Pediatric Rheumatology Online Journal → June 2003 → Epidemiology, Classification, Immunology and Immunogenetics → Abstract #6

USE OF MICROARRAYS TO CATEGORIZE CHILDHOOD ARTHRITIS

S. D. Thompson,¹ M. G. Barnes,¹ B. J. Aronow,² L. L. Luyrink,¹ A. A. Grom,¹ M. B. Moroldo,¹ E. H. Gianinni,¹ R. A. Colbert,¹ D. N. Glass.¹

¹Rheumatology, Cincinnati Childrens Hospital Medical Center, Cincinnati, OH; ²Pediatric Informatics, Cincinnati Childrens Hospital Medical Center, Cincinnati, OH

We initiated a large-scale study to evaluate the power of microarray-based differential gene expression analyses to identify biological markers predictive of juvenile arthritis disease subtype, course and response to treatment. We initially evaluated gene expression patterns in peripheral blood mononuclear cells (PBMC) from patients with long-standing disease. Samples were collected from 21 patients with juvenile rheumatoid arthritis (pauciarticular, polyarticular, or systemic onset), 6 with juvenile spondyloarthropathy (JSpA), and 11 healthy controls. Samples were collected from patients (median 8.4 yr after diagnosis) with active arthritis (ACR criteria) and an established course. RNA prepared from PBMC (Ficoll) was hybridized to Affymetrix U95Av2 GeneChips. The resultant data were analyzed by ANOVA for genes exhibiting significant expression differences related to disease course (pauci n=5; poly n=15; systemic n=1; JSpA n=6). Polyarticular course patients were readily distinguished from controls by approximately 200 differentially expressed genes (p0.0001). This list included genes both previously associated with polyarticular JRA and not associated. Several differentially expressed genes were cell type-specific (T, B, NK, Dendritic or macrophage/monocyte). To obtain deeper insight into cell-type and activation-state-specific patterns, we systematically searched for immuno-inflammatory genes that exhibited coordinate regulation across the patient series. Small sets of functionally related index genes were identified that served as kernels to identify genes that shared a functional group-specific pattern. The results demonstrated that many immuno-inflammatory genes are strongly and coordinately modulated in JRA patients. Furthermore, patient-specific variations among these groups of genes strongly suggest that new subclasses of JRA patients may be identifiable using this technology and may provide a basis for the pursuit of novel, group-specific therapeutic approaches.