Distinguished Research Fellow & Deputy Director, Genomics Research Center, Academia Sinica, Taiwan
Aberrant glycosylation is a feature of cancer cells. GD2, a disialoganglioside, is highly expressed in neuroblastoma, melanoma and small cell lung cancer and some sarcomas. I have pursued immunotherapy of neuroblastoma with a monoclonal anti-GD2 antibody, ch14.18, from preclinical studies to IND, phase I and II studies. The promising results in early clinical trials paved the way to the pivotal international phase III randomized trial in high risk neuroblastoma. In the phase III study, newly diagnosed high-risk NB patients who achieved a CR or PR to induction therapy and received myeloablative consolidation with stem cell rescue were randomized to 13cisRA x 6 cycles (standard therapy) or 13cisRA x 6 with 5 concomitant cycles of ch14.18 combined with GM-CSF or IL2 in alternating cycles (immunotherapy). Analysis of 226 randomized patients (113 to standard therapy and 113 to immunotherapy) revealed immunotherapy to improve event free survival significantly from 46%±5% to 66%±5% vs (p=0.0115). Preliminary OS was also significantly higher for immunotherapy group (86%±4% vs 75%±5% at 2 yrs, p=0.0223). This is the first time an antibody targeting a glycolipid is shown to be effective for cancer immunotherapy.
Another potential glycan target for cancer immunotherapy is Globo H, a hexasaccharide which is overexpressed on a variety of epithelial cell tumors such as colon, ovarian, gastric, pancreatic, lung, prostate and breast cancers, but only limited weak expression in normal tissues. To address the issue whether Globo H is expressed in breast cancer stem cells (BCSCs), we examined 53 human breast cancer specimens and found Globo H to be present in BCSCs, although to a much lesser extent than non-BCSCs. We also showed for the first time the expression of Gb5, the pentasaccharide precursor of Globo H, in BCSCs from >60% of tumors. Immunization of mice with Globo H-KLH with NKT stimulatory glycolipids as adjuvants induced antibody reactive with not only Globo H but also Gb5, suggesting that Globo H-based immunotherapy will target Globo H and Gb5-expressing tumor cells. These findings provide further impetus for our development of a phase II/III clinical trial of Globo H vaccine for metastatic breast cancer, which is now ongoing in Taiwan, Hong kong and USA. In addition, we examined the plasma samples of 58 breast cancer patients and 47 healthy blood donors for autoantibodies bound to Globo H and its truncated fragments on highly sensitive Globo H microarray. When the antibody reactivity to each glycan was normalized to Gb5, both the levels of IgG and IgM against Globo H were significantly higher in breast cancer patients than in normal individuals (P <0.0001). These data suggested that the profiles of glycan-binding antibodies in peripheral blood might serve as a biomarker for cancer diagnosis.