LSO-045 Multiple molecular biomarkers that predict response to treatment in lupus nephritis

BackgroundLittle has been known regarding the biomarker to predict the treatment response in lupus nephritis. This study aimed to identify potential biomarkers that predict treatment response in lupus nephritis (LN).MethodsIn this prospective longitudinal study, 66 active LN patients were included and underwent renal biopsy at the time of enrollment. Patients were divided into two groups according to one-year response: 50 responders and 16 non-responders. Serum and urine samples were collected at 0, 12, 24, and 48 weeks after induction therapy. Twelve serum and urine biomarkers were measured by the multiplex immunofluorescence assay.ResultsUrine (VDBP, MCP-1, IL-6, and IP-10) and serum (IP-10 and IL-23) levels of 12 biomarkers sampled one year after treatment differed significantly between responders and non-responders. Compared with the baseline, their levels in one year were significantly higher in non-responders than in responders. Urine VDBP was significantly correlated with proteinuria (rho=0.62, p< 0.0001), creatinine (rho=0.37, p=0.0025), and renal activity index (rho=0.35, p=0.0042).The change in urine IL-6 and IL-23 levels during three months after induction treatment could predict the treatment response in lupus nephritis with an AUC of 0.70 (p=0.025) and 0.71 (p=0.018), respectively. A model incorporating these two predictors into complement C3 and C4, which are significant clinical factors for treatment response, showed increased predictive value with an AUC of 0.78.ConclusionsUrine VDBP, MCP-1, IL-6, IP-10, and serum IP-10 and IL-23 after one year of treatment differed significantly between responders and non-responders. Moreover, our predictive model composed of urine IL-6, 23, and complement showed increased discriminative ability between responders and non-responders in patients with LN.