Anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) encompasses a heterogeneous disease group characterized by autoimmune necrotizing small vessel vasculitis. A severe type of renal involvement is common in AAV, known as relapsing pauci-immune crescentic and necrotizing glomerulonephritis (AAV-GN). AAV-GN is an important determinant of disease-specific morbidity and mortality. While detectable ANCAs provide some correlation with ongoing renal involvement, the ‘ANCA-centric’ approach has limitations, including the presence of ANCAs in non-AAV diseases and the fact that rising titers do not always relate to AAV-GN activity/flare (1). An accurate noninvasive biomarker reflecting active AAV-GN thus remains elusive. Consequently, a renal biopsy, an invasive procedure with inherent risks, remains the diagnostic and prognostic gold standard in cases clinically suspected of AAV-GN.

Due to observed phenotypic and genotypic differences, AAV is best classified by the antigen specificity of ANCAs. The two primary antigens of pathogenic ANCAs are myeloperoxidase (MPO) and proteinase-3 (PR3). Notably, MPO-AAV and PR3-AAV are increasingly considered as entities with distinct pathogeneses, clinical features and management strategies (2, 3).

Altered macrophage polarization (4) and differentiation of CD4+ T-cells towards Th17 subsets (5) are two recognized biological processes implicated in the pathogenesis of AAV; however, the precise mechanisms driving their activation remain poorly understood. Epigenetic regulation through microRNAs (miRNAs) has been implicated in both macrophage polarization (6) and Th17 differentiation (7), but not in the specific context of AAV-GN. While miRNAs have been proposed as candidate biomarkers in various autoimmune diseases (8), validated data in AAV-GN are lacking. This study aimed to validate a renal miRNA expression profile identified as specific to AAV-GN (9) and investigate its potential as a circulating biomarker for active AAV-GN. Additionally, it aimed to identify miRNA expression differences between MPO- and PR3-positive AAG-GN cases and correlate the molecular data with clinical and histologic markers of disease activity to improve diagnostic accuracy.

A schematic workflow of this research is depicted in Figure 1 . Briefly, we identified 17 candidate miRNAs that potentially distinguish AAV-GN as a group from non-AAV GN controls, as well as from individuals without renal disease (living donors) in the previous study (9). In this study, we performed a similar analysis of screening data and identified 13 candidate miRNAs that could distinguish MPO- from PR3-positive AAV-GN. The 21 out of 30 candidate miRNAs selected for validation in renal tissue samples and their specific inclusion criteria are presented in Supplementary Table S1 .

Our study validated significant renal tissue expression differences in 11 out of 21 candidate miRNAs selected through screening. This suggests that miRNA expression profiling in renal tissue can effectively distinguish AAV-GN from various non-AAV RD, including cases with detectable ANCA without clinical and histologic evidence of AAV-GN. Although tissue-based analysis provides valuable insights, it requires a renal biopsy to obtain sample material, which limits its applicability. Therefore, we also analyzed the expression of validated miRNAs in serum samples. Despite the inherent limitations of biofluid-based miRNA analysis (14), five validated miRNAs distinguished AAV-GN from non-AAV RD using matched serum samples. Notably, three of these five miRNAs also differentiated AAV-GN from both non-AAV RD and HC, consistent with a circulating miRNA expression profile specific to AAV. Additionally, miR-181a-5p also stratified AAV-GN into MPO- and PR3-positive.

