Rituximab is an important second line therapy in difficult nephrotic syndrome (NS), especially given toxicity of long-term glucocorticoid or calcineurin inhibitor (CNI) use. However, clinical response to rituximab is heterogenous. We hypothesized that this was underpinned by immunological differences amongst patients with NS. We recruited a cohort of 18 subjects with glucocorticoid-dependent or glucocorticoid-resistant childhood-onset minimal change NS who received rituximab either due to CNI nephrotoxicity, or due to persistent glucocorticoid toxicity with inadequate response to cyclophosphamide or CNIs. Immunological subsets, T-cell activation assays and plasma cytokines were measured at baseline and 6-months post-rituximab. Time to relapse was bifurcated: 56% relapsed within one year ("early relapse"), while the other 44% entered remission mainly lasting ≥3 years ("sustained remission"). At baseline, early relapse compared to sustained remission group had lower regulatory T-cells (Tregs) [2.94 (2.25, 3.33)% vs 6.48 (5.08, 7.24)%, P<0.001], PMA-stimulated IL-2 [0.03 (0, 1.85)% vs 4.78 (0.90, 9.18)%, P=0.014] and IFNγ [2.22 (0.18, 6.89)% vs 9.47 (2.72, 17.0)%, P=0.035] levels. Lower baseline Treg strongly predicted early relapse (ROC-AUC 0.99, 95% CI 0.97-1.00, P<0.001). There were no differences in baseline plasma cytokine levels. Following rituximab, there was significant downregulation of Th2 cytokines in sustained remission group (P=0.038). In particular, IL-13 showed a significant decrease in sustained remission group [-0.56 (-0.64, -0.35)pg/ml, P=0.007)], but not in the early relapse group. In conclusion, early relapse following rituximab is associated with baseline reductions in Treg and T-cell hyporesponsiveness, which suggest chronic T-cell activation and may be useful predictive biomarkers. Sustained remission, on the other hand, is associated with downregulation of Th2 cytokines following rituximab.