Identification of inflammatory biomarkers in IgA nephropathy using the NanoString technology: a validation study in Caucasians

All patients with biopsy-proven IgAN and antineutrophil cytoplasmic antibody (ANCA)-associated pauci-immune glomerulonephritis between 2005 and 2021 diagnosed at Maisonneuve-Rosemont Hospital, a single tertiary care hospital in Montreal, were considered for this retrospective study. Only patients with formalin-fixed paraffin-embedded (FFPE) biopsy samples available for gene expression analysis were included. Patients with known secondary causes of IgAN were excluded. For the immunofluorescence experiments, controls with no glomerular disease (after pathology review) were recruited among patients who underwent a nephrectomy in the context of a kidney lesion. All subjects provided written informed consent. This study was approved by our institution’s Research Ethics Committee (No MP-12-2014-525) in agreement with the Declaration of Helsinki.

Clinical data were collected by reviewing medical records, including demographics, kidney presentation, comorbidities, history of familial kidney disease, and medications. Specifically, serum creatinine, eGFR, and proteinuria were noted at the time of the biopsy, at 6 months from biopsy time and then annually to the last available follow-up visit and/or initiation of renal replacement therapy. The Oxford MEST-C score [6] was retrieved from the biopsy report. Baseline eGFR was estimated using the Chronic Kidney Disease Epidemiology Collaboration equation [7]. Urinary protein excretion was estimated using either total 24-h proteinuria (in g/day), spot urinary protein–creatinine ratio (in mg/mmol), or spot urinary albumin–creatinine ratio (in mg/mmol). For IgAN, the risk of a 50% decline in eGFR or progression to end-stage kidney disease (ESKD) 5 years after kidney biopsy was calculated using the New International Risk-Prediction Tool in IgAN [8], a prediction model that includes baseline clinical and pathological characteristics, demographics, and medication use. The studied primary outcomes were ESKD (eGFR < 15 ml/min/1.73m², dialysis or kidney transplantation) or death.

For our NanoString analysis, a total of 30 biopsy-confirmed IgA and 7 ANCA-positive pauci-immune glomerulonephritis patients were used. Total RNA from FFPE kidney biopsy specimens from patients with IgAN and ANCA-positive pauci-immune glomerulonephritis (control group) were extracted and purified using RNeasy FFPE kit (Qiagen) according to the manufacturer’s protocol. A total of 100 nanograms of RNA from each sample was hybridized to a customized 48-plex human gene CodeSet and processed on the nCounter platform. Purified RNA samples were analyzed using the nCounter gene expression system from NanoString Technology. A total of 29 genes implicated in different biological pathways were studied; those genes have been selected for their role in complement activation (C2, C3, C1QA, C1QB), apoptosis (BAX, BAD, Caspase 3), cell activation, and/or leucocytes stimulation (AGTR1, AGTR2, CXCL1, CD89S, ICAM1, MIF, PTEN, CD71, TGM2, TNFR1B, TNFRSF1B, TNFSF13, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-13, IFN-γ, TGF-β1, TNF). Housekeeping gene (ACTB, GAPDH, HPRT1, LDHA) normalization was performed to adjust counts of all probes relative to a probe that is not expected to vary between samples; the factor should be in the range between 0.1 and 10. All normalization procedures were performed according to the manufacturer’s protocol.

Data collection for gene expression was carried out in the nCounter Digital Analyzer. At the highest standard resolution, 555–1155 fields of view were collected per flow cell using a microscope objective and a CCD camera, yielding data of hundreds of thousands of target molecule counts. Every RNA target is identified by the color code generated by the ordered fluorescent segments present on the reporter probe. The expression level of each gene is determined by scoring the number of times its corresponding color code is detected. Data were imported into nSolver Analysis Software for downstream analysis.

Comparisons in gene expression were made between IgAN patients and those with another inflammatory glomerulonephritis (ANCA-associated pauci-immune glomerulonephritis) to depict an expression profile specific to IgAN. Subsequently, among individuals with IgAN, gene expression was compared between patients with a low chance of progression to ESKD (< 10%) and those at mild to moderate risk of progression (10–20%) with relatively minimal fibrosis on kidney biopsy. Genes for which RNA expression was associated with disease progression were analyzed in 6 patients for protein expression by immunofluorescence and compared with controls.

