Quantitative trait loci disposing for both experimental arthritis and encephalomyelitis in the DA rat; impact on severity of myelin oligodendrocyte glycoprotein-induced experimental autoimmune encephalomyelitis and antibody isotype pattern

Quantitative trait loci (QTL) controlling inflammatory diseases with different organ specificity may hypothetically either be unique for one disease or shared among different diseases. We have investigated whether five non-MHC QTL controlling susceptibility to experimental arthritis in the DA rat also influence myelin oligodendrocyte glycoprotein (MOG)-induced experimental autoimmune encephalomyelitis (EAE) in an F2 intercross between inbred DA and PVG.RT1a rats. Two of the five chromosome regions affecting arthritis in the DA rat also regulate phenotypes of EAE. The DA allele at markers in Cia3 (collagen-induced arthritis QTL) on chromosome 4 is associated with more severe EAE and high levels of anti-MOG antibodies of the IgG2c subclass. Since production of antibodies of the IgG2c subclass may be stimulated by Th1 cells, and there is previous evidence that such cells promote EAE, it is possible that both of the studied phenotypes are controlled by the same gene or genes regulating Th1/Th2 cell differentiation. Furthermore, we show that Oia2 (oil-induced arthritis QTL) on chromosome 4 regulates levels of anti-MOG antibodies of the IgG1 subclass and of anti-MOG IgE, but that this gene region does not affect clinical disease severity in our study. Since production of IgE and IgG1 may be stimulated by Th2 cells, this QTL may also control Th1/Th2 bias. We conclude that Cia3 and Oia2 regulate MOG-induced EAE in rats. Furthermore, since both EAE and arthritis phenotypes co-localize to these gene regions, they may harbor genes which are key regulators of pathogenic immune responses.     

Genes on the X chromosome affect development of collagen-induced arthritis in mice

Susceptibility to collagen-induced arthritis (CIA) in mice is associated with a class II gene in MHC (Aq) but also with unknown genes outside MHC. Investigated here is the influence of genes on the X chromosome as well as the role of the X-linked immunodeficiency (xid) mutation. Reciprocal male F1 hybrids, bred to be heterozygous or homozygous for Aq, showed a genetic influence in their susceptibility to develop CIA. Crosses were made between B10.G, B10.Q, DBA/1, SWR/J, C3H.Q and CBA/Ca, and all F1 mice were castrated to avoid sex hormone modulation of the susceptibility. A differential timing of arthritis onset and severity were seen in the reciprocal F1 males. An exception was the reciprocal F1 male offspring from SWR/J and DBA/1 crosses which differed only in disease severity late in the course of the disease. The female F1 crosses did not show the same pattern of differential susceptibility to CIA as the F1 males. To exclude the possible influence of the Y chromosome, F1 males of reciprocal crosses were back-crossed to the parental strains creating offspring with equal X chromosomes but divergent Y chromosomes. No difference in development of arthritis was observed in these. The influence of the xid mutation was investigated next. The xid loci from the CBA/N mouse was bred into DBA/1 strain which is highly susceptible to CIA. The resulting congenic DBA/1-xid strain was resistant to induction of CIA and did not develop an antibody response to type II collagen. We conclude that polymorphic genes on the X chromosome modulate susceptibility to CIA. The results from the experiments with mice carrying xid mutations confirm that such immune modulating genes exist on the sex chromosomes.     

Genetic linkage analysis of collagen-induced arthritis in the mouse

The genetic susceptibility to collagen-induced arthritis (CIA) in mice, the most commonly used model for rheumatoid arthritis, has been analyzed. The highly susceptible B10.RIII strain was crossed with the resistant RIIIS/J strain and the F2 intercross mice were subjected to genomic screening using microsatellite markers. These strains share the MHC region on chromosome 17, known to be of importance in CIA (this locus is named Mcia1). The same cross has earlier been used to map the major genes outside the MHC controlling chronic relapsing experimental allergic encephalomyelitis (EAE). It was found that the major locus controlling CIA (Mcia2; lod 4.12) was located on chromosome 3 in the same region as one of the major loci controlling EAE (Eae3). The linkage was reproduced in a mouse strain in which the locus was isolated on the B10.RIII background at the N4I2 generation. A second putative locus was identified on chromosome 13 (lod 3.13). The present finding identifies new loci outside the MHC controlling CIA and provides evidence that mouse CIA is controlled by polymorphic genes.     

