GLOMERULAR TRANSCRIPTOMICS IN IGA NEPHROPATHY DIFFERENTIATES BETWEEN DISEASE PROGRESSION AND STABILITY IN LOW-RISK PATIENTS AFTER PROLONGED FOLLOW-UP

Abstract BACKGROUND AND AIMS IgA nephropathy (IgAN) is the most common primary glomerulonephritis worldwide. We have previously shown that patients with assumed benign IgAN can develop progressive kidney failure, including end-stage kidney disease, after a sufficiently long follow-up period [1]. In the reported patient cohort [1], histological, clinical and laboratory findings at the time of diagnosis were unable to predict a subsequent stable or progressive disease course. Thus, we hypothesized that glomerular transcriptomics from the diagnostic kidney biopsy could help make this distinction. METHOD We included all progressive patients (n = 27) and patients with the stable or remitting disease (non-progressors, n = 42) from our previously reported cohort of adult patients with biopsy-proven and assumed benign IgAN (n = 192). Progression was defined as a ≥ 50% decline in eGFR from the diagnostic kidney biopsy, performed between 1988 and 1999 until follow-up examination [1]. The median follow-up time was 22 years. Glomerular cross-sections were obtained through laser-capture microdissection from archival kidney biopsy sections for RNA extraction and sequencing, using NovaSeq 6000 (Illumina, USA) at Functional Genomics Centre Zurich, Switzerland. Samples yielding insufficient sequencing quality were excluded, leaving n = 8 progressors and n = 9 non-progressors for analysis, using limma [2] and edgeR [3] in R Bioconductor. RESULTS In the first round of analysis, we identified 1818 differentially expressed genes (P ≤ 0.05, absolute fold change ≥ 2), of which 1562 genes were overrepresented in progressors and 256 genes were overrepresented in non-progressors. Principal component analysis and hierarchical cluster analysis revealed a separation between the two study groups, indicating that underlying transcriptomic differences are present many years prior to the overt manifestation of disease progression. Interestingly, in progressors, the nuclear factor-kappa B complex, linked to IgAN pathogenesis [4], was the most overabundant transcription factor and Fc Fragment of IgA Receptor (FCAR) was the most overrepresented differentially abundant mRNA. Ingenuity Pathway Analysis (Qiagen, USA) suggested an overrepresentation of the phagosome formation pathway (P = 3.09E-08, Z-score = 8.43), indicating phagosome activation in progressors, possibly a response to the observed FCAR overexpression. To identify potential markers for disease progression, we used an unsupervised K Nearest Neighbour analysis of 1818 differentially expressed genes, allowing combinations of ≤ 10 genes. A two-component classifier (APOL5 and ZXDC) performed best, classifying 15/17 samples correctly (sensitivity 75%, specificity 100%, accuracy 88.24%) on average 21 years prior to a manifested decrease in eGFR. Further refinements of the statistical analyses and confirmation studies are planned to substantiate our initial findings. CONCLUSION Glomerular mRNA sequencing performed at the time of diagnosis of assumed benign IgAN can differentiate between subsequent stable and progressive disease courses in the distant future. In our cohort, combining APOL5 and ZXDC can predict subsequent disease course with 88.24% accuracy already 21 years prior to the discovery of progression with conventional means.