Critical role for endogenous G-CSF during CIA. (A) Elevated serum G-CSF levels in WT mice during CIA. Data represent mean plus or minus SEM; n > 12 mice per time point from 2 to 3 independent experiments. *P < .001 compared with naive mice. Dashed line indicates typical profile of clinical features developing in CIA. (B) Reduced incidence of CIA in G-CSFR^−/− mice. Data represent percentage of total mice. P < .001 at day 45. (C) Reduced clinical scores in G-CSFR^−/− mice. Data represent mean plus or minus SEM; P < .01 during days 21 to 45. n = 21 to 32 mice per group from 2 independent experiments. (D) Reduced histologic arthritis in joints from G-CSFR^−/− mice. Data represent percentage of total joints. n > 93 interphalangeal and metacarpophalangeal/metatarsophalangeal joints scored per treatment. *P < .005. Histologic sections of (E) an arthritic joint from a CIA WT mouse and (F) a typical unaffected joint from a CIA G-CSFR^−/− mouse. E indicates exudate; JS, joint space; C, cartilage; B, bone; and S, synovium. Magnification, ×100 (H&E stained). (G) Reduced clinical severity in G-CSF^−/− mice reconstituted with WT bone marrow ([WT] → [G-CSF^−/−]) compared with [G-CSF^−/−] → [WT] mice. Data represent mean plus or minus SEM; P < .05 during days 56 to 63. n = 12 to 21 mice per group from 2 independent experiments. (H) Reduced clinical severity in WT mice reconstituted with G-CSFR^−/− bone marrow ([G-CSFR^−/−] → [WT]) compared with [WT] → [G-CSFR^−/−] mice. P < .05 during days 21 to 63. n = 15 to 17 mice per group from 2 independent experiments.