Subtype-dependent LOX-family transcript changes in aging murine thymic stroma

This preprint reports a computational reanalysis of public thymic stromal transcriptomic datasets focused on age-associated changes in the lysyl oxidase (LOX) gene family. Biological-sample-aware pseudobulk analysis of 22,932 CD45-negative stromal cells from the GSE240016 murine thymus aging dataset identified significant, subtype-dependent LOX-family transcript changes rather than a uniform fibroblast-wide response. Capsular fibroblasts showed lower Lox in aged samples (log2FC = -1.46, adjusted p = 7.5 × 10^-5). Medullary fibroblasts showed divergent family-member behavior, with higher Loxl1 (log2FC = +0.75, adjusted p = 0.023) and lower Loxl2 (log2FC = -1.00, adjusted p = 0.0067). In annotated mTEC1 cells, Loxl2 showed a marked aged-lower pseudobulk estimate (log2FC = -3.29, adjusted p = 9.6 × 10^-4). All four focal transcript directions retained the same sign in stricter marker-positive subsets. Sample-level comparisons and depth-matched downsampling preserved the young-higher/aged-lower mTEC1 Loxl2 direction, although sparse detection, lower aged sequencing depth, and the n=2 young versus n=2 aged design keep dropout/depth contribution partially unresolved. Independent datasets added external transcript-level context. GSE223049 showed directionally similar aged-lower broad thymic fibroblast Lox and Loxl2 and broad epithelial Loxl2. E-MTAB-8560 provided independent mTEC-focused context, with the clearest aged-lower Loxl2 patterns in mTEClo and cTEC and weaker support in mTEChi. A controlled GSE231906 pilot successfully linked human thymus expression data to metadata, recovered all five LOX-family genes and TEC/mTEC-like compartments, and made donor/sample-aware aged-human transcript context technically evaluable; the LOXL2 mTEC-like age pattern was mixed. Together, the analyses identify subtype-dependent LOX-family transcript remodeling as a reproducible feature of murine thymic stromal aging and nominate Loxl2 as a recurrent candidate transcript across fibroblast and epithelial compartments. Direct LOXL2 protein or spatial testing by IHC/IF, RNA-ISH, or related assays remains the key experimental next step.