Bruce Beutler1 Department of Immunology The Scripps Research Institute 10550 North Torrey Pines Road of susceptibility loci. The strongest genetic association with Crohn’s involves a chromosome 16 locus, desig-La Jolla, California 92037 nated IBD1 (Curran et al., 1998). Loci on chromosomes 1p, 3q, 4q, and 12q have also been described (Cho et The Fundamental Question: Autoimmune al., 1998; Curran et al., 1998). In a recent issue of Nature, Hugot et al.(2001) and Ogura et al.(2001a) have reported versus Infectious

A protean malady of unknown cause, Crohn’s disease that mutations of the gene encoding NOD2 are strongly associated—and almost certainly etiologically linked—classically occurs as a relapsing/remitting “regional ileitis” but can affect any part of the gastrointestinal tract, to the development of Crohn’s disease. While not all Crohn’s disease can be explained by the NOD2 mutanot merely the terminal ileum. On rare occasions, lesions are observed even in the oropharynx or esophagus. His- tions found so far, the bulk of the chromosome 16 association with Crohn’s disease is explained by the findings. tologically, the lesions are intensely inflammatory, and they may lead to the formation of fistulae between en- In fact, pooling all variant alleles from the study of Hugot et al., it would appear that 201/468 patients and only teric (and sometimes nonenteric) structures. Crohn’s disease and ulcerative colitis are the two major forms 15/103 controls displayed coding changes, a figure that is overwhelmingly significant in itself and takes no acof inflammatory bowel disease (IBD). Crohn’s disease can be distinguished from ulcerative colitis on both clini- count of deep intronic mutations, promoter mutations, large deletions, and the like. We shall come to the funccal and histopathologic grounds. Ulcerative colitis is not associated with fistulae, it affects only the colon, tion of NOD2 presently. But first, let us consider precisely what was done. producing characteristic shallow, ulcerative lesions, and it rather frequently presages the development of adeno- Hugot et al. used an unbiased genetic search based first on linkage disequilibrium mapping, then on pedicarcinoma of the colon (whereas Crohn’s disease does not). The two diseases can also be distinguished on gree disequilibrium testing, and finally on very extensive sequence analysis of the best candidate gene. This is the basis of their responses to therapy: TNF blockade usually yields an excellent response in Crohn’s disease to say that they built their case in a rigorous and painstaking manner. They ultimately identified a multitude (Van Dullemen et al., 1995) but rarely or never does so in ulcerative colitis. of mutations (more than 30 in all), the great majority of which were observed only in Crohn’s patients. This is a Crohn’s disease has widely been considered an autoimmune disorder, although an infectious cause cannot key point in the establishment of a causal relationship, since an isolated mutation might be dismissed as a be excluded. Recurrent hypotheses have held that microbial flora contribute initiators that trigger the disease, nonetiologic marker in linkage disequilibrium with the authentic defect. or, in some cases, prevent disease. Several knockout mutations produce IBD, and, in at least one case (the Ogura et al. appear to have proceeded in a more intuitive fashion, declaring NOD2 as a candidate locus TCRα knockout mouse), germ-free animals fail to develop colitis, suggesting the pathogenic involvement of from the very outset. They identified three mutations in