Therapeutic effects of monoclonal antibodies to alpha beta TCR but not to CD4 on collagen-induced arthritis in the rat

The role which T cells play in the pathogenesis of the collagen-induced arthritis (CIA) model is not yet fully understood. Although CIA is most likely dependent on the activity of class II-restricted CD4+ T cells, only prophylactic but not therapeutic anti-CD4 treatments have been successful. The lack of therapeutically effective anti-T cell monoclonal antibody treatments has questioned the importance of T cells in ongoing CIA. However, recently we found that ongoing CIA in DA rats induced with homologous CII can be suppressed by injections with an anti-alpha beta TCR antibody. Having a CIA model where ongoing disease was clearly dependent on T cells, we addressed in the present work whether also an anti-CD4 treatment could suppress ongoing arthritis in this model. Although no CD4hi lymph node cells were seen after an anti-CD4 injection, the arthritis was suppressed only after treatment at immunization but not after treatment just before onset of disease. In comparison, the anti-TCR treatment at the time of onset was clearly suppressive even though a large fraction of the T cells was not depleted. This indicates that the different outcome of the anti-TCR and anti-CD4 treatment was not due to a different capacity to deplete T cells in vivo.     

Identification of quantitative trait loci governing arthritis severity and humoral responses in the murine model of Lyme disease

A spectrum of disease severity has been observed in patients with Lyme disease, with approximately 60% of untreated individuals developing arthritis. The murine model of Lyme disease has provided strong evidence that the genetic composition of the host influences the severity of arthritis following infection with Borrelia burgdorferi: infected C3H mice develop severe arthritis while infected C57BL/6N mice develop mild arthritis. Regions of the mouse genome controlling arthritis severity and humoral responses during B. burgdorferi infection were identified in the F2 intercross generation of C3H/HeNCr and C57BL/6NCr mice. Rear ankle swelling measurements identified quantitative trait loci (QTL) on chromosomes 4 and 5, while histopathological scoring identified QTL on a unique region of chromosome 5 and on chromosome 11. The identification of QTL unique for ankle swelling or histopathological severity suggests that processes under distinct genetic control are responsible for these two manifestations of Lyme arthritis. Additional QTL that control the levels of circulating Igs induced by B. burgdorferi infection were identified on chromosomes 6, 9, 11, 12, and 17. Interestingly, the magnitude of the humoral response was not correlated with the severity of arthritis in infected F2 mice. This work defines several genetic loci that regulate either the severity of arthritis or the magnitude of humoral responses to B. burgdorferi infection in mice, with implications toward understanding the host-pathogen interactions involved in disease development.     

Genetic mapping of two new blood pressure quantitative trait loci in the rat by genotyping endothelin system genes

The endothelin system, consisting of a series of potent vasoconstrictor peptides and their receptors, is potentially important in the control of blood pressure. We found that the gene coding for endothelin-2 (ET2), also known as vasoctive intestine peptide, cosegregated strongly with systolic blood pressure in a F2 population [F2(S x LEW)] derived from a cross of the Dahl salt-sensitive (S) rat and the Lewis (LEW/NCrlBR) (LEW) rat. The ET2 locus was assigned to rat chromosome 5. The testis-specific histone (HITH) locus also strongly cosegregated with blood pressure in the F2(S x LEW) population and was assigned to rat chromosome 17. Genetic maps of the regions containing the quantitative trait loci (QTL) for blood pressure on chromosomes 5 and 17 were constructed and the QTL were localized using the MAPMAKER/QTL program. The rat genes for endothelin-1, endothelin-3, and endothelin receptor A did not cosegregate with blood pressure in several F2 populations tested and were assigned to rat chromosomes 17, 3, and 19, respectively. Endothelin receptor B cosegregated weakly with blood pressure and was provisionally assigned to rat chromosome 15. We conclude that, in the rat, one new blood pressure QTL is located on chromosome 5 marked by the ET2 locus and another new QTL is located on chromosome 17 near the HITH locus.     

Is the nitric oxide system involved in genetic hypertension in Dahl rats?

