BACKGROUND Adjunctive inflammatory modulation improved remission rates in treatment-resistant bipolar depression (TRBDD), but reliable biomarkers must be established to characterize the biosignature of TRBDD and the mechanisms underlying treatment response. In this molecular profiling study, we describe TRBDD and treatment response from the standpoint of interleukin-1 Beta (IL-1β) and KYN/TRP. METHODS 47 TRBDD patients with moderately severe HAMD-17 scores were randomized to receive either escitalopram (ESC) (10mg - 40mg daily dose range) + celecoxib (CBX) (200mg twice daily), or ESC (10mg - 40mg daily dose range) + placebo (PBO) (twice daily). Plasma cytokine levels were measured in both treatment arms at baseline and week 8, and in a healthy control (HC) group of subjects (N=43) once. A linear mixed model (LMM) was applied to evaluate whether clinical outcome is related to CBX and changes to biomarkers throughout treatment. A binary logistic regression model was formulated from this series to predict both the primary outcome of treatment response to CBX, and the secondary outcome of diagnosis of TRBDD using age, BMI, gender, and IL-1β at baseline. RESULTS Patients receiving ESC + CBX had 4.278 greater odds of responding (p=0.021) with NNT=3, and 15.300 times more likely to remit (p<0.001) with NNT=2, compared with ESC + PBO patients. Patient BMI (p=0.003), baseline IL-1β (p=0.004), and baseline KYN/TRP (p=0.001) were most predictive of TRBDD diagnosis. By Week 8, responders showed a downtrend in IL-1β compared to non-responders in the ESC+CBX treatment arm. However, there was no statistical difference in the IL-1β or KYN/TRP change after treatment between placebo and ESC+CBX group responders/non-responders (p=0.239, and p=0.146, respectively). While baseline IL-1β was elevated in TRBDD compared to HC (p<0.001), there was no difference in IL-1β between treatment responders at Week 8 compared to HC (p=0.067). CONCLUSIONS Elevated IL-1β and low KYN/TRP at baseline are components of the TRBDD molecular signature. CBX but not baseline IL-1β or KYN/TRP predict treatment response. Change in IL-1β and KYN/TRP did not predict treatment response.

中文翻译：

---

辅助性炎症调节对难治性双相抑郁症中 IL-1β 的影响

背景辅助炎症调节提高了难治性双相抑郁 (TRBDD) 的缓解率，但必须建立可靠的生物标志物来表征 TRBDD 的生物特征和治疗反应的潜在机制。在这项分子分析研究中，我们从白细胞介素 1 β (IL-1β) 和 KYN/TRP 的角度描述了 TRBDD 和治疗反应。方法 47 名具有中重度 HAMD-17 评分的 TRBDD 患者随机接受艾司西酞普兰 (ESC)（10mg - 40mg 每日剂量范围）+塞来昔布（CBX）（200mg 每日两次）或 ESC（10mg - 40mg 每日剂量范围）+安慰剂（PBO）（每天两次）。血浆细胞因子水平在基线和第 8 周时在两个治疗组中以及在健康对照 (HC) 受试者组 (N = 43) 中测量一次。应用线性混合模型 (LMM) 来评估临床结果是否与 CBX 和整个治疗过程中生物标志物的变化有关。从该系列中制定了二元逻辑回归模型，以预测对 CBX 治疗反应的主要结果和使用基线年龄、BMI、性别和 IL-1β 诊断 TRBDD 的次要结果。结果 与 ESC + PBO 患者相比，接受 ESC + CBX 的患者在 NNT=3 时的反应几率高 4.278（p=0.021），在 NNT=2 时缓解的可能性高 15.300 倍（p<0.001）。患者 BMI (p=0.003)、基线 IL-1β (p=0.004) 和基线 KYN/TRP (p=0.001) 最能预测 TRBDD 诊断。到第 8 周，与 ESC+CBX 治疗组中的无反应者相比，有反应者的 IL-1β 呈下降趋势。然而，安慰剂组和 ESC+CBX 组应答者/非应答者治疗后 IL-1β 或 KYN/TRP 变化没有统计学差异（分别为 p=0.239 和 p=0.146）。虽然与 HC 相比，TRBDD 中的基线 IL-1β 升高（p<0.001），但与 HC 相比，第 8 周的治疗反应者之间的 IL-1β 没有差异（p=0.067）。结论 基线时升高的 IL-1β 和低 KYN/TRP 是 TRBDD 分子特征的组成部分。CBX 但不是基线 IL-1β 或 KYN/TRP 预测治疗反应。IL-1β 和 KYN/TRP 的变化不能预测治疗反应。与 HC 相比，第 8 周的治疗反应者之间的 IL-1β 没有差异（p = 0.067）。结论 基线时升高的 IL-1β 和低 KYN/TRP 是 TRBDD 分子特征的组成部分。CBX 但不是基线 IL-1β 或 KYN/TRP 预测治疗反应。IL-1β 和 KYN/TRP 的变化不能预测治疗反应。与 HC 相比，第 8 周的治疗反应者之间的 IL-1β 没有差异（p = 0.067）。结论 基线时升高的 IL-1β 和低 KYN/TRP 是 TRBDD 分子特征的组成部分。CBX 但不是基线 IL-1β 或 KYN/TRP 预测治疗反应。IL-1β 和 KYN/TRP 的变化不能预测治疗反应。