without the addition of PyMT cancer cell conditioned media (CM) and transforming growth factor β receptor (TGF-βR) inhibitor. BMMs are stained with CXCR4 (red) and DAPI (blue). Scale bar is 10 μm. (C) qPCR analysis for CXCR4 RNA expression performed on the RNA isolated from BMMs treated with PyMT cancer cell CM with (blue) or without (burgundy) TGF-βR inhibitor (TGF-βi). (D) IF of BMMs cultured in BMM media without or with the addition of TGF-β (2 ng/mL). BMMs are stained for CXCR4 (red) and DAPI (blue). Scale bar is 10 μm. (E) IF of PyMT sections stained with CXCL12 (cyan), endomucin (red), and DAPI (gray). Scale bar is 20 μm. (F) IF of a PyMT tumor section stained with vimentin (magenta), α-smooth muscle actin (αSMA, yellow), CXCL12 (cyan), and DAPI (gray). Scale bar is 20 μm. Data show mean ± SEM, and each data point represents an individual animal (in C, D, F, and G).