To address this issue, a series of genetic tests were carried out. Linkage studies showed that the inducible nitric oxide synthase (Nos2) locus cosegregated with blood pressure in three F2 populations originated from crosses of Dahl salt-sensitive (S) rats with rats of various normotensive strains. However, the brain nitric oxidase synthase (Nos1) and endothelial nitric oxide synthase (Nos3) loci did not cosegregate with blood pressure in five F2 populations. Thus, only Nos2, but not Nos1 and Nos3, was considered as a candidate gene for being a quantitative trait locus (QTL) for blood pressure in the S rat. To further test this hypothesis, congenic strains were constructed by substituting regions on Chromosome 10 of the S rat with the homologous regions of the Milan normotensive (MNS) rat. Results showed that the chromosome region including Nos2 did not contain a blood pressure QTL. In consequence, Nos2 per se is not supported as a candidate QTL capable of causing a blood pressure difference between the S and MNS rats. Nevertheless, the nitric oxide system appears to be involved secondarily in the regulation of blood pressure in the S rat, as evidenced by physiological data.     

Identification of rat susceptibility loci for adjuvant-oil-induced arthritis

One intradermal injection of incomplete Freund's adjuvant-oil induces a T cell-mediated inflammatory joint disease in DA rats. Susceptibility genes for oil-induced arthritis (OIA) are located both within and outside the major histocompatibility complex (MHC, Oia1). We have searched for disease-linked non-MHC loci in an F2 intercross between DA rats and MHC-identical but arthritis-resistant LEW.1AV1 rats. A genome-wide scan with microsatellite markers revealed two major chromosome regions that control disease incidence and severity: Oia2 on chromosome 4 (P = 4 x 10(-13)) and Oia3 on chromosome 10 (P = 1 x 10(-6)). All animals homozygous for DA alleles at both loci developed severe arthritis, whereas all those homozygous for LEW.1AV1 alleles were resistant. These results have general implications for situations where nonspecific activation of the immune system (e.g., incomplete Freund's adjuvant-oil) causes inflammation and disease, either alone or in conjunction with specific antigens. They may also provide clues to the etiology of inflammatory diseases in humans, including rheumatoid arthritis.     

Contribution of abdomen and legs to central blood volume expansion in humans during immersion

Collagen-induced arthritis (CIA) is a T cell-dependent disease in which susceptibility is controlled by genes both within and outside the major histocompatibility complex (MHC). In the present study, we compared the humoral responses and kinetics of cytokine secretion patterns in the draining lymph nodes of arthritis-susceptible DA rats and arthritis-resistant F344 and DA MHC congenic PVG.1AV1 rats immunized with rat type II collagen (RCII) in incomplete Freund's adjuvant. The results demonstrate a marked humoral RCII response and a Th1 cytokine profile, with expression of interferon-gamma and interleukin (IL)-2 mRNA in DA rats; a limited humoral RCII response and a Th2 cytokine profile, with expression of IL-4 mRNA in arthritis-resistant F344 rats; and a marked humoral RCII response in arthritis-resistant PVG.1AV1 rats. However, in contrast to DA rats, PVG.1AV1 rats produce IgG1 autoantibodies which, together with strong expression of IL-4 mRNA, indicates the involvement of Th2 subsets. From these data, we conclude that non-MHC gene(s) determines the direction of the anti-RCII response towards a Th1 disease-promoting, or a Th2 disease-limiting response.     

Can health care assistants replace student nurses?

Quantitative trait loci (QTLs) affecting body weight were investigated in the backcross population derived from diabetic BB/OK and spontaneously hypertensive rat (SHR). The F1 hybrids were backcrossed onto BB/OK rats, and QTL analysis was performed with the resulting backcross population on chromosomes 1, 3, 4, 10, 13 and 18. According to the stringent threshold for a lod score of 3.0, markers on chromosomes 1 and 4 were found to be linked with body weight. The QTL with a peak lod score (3.3) on chromosome 1 for a male population was located within the region flanked by loci Igf2 and D1Mgh12. On chromosome 4, linkage between the body weight and the region around the Npy locus was observed (lod score 3.1). The existence of the QTL on chromosome 4 affecting body weight was confirmed by congenic BB.LL rats, carrying chromosomal region of SHR (D4Mit6-Npy-Spr) on the genetic background of the BB/OK rat.     