Noteworthy observations apply to ‘bystander ANCA’ cases, ie. patients who underwent kidney biopsy, typically due to asymptomatic urinary abnormalities, and positive MPO- or PR3-ANCA but no histologic evidence of AAV-GN. Clinicians often struggle to distinguish between harmless (bystander) and pathogenic, disease-driving ANCAs in ANCA-positive patients. A reliable biomarker would be invaluable for guiding clinical decision-making in such cases. Comparative miRNA expression profiling in the renal tissue and serum revealed that bystander ANCA cases clustered separately from AAV-GN and non-AAV RD. While overall more aligned with non-AAV RD, bystander ANCA samples exhibited miR-142-5p, miR-150-5p, and miR-181a-5p expression patterns indistinguishable from AAV, suggesting a partial AAV-like epigenetic profile. This finding bears potential clinical implications as it raises questions about the ‘bystander’ nature of ANCAs in cases lacking clinical or histologic evidence of AAV-GN. Notably, miR-21-3p expression effectively differentiated AAV-GN from bystander ANCA cases in renal tissue or serum samples, highlighting its potential as a key biomarker in resolving this diagnostic uncertainty.

To refine the validated miRNA signature, the molecular data was correlated with established scoring systems for renal involvement in AAV-GN In the renal tissue, miR-142-5p, miR-150-5p and miR-181d-5p correlated with the AKRiS, while miR-150-5p and miR-181d-5p associated with the Berden classes. Accordingly, renal tissue expressions of miR-142-5p and miR-150-5p also correlated with %NG, supporting the importance of %NG in estimating risk in AAV-GN and corroborates the need for early diagnosis of kidney involvement in AAV (15–17).

Furthermore, the miRNA signature was analyzed in the context of clinical variables of disease activity. Notably, renal tissue expression of miR-21-3p and miR-142-5p differentiated AAV-GN cases presenting as RPGN, while the expression of miR-30b-5p in the renal tissue and the expression of miR-181b-5p in the serum associated with the increasing number of affected organs. Accordingly, BVAS correlated with the expressions of miR-30b-5p and miR-181b-5p on renal tissue and serum samples, respectively. Additionally, the expression of miR-30d-5p in the serum differentiated cases presenting with pulmonary-renal syndrome. Interestingly, no significant miRNA expression differences in the renal tissue or serum were identified comparing samples of AAV-GN labelled as RLV to those with multiorgan involvement.

Importantly, we could not directly translate any individual finding observed between variables of disease activity and validated miRNAs in the renal tissue to serum samples. This may be due to inadequate size of technically adequate serum samples, a lack of expression correlation between matched renal tissue and serum samples, or to expression levels of certain miRNAs in the serum falling below the detection threshold. Furthermore, some miRNAs may preferentially be expressed and thus detectable only in renal tissue. Conversely, aggregated miRNA expression in serum may also reflect influences from related or unrelated conditions in other organs or homeostatic mechanisms.

We would like to highlight miR-21-3p, miR-181a-5p and miR-181d-5p as the three most promising serum-expressed miRNAs. These miRNAs effectively differentiated AAV-GN from both non-AAV RD and HC cases. Moreover, the expression of miR-21-3p and miR-181d-5p differentiated the presence of RPGN and correlated with AKRiS in renal tissue samples, respectively, while miR-181a-5p differentiated MPO- from PR3-positive AAV-GN in renal tissue and serum samples. Furthermore, miR-21-3p distinguished AAV-GN from bystander ANCA cases in renal tissue and serum samples.

Importantly, miR-21-3p, miR-181a-5p and miR-181d-5p were the only validated miRNAs showing significant expression correlation between matched renal tissue and serum samples, underscoring their potential as robust noninvasive biomarkers for active AAV-GN. Additionally, ROC curve analyses demonstrated that combining serum levels of miR-21-3p and miR-181a-5p yielded a strong ability to differentiate AAV-GN from other groups. Beyond statistical analyses, functional annotations of miR-21-3p, miR-181a-5p and miR-181d-5p to biological processes implicated in AAV pathogenesis support their relevance. Namely, miR-21-3p and miR-181a-5p are implicated in altered macrophage polarization (6, 18). Indeed, alternatively activated (i.e., M2-polarized) macrophages are considered the dominant macrophage phenotype in AAV and are associated with a higher end-stage renal disease risk in AAV (4, 19).

