Immunoprofiling in Neuroendocrine Neoplasms Unveil Immunosuppressive Microenvironment

Abstract

Checkpoint inhibitors have shown promising results in a variety of tumors; however, in neuroendocrine tumors (NET) and neuroendocrine carcinomas (NEC), low response rates were reported. We aimed herein to investigate the tumor immune microenvironment in NET/NEC to determine whether checkpoint pathways like programmed cell death protein 1 (PD-1)/programmed death-ligand 1 (PD-L1) might play a role in immune escape and whether other escape mechanisms might need to be targeted to enable a functional antitumor response. Forty-eight NET and thirty NEC samples were analyzed by immunohistochemistry (IHC) and mRNA immunoprofiling including digital spatial profiling. Through IHC, both NET/NEC showed stromal, but less intratumoral CD3+ T cell infiltration, although this was significantly higher in NEC compared to NET. Expression of PD1, PD-L1, and T cell immunoglobulin and mucin domain-containing protein 3 (TIM3) on immune cells was low or nearly absent. mRNA immunoprofiling revealed low expression of IFNγ inducible genes in NET and NEC without any spatial heterogeneity. However, we observed an increased mRNA expression of chemokines, which attract myeloid cells in NET and NEC, and a high abundance of genes related to immunosuppressive myeloid cells and genes with immunosuppressive functions like CD47 and CD74. In conclusion, NET and NEC lack signs of an activation of the adaptive immune system, but rather show abundance of several immunosuppressive genes that represent potential targets for immunomodulation.

Keywords: ileal neuroendocrine tumors; immunoprofiling; neuroendocrine carcinoma; programmed cell death protein 1; programmed death-ligand 1; tumor infiltrating lymphocytes; tumor-associated macrophages.

Figure 1

Immunohistochemical staining. (A) CD3 in neuroendocrine tumors (NET) G1/G2 with stromal T cell infiltration 8/1 high-power fields (HPF); (B) CD3 in NET G3/neuroendocrine carcinomas (NEC) with intratumoral T cell infiltration 10/1 HPF; (C) programmed death-ligand 1 (PD-L1) in NET G1/G2, weak staining in >1% tumor cells; (D) PD-L1 in NET G3/NEC; strong staining >1% of tumor cells original magnification × 300.

Figure 2

Differential expression at the gene set level in NET G1/G2 and NET G3/NEC. For each gene set the most differentially expressed genes were compared to healthy tissue and the extent of differential expression in each gene set was summarized using a global significance score.

Figure 3

Differential expression of genes associated with cell types in NET G1/G2 and NET G3/NEC. Cell scores (cell type/tumor infiltrating immune cells) of different immune cell subsets were calculated by the expression of mRNA markers specific for each cell type relative to CD45 mRNA levels: T cells, CD3D, CD3E; CD8+ T cells, CD8A; cytotoxic cells, PRF1, KLRK1, KLRB1, GNLY, GMZA; macrophages, CD84, CD163, CD68; DC-SIGN positive cells, CD209/DC-SIGN.

Figure 4

Digital spatial profiling (DSP) tumor regions were scanned for detection of immune cells in NET G3/NEC tumors. (A) Fluorescent labeling aided in characterizing tumor versus immune cells that included hematopoietic marker CD45 (cyan) and monocyte/macrophage marker CD68 (green), NET/NEC tumor marker, synaptophysin (red), and DNA (blue). Representative images portray the varying levels of leukocytes (CD45 and CD68) within the tumors (gastrointestinal (GI) and prostate) and show staining for the immune cells within the regions of interest (ROIs) in a selected 300-µm width (circled). (B) The distribution of RNA abundance in key genes for each patient surveyed in this study. Four to six ROI were selected for each patient. The geometric mean of 5 housekeeping proteins (HKG) was used as normalization and as a baseline reference for low and high RNA abundance. Highlighted are the low DSP RNA abundance for IFNγ and CD3E genes and high DSP RNA abundance for CD74, VEGFA, and CTNNB1. (C) Swarm blot to stratify the RNA abundance of high and low expressed genes relative to the scoring of ROIs with high CD45 positive cells (indicated by circle size). To determine the relative amount of CD45 positive cells within an ROI, the immunofluorescent 300-µm scanned regions were qualitatively scored. In the legend, the colors represent the tumor site.

Figure 5

RNA abundance heatmap of genes related to the immune response in NET G3 and NEC. The columns represent the log10 transformation of 85 gene probes. The red to green hues represent the gradient high to low RNA abundance. The values used are normalized to the geomean of 5 housekeeping proteins. Rows represent ROI ID and the bar on the right is color coded for NEC (yellow) and NET G3 (blue). The hierarchical gene clustering shows the high to low RNA abundance.