Genetic determinants of autoimmune disease and coronary vasculitis in the MRL-lpr/lpr mouse model of systemic lupus erythematosus

MRL-lpr/lpr (MRL/lpr) mice are a model of human autoimmune disease. They exhibit a number of characteristics of systemic lupus erythematosus, including anti-DNA Abs, anti-cardiolipin Abs, immune complex-mediated vasculitis, lymphadenopathy, and severe glomerulonephritis. Although the autoimmune disorder is mediated primarily by mutation of the Fas gene (lpr), which interferes with lymphocyte apoptosis, MRL/lpr mice also have other predisposing genetic factors. In an effort to identify these additional factors, we have applied quantitative trait locus (QTL) mapping using an intercross between MRL/lpr mice and the nonautoimmune inbred strain BALB/cJ. A complete linkage map spanning the entire genome was constructed for 189 intercross progeny, and genetic loci contributing to features of the autoimmunity were identified using statistical analytic procedures. As expected, the primary genetic determinant of autoimmune disease in this cross was the Fas gene on mouse chromosome 19, exhibiting a lod score of 60. In addition, two novel loci, one on chromosome 2 (lod score, 4.3) and one on chromosome 11 (lod score, 3.1), were found to contribute to levels of anti-DNA Abs. Interestingly, the chromosome 19 and chromosome 11 QTLs, but not the chromosome 2 QTL, also exhibited associations with anti-cardiolipin Abs (lod scores, 38.4 and 2.6). We further examined the effects of these QTLs on the development of coronary vasculitis in the F2 mice. Our results indicate that the QTLs on chromosomes 11 and 19 also control the development of vasculitis, demonstrating common genetic determinants of autoantibody levels and vasculitis.     

Mapping of selective rat chromosome regions using mouse microsatellite markers

Four hundred eighty mouse microsatellite markers distributed in discrete regions on five mouse chromosomes were screened for producing PCR products in the rat. Ninety-eight of these markers or 20% give distinctive PCR products. Among these ninety-eight markers, twenty-three are polymorphic between the inbred hypertensive Dahl salt-sensitive (S) rat strain and several normotensive rat strains of interest. Fourteen of these polymorphic markers have been mapped to the homologous chromosome regions of the rat, and have further been utilized to localize quantitative trait loci (QTL) for blood pressure in the S rat.     

Genetic modifiers of Leprfa associated with variability in insulin production and susceptibility to NIDDM

In an attempt to identify the genetic basis for susceptibility to non-insulin-dependent diabetes mellitus within the context of obesity, we generated 401 genetically obese Leprfa/Leprfa F2 WKY13M intercross rats that demonstrated wide variation in multiple phenotypic measures related to diabetes, including plasma glucose concentration, percentage of glycosylated hemoglobin, plasma insulin concentration, and pancreatic islet morphology. Using selective genotyping genome scanning approaches, we have identified three quantitative trait loci (QTLs) on Chr. 1 (LOD 7.1 for pancreatic morpholology), Chr. 12 (LOD 5.1 for body mass index and LOD 3.4 for plasma glucose concentration), and Chr. 16 (P < 0.001 for genotype effect on plasma glucose concentration). The obese F2 progeny demonstrated sexual dimorphism for these traits, with increased diabetes susceptibility in the males appearing at approximately 6 weeks of age, as sexual maturation occurred. For each of the QTLs, the linked phenotypes demonstrated sexual dimorphism (more severe affection in males). The QTL on Chr. 1 maps to a region vicinal to that previously linked to adiposity in studies of diabetes susceptibility in the nonobese Goto-Kakizaki rat, which is genetically closely related to the Wistar counterstrain we employed. Several candidate genes, including tubby (tub), multigenic obesity 1 (Mob1), adult obesity and diabetes (Ad), and insulin-like growth factor-2 (Igf2), map to murine regions homologous to the QTL region identified on rat Chr. 1.     

A quantitative trait locus for alcohol consumption in selectively bred rat lines

Selective breeding for high and low alcohol consumption led to the establishment of alcohol-preferring (P) and alcohol-nonpreferring (NP) rat lines that differ greatly in their alcohol consumption. These lines were inbred and F2 intercross progenies were generated to detect quantitative trait loci (QTLs) influencing alcohol consumption. A QTL on chromosome 4 was identified with a maximum lod score of 8.6. This QTL acts in an additive fashion and accounts for 11% of the total phenotypic variability and approximately one-third of the genetic variability. Neuropeptide Y, an endogenous anxiolytic and neuromodulator, has been mapped to this same region of chromosome 4. This study is an advance in genome analyses, demonstrating that crosses between divergent, selectively bred rat lines can be used to identify QTLs. Localization of a gene influencing alcohol consumption may have important implications for the etiology of alcohol abuse and alcoholism in humans.     

