Molecular Identity Changes of Tumor-Associated Macrophages and Microglia After Magnetic Resonance Imaging-Guided Focused Ultrasound-Induced Blood-Brain Barrier Opening in a Mouse Glioblastoma Model

Abstract

An orthotopically allografted mouse GL26 glioma model (Ccr2^RFP/wt-Cx3cr1^GFP/wt) was used to evaluate the effect of transient, focal opening of the blood-brain barrier (BBB) on the composition of tumor-associated macrophages and microglia (TAMs). BBB opening was induced by magnetic resonance imaging (MRI)-guided focused ultrasound (MRgFUS) combined with microbubbles. CX3CR1-GFP cells and CCR2-RFP cells in brain tumors were quantified in microscopic images. Tumors in animals treated with a single session of MRgFUS did not exhibit significant changes in cell numbers when compared with tumors in animals not receiving FUS. However, tumors that received two or three sessions of MRgFUS had significantly increased amounts of both CX3CR1-GFP and CCR2-RFP cells. The effect of MRgFUS on immune cell composition was also characterized and quantified using flow cytometry. Glioma implantation resulted in increased amounts of lymphocytes, monocytes and neutrophils in the brain parenchyma. Tumors administered MRgFUS exhibited increased numbers of monocytes and monocyte-derived TAMs. In addition, MRgFUS-treated tumors exhibited more CD80+ cells in monocytes and microglia. In summary, transient, focal opening of the BBB using MRgFUS combined with microbubbles can activate the homing and differentiation of monocytes and induce a shift toward a more pro-inflammatory status of the immune environment in glioblastoma.

Keywords: Focused ultrasound; Glioblastoma; Magnetic resonance imaging–guided focused ultrasound; Microbubbles; Microglia and macrophage activation; Sterile inflammation; Tumor-associated macrophages and microglia.

Figure 1.. Experimental apparatus and post contrast T1-weighted image immediately after sonication.

A: MRgFUS system consisting of a 1.5-MHz transducer (T) that rests on the top of the mouse head and can move in X-Y planes and can be focused in the Z axis. B: A 3T MRI scanner was used to detect BBB-opening after sonication. C: Post-contrast T1-weighted image acquired immediately after sonication. Enhancement of the tumor (white arrow) and the brain tissue along the acoustic beam below the tumor (black arrow) indicates the area of BBB opening.

Figure 2.. Examples of MR images of GL26 allograft acquired before and immediately post MRgFUS.

A) T2-weighted images did not show edema after MRgFUS. T2*-weighted gradient echo image (GRE) did not show hemorrhage from MRgFUS. Post-contrast T1-weighted images showed the enhancement of the tumor both before and post MRgFUS (white arrows), with a more pronounced enhancement on the post MRgFUS images. Post-contrast T1-weighted images in the axial plane also showed the enhancement in the brain tissue below the tumor along the path of the ultrasound beam (black arrow). Ktrans map from horizontal direction showed elevated Ktrans in the tumor prior to MRgFUS and an increase in Ktrans signal upon MRgFUS induction (white arrows). B showed the comparison of Ktrans values for Pre_FUS versus Post_FUS from the tumors of FUSX1 group. The Post_FUS Ktrans values increased significantly from Pre_FUS ones (p=0.03) (n=5 in each group)..

Figure 3.. Infiltration of brain tumors by CCR2-RFP and CX3CR1-GFP immune cells.

A) Immunohistochemical staining for CX3CR1-GFP and CCR2-RFP cells is shown from an animal in the tumor-implanted group that did not receive MRgFUS. The fluorescent immune cells are mainly located at the edges of the tumor and in the peritumoral area with only a sparse distribution of cells in the core. Both in the tumor core and at the edges, dual-positive cells (First and second rows, triangles) are the predominant type, with occasional single-positive CX3CR1-GFP (First row, arrows) and CCR2-RFP (First and second rows, arrow heads) cells. In the peritumoral area (i.e. the areas below the dotted lines in the lower panels), there are more single-positive CX3CR1-GFP cells than single-positive CCR2-RFP cells, although dual-positive cells are seen in this area as well. B) Lower magnification images encompassing the tumor and peritumoral area for four experimental groups are shown. As described in A, stained cells in the Tumor_No FUS group exhibit stained cells toward the periphery of the tumor and in the peritumoral area. One session of FUS (FUSX1) did not appear to alter the distribution of cells. In contrast, two or three sessions of FUS (FUSX2 and FUSX3, respectively) resulted in increased numbers of immune cells infiltrating the tumor. C) Quantification of the cell density across the core and periphery of tumors demonstrated significantly greater infiltration of immune cells into tumors in animals receiving 2 or 3 sessions of FUS.

Figure 3.. Infiltration of brain tumors by CCR2-RFP and CX3CR1-GFP immune cells.

A) Immunohistochemical staining for CX3CR1-GFP and CCR2-RFP cells is shown from an animal in the tumor-implanted group that did not receive MRgFUS. The fluorescent immune cells are mainly located at the edges of the tumor and in the peritumoral area with only a sparse distribution of cells in the core. Both in the tumor core and at the edges, dual-positive cells (First and second rows, triangles) are the predominant type, with occasional single-positive CX3CR1-GFP (First row, arrows) and CCR2-RFP (First and second rows, arrow heads) cells. In the peritumoral area (i.e. the areas below the dotted lines in the lower panels), there are more single-positive CX3CR1-GFP cells than single-positive CCR2-RFP cells, although dual-positive cells are seen in this area as well. B) Lower magnification images encompassing the tumor and peritumoral area for four experimental groups are shown. As described in A, stained cells in the Tumor_No FUS group exhibit stained cells toward the periphery of the tumor and in the peritumoral area. One session of FUS (FUSX1) did not appear to alter the distribution of cells. In contrast, two or three sessions of FUS (FUSX2 and FUSX3, respectively) resulted in increased numbers of immune cells infiltrating the tumor. C) Quantification of the cell density across the core and periphery of tumors demonstrated significantly greater infiltration of immune cells into tumors in animals receiving 2 or 3 sessions of FUS.

