MUC1 is a tumor associated antigen expressed on all human adenocacinomas. Antigenicity of MUC1 is derived from its overexpression and abnormal glycosylation on tumor cells compared to normal epithelial cells. Tumor associated MUC1, but not normal MUC1, is processed and presented to the immune system leading to both humoral and cellular responses that are progressively suppressed as the tumor grows leading to tumor escape. Improving MUC1-specific immunity in the therapeutic setting in cancer patients has been only marginally effective. A discovery that MUC1 is also expressed on premalignant lesions provided an opportunity to test MUC1 vaccine immunogenicity in the prophylactic setting in patients at risk for developing cancer. We completed a clinical trial of a MUC1 vaccine in patients with a history of advanced colonic adenomas and thus a highly increased risk for adenoma recurrence and progression to colon cancer. Patients were vaccinated post adenoma removal with a synthetic MUC1 peptide (100aa long representing five 20aa-long repeats) admixed with the TLR-3 agonist Poly-LCIC (Hiltonol®) as adjuvant. Response was determined by measuring MUC1 specific IgG two weeks after each injection given at W0, 2, 10 and 52. High levels of IgG were observed in 43% of patients after the second and third injection. By W28, IgG titers had decreased but were very effectively boosted by the fourth injection at W52. 53% of vaccinated patients did not respond to the vaccine. The ability to respond or not did not correlate with the standard predictive factors such as age, gender or HLA-Class II type. We did find, however, that non-responders had increased numbers of circulating myeloid derived suppressor cells (MDSC).
To determine what else could have predicted either a response to the vaccine or a failure to respond, we performed microarray and bioinformatic analysis to identify gene signatures and pathway activity differentially expressed in whole blood collected pre-vaccination from 8 responders and 8 non-responders. Linear Regression analysis was used to correlate gene expression with immune response data. Statistical Inference was assessed using Wilcoxon rank sum test.
Over 150 genes were found differentially expressed between the responders and the non-responders. Within the responder group, levels of expression of these genes correlated with the IgG titer, the three responders with the lowest IgG titers clustering more closely with the non-responders. Among the genes and pathways that distinguished the two groups were inflammatory genes characterizing the non-responders while high expression of monocyte/macrophage and chemokine genes characterized the responders. Similarly, non-responders showed decreased expression of leukocyte migration and HLA Class II genes and deregulated expression of apoptosis genes while responders showed increased expression of T and B cell activation and motility genes.
The goal of this analysis was to define a genomic signature that can predict a response to this cancer vaccine and potentially other vaccines, thus helping in the selection of eligible patients. Clear differences we found in pre-vaccination gene expression between the responders and the non-responders is allowing us to create such a signature. This signature, we hypothesize, might also correlate with a predisposition to MDSC accumulation in response to a development of a premalignant lesion, which suppress adaptive immunity and increase the risk of cancer.
Citation Format: Olivera Finn, Robert Schoen, John McKolanis, Courtney Steele, Petra Stafova, Stephanie Richards, Alejandro Quintero, Mark Cameron. MUC1 vaccine for colon cancer prevention: Microarray and bioinformatic analysis of pre-vaccination gene signatures distinguishing responders from non-responders. [abstract]. In: Proceedings of the AACR Special Conference: Tumor Immunology and Immunotherapy: A New Chapter; December 1-4, 2014; Orlando, FL. Philadelphia (PA): AACR; Cancer Immunol Res 2015;3(10 Suppl):Abstract nr PR02.
- ©2015 American Association for Cancer Research.