Background Type I interferon (IFN) has been implicated in systemic lupus erythematosus (SLE) pathogenesis, and the majority of patients with SLE have elevated expression of type I IFN-inducible genes in their blood. Anifrolumab, a fully human, IgG₁κ monoclonal antibody against the type I IFN receptor, is in Phase III development for the treatment of moderate to severe SLE (NCT02446912 and NCT02446899).

Objectives We sought to understand other molecular pathways (either dependent on or independent of type I IFN signaling), to elucidate heterogeneous mechanisms in SLE, and to identify patient subsets for personalized disease management.

Methods Baseline blood samples from adult patients with moderate to severe SLE from two Phase IIb clinical studies (NCT01438489, N=265; NCT01283139, N=416) were profiled with whole genome array analyses. Type I IFN gene signature (IFNGS) test status was determined by a central laboratory utilizing an analytically validated four gene (IFI27, IFI44, IFI44L, RSAD2) quantitative polymerase chain reaction-based test from patients' whole blood. A predetermined, delta Ct-based cut-off point, in the trough of the bimodal distribution, was utilized to segregate type I IFNGS test–high from –low patients at baseline. Blood from healthy controls was stimulated ex vivo with IFN-β, IFN-γ, IFN-λ, IFN-ω, or a pool of all IFN-α subtypes, with or without blocking antibodies for each IFN type, to develop IFN-type-specific signatures. Cell type- and cytokine pathway-specific gene signatures derived from the literature were also evaluated with the Phase IIb sample data. A Fisher's exact test was used for enrichment calculations (signatures cut at median), and comparisons were adjusted for multiplicity through false discovery rate.

Results A total of 79% of SLE patients in the combined study population had a type I IFNGS test–high status. From the type I IFNGS test–high patients, 29/95 signatures evaluated had significant enrichment, including those for B cells (q=1.17E-17, odds ratio [OR]=6.4), plasma cells (q=6.96E-11, OR=3.9), and CD40L signaling (q=1.07E-08, OR=3.3), relative to type I IFNGS test–low patients. In contrast, type I IFNGS test–low patients had enrichment for eosinophils (q=5.4E-6, OR=0.39) and type II IFN (IFN-γ) specifically inducible gene signatures (q=4.6E-3, OR=0.47). These findings were significant for the combined study population, as well as for the NCT01438489 study population, and were either significant or trending for the NCT01283139 population (q<0.05).

Conclusions SLE patients who are type I IFNGS test–high had elevated concentrations of B cells, plasma cells, and other inflammatory cytokine pathways. Type I IFNGS test–low patients, by contrast, were enriched for eosinophil and type II IFN pathways. These observations provide new insights into the molecular heterogeneity underlying SLE and suggest new therapeutic approaches, particularly for type I IFNGS test–low patients.

Acknowledgements Funded by MedImmune. Medical writing support was provided by R. Plant, QXV Comms, an Ashfield business, UK.

Disclosure of Interest H. Liu Employee of: MedImmune LLC, B. Higgs Shareholder of: AstraZeneca, Employee of: MedImmune LLC, W. Rees Employee of: MedImmune LLC, C. Morehouse Shareholder of: AstraZeneca, Employee of: MedImmune LLC, K. Streicher Employee of: MedImmune LLC, P. Brohawn Shareholder of: AstraZeneca, Employee of: MedImmune, G. Illei Shareholder of: AstraZeneca, Employee of: MedImmune LLC, K. Ranade Employee of: MedImmune LLC