Figure 3.. Infiltration of brain tumors by CCR2-RFP and CX3CR1-GFP immune cells.

A) Immunohistochemical staining for CX3CR1-GFP and CCR2-RFP cells is shown from an animal in the tumor-implanted group that did not receive MRgFUS. The fluorescent immune cells are mainly located at the edges of the tumor and in the peritumoral area with only a sparse distribution of cells in the core. Both in the tumor core and at the edges, dual-positive cells (First and second rows, triangles) are the predominant type, with occasional single-positive CX3CR1-GFP (First row, arrows) and CCR2-RFP (First and second rows, arrow heads) cells. In the peritumoral area (i.e. the areas below the dotted lines in the lower panels), there are more single-positive CX3CR1-GFP cells than single-positive CCR2-RFP cells, although dual-positive cells are seen in this area as well. B) Lower magnification images encompassing the tumor and peritumoral area for four experimental groups are shown. As described in A, stained cells in the Tumor_No FUS group exhibit stained cells toward the periphery of the tumor and in the peritumoral area. One session of FUS (FUSX1) did not appear to alter the distribution of cells. In contrast, two or three sessions of FUS (FUSX2 and FUSX3, respectively) resulted in increased numbers of immune cells infiltrating the tumor. C) Quantification of the cell density across the core and periphery of tumors demonstrated significantly greater infiltration of immune cells into tumors in animals receiving 2 or 3 sessions of FUS.

Figure 4.. Immune micro-enviroment of naïve brain and the brain tissue implanted with GL26 glioma

A: Dot plots of the CD45+CD11b− lymphocytes and CD45+CD11b+ myeloid cells, Compared to the naïve brain, both ipsilateral and contralateral parenchyma of the brains allografted with tumor showed more lymphocytes.B: Dot plots gated on CX3CR1-GFP and CCR2-RFP from the CD45+CD11b+ myeloid cells. Monocytes were increased in both ipsilateral and contralateral parenchyma. C: Histogram plots from the staining of F4/80 showing the differentiation of monocytes and microglia identified from B; from the staining of CD80 and CD206 indicating the CD80+ proinflammation and CD206+ anti-inflammation polarization. D-I: Quantification of the immune cells identified from A, B, and C. The ipsilateral parenchyma showed more lymphocytes (p=0.001, D), monocytes(p=0.001, E), and neutrophils (p=0.04, E), and relatively lower proportion of microglia (p=0.001, E). The contralateral parenchyma displayed the same changes (lymphocytes: p=0.001, monocytes: p=0.002, neutrophils: P=0.02, microglia: p=0.002). The proportions of the immune cells between the ipsilateral and contralateral parenchyma did not show significant difference. Compared with naïve brain, the allografted parenchyma showed more monocyte-derived TAMs (ipsilateral parenchyma: p=0.003, contralateral parenchyma: p=0.03, F), and more microglia-derived TAMs (ipsilateral parenchyma: p=0.005, contralateral parenchyma: p=0.02, G). The bilateral parenchyma of the brain with GL26 glioma did not show significant difference in the differentiation of the monocytes (p=0.88) and microglia (p=0.44). H-I: There were more proinflammation CD80+ cells in monocytes (ipsilateral parenchyma: p=0.01, contralateral parenchyma: p=0.04) and microglia (ipsilateral parenchyma: p=0.04, contralateral parenchyma: p=0.002) of the ipsilateral and contralateral parenchyma compared with the naïve brain. The proportion of CD206+ cells in the contralateral parenchyma (p=0.03) was lower than the one from naïve brain. Abbreviation: naïve-Mo: naïve monocytes; Mo: monocyte; Mg: microglia; Mo-TAMs: monocyte-derived tumor associated macrophages; naïve-Mg: maive microglia; Mg-TAMs: microglia-derived tumor associated macrophages.

Figure 5.. Effects of MRgFUS on the immune micro-enviroment of tumor tissue

A: Dot plots of the CD45+Cd11b− lymphocytes and CD45+CD11b+ myeloid cells. B: Dot plots of the cells gated on CX3CR1-GFP and CCR2-RFP from myeloid cells, the tumor treated with FUS showed more cells in the monocytes quardrant. C: Histograms of the monocytes and microglia cells identified through markers of F4/80, CD80, and CD206. D-H: Quantification of the cell populations identified from A, B, and C. The animals treated with MRgFUS showed increased monocytes (p=0.01) and decreased proportion of neutrophils (p=0.03) compared to the Tumor_No FUS group. The proportions of lymphcytes (p=0.69), myeloid cells (p=0.67), and microglia (p=0.76) did not show any significant difference between the gliomas from the Tumor_No FUS and Tumor_FUS groups (D and E). Tumor_FUS group showed more monocytes differentiation: higher proportion of monocytes-derived TAMs (p=0.04), whereas microglia did not show changes in differentiation (p=0.19) (F). In the tumors treated with FUS, there were more proinflammation CD80+ cells both in CCR2+ monocytes (p=0.03) and CX3CR1+CCR2− microglia cells (p=0.05). (G, H). Abbreviation: Abbreviation: naïve-Mo: naïve monocytes; Mo: monocyte; Mg: microglia; Mo-TAMs: monocyte-derived tumor associated macrophages; naïve-Mg: maive microglia; Mg-TAMs: microglia-derived tumor associated macrophages.