Subsets of Patients With Lupus Identified by Gene Expression Profiles Exhibit Differential Clinical Responsiveness to Baricitinib.

To understand this discrepancy and identify patients with SLE who are responsive to baricitinib, we conducted baseline and longitudinal transcriptomic analysis in patients enrolled in the phase 2 trial.

为了理解这种差异并识别对baricitinib有反应的SLE患者，我们在2期试验中对入组患者进行了基线和纵向转录组分析。

Methods

Whole-blood samples from 272 patients with SLE were analyzed.

分析了272名SLE患者的全血样本。

Clinical response was assessed using SLE Responder Index-4 after 24 weeks of placebo or baricitinib (2 or 4 mg daily).

使用系统性红斑狼疮反应指数-4在24周后评估了安慰剂或baricitinib（每天2或4毫克）的临床反应。

Gene expression profiling was conducted using Gene Set Variation Analysis (GSVA) of 32 immune-related gene modules.

使用32个免疫相关基因模块的基因集变异分析（GSVA）进行了基因表达谱分析。

Results

Eight distinct molecular endotypes (A-H) were identified from baseline GSVA scores, with progressively increasing immune disturbances.

从基线GSVA评分中识别出8种不同的分子表型（A-H），免疫紊乱逐渐增加。

Baseline demographics were similar across endotypes, with modest but significant differences in anti-double-stranded DNA and complement levels but no SLE Disease Activity Index differences.

各表型的基线人口统计学特征相似，抗双链DNA和补体水平存在适度但显著的差异，但SLE疾病活动指数无差异。

Significant clinical responses to baricitinib were confined to endotypes D and G (P = 0.004 and 0.048 vs placebo, with effect sizes of 41.54% and 31.89%, respectively).

对baricitinib的显著临床反应仅限于表型D和G（与安慰剂相比，P=0.004和0.048，效应大小分别为41.54%和31.89%）。

Importantly, patient clustering based on clinical features failed to identify treatment-responsive subsets.

重要的是，基于临床特征的患者分型未能识别出治疗反应性亚组。

Endotype G demonstrated the most pronounced treatment-related transcriptional modulation, with dose-dependent suppression of interferon, JAK1/JAK2/TYK2 signaling, immunoproteasome, and inflammatory pathways evident as early as week 2 and sustained through week 24.

Endotype G 显示出最显著的治疗相关转录调控，剂量依赖性的干扰素、JAK1/JAK2/TYK2信号通路、免疫蛋白酶体和炎症途径的抑制在第2周就很明显，并持续到第24周。

Transcriptional changes in response to baricitinib were largely confined to clinical responders in endotype G.

对巴瑞替尼的反应导致的转录变化主要限于Endotype G的临床反应者。

Feature importance analysis identified the interferon and Treg cell signatures as major predictors of response.

特征重要性分析确定干扰素和调节性T细胞标志物是反应的主要预测因子。

Conclusions

Transcriptomic analysis identified subsets of patients with SLE responsive to baricitinib.

转录组分析确定了对巴瑞替尼有反应的系统性红斑狼疮患者亚组。