OBJECTIVE: To investigate the genetic control of chronic arthritis and collagen epitope specific antibody responses in an experimental model for rheumatoid arthritis. METHODS: The chronic collagen induced arthritis (CCIA) model was used, induced with collagen type II (CII) in mineral oil lacking mycobacterium in BALB/c (n=24), B10.Q (n=44), (BALB/c x B10.Q) F1 (n=85) and B10.Q x (BALB/c x B10.Q) N2 (n=684) mice. Genome-wide genotyping for 190 N2 mice was performed with extreme phenotypes: chronic arthritis that persisted for 4 months or non-affected. Statistical and linkage analysis were performed with R/qtl software using arthritis and serum subphenotypes. RESULTS: (BALB/c x B10.Q) F1 mice were highly prone to develop a chronic relapsing arthritis (66%), whereas both parental strains were relatively resistant: BALB/c (H-2(d); 0%) and B10.Q (H-2(q); 4.5%). CCIA experiments were performed on 684 mice backcrossed to B10.Q; 38% of the mice developed arthritis and more than half of them developed chronic arthritis phenotype. Genome-wide genotyping revealed mainly the major histocompatibility complex (MHC) locus that had an independent and dominant influence on the chronicity. Interestingly, the H2(d) allele had a dominant suppressive effect. This effect overrode the role of other loci as interaction analysis, after conditioning MHC, revealed additional loci, controlling arthritis and autoantibody phenotypes. CONCLUSIONS: A dominant negative influence of specific MHC haplotype (H2(d)) on CCIA was identified. Further, loci controlling the autoantibody response to different CII epitopes were also identified, and it has been shown that these are dependent on MHC and non-MHC genes.