[Is type I diabetes a virus-induced disease?].

The conditions are critically reviewed under which diseases might be aetiologically related to infection by a certain virus. Such a causal correlation has to obey Koch's postulates, but may be very difficult to prove in the case of a "hit and run" process. This will be exemplified in the case of insulin-dependent diabetes mellitus (IDDM). Observations in humans and the results of experiments on laboratory animals are reported, whereby the Coxsackie B and mumps viruses are of particular interest. Furthermore, the mechanisms by which viruses may produce autoimmune diseases are discussed, including virological and immunological aspects. The hypothesis of "molecular mimicry" by Oldstone is quoted as a unifying one, allowing the combination of both aspects. His main assumption is oligopeptide homology between certain virus proteins and some cell proteins and some examples are given. In the presence of such homologies the immune system is first stimulated by the parasitic virus protein, but somewhat later this reaction switches against the host structures, causing continuing cellular damage with the development of autoimmune disease. It is concluded that Koch's postulates in such cases have to be supplemented by assaying for amino acid homologies in viruses and certain cell types.