The same nephropathologist reviewed histologic slides for all patients with IgAN to provide MEST-C score: mesangial hypercellularity (with M0 and M1 corresponding to ≤ 50% and  > 50% of glomeruli with hypercellularity, respectively), segmental glomerulosclerosis (with S0 if absent and S1 if present), endocapillary hypercellularity (with E0 if absent and E1 if present), tubular atrophy/interstitial fibrosis (with T0, T1, and T2 corresponding to ≤ 25%, 26%–50%, and > 50% of cortical area involvement, respectively), and cellular/fibro cellular crescents (with C0, C1, and C2 corresponding to their absence, presence in ≥1 and < 25% of glomeruli, and presence in ≥25% of the glomeruli, respectively). ANCA-associated pauci-immune glomerulonephritis was also graded based on its degree of tubular atrophy/interstitial fibrosis.

The kidneys were fixed in 10% formaldehyde, dehydrated, and embedded in raffin. The tissue was sectioned (5 µM) and was subjected to antigen retrieval in citrate solution at pH 6. The sections were blocked with anti-donkey serum 5% and labeled with anti-Complement C3 (Catalog #PA1-29715 from Life Technologies/ThermoFisher) and TNFRSF1B (Catalog #MA5-31661 from Life Technologies/ThermoFisher). The slides were subsequently exposed to donkey anti-rabbit AF647-conjugated (1:400, Catalog # 711 605 152 Jackson ImmunoResearch Laboratories), donkey anti-mouse AF568 (Catalog # A10037 Life Technologies/ThermoFisher), and donkey anti-goat AF488 (Catalog # A11055 Life Technologies/ThermoFisher). Fluoroshield with DAPI (Millipore-Sigma) was used for nuclear staining and mounting. Slides were imaged using a Zeiss AxioObserver.Z1 inverted microscope coupled to an X-Cite 120LED Boost High-Power LED illumination system. Images for quantitative analysis were captured with a 20X objective and the number of positive cells was determined as the average of positive cells in at least 8 fields per kidney section.

Baseline characteristics, clinical and pathological parameters, and gene expression were compared between patients with IgAN and those with ANCA-associated pauci-immune glomerulonephritis (controls). Among patients with IgAN, gene expression was compared between progression score groups (< 10%, 10–20%, 21–40%, and > 40%), and more specifically between patients with scores < 10% and those with 10–20% to delineate markers of early disease. Wilcoxon or t-tests for continuous variables and the χ² test or the Fisher exact tests were used for categorical variables. The Kruskal–Wallis test was also used for comparison between more than two groups. Differences were considered statistically significant if p-value < 0.05. Statistical analysis was performed by N.E. using SAS 9.4 (SAS Institute, Cary, NC).

Patients with IgAN with gene expression analyzed by NanoString were mostly Caucasians (76.7%) and females (73.3%), as shown in Table 1. Mean age at biopsy was 44.7 ± 15.3 years. Mean baseline eGFR and proteinuria were 60.9 ± 35.8 ml/min/1.73m², and 3.2 ± 2.9 g/d respectively; most patients (96.5%) had a stage 0 or 1 of fibrosis at diagnosis. At baseline, 60% of patients had chronic kidney disease stage 3 or 4. Treatments were highly variable, but a significant proportion (80%) were on renin–angiotensin–aldosterone (RAA) system inhibitors, and 45% received glucocorticoids at some point in their disease course. Patients with ANCA-associated pauci-immune glomerulonephritis were all Caucasians (100%), with a majority of females (57.1%). Mean age at biopsy was 62.6 ± 9.4 years. At baseline, mean eGFR and proteinuria were 36.6 ± 33 ml/min/1.73m² and 2.4 ± 2.2 g/d, respectively; all patients had stage 0 or 1 of fibrosis at diagnosis. Patients with ANCA-associated pauci-immune glomerulonephritis were significantly older and had more kidney impairment at baseline than patients with IgAN.

The 5-year risk of a 50% decline in eGFR or ESKD after biopsy was variable in IgAN patients: 9, 5, 8, and 6 patients (2 missing) had a 5-year risk of progression of  < 10%, 10%–20%, 21%–40%, and > 40%, respectively. The median follow-up time for the IgAN group was 4.62 years (interquartile range 2.45 years). Among patients with IgAN, 8 patients (26.7%) reached ESKD and 3 (10%) died.

