Transcriptomic evidence of an IFN-gamma-associated inflammatory state in active visceral leishmaniasis: an exploratory comparison with pediatric HLH

Background Visceral leishmaniasis (VL) can progress to secondary hemophagocytic lymphohistiocytosis (HLH), but transcriptomic similarities between active VL and pediatric HLH remain uncertain. Methods We integrated VL whole-blood expression data (GSE125992), sorted CD68-positive monocyte bulk-expression data from pediatric HLH and severe sepsis (GSE150707), and supportive Leishmania-infected mouse spleen single-cell data (GSE240385). Differential expression, pathway enrichment, an exploratory 132-gene HLH-versus-sepsis-derived ssGSEA score, FARDEEP-LM22 deconvolution, signature-excluded WGCNA, CellChat, and a refitted three-gene classifier were evaluated. Results Active VL showed IFN-gamma-associated expression changes, including higher CXCL9 and CXCL10. The exploratory 132-gene score was higher in ActiveCase than EHC samples (Mann-Whitney P = 0.000637). Across 22 deconvolved cell types, only M1 macrophages showed an FDR-supported positive correlation with this score (rho = 0.787, BH q = 0.000503). After all 132 score genes were excluded before network construction, a 124-gene magenta module remained associated with the score (rho = 0.837, BH q = 4.29 × 10^-5). A refitted IFNG/GBP5/MX2 classifier achieved a same-study validation AUC of 0.917 (bootstrap 95% CI 0.733-1.000), but the validation set contained only 16 samples. Conclusion These public datasets support an IFN-gamma-associated inflammatory state in active VL and exploratory cross-compartment convergence with pediatric HLH. The 132-gene score, deconvolution, network, and classifier results are hypothesis-generating and do not establish HLH specificity, cellular causality, independent external validity, or clinical HLH prediction.