Medulloblastoma recurrence and metastatic spread are independent of colony-stimulating factor 1 receptor signaling and macrophage survival

Abstract

Purpose: Tumor infiltration by immunosuppressive myeloid cells or tumor-associated macrophages (TAMs) contributes to tumor progression and metastasis. In contrast to their adult counterparts, higher TAM signatures do not correlate with aggressive tumor behavior in pediatric brain tumors. While prominent TAM infiltrates exist before and after radiation, the degree to which irradiated macrophages and microglia support progression or leptomeningeal metastasis remains unclear. Patients with medulloblastoma often present with distant metastases and tumor recurrence is largely incurable, making them prime candidates for the study of novel approaches to prevent neuroaxis dissemination and recurrence.

Methods: Macrophage depletion was achieved using CSF-1 receptor inhibitors (CSF-1Ri), BLZ945 and AFS98, with or without whole brain radiation in a variety of medulloblastoma models, including patient-derived xenografts bearing Group 3 medulloblastoma and a transgenic Sonic Hedgehog (Ptch1^+/-, Trp53^-/-) medulloblastoma model.

Results: Effective reduction of microglia, TAM, and spinal cord macrophage with CSF-1Ri resulted in negligible effects on the rate of local and spinal recurrences or survival following radiation. Results were comparable between medulloblastoma subgroups. While notably few tumor-infiltrating lymphocytes (TILs) were detected, average numbers of CD3+ TILs and FoxP3+ Tregs did not differ between groups following treatment and tumor aggressiveness by Ki67 proliferation index was unaltered.

Conclusion: In the absence of other microenvironmental influences, medulloblastoma-educated macrophages do not operate as tumor-supportive cells or promote leptomeningeal recurrence in these models. Our data add to a growing body of literature describing a distinct immunophenotype amid the medulloblastoma microenvironment and highlight the importance of appropriate pediatric modeling prior to clinical translation.

Keywords: CSF-1R; Colony stimulating factor receptor; M-CSF; Macrophage; Medulloblastoma.

Figure 1.. CSF-1R Inhibition Depletes Medulloblastoma-Associated Macrophage and Microglia.

Murine macrophage proliferation after exposure to selected CSF-1R inhibitors detected by Cell-Titer Glo® (CTG). (a) Survival of murine macrophage derived from PBMCs following a 3-day culture with CSF-1R tyrosine kinase inhibitors, PLX3397 (IC₅₀ 24 nM), BLZ945 (IC₅₀ 171 nM), and GW2580 (IC₅₀ 414 nM). (b) Survival of murine macrophage in vitro following a 3-day exposure to the anti-mCSF-1R antibody, AFS98, and compared to an isotype control antibody. (c) H&E and albumin IHC stains of representative whole brain section from a Med-411FH xenograft with metastatic tumor spread. Diffuse albumin staining beyond the boundaries of the tumor indicate the ability of proteins up to 70 kDa to access surrounding brain parenchyma. (d) IHC showing lack of albumin staining and vascular integrity in non-tumor bearing nude mice. (e -h) CSF-1R targeting antibody, AFS98, specifically depletes tumor-associated macrophages and microglia in a G3 MB PDX model, Med-411FH. Representative immunohistochemistry images of brain tumor sections from mice treated with 21 days of AFS98 (50 mg/kg/dose three times weekly) or an isotype-matched control antibody and stained for F4/80 to detect all macrophage(e,f; p = 0.0002) or Tmem119 to detect resting microglia(g,h; p = 0.0008). Scale bars 25–100 μm.

Figure 2.. CSF-1R inhibition combined with RT in a Group 3 medulloblastoma PDX model.

(a) G3 MB Med-411FHTC tumor cells treated in vitro with a single dose of 10 Gy radiation and analyzed for gene expression of CSF-1 and SDF-1 by RT-qPCR at subsequent time points following exposure and compared to untreated controls. Statistical significance was calculated by 2-way ANOVA with Tukey’s multiple comparisons test. Data shown as mean ± s.e.m, ****p<0.0001, *p<0.05). (b-c) Representative IHC stains and quantification for F4/80+ macrophage in Med-411FH models treated in vivo with or without 10 Gy fractionated radiation and showing equivalency in TAM prevalence in tumor tissue (p = 0.52). (d) Schematic of combination study in Med-411FH with RT administered on day 11–15 and antibody delivered by intraperitoneal injection before, during, and after RT (n = 9 per group). (e) Representative IVIS® images of mice treated with RT alone, AFS98 +RT, or control + RT, showing tumor distribution at site of injection along with disseminated spinal metastases. (f) Mean cerebellar tumor volume as measured by BLI. (g) Kaplan-Meier estimate of survival among 5 treatment groups and showing a survival benefit from RT (45 days vs. 26 days, p < 0.0001), but no survival difference in AFS98-treated animals. Survival curve comparison made by log-rank (Mantel-Cox) test.

Figure 3.. Spinal dissemination following AFS98 + RT in a G3 medulloblastoma metastatic model.

(a) Fold change in spinal tumor volume as measured by BLI in a Med114-FH PDX model. RT was delivered on day 11–15 and antibodies (AFS98 or control) delivered by i.p. injection (n = 15 per group). (b) Graphs of individual spinal tumor burden from the same cohort of 15 mice per group. (c) Representative IHC images of spinal cord and leptomeningeal tumor sections from mice treated with AFS98 or control antibody and stained for F4/80+ macrophage. Scale bars 25–100 μm. Black box highlights leptomeningeal tumor distribution. (d) Quantification of F4/80+ macrophage depletion in spinal cord sections (p = 0.014). (e) Kaplan-Meier estimate of survival among 2 treatment groups with curve comparison showing no statistically significant difference by log-rank (Mantel-Cox) test.

Figure 4.. CSF-1R inhibition in a SHH GEMM medulloblastoma model.

(a) Expression analysis of CSF-1R obtained from the Pediatric Cancer Genome Project. (b) Immunophenotype of tumor infiltrating CD45+ cells as detected by flow cytometry on untreated SHH MB tumor specimens. PMN-MDSCs represented by CD11b+Ly6G+ cells, “M2” or alternatively-activated macrophage by F4/80+CD163+ cells, “M1” or classically-activated macrophage by F4/80+CD80+ cells, and cytotoxic lymphocytes by CD3+CD8+ cells. (c - f) Representative IHC stains for F4/80+ macrophage (c,d; p = 0.0028) and IBA1+ microglia (e,f; p<0.0001) in SHH MB spontaneous model treated with or without 200 mg/kg BLZ-945 by daily oral gavage. (g) Average body weight measurements in each group over the study time course. (h) Kaplan-Meier estimate of survival among 2 treatment groups (n = 4 treated with BLZ-945 and n = 2 with vehicle, n = 8 historical untreated).

Figure 5.. Immune infiltrate following CSF-1R inhibition in SHH medulloblastoma.

Representative IHC images from (Ptch1^+/−, Trp53^−/−) mice who received treatment with 200 mg/kg BLZ-945 by daily oral gavage compared with control mice dosed with a vehicle or untreated. Graphs show average cells counts per animal with sparse CD3+ T cells (a,b) and FoxP3+ T regulatory cells (c,d) detected in both treatment groups. Ki67+ cells were analyzed in the tumor only (e,f) with the Ki67 proliferation index (%) calculated by positive cell area/tumor cell area. CD68 (g) staining provided to show macrophage depletion by additional markers in BLZ-945 treated animals (n = 4–5 per group).