Cutting edge: AIM2 and endosomal TLRs differentially regulate arthritis and autoantibody production in DNase II-deficient mice

Abstract

Innate immune pattern recognition receptors sense nucleic acids from microbes and orchestrate cytokine production to resolve infection. Inappropriate recognition of host nucleic acids also results in autoimmune disease. In this study, we use a model of inflammation resulting from accrual of self DNA (DNase II(-/-) type I IFN receptor [Ifnar](-/-)) to understand the role of pattern recognition receptor-sensing pathways in arthritis and autoantibody production. Using triple knockout (TKO) mice deficient in DNase II/IFNaR together with deficiency in either stimulator of IFN genes (STING) or absent in melanoma 2 (AIM2), we reveal central roles for the STING and AIM2 pathways in arthritis. AIM2 TKO mice show limited inflammasome activation and, similar to STING TKO mice, have reduced inflammation in joints. Surprisingly, autoantibody production is maintained in AIM2 and STING TKO mice, whereas DNase II(-/-) Ifnar(-/-) mice also deficient in Unc93b, a chaperone required for TLR7/9 endosomal localization, fail to produce autoantibodies to nucleic acids. Collectively, these data support distinct roles for cytosolic and endosomal nucleic acid-sensing pathways in disease manifestations.

Figure I. Arthritis in DKO mice is regulated by distinct DNA sensing pathways

A) Representative images of clinical arthritis in forepaws (top) and hindpaws (bottom) from 10 month-old female mice demonstrating significant swelling in DKO mice, absence of swelling in STING TKO mice, and an intermediate arthritic phenotype in AIM2 TKO mice. B) Clinical inflammation scores (n=14–24/genotype) showing a statistically different mean inflammation score in STING TKO and AIM2 TKO compared with DKO mice. 10 month-old mice: Het (14 female, 5 male), DKO (11 female, 14 male), STING TKO (9 female, 6 male), AIM2 TKO (9 female, 13 male). C) Histologic inflammation at the ankle (upper panel) and midfoot (lower panel) and D) quantitation of histologic inflammation, confirming differences in inflammation in STING TKO and AIM2 TKO mice compared with DKO mice. Histological analysis performed on 10 month-old female mice. Values are the mean +/− SEM; *** = p<0.001 compared to DKO. Arrow and (*) designate two sites of inflammation.

Figure II. AIM2 TKO mice demonstrate a significant decrease in IL-18 expression

A) Joint cytokine mRNA levels show a significant decrease in TNF in STING TKO mice compared with DKO mice, whereas a trend toward a decrease in TNF is seen in AIM2 TKO mice. MMP3 levels are significantly decreased in both STING TKO and AIM2 TKO mice compared with DKO mice, consistent with attenuation of clinical arthritis (n=4–6/genotype). 10 month-old mice: Het (4 female, 2 male), DKO (2 female, 4 male), STING TKO (3 female, 2 male), AIM2 TKO (2 female, 3 male). B) Western blot confirms a decrease in IL-18 protein expression in the joints of AIM2 TKO mice compared with DKO and STING TKO mice (n=1–4/genotype). n.s. designates non-specific protein staining (Ponceau S.). 10 month-old male mice. C) Serum ELISA assay demonstrates that IL-18 levels are significantly decreased in AIM2 TKO mice compared with DKO and STING TKO mice (n=6/genotype, 3 males, 3 females; 10 month-old mice). Values are the mean +/− SEM; * = p<0.05, ** = p<0.01, *** = p<0.001 compared to DKO.

Figure III. The Unc93b pathway uniquely regulates autoantibody production to nucleic acid

Heat map of an array of 125 autoantigens, including DNA- and RNA-associated antigens and common joint associated antigens. Autoantibody production is minimal in DKO and STING TKO mice at 3 months of age (left panel), prior to the onset of arthritis. Despite the significant decrease in arthritic inflammation in STING TKO and AIM2 TKO mice at 10 months, heat maps demonstrate autoantibody production, as in DKO mice. Marked attenuation of autoantibody production is seen in Unc93b TKO mice, confirming the essential role of endosomal TLRs in sensing nucleic acid for autoantibody production. Female and male mice were used in this analysis.

Figure IV. STING and AIM2 do not regulate arthritic inflammation in an immune complex-mediated model

A) Clinical inflammation scores and measurements of change in ankle thickness demonstrate no difference in inflammation in STING deficient (KO) mice compared with controls. 11 week-old male mice. B) H&E images (left) and histologic scoring (right) confirm equivalent inflammation. C) Clinical inflammation scores and measurements of change in ankle thickness demonstrate no difference in inflammation in AIM2 deficient (KO) mice compared with controls (Wt). 8 week-old male mice. Values are the mean +/− SEM. (*) designates synovial inflammation.