Abstract
Purpose:
NSOI is defined by an inflammatory infiltrate in the orbit in the absence of a malignancy such as lymphoma and in the absence of an associated systemic illness such as thyroid disease (TED), granulomatosis with polyangiitis (GPA), or sarcoidosis. We tested the hypotheses that analysis of gene expression by microarray could more clearly define NSOI, identify potential subsets, and possibly clarify pathogenesis and/or prognosis.
Methods:
82 formalin-fixed, anterior orbit biopsies (19 from normal, 24 NSOI, 22 with TED, 6 GPA, and 11 sarcoid) were submitted by 10 centers in the US, Canada, Saudi Arabia, and Australia. Biopsies were roughly divided into an initial discovery set and a subsequent validation set. In addition, a discovery set of lacrimal gland biopsies (7 normal and 21 NSOI) was studied. Gene expression was quantified by hybridization to Affymetrix HG-U133 plus 2.0 microarray chips. After quantile normalization, statistical analysis was with an empirical Bayes F-test and p-value adjusted for the Benjamini and Hochberg false discovery rate.
Results:
Transcripts for 449 known genes were more abundant in both the orbit discovery and validation sets by at least 1.5 fold and with a false discovery rate of p<0.05 comparing NSOI to normal tissue; transcripts for 1735 genes were less abundant by similar criteria. The mRNAs with greatly increased levels in the orbit frequently had an obvious role in inflammation, in marked contrast to genes with increased expression in TED orbit or NSOI lacrimal gland. Principal coordinate analysis indicates that NSOI is a heterogeneous collection of diseases with some biopsies resembling GPA, some sarcoid, and some TED. Gene expression in TED overlapped more consistently with gene expression in normal tissue.
Conclusions:
The analysis of transcripts expressed in biopsies from patients with orbital inflammation provides information not assessable by light microscopy. Gene expression distinguishes TED, GPA, and sarcoid and indicates that NSOI is a heterogeneous collection of diseases with a transcriptional profile that can resemble TED or GPA or sarcoid.
Keywords: 631 orbit •
557 inflammation •
535 gene microarray