Background: Globo H, a hexasaccharide linked to ceramide, is a glycolipid over-expressed on the surface of many common cancers. Although globo H has been identified as a potential target for cancer vaccine and clinical trials of globo H vaccine in breast cancer are ongoing, its role in cancers is unknown. Here we investigated the effects of globo H ceramide (GHCer) on immune functions and the underlying mechanisms involved.

Methods: The incorporation of globo H by human peripheral blood mononuclear cells (PBMCs) was evaluated by co-culturing PBMCs with globo H expressing breast cancer cell line, MCF-7, and by immunohistochemical analysis of tumor infiltrating lymphocytes in breast cancer specimens. The effects of GHCer on T cells in response to anti-CD3/CD28 activation were assessed by pre-incubation of mouse splenocytes or purified CD4+ T cells or PBMSs with synthetic GHCer. The impacts of GHCer on the mouse CD19+ B cell upon activation with LPS or LPS+IL-4+CD40L were also examined. The effects of GHcer on the protein and gene expression level were determined by flow cytometry and real-time quantitative PCR, respectively.

Results: PBMCs co-cultured with MCF-7were found to express GHCer as detected by flow cytometry. Importantly, among 98 breast cancer specimens examined, globo H was detected on tumor-infiltrating lymphocytes in 61% of 61 globo H positive breast cancers. Addition of GHCer to human PBMCs, mouse splenocytes or purified CD4+ T cells inhibited their proliferative response to anti-CD3/CD28 to 60±1%, 50±7% and 62±5% of control, respectively, and significantly reduced the secretion of IL-2, IFN-g and IL-4. GHCer also suppressed the proliferation of splenocytes or purified CD19+ B cells to 45±10% and 26±3% of control in response to LPS or LPS + IL4 +CD40 ligand and reduced their IgM and IgG production, along with negligible induction of plasma cells. Ceramide displayed no such inhibitory effects. On the other hand, GHCer failed to raise regulatory T cells, or their expression of FOXP3/CTLA4, nor did it increase apoptosis. Notably, GHCer significantly suppressed the Notch1 signaling during activation of human PBMC and murine T and B cells. Furthermore, GHCer upregulated the expression of ID3 and itch by 2.1±0.2 and 4.7±0.4 fold, respectively, leading to ID3-dependent downregulation of Notch1 transcription and itch-mediated Notch1 degradation. The latter was associated with increased expression of egr2 and egr3, preceding itch upregulation.

Conclusions: Our study uncovered a new aspect of immunosuppressive effects of GHCer which facilitates the escape of cancer from immune surveillance. We also elucidated its molecular processes involving impaired Notch1 signaling through transcriptional repression by ID3, and protein degradation via egr2/3 controlled itch expression. Together with our previous report of the expression of globo H in breast cancer stem cells, these data indicate that GHCer is not merely a cancer-associated antigen, but also acts as an immune checkpoint. Such dual role of GHCer provides further impetus for the development of globo H-targeted cancer immunotherapy and strengthened the rationales for our ongoing phase II/III clinical trial of globo H vaccine in breast cancer.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P5-01-14.