IgAN patients had a different mRNA expression profile than patients with ANCA-associated pauci-immune glomerulonephritis. IL-6, INFG, and C1QB were significantly transcribed at higher levels in patients with ANCA-associated pauci-immune glomerulonephritis (Table 2). AGTR1 was more expressed in patients with IgAN compared to ANCA controls (p = 0.05).

The distribution of biomarkers in function of the IgA nephropathy prediction tool score range was analyzed. Markers known as CIQA, ICAM1, TNF, and TNFRSF1B did not show a statistically significant distribution, but their distribution difference was near the significance level (Table 3). As shown in Table 4, CIQB and TNFRSF1B expressions were higher among patients with an IgA prediction tool score range of 10–20% compared to a score < 10% (p-value = 0.01342 and p = 0.03337, respectively). C3 expression was also higher in patients with a score range of 10–20% (vs. with a score < 10%) with a p-value of 0.04911.

To ensure the accuracy and reliability of our findings obtained through NanoString mRNA profiling, we devised a validation strategy that involved the examination of protein expression levels of candidate genes. To conduct our protein expression analyses, we selected Complement C3 as a marker of mild to moderate risk of ESKD progression with relatively minimal fibrosis on kidney biopsy, and TNFRSF1B to serve as a marker expressed during stage 2 fibrosis. To this end, we obtained biopsies from patients diagnosed with IgA nephropathy that had high degree of fibrosis compared to the controls (Fig. 1A). Our findings revealed that both target genes identified in gene expression analysis (C3 and TNFRSF1B) exhibited a significant upregulation of protein expression in the biopsies obtained from patients with IgA nephropathy when compared to control tissue (Fig. 1B). Importantly, the expression patterns of Complement C3 and TNFRSF1B were distinct, further underscoring the significance of our observations. Complement C3 was found to be distributed around the tubular epithelial cells and glomerulus, with a modest signal, suggesting a possible role in the regulation of inflammation and immune responses in such regions. On the other hand, the signal for TNFRSF1B was more pronounced and concentrated in distinct regions that differed from those observed for Complement C3, highlighting its potential role in cellular proliferation and differentiation. Overall, our findings validate our NanoString mRNA profiling by examining protein expression levels of two candidate genes, C3 and TNFRSF1B, in IgAN patients and controls.

Fibroblasts are responsible for producing and depositing extracellular matrix components and may be stimulated by persistent inflammation and immunological responses, such as those during IgAN. The excessive production and accumulation of collagen and other matrix proteins lead to scar tissue formation, known as fibrosis. Therefore, we analyzed our IgAN patient samples for fibrosis formation and compared the severity of fibrosis with the NanoString mRNA transcript expression of our targets. From our histological slides, we observed fibrosis formation in IgAN patients that resulted in structural changes within the kidney, including thickening of the glomerular basement membrane, obliteration of the glomerular capillaries, and an increase in the volume of the interstitial compartment (Fig. 1A). From our NanoString mRNA expression profiles, we observed a significant increase in PTEN, CASPASE 3, TGM2, TGFB1, IL2, and TNFRSF1B from IgAN patients with level 2 fibrosis compared to those with none (Table 5). To gain insights into such significantly upregulated mRNA from IgA fibrotic patients, we compiled them into the STRING platform (https://​string-db.​org/​) to analyze any potential biological processes and interactions in the form of an interaction network. Set at a high confidence of 0.7, our identified “fibrotic” genes did not show any strong correlation among themselves. However, when adding 5 more relevant nodes to the network, IL2RB, SLC9A3R1, TP53, TGFBR2, and XIAP, we were able to establish a more complete network of associating genes, with a total of 11 nodes and 14 edges (Fig. 2). The protein–protein interaction networks (PPI) enrichment p-value of 0.0393 indicates that the nodes are not random and that the observed number of edges is significant. When studying the various descriptions in the “Biology Process” category, we identified that such genes are primarily involved in apoptotic pathways, regulation of apoptosis, wound healing, response to growth factors, and regulation of lymphocyte activation, while the meaningful “Pathways” included Th17 cell differentiation, cellular senescence, MAPK signaling pathway, and PI3K-Akt signaling pathway. Therefore, our study identified several upregulated mRNA transcripts that were upregulated during the development of fibrosis and may be considered fibrotic markers within IgAN patients.