Background/Purpose

CANDLE syndrome is a novel autoinflammatory disease with strong IFN response signature.  We hypothesize that IFN dysregulation may drive clinical manifestations in CANDLE and treatment with baricitinib, a JAK 1 /2 inhibitor, will reduce the IFN-regulated genes (IRG) signature (Liu, 2012). Clinical improvement has been seen in CANDLE patients on baricitinib with significantly decreased steroid requirement and symptom scores (Montealetre, 2013).  We assess IRG expression in CANDLE patients on increasing doses of bariticinib and validate an IRG score as a potential biomarker. 

Methods

12 CANDLE patients (1.8-24.7yo, 9 male) enrolled in a NIH compassionate use protocol for baricitinib were assessed at baseline and on increasing doses (2-11mg/day).  Initial IRG list included all genes at least 2x upregulated in a chronic hepatitis patient and healthy peripheral blood mononuclear cells (PBMCs) exposed to IFN alpha was selected for IFN pathway genes in Ingenuity Pathway Analysis (IPA, Ingenuity® Systems).  IFN alpha, beta, and gamma and IFN receptor genes were added. IRGs with lowest Z-scores and little variability from initial studies on 10 CANDLE patients were cut to reduce list to 31 IRGs.  These and 4 housekeeping genes, were analyzed through RNA extracted from PAX gene tubes by the NanoString nCounter gene expression system (Seattle, WA). 3 healthy pediatric and 1 healthy adult were used as controls (HCs).  IRG scores were calculated by a) summing Z-score IRGs (summary score) and b) summing normalized value for IRGs on 0-1 scale (normalized score) for the a) 31 genes selected above and b) 6 most highly expressed IRGs at baseline in 9 CANDLE patients from previous RNAseq data.   The 6 and 31 gene IRG scores were assessed for correlation.   Mean of 6 gene IRG scores from all baselines were compared to mean IRG scores on treatment via t-test.  Paired t-test of 6 gene IRG scores in patients on the lowest and highest dose of treatment was also done excluding flaring patients.

Results

IRGs scores at 31 genes versus 6 genes highly correlate (linear model R ²: 0.95-0.99). Summary scores versus normalized scores strongly correlate (linear model R²: 0.72-0.76). Normalized 6 gene IRG score significantly decreased (p=0.003) on treatment (mean 0.9, 5mg/day, 0.22mg/kg/day) from baseline (mean 1.9).  With mean summary score (4828 vs. 2124), p value was 0.02.  Paired analysis of 8 CANDLE patients showed significant decrease in normalized 6 gene IRG score (p=0.03, mean score 1.01 vs. 0.38) at higher dose versus lower dose (0.25 vs. 0.12 mg/kg/day) but summary score was not significant (mean score 1915 vs. 946).

Conclusion

6 and 31 gene IRG score highly correlate.  6 gene normalized and summary IRG score significantly decrease in CANDLE patients on barcitinib treatment versus baseline with dose-dependent decrease with 6 gene normalized IRG.    Preliminary analysis indicates normalized score may be a better biomarker than summary score.  The IRG score may be useful as a biomarker in CANDLE syndrome and other autoinflammatory conditions with IFN-driven pathology, particularly for JAK inhibitor treatment.  Further pharmacodynamics studies to correlate pharmacokinetics, IRG score, and clinical status are needed.

---

« Back to 2014 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/validation-of-a-novel-ifn-regulated-gene-score-as-biomarker-in-chronic-atypical-neutrophilic-dermatosis-with-lipdoystrophy-and-elevated-temperature-candle-patients-on-baricitinib-a-janus-kinase-1/