Background/Purpose: Inflammatory cytokines influence steroid hormone production in tissue culture and affect serum levels and synovial fluid of active RA patients.  Inflammatory biomarker and hormonal network correlations were compared in this nested prospective study of pre-RA and CN subjects, to identify differences in their associations prior to clinical RA onset.  

Methods: Residents of Washington County, MD (21,061: 12,381 F, 8,680 M) enrolled in the Project CLUE cohort in 1974 and donated serum samples. After 3 to 20 (median 11) years, 54 cohorts were diagnosed ACR-definite RA (36 F, 18 M). Each pre-RA was matched on entry features with 4 non-RA (216 total: 144 F, 72 M) cohort members. Stored (-70 °C) sera were available on most subjects for assays of a comprehensive panel of 8 inflammatory markers, 15 steroids in women, and 8 in men. CRP and ASAA were assayed (mg/L) by high sensitivity ELISA at Boston University using reagents provided by Hemagen. A subset of female CRP assays was done by a modified-membrane ELISA method (mg/L) at Northwestern University (NWU). Quantikine cytokine assays (pg/mL) were done at Specialty Laboratories, Inc., Santa Monica, CA, using reagents from R & D Systems, Inc. Cytokine assays included: IL-1β; IL-6 (F only); TNF-α; IL-1rα; sTNF-R1 and IL-2sRα. A panel of 15 steroids in women and 8 in men were tested by immunoassay methods in a hormone research laboratory at NWU.  First, histograms of pre-RA vs CN immunomarkers and hormonal values, stratified by entry pre- vs post-menopausal and sex status, were compared to identify range differences (p<0.050). Next, matrix correlation tables of 8 inflammatory biomarkers and 18 hormones, and their P-P ratios (18 F, 8 M) were constructed using log-transformation and analyzed by age- and sex-adjusted partial Pearson (r_p) and unadjusted rank order Spearman (rho) methods to identify subject group differences (p<0.050). Principal component analysis (PCA) was performed on 148 subjects having values on 8 hormones, 6 immunologic markers, sex, entry age, and the study group variable (CN vs pre-RA).

Results: Among 225 total subjects with at least one paired immunologic and hormonal (I-H) test (46 pre-RA, 179 CN), 6 significant differences in correlations (deltas) were found, all stronger in the pre-RA vs CN (Table). PCA on 148 total subjects yielded 6 components, explaining 71.7% variance, which may be labeled as: (1) male sex-related hormones and sTNF-R1 (22.9%); (2) precursor hormones and estradiol (18.0%); (3) IL-1β and IL-1rα (10.2%); (4) cortisol, TNF-α, and entry age (7.7%); (5) CN vs pre-RA (6.9%), and (6) IL-2sRα (6.1%).