Myeloid-targeting immunotherapies overcome inhibitory barriers in immune-evasive neuroblastoma

Abstract

Neuroblastomas are highly heterogeneous tumors originating from neural crest-derived cells destined to form the sympathetic nervous system. Nearly half of high-risk tumors present with amplification of the MYCN proto-oncogene. Here, we describe a Mycn-driven, transplantable, non-germline, genetically engineered mouse model (Mycn-nGEMM). Mycn-nGEMM tumors recapitulate the immune-evasive, macrophage-rich tumor microenvironment of high-risk, MYCN-amplified human neuroblastoma. Treatment of tumor-bearing mice with anti-PD-L1, but not anti-PD-1 or anti-CTLA-4, inhibited tumor growth, profoundly remodeling the tumor microenvironment by depleting anti-inflammatory macrophages and increasing T cell infiltration. Surprisingly, while tumor cells showed low expression of PD-L1, anti-inflammatory macrophages from both murine and human neuroblastoma expressed PD-L1. We identified cytokines, including macrophage migration inhibitory factor, secreted by the Mycn-nGEMM cancer cells that drive expression of PD-L1 on macrophages. Combining anti-PD-L1 with CD40 agonist antibodies further improved survival in Mycn-nGEMM mice, demonstrating the potential for myeloid-targeting immunotherapies to overcome inhibitory barriers in immune-evasive neuroblastoma.

Figure 1:. Expression of Mycn in migrating mouse neural crest cells leads to tumor formation in immunocompetent, syngeneic C57BL/6J hosts

(A) Neural tubes (NT) from E9.5 C57BL/6J embryos were dissected between somite 16 and 24 (indicated by red lines) and cultivated in vitro on laminin-coated plates for 24h. The migrating neural crest cells (NCC) were transduced with a Mycn-IRES-GFP retroviral construct. GFP-positive cells were sorted and formed spheres. (B) Mycn-expressing NCC (green, n=13) and NCC transduced with an empty vector (black, n=10) were injected in the flank of C57BL/6J mice. Tumor formation was monitored over time until endpoint was reached. (C) H&E and IHC staining for N-Myc, Phox2B, Tyrosine Hydroxylase (TH), Synaptophysin and NCAM on Mycn-nGEMM tumor shows high expression of all markers. Scale bar = 50 μm. (D) Mycn expression relative to HPRT as assessed by qRT-PCR in Mycn-nGEMM-high and Mycn-nGEMM-low cell lines. Data represent mean values ± SD (n=2 independent experiments). (E) N-Myc expression assessed by western blot in Mycn-nGEMM-high and Mycn-nGEMM-low tumors. (F) Growth curves for Mycn-nGEMM-high (n=6 mice) and Mycn-nGEMM-low (n=7 mice) tumors as assessed by bioluminescent imaging. Data represent mean values ± SEM. (G) Kaplan-Meier curves for C57BL/6J mice injected with Mycn-nGEMM-high or Mycn-nGEMM-low cell lines in the renal capsule (n=5 mice per group). Kaplan–Meier survival analyses were performed by log-rank test. (H) t-SNE clustering including TH-MYCN, Mycn-nGEMM, NCC, Mycn-NCC samples and indicated human cancer datasets. (I) Umap plot using GMKF dataset depending on MYCN-amplification.

Figure 2:. Immune-profiling of Mycn-nGEMM tumors reveals a highly immunosuppressive TME dominated by TAMs

(A) t-SNE representation of GFP and CD45 signals in Mycn-high and Mycn-low nGEMM tumors (n=3 tumors per cell line). (B) Average proportion of tumor cells (GFP+), immune cells (CD45+) and other cells in Mycn-nGEMM tumors (n=3 tumors per cell line). (C) Heatmap indicating levels of expression for each markers for all CD45+ meta clusters identified with the mouse mass cytometry panel. (D) t-SNE representation of clustering for CD45+ cells in Mycn-nGEMM tumors (n=3 tumors per cell line). (E) Percentage of the different immune subsets in Mycn-nGEMM tumors (n=3 tumors per cell line) (F) Expression of cytokines assessed by RNAseq analysis on Mycn-nGEMM-high and Mycn-nGEMM-low cells, represented as normalized read counts. (G) Secreted Mif expression from conditioned cell media of Mycn-nGEMM, MC38 murine colon adenocarcinoma and TCMK1 mouse kidney epithelial cells determined by ELISA assay. Data represent mean ± SD from three independent experiments. Statistical significance was determined by one-way ANOVA multiple comparisons test (G). ***P < 0.001, ns = no significance.

