Hierarchical clustering uncovered disease patterns and further untangled complexities in immune complex-mediated...

Membranoproliferative glomerulonephritis (MPGN) is currently stratified into complement C3 glomerulopathy (C3G) and immune complex-mediated MPGN (IC-MPGN). However, classification is subject to continued debate.

Here, we applied hierarchical clustering to a much larger cohort of patients with C3G/IC-MPGN (295 individuals), extensively characterized for genetic and autoimmune complement abnormalities, with the goal of unraveling specific disease patterns. We also designed a user-friendly web application that with input of data at diagnosis could make cluster classification clinically applicable.

Five clusters with unique phenotypic and complement profiles were identified. Cluster 1 and 2 patients showed systemic complement activation until C5. Consistently, C5 nephritic factor and anti-factor B antibodies were prevalent in these clusters. Cluster 2 was distinguished from cluster 1 for classical pathway activation markers in biopsy. Cluster 3 showed C3-restricted systemic complement activation associated with the prevalence of C3 nephritic factor. Cluster 4 and 5 patients shared a normal complement profile and intense glomerular C3 staining, consistent with solid-phase complement activation, but cluster 5 distinguished for the higher prevalence of genetic abnormalities. Cluster 4 patients had the highest incidence of kidney failure during follow-up, while cluster 1 had the best kidney prognosis. However, clusters 1 and 2 showed a high risk of post-transplant recurrence. Through our web application, we could visually compare the predicted profile of new patients with those of patients included in clustering analysis and assign these patients to different clusters. The cluster-based classification allows etiologic diagnosis of C3G/IC-MPGN and had better prognostic value than current approaches.

Our proposed strategy may possibly guide anti-complement treatment.