The nuclear factor-kappa B RelA transcription factor is constitutively activated in human pancreatic adenocarcinoma cells

To assess the efficiency of nasally administered cartilage-specific collagens as vaccination against development of arthritis and to ameliorate already established chronic arthritis, experimental models which develop chronic arthritis, pristane-induced arthritis (PIA), and homologous collagen-induced arthritis (CIA) in the rat were selected. Cartilage-specific collagens type IX (CIX) and type II (CII) were used for vaccination intranasally. A single dose of 250 microg CII instilled intranasally in rats with established PIA ameliorated the disease. For the prevention of disease, the same dose given before immunization was found to be most effective. Most importantly, the disease was more severe if this dose was given three times. For treatment of PIA, CIX was found to be more effective than CII, whereas for treatment of CIA only CII was effective. The amelioration of CIA was associated with a marked suppression of delayed type hypersensitivity and the flare reaction to CII and lower levels of IgG2b anti-CII antibodies in serum, i.e., with suppression of the TH1 rather than the TH2 response to CII. These findings, that cartilage proteins, if given intranasally, can both prevent and ameliorate established chronic arthritis in rats, are of significant importance for possible use in rheumatoid arthritis. The identification of two different cartilage-specific proteins (CII and CIX) effective against a disease induced with a well-defined nonimmunogenic adjuvant such as pristane will be of value for enhancing the effectiveness of the treatment.     

Collagen-induced arthritis in CD4- or CD8-deficient mice: CD8+ T cells play a role in initiation and regulate recovery phase of collagen-induced arthritis

Collagen-induced arthritis (CIA) is an experimental autoimmune disease induced by immunization with collagen type II (CII). We studied CIA in CD4- or CD8-deficient DBA/1 mice to further define the roles of CD4+ and CD8+ T cells in the disease. CD4-deficient mice developed severe arthritis, and no differences in incidence, clinical course, and severity were observed between CD4 -/- and CD4 +/- mice. Proliferative responses of lymph node T cells to CII was, however, reduced in CD4 -/- mice, and inflamed joints revealed relative accumulation of CD4-CD8-TCR(alpha)(beta)+ cells. A CII-specific T cell line generated from CD4-deficient mice responded to CII in a MHC-restricted fashion and had a CD4-CD8-TCR(alpha)(beta)+ phenotype. Disease incidence in CD8 -/- mice was significantly decreased compared with CD8 +/- mice, even though the severity of arthritis in arthritic mice was not different. These results suggests a role for CD8+ T cells in initiating CIA. Interestingly, CD8-deficient mice were more susceptible to a second induction of arthritis after remission of initial disease, pointing towards an immunoregulatory role for CD8+ T cells. CD8-deficient mice did not, however, show any defect in oral tolerance induction using CII. Taken together, our findings demonstrate that CD4-CD8-TCR(alpha)(beta) cells can trigger systemic arthritis in CD4-deficient mice and that CD8+ T cells can play dual and opposing roles, important both in initiation of CIA and in providing resistance to reinduction of CIA after recovery from initial disease.     

Identification of a quantitative trait locus influencing plasma insulin levels after 70% pancreatectomy using the OLETF rat

The Otsuka Long-Evans Tokushima Fatty (OLETF) rat is an animal model for obese-type non-insulin-dependent diabetes mellitus (NIDDM) in humans. The OLETF rat has poor capacity for proliferation of pancreatic beta-cells after partial pancreatectomy, which may be the critical pathogenetic event in NIDDM development. The poor pancreatic beta-cell proliferation in this model is characterized by reduction in beta-cell mass and decrease in insulin content in the remnant pancreas. Our investigation was designed to identify quantitative trait loci (QTLs) responsible for beta-cell mass and plasma insulin levels after partial pancreatectomy by performing a genome-wide scan in an F2 intercross obtained by mating the OLETF and the Fischer-344 (F344) rats. We have identified a suggestive QTL for the plasma insulin levels, near D20Mgh5 on rat chromosome 20, with a maximum lod score of 3.75 which accounts for 20% of the total variance, while no QTLs were detected for beta-cell mass. This chromosome 20 QTL, whose OLETF allele is associated with low plasma insulin levels through acting in an incompletely recessive manner, may affect insulin secretion itself rather than beta-cell proliferation.     

