Neutrophilic skin diseases, including Behcet disease, Sweet syndrome, pyoderma gangrenosum (PG), and epidermolysis bullosa acquisita (EBA), are characterized by an exaggerated inflammatory response following mechanical skin stimulation, yet the underlying mechanisms remain unclear. We identify adenosine triphosphate (ATP) released from keratinocytes as a key mediator of this phenomenon, promoting neutrophil extracellular trap (NET) formation. Using an EBA murine model as a model of neutrophilic skin disease, where scratching (a prototypic mechanical stimulation) exacerbates lesional severity, we observed abundant NET deposition in lesional skin. Degradation of these NETs with DNase1 reduced clinical and histopathological severities. In vitro, purified NET components increased IL-8 secretion from keratinocytes and fibroblasts, suggesting that NETs amplify inflammation via a self-amplifying loop of neutrophil recruitment. In the EBA mouse, scratch restriction with neck collars not only attenuated clinical and histological disease severities but also decreased lesional NETosis and neutrophils. Mechanistically, keratinocytes released ATP in response to mechanical stress in vitro, and pharmacologic purinergic blockade in the EBA mice with suramin phenocopied the protective effects of scratch restriction. While ATP alone did not induce NETosis, ATP enhanced complement component 5a (C5a)-induced NET formation in vitro. These findings indicate that keratinocyte-derived ATP, