Spatial transcriptomics identifies a translayer architecture of pyroptosis-related transcription in systemic sclerosis skin

ObjectiveTo determine whether pyroptosis-related transcription in systemic sclerosis skin forms a translayer spatial architecture rather than a single coextensive inflammatory program. MethodsWe reanalyzed public Visium formalin-fixed paraffin-embedded skin sections (4 healthy controls, 9 systemic sclerosis) from a discovery cohort and tested prespecified endpoints in 10 independent systemic sclerosis sections. The tissue section was the inferential unit. Epidermal versus dermal contrasts within each section were evaluated for inflammasome-related and gasdermin genes, followed by cell2location spatial deconvolution and partial correlation adjusted for endothelial context in the dermis. ResultsNLRP1, PYCARD, and CASP4 displayed epidermal bias in all 13 discovery sections, whereas GSDMD displayed dermal bias in all 13. This spatial separation was detectable in healthy skin and appeared stronger in systemic sclerosis. A tier 1 triad captured the epidermal signal better than broader composites (dilution 35.5%; P = 0.0002). In an independent systemic sclerosis cohort, the dermal gasdermin endpoint retained its direction in 8 of 10 sections and the epidermal inflammasome-related endpoint in 10 of 10. Spatial deconvolution indicated that dermal GSDMD associated most strongly with estimated endothelial abundance in both healthy and systemic sclerosis skin. The IFN{gamma}-GSDMD association remained positive after endothelial adjustment across sections, compatible with an additional IFN{gamma} component. ConclusionSystemic sclerosis skin harbors a reproducible translayer pyroptosis-related transcriptional architecture in which upstream epidermal inflammasome-related transcription and dermal GSDMD expression are spatially dissociated. This organization, detectable in healthy skin and often stronger in SSc, may warrant future mechanistic and therapeutic interrogation by compartment.