It is inevitable that in any biopsy/excision of malignant tissue, normal cells and tissue is included with the abnormal. If dendrocytes from a non-identical "donor" are used the dendrocytes would be sensitised not only to malignant (therefore "foreign") epitopes / peptides but also to the patient's normal tissue which will be recognise as being non-self (ie foreign to the dendrocytes). Injecting those sensitised dendrocytes into the recipient will result in what is essentially a graft-vs-host reaction, potentially causing multi-organ damage should the injected clones of dendrocytes become established in the recipients immune system. .
I am not really clear enough on how the MH! and MHII work to identify tissue as foreign to really argue this point intelligently. But I do assume that the recognition of self depends on having a large enough sample of the cell for that to happen. Thus if only the antigen itself and no more is given the DC, no such self v non-self determination is made and no harmful autoimmune consequences will follow. So for example when a DC samples a single tumorous cell, now presumably non-self, certainly only a small part of the cell undergoes mutation while the large majority of the cell remains unchanged and is no different than in the non tumorous cell. so how does the DC recognize that the unchanged portions of the cell are not non-self and refrain from attacking them as non-self? I'm not clear on how that works. Care to explain?
The answer is contained within the question. “Self” identifies “self” in normal tissue. You essentially can not sensitize your own dendritic cells to attack other normal cells and tissue in your own body. Dendritic cells are designed to recognize and properly detect and respond to non self which displays non-self antigens.