Pre-transplant microRNA serum profiles and association with acute graft-versus-host disease in allogeneic hematopoietic stem cell transplantation.

Background: Allogeneic hematopoietic stem cell transplantation (allo-HSCT) remains associated with significant morbidity and mortality, with acute graft-versus-host disease (aGVHD) being one of the leading contributors. Early prediction of aGVHD is essential for timely intervention, however, current diagnostic tools only detect disease after tissue damage has occurred. Circulating microRNAs (miRNAs) are potential non-invasive biomarkers for early detection or prediction of aGVHD. Objective: The objective of this study was to evaluate whether serum miRNA expression profiles obtained prior to transplantation can serve as a predictive tool for aGVHD in allo-HSCT recipients, with a focus on identifying signatures that support early detection and targeted management of post-transplant complications. Methods: We investigated the association with serum miRNA profiles and the development of early complications in 78 allo-HSCT recipients. Serum samples were collected prior to transplantation, and miRNA expression was quantified using next-generation sequencing. Statistical and bioinformatic analyses were applied to identify miRNAs associated with aGVHD (grade II-IV). Associations with the Endothelial Activation and Stress Index (EASIX) were also explored. Results: Among 78 allo-HSCT recipients, 18 patients (23%) developed aGVHD grade II-IV. Early mortality occurred in nine patients within four months post-transplantation, and seven patients relapsed within one year post-transplantation. Receiver operating characteristic (ROC) analysis identified eight pretransplant miRNAs (miR-664a-5p, miR-20b-5p, miR-93-5p, miR-25-3p, miR-1224-5p, miR-106b-5p, miR-454-3p, and miR-3679-5p) associated with subsequent aGVHD development. A predictive model based on these miRNAs achieved an AUC of 0.855, which decreased to 0.692 after bootstrap validation. Hierarchical clustering using these miRNAs separated patients into two distinct clusters with markedly different aGVHD risks: cluster 1 showed a significantly higher incidence (42% vs. 9%) and greater severity, including all grade IV cases and most miRNAs were upregulated in patients developing aGVHD. Additionally, eight other miRNAs correlated with the EASIX score, reflecting endothelial stress, although these did not overlap with aGVHD-associated miRNAs. Conclusion: Pre-transplant serum miRNA signatures can predict aGVHD risk and severity, offering a novel approach for early patient stratification. These findings support the integration of miRNA profiling into pre-transplant assessment to guide personalized GVHD prophylaxis and monitoring. Further validation in independent cohorts is warranted.