Figure 3:. Anti-PD-L1 treatment of Mycn-nGEMM tumors depletes CD206+ TAMs and remodels the intra-tumoral immune landscape

(A) Mycn-nGEMM low cells were transplanted in the renal capsule of C57BL/6J hosts on day 0. After BLI imaging and randomization, treatments started on day 7 with 200ug of anti-PD-L1 (or corresponding IgG) injected IP every 3 days for a total of 5 doses. (B) Growth curves of Mycn-nGEMM low tumors treated with IgG or anti-PD-L1 antibodies, as assessed by bioluminescent imaging. Grey area indicates treatment window. Data represent mean values ± SEM (n=5 mice per group). (C) Kaplan-Meier curves for C57BL/6J mice injected with Mycn-low cell line in the renal capsule and treated with 5 doses of IgG or anti-PD-L1 antibodies (n=10 mice per group). (D) Volcano plot comparing abundance of tumor-infiltrating leukocyte (TIL) subpopulations in anti-PD-L1 treated tumors (red) and IgG-treated tumors (blue) using the mouse mass cytometry immune cell panel (n=3 tumors per group). Statistically significant clusters in volcano plots are highlighted in opaque color and indicated with a cell type label. (E) Mass cytometry data from Mycn-nGEMM tumors was filtered on CD45+CD11b+F4/80+ TAMs, then gated manually to identify cells positive or negative for CD206 or MHC-II. Frequencies of TAMs expressing CD206 and MHCII was quantified for IgG-treated and anti-PD-L1 treated tumors (n=3 tumors per group). (F) Percentage of CD11b+F4/80+ TAMs expressing MHC-II, CD206 or PD-L1 markers in Mycn-nGEMM tumors after treatment with IgG or anti-PD-L1 antibodies (n=3 tumors per group). (G) Mass cytometry data from Mycn-nGEMM tumors was filtered on CD45+ cells, then gated manually to identify cells positive or negative for CD3e, CD4 or CD8. Frequencies of T cells expressing CD4 and CD8 was quantified for IgG-treated and anti-PD-L1-treated tumors (n=3 tumors per group). (H) Percentage of CD4+ T cells and CD8+ T cells in Mycn-nGEMM tumors after treatment with IgG or anti-PD-L1 antibodies (n=3 tumors per group). (I) After CD4+ T and CD8+ T cell depletion by neutralizing antibodies and treatment of IgG and anti-PD-L1, tumor volume was measured by ultrasound on day 21 post injection (n=5 mice per group). (J) Kaplan-Meier curves representing percentage of tumor-free mice for naïve (n=10 mice) and anti-PD-L1-cured mice (n=7 mice) re-challenged with Mycn-nGEMM low cells in the right flank. Statistical significance was determined by two-way ANOVA (B), unpaired t-test (F and H), and one-way ANOVA multiple comparisons test (I). Kaplan–Meier survival analyses were performed by log-rank test (C and J). *p < 0.05, **P < 0.01, ***P < 0.001, ns = no significance.

