Background/Purpose: Data from the NIH Roadmap Epigenomics and ENCODE projects have revealed unexpected richness in mammalian transcriptomes. Large regions the genome that do not encode proteins have now been shown to be functional. These studies were performed to examine the JIA neutrophil transcriptome in detail.

Methods: We studied neutrophils in 35 children with newly-diagnosed, untreated polyarticular, RF-negative JIA using Affymetrix exon and miRNA arrays.  Labeling, hybridization, and scanning were undertaken using standard procedures. Patients were compared with 43 healthy controls (HC) and 15 children with cystic fibrosis (CF), a control group added to distinguish transcriptional features specific to JIA and those generic to chronic inflammation in soft tissues. Both exon and miRNA expression data were processed using RMA software in the Affy package of Bioconductor and analyzed using t-tests to detect differentially expressed genes (DEGs) between JIA and HC as well as between CF and HC. Exon arrays were also processed using APT program for Exon-level analysis. Splicing analyses were performed using APT program MiDAS and in-house developed PERL scripts.

Results:  We found significant differences in the transcriptomes of children with JIA compared with children with CF, which was evidenced by the number of DEGs from JIA and CF when compared to HC.  Whereas the analyses detected 5965 DEGs from CF and 216 DEGs from JIA, 148 of them are common to both groups (p = 7.5E-28). Similar results were obtained from both exon-level and miRNA gene expression data analysis. While 159 and 73 differentially expressed miRNA genes were detected in CF and JIA, respectively, 27 of them are common to both groups (p <0.001).  At  the isoform level, we found 5617 isoforms from CF and 446 isoforms from JIA displaying differential exon splicing, 169 of them are common to both groups (p=1.56E-129).  Functional analyses indicated that those 148 common DEGs were enriched for lysosome and proteasomal ubiquitin-dependent protein catabolic processes, while those DEGs  unique to JIA (68) are involved in intracellular signaling cascades, hormone-mediated signaling, and cell motion.  DEGs to unique to JIA were able to clearly separate JIA from healthy controls in a clustering analysis, and therefore these DEGs could serve as a gene expression signature for the disease. Functional associations from differential exon splicing analyses indicated that the common 169 isoforms were enriched in functions that included actin cytoskeleton organization, regulation of transcription, ceramide metabolic process, leukocyte activation, and cell activation during immune response. Those DEGs unique to JIA (277) were associated with functions that included as spliceosome regulation, B and T cell receptor signaling pathways, and apoptosis and MAPK signaling pathways. 

Conclusion: These findings reveal the complexity and specificity of the JIA neutrophil transcriptome.  Furthermore, they demonstrate the importance of filtering JIA transcriptome results with controls that include other chronic inflammatory conditions.  Finally these results indicate that a set of DEGs unique to JIA can serve as gene expression signature for the disease.

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/exon-and-mirna-arrays-reveal-complexity-and-specificity-of-the-juvenile-idiopathic-arthritis-neutrophil-transcriptome/