miR-181a-5p also regulates the processes of B- and T-cell differentiation, namely the differentiation towards pathogenic autoantigen specific Th17 subsets in CD4+ T-cells and is involved in T-cell tolerance and selection stringency (7, 20). These processes are implicated in several autoimmune diseases, including AAV (5, 21, 22). Similarly, miR-21-3p overexpression promotes Th17 differentiation and activates the IL-23/IL-17 axis through MAP3K14, exacerbating inflammation in a psoriasis model (23). Elevated IL-23 and IL-17 levels were also observed in AAV (5). Interestingly, increased IL-17 levels in the context of Th17-predominant inflammation skew macrophage polarization toward the M2 phenotype (24). These insights not only align with established aspects of AAV pathogenesis but also suggest potential therapeutic implications, either by miRNA ago-/antagomirs and sponges, or by targeting implicated biologic processes with existent modalities, such as anti-IL23/IL-17 agents or abatacept (CTLA4-Ig).

It would be valuable to explore the exact function and target genes of miR-21-3p, miR-181a-5p and miR-181d-5p as the three most promising miRNAs. This can be achieved by experimentally validating the target genes of these miRNAs in appropriate cell line models. A commonly employed method involves cloning the 3’-untranslated region (3’-UTR) of the predicted miRNA target gene into a luciferase reporter construct and analyzing the change in luciferase expression as evidence of binding and regulatory activity of a particular miRNA. One such approach involves transfecting a cell line that expresses the miRNA of interest with a reporter containing either the wild-type UTR of a target gene or the UTR in which the miRNA binding site is mutated. If the construct containing the wild-type UTR shows a reduction in luciferase expression and this decrease is absent in the mutated version, this likely indicates that the endogenously expressed miRNA is capable of regulating that UTR. In this same setup, a miRNA inhibitor to block miRNA function could be used, which should lead to an increase in luciferase. Another approach is to transfect cells with both the luciferase reporter construct and increase amounts of the miRNA of interest. The advantage of introducing a miRNA is that if the miRNA really targets the binding site in the UTR, a dose-dependent effect on luciferase readout can be determined, with this response being absent with the mutated construct (25).

Several important study limitations should be acknowledged. The inherent variability of biofluid-based miRNA analysis, influenced by low isolation yields, expression variability, and the lack of standardized normalization methods, remains a challenge. This issue could be addressed by adjusting the sample volume or protocol (i.e., employing a more sensitive detection method, such as digital PCR). Additionally, the heterogeneity of renal tissue may lead to the omission or misidentification of compartment- or cell-specific miRNAs, which could be mitigated by spatial transcriptomics. Furthermore, the proposed miRNA signature relates specifically to AAV-GN and our findings may not apply to all AAV patients without renal involvement or ANCA-negative patients. Our study was also limited by small size of technically adequate samples (especially for bystander ANCA group), resulting in limited statistical power, incomplete correlation between renal tissue and serum miRNA expression for some miRNAs, and the exclusion of miR-181d-5p from the ROC curve analysis. Therefore, an independent validation in larger multicenter cohorts, using an alternative platform such as next-generation sequencing to reduce platform-dependent bias, is essential to confirm the proposed AAV-specific miRNA signature.

In conclusion, this study establishes miR-21-3p, miR-181a-5p, and miR-181d-5p as constituents of a circulating miRNA signature related to AAV-GN, providing a potential noninvasive biomarker for active AAV-GN in the appropriate clinical setting. By integrating renal tissue and serum miRNA expression profiles, we demonstrate that these miRNAs might effectively differentiate AAV-GN from non-AAV RD and correlate with disease activity markers. Future research, ideally as a multi-center study to increase sample sizes, should focus on refining presented findings to establish the role of identified miRNAs in the pathogenesis of AAV-GN and support the clinical application of miRNA-based biomarkers in AAV-GN diagnostics and disease monitoring.