Figure 4:. PD-L1 is expressed on TAMs in Mycn-nGEMM tumors

(A) t-SNE representation of GFP, CD11b and PD-L1 signals in Mycn-nGEMM-high and Mycn-nGEMM-low tumors (n=3 tumors per cell line) shows that PD-L1 is expressed mostly on myeloid cells in Mycn-nGEMM tumors. (B) Representative Serial IHC of Human MYCN Amplified NB Tumor and Pseudofluorescence Visualization. Scanned images (10X) of serial IHC of Phox2b (left, nuclear staining, red), PDL1 (middle, membrane staining, brown), and CD163 (right, membrane staining, red) of a representative formalin-fixed paraffin embedded (FFPE) section of a human MYCN amplified NB obtained post chemotherapy. Scale bar = 200 μm. (C) Magnified region from (b) (dotted black lines) of deconvoluted images stacked to demonstrate IHC of Phox2b, Phox2b with PDL1, Phox2b with CD163, and Phox2b with CD163 and PDL1. Scale bar = 100 μm. (D) Quantitative Image Analysis of MYCN Amplified tumors for CD163, Phox2b, and PDL1. Boxplot of the percent positivity of Phox2b, CD163, and PDL1 in MYCN amplified NB tumors. Percent positivity is also shown for Phox2b+PDL1+ cells (mean 2.2%) and CD163+PD-L1+ cells (mean 34.1%). (E) PD-L1 expression assessed by flow cytometry on BMDM (Bone Marrow-derived macrophages) co-cultured with indicated cell lines (Mycn-high, Mycn-low, SB28, MC38 or B16F10) or treated with IFNg for 72h. Mean fluorescent intensity (MFI) for the PD-L1 signal is indicated for each condition. Data represent mean ± SD from two independent experiments. Statistical significance was determined by one-way ANOVA multiple comparisons test (D), ***P < 0.001.

Figure 5:. Single cell RNA sequencing identifies heterogeneity among TAMs in Mycn-nGEMM tumors, which correlates with TAMs in human NB.

(A) UMAP of 11,943 CD45+ cells and neuronal cells obtained from 3 replicates of Mycn-nGEMM tumors, after integration. (B) Violin plot of Mif mRNA expression by cell type of the Mycn-nGEMM tumors. (C) UMAP of 1,780 Mycn-nGEMM TAMs. Three major clusters were identified and annotated according to marker genes and pathway analysis (See text for details). (D) Dotplot presenting the marker genes for each cluster of (C) (E) Distributions of M2 signature (left) and M1 signature (right) scores for each TAM subtypes in (C). (F) Heatmap of correlation scores between Mycn-nGEMM TAMs and Human TAMs subtypes defined by Wienke et al (Wienke et al., 2024). Columns represent Mycn-nGEMM TAM cells and rows represent Human TAM subtypes. The colored bar at the top represents the Mycn-nGEMM macrophage clusters (subtypes).

Figure 6:. Combining anti-PD-L1 with anti-CD40 significantly impairs tumor growth and increases survival in the Mycn-high nGEMM model

(A) Mycn-nGEMM high cells were transplanted in the renal capsule of C57BL/6J hosts on day 0. After BLI imaging and randomization, treatments started on day 7 with 200ug of anti-PD-L1 (or corresponding IgG), CD40 agonist or a combination of both injected IP every 3 days for a total of 5 doses. (B) Kaplan-Meier curves for C57BL/6J mice, transplanted with Mycn-nGEMM high cells in the renal capsule on day 0, and treated with indicated antibodies (n=10-12 mice per group). (C) Pictures of three tumors harvested at day 21 (5 treatment doses) for indicated treatment group. (D) Tumor volumes were measured for tumors harvested at day 21 for each treatment group. Indicated values are mean of n=3 tumors per group ± SD. (E) Proportion of CD45+ cells in tumors from each treatment group as a percentage of total cells in the tumors (mean values of n=3 tumors) ± SD. (F) Proportion of GFP+H-2Kd/H2Dd+ cells in tumors from each treatment group as a percentage of total cells in the tumors (mean values of n=3 tumors) ± SD. (G) Proportion of CD38+ macrophages as a percentage of total cells in the tumors (mean values for n=3 tumors) ± SD. (H) Proportion of CD11c+ dendritic cells (DC) as a percentage of total cells in the tumors (mean values for n=3 tumors) ± SD. (I) Proportion of CD4+ T cells as a percentage of total cells in the tumors (mean values for n=3 tumors) ± SD. (J) Proportion of CD8+ T cells as a percentage of total cells in the tumors (mean values for n=3 tumors) ± SD. (K) Proportion of B220+ B cells as a percentage of total cells in the tumors (mean values for n=3 tumors) ± SD. (L) Cartoon summarizing effect of anti-PD-L1 and CD40 agonist therapies on the TME and growth of Mycn-nGEMM tumors. Statistical significance was determined by ordinary one-way Anova test (D-K), Kaplan–Meier survival analyses were performed by log-rank test (B). *p < 0.05, **P < 0.01, ***P < 0.001, ns = no significance.