Article Text

FRI0003 Determination of interferon (IFN) signatures for sle patients may be critical for optimal treatment selection but depends on the genes chosen: report from the bold (biomarkers of lupus disease) study
  1. A. A. Hill1,
  2. F. W. Immermann2,
  3. Y. Zhang3,
  4. P. S. Reddy1,
  5. T. Zhou4,
  6. M. O’Toole3,
  7. M. Z. Whitley1,
  8. J. Masferrer3,
  9. P. W. Wu3,
  10. T. Paradis3,
  11. S. Sridharan3,
  12. J. T. Merrill5,
  13. J. A. James5
  1. 1Research Business Technologies, Pfizer, Cambridge
  2. 2Biostatistics, Pfizer, Pearl River
  3. 3BioTx Clinical Research, Pfizer, Cambridge
  4. 4GBS Clinical Discovery, BMS, Lawrensville
  5. 5Rheumatology, OMRF, Oklahoma City, United States


Background Elevated expression of IFN pathway genes, known as the IFN signature (IS) is found in roughly half of SLE patients1. Various studies have used different IFN genes to define the IS, and different methods to assign patients to “IFN high” or “low” expression groups. Importantly a recent report from the Genentech ROSE study found higher responsiveness to an interferon antagonist in patients with lower IFN expression using a 7 gene set (G-7), suggesting that IS might inform optimal selection of treatments in the future2.

Objectives As a step toward understanding the clinical relevance and molecular drivers of the IS, we used data from the BOLD study to test the difference between IS-derived patient groupings derived from different sets of IFN genes.

Methods The BOLD study enrolled SLE patients with active disease (minimum SLEDAI of 6 or two BILAG B scores). 41 patients had background immune suppressives withdrawn, were given brief steroids until improvement, then followed with serial samples until disease flare. 62 additional active SLE patients were evaluated only once. RNA expression levels for 329 genes were assayed using TaqMan® Low Density Arrays. 238 SLE patient-visits were partitioned into two groups by unsupervised hierarchical clustering of the expression levels of IFN-related genes.

Results Clustering was done separately for 5 gene sets, denoted B-30, P-11, M-20, G-7, and O-3. Two of the gene sets (M-20 and G-7) were 95% and 100% matched to sets used in clinical trials of IFN-inhibiting treatments. Consistent with previous reports, all of the gene sets assigned about 40-50% of visits (96 to 118 patient-visits) to an “IFN high” group characterized by IFN pathway expression that was higher than healthy volunteers. In a cross-sectional analysis using the P-11 gene set, membership in the IFN high group was associated with higher SLEDAI score than IFN low assignment (9.73 vs. 7.70, p<0.0001) and higher TNFSF13B (BLYS/BAFF) gene expression (2.4 fold increase, p=0.0005). Most of the patients followed longitudinally remained in the same group over the course of the study, despite changes of gene expression with disease improvement and flare. IFN group assignments were similar for the B-30 and P-11 gene sets, and also similar among the M-20, G-7, and O-3 gene sets, but differed more across these groups. For instance, 21 patient-visits (9%) were assigned to the IFN high group by the G-7 gene set, but placed in the IFN low group by the P-11 gene set.

Conclusions The comparison of IFN group assignments for the same samples, derived using different IFN gene sets, highlights the need for more investigation of the molecular and cellular drivers of IFN gene expression in SLE blood. An optimal IFN gene set, once determined, might improve how patients are selected for targeted therapies in the future.


  1. Baechler et al., PNAS 100(5) 2610-2615 (2003);

  2. Kalunian et al., ACR 2012 Washington D.C, Abstract 2622.

Disclosure of Interest None Declared

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.