Sexual dimorphism in keratoconus: transcriptomic and hormonal mechanisms underlying stromal remodelling

Keratoconus (KC) is a vision-threatening condition with a higher prevalence and earlier onset in males than in females. The study aims to investigate sex-associated transcriptomic features of KC and assess whether sex hormones modulate stromal remodelling.

We performed whole-transcriptome sequencing of human corneal tissues from patients with KC and matched controls (n = 20; five males and five females per group), followed by weighted gene co-expression network analysis, Gene Ontology, Kyoto Encyclopaedia of Genes and Genomes enrichment, and competitive endogenous RNA (ceRNA) network construction. Our findings were functionally probed in primary human corneal stromal fibroblasts (HCSFs) using testosterone, β-oestradiol, and their antagonists. Key nodes were validated by reverse transcription-quantitative polymerase chain reaction analysis, Western blotting, and immunofluorescence.

Across 17,496 genes, we identified 3,345 differentially expressed genes (adjusted P < 0.01, |log₂ fold-change|≥ 2). Module analyses highlighted pathways related to intracellular transport, energy metabolism, and hormone responses. Sex-stratified analyses revealed male-specific up-regulation of gonadal development programs and female-specific down-regulation of immune and hormonal processes. In HCSFs, testosterone down-regulated type I collagen but up-regulated type III collagen and α-smooth muscle actin; these effects were mitigated by flutamide. Conversely, oestrogen inhibition reproduced altered stromal remodelling, which was rapidly reversed by β-oestradiol. A female-specific ceRNA axis (circEPB41L2_0001-miR-942-5p-DCP1A) was identified and validated by performing perturbation experiments.

KC exhibited sex-biased molecular programs consistent with androgen-driven and oestrogen-deficiency-related stromal remodelling. These findings elucidate hormone-driven mechanisms underlying KC and suggest that sex-associated hormonal regulation could inform future personalised therapeutic strategies.