Background/Purpose: Systemic Lupus Erythematous (SLE) is a clinically and biologically diverse disease, for which only one new therapy has been approved in the past 60 years. In a phase 2 trial of mild-to-moderate SLE patients, ustekinumab (UST) improved clinical and laboratory measures of disease activity compared with placebo (PBO).¹ We previously reported an association of IFN-g reduction with response to UST,² suggesting an impact on the IL12/Th1 axis. To extend these findings, we performed unbiased transcriptomic analysis from baseline whole blood samples to identify genes that discriminate UST responders (UST-R) from non-responders (UST-NR) using the primary endpoint of Systemic Lupus Erythematosus Responder Index (SRI)-4 at week 24 to define response.

Methods: UST was studied in a Ph2 PBO-controlled study of 102 patients with seropositive SLE and active disease despite standard therapy. Patients were randomized 3:2 to receive IV UST 6 mg/kg or placebo followed by subcutaneous injections of UST 90mg or placebo every 8 weeks. Whole blood gene expression at baseline was measured via microarray using RNA samples from 100 patients, as samples from 2 patients failed quality control. An unbiased approach was used to identify gene signatures present at baseline that associate with UST response. Recombinant IL-12 or IL-23 was incubated in vitro with whole blood from 6 healthy donors for 24h and RNA-Seq was performed to determine the effect of these treatments on representative genes comprising the UST response signature.

Results: A non-biased machine learning algorithm identified a 9-gene whole blood signature composed primarily of cytotoxic cell-associated transcripts (PRF1, KLRD1, GZMH, NKG7, GNLY, FGFBP2, TRGC2, TARP, TRGV2) that was enriched at baseline in UST-R vs UST-NR. By Gene Set Variation Analysis, the cytotoxic signature enrichment in UST-NR was less at baseline than both UST-R and a healthy control cohort (P=0.0087, P=0.056, respectively), whereas UST-R cytotoxic gene enrichment was similar to healthy controls (P=0.31). No significant difference in cytotoxic signature enrichment was observed at baseline between PBO responders and PBO non-responders or healthy controls (Figure). Enrichment levels of the cytotoxic gene signature remained stable over time in PBO and UST-NR groups while a trend of decreased cytotoxic signature was observed in UST-R, although never reaching levels seen in UST-NR. To begin to understand the relationship between IL-12 and IL-23, the targets of UST, and the cytotoxic signature, whole blood was stimulated with these cytokines in vitro. Recombinant IL-12, but not IL-23, resulted in increased expression of representative members of this cytotoxic gene signature.

Conclusion: We identified a novel cytotoxic signature in baseline blood samples that associated with UST response in SLE. The observation that IL-12 can increase this signature in vitro and that IL-12 is a robust inducer of cytotoxic cell activity³ as well as IFN-γ³ suggests an important role of IL-12 blockade in the mechanism of action of UST in SLE.

1. van Vollenhoven RF. Lancet. 2018;392:1330-39
2. Jordan. ACR 2018 Abstract # 2951
3. G. Trinchieri. Nat Rev Immunol. 2003;3:133-46

« Back to ACR/ARP Annual Meeting 2019

ACR Meeting Abstracts - https://acrabstracts.org/abstract/baseline-cytotoxic-gene-expression-associates-with-ustekinumab-response-in-systemic-lupus-erythematosus/