Increased serum levels of cartilage oligomeric matrix protein in chronic erosive arthritis in rats

OBJECTIVE: To investigate the utility of serum cartilage oligomeric matrix protein (COMP) for disease monitoring in rats with chronic pristane-induced arthritis, and to examine the influence of age, sex, and genes on COMP concentrations in rat serum. METHODS: Serum COMP levels were quantified by immunoassay. Sera were obtained from DA, E3, and (E3 x DA)F1 rats each week between the ages of 4 and 30 weeks. The (E3 x DA)F2 (second generation after intercrossing) rats were injected intradermally with the synthetic oil pristane. Arthritis was monitored by a macroscopic scoring system, and serum levels of COMP were measured on days 6, 35, and 49 after immunization. RESULTS: Serum levels of COMP decreased during growth, and reached a plateau after the age of 12 weeks. The DA rats had higher COMP levels than the E3 rats, and the (E3 x DA)F1 rats had intermediate levels. No differences were observed in these levels when the rats were grouped by sex. Arthritic (E3 x DA)F2 rats had increased serum concentrations of COMP on days 35 and 49 after pristane injection (P < 0.0001 versus the nonarthritic animals). COMP levels correlated with the severity of macroscopically detectable arthritis at both time points (r > 0.9). Rats with a chronic disease course were distinguished by higher serum concentrations of COMP during the acute stage than animals with similar clinical scores but with resolving arthritis (P < 0.01). CONCLUSION: Serum analyses of COMP offer promise for monitoring tissue involvement in experimental arthritis. Such analyses should be useful for monitoring therapeutic interventions aimed at retarding tissue destruction as well as for studies of the genetically determined regulation of COMP turnover and susceptibility to arthritis. The application of molecular marker measurements to experimental arthritis offers a new possibility for verifying the utility of such measurements in human arthritis.     

New genetic loci that control susceptibility and symptoms of experimental allergic encephalomyelitis in inbred mice

Experimental allergic encephalomyelitis (EAE), the principal animal model of multiple sclerosis, is a genetically determined phenotype. In this study, analyses of the cumulative disease frequencies in parental, F1 hybrid, and F2 mice, derived from the EAE-susceptible SJL/J strain and the EAE-resistant B10.S/DvTe strain, confirmed that susceptibility to EAE is not inherited as a simple Mendelian trait. Whole genome scanning, using 150 informative microsatellite markers and a panel of 291 affected and 390 unaffected F2 progeny, revealed significant linkage of EAE susceptibility to marker loci on chromosomes 7 (eae4) and 17, distal to H2 (eae5). Quantitative trait loci for EAE severity, duration, and onset were identified on chromosomes 11 (eae6, and eae7), 2 (eae8), 9 (eae9), and 3 (eae10). While each locus reported in this study is important in susceptibility or disease course, interactions between marker loci were not statistically significant in models of genetic control. One locus, eae7, colocalizes to the same region of chromosome II as Orch3 and Idd4, susceptibility loci in autoimmune orchitis and insulin-dependent diabetes mellitus, respectively. Importantly, eae5 and eae7 are syntenic with human chromosomes 6p21 and 17q22, respectively, two regions of potential significance recently identified in human multiple sclerosis genome scans.     

Dual role of IL-12 in early and late stages of murine collagen type II arthritis

IL-12 can promote Th1 responses, and early administration of IL-12 during immunization was shown to enhance expression of autoimmune collagen-induced arthritis (CIA). We now studied the impact of IL-12 at the stage of disease expression and during established CIA in DBA-1 mice. Accelerated onset and enhanced severity were provoked when i.p. injections of 100 ng of murine IL-12 (mIL-12) were given around the time of arthritis onset. Moreover, the onset of CIA could be ameliorated with anti-mIL-12 Abs, indicating that IL-12 is a pivotal mediator in the expression of CIA. In addition, the effect of anti-mIL-12 treatment was analyzed in established CIA. Continued treatment did not suppress established arthritis. Instead, these mice showed an impressive exacerbation of arthritis shortly after cessation of anti-mIL-12 treatment, indicative of impairment of endogenous control. Exaggerated disease was characterized by massive granulocyte influx and enhanced expression of IL-1 beta and TNF-alpha mRNA in the synovial tissue. Subsequently, we treated established collagen arthritis with recombinant mIL-12 for 7 days. Profound suppression of the arthritis score was noted, including reduced influx of cells and diminished cartilage damage. Tenfold enhanced levels of IL-10 were detected in sera of mIL-12-treated mice, and up-regulated mRNA levels of IL-10, IFN-gamma, and IL-12 were measured in synovial tissue. Finally, the anti-inflammatory effect of IL-12 on CIA could be reversed by coadministration of anti-IL-10 Abs. This study indicates that IL-12 has a stimulatory role in early arthritis expression, whereas it has a suppressive role in the established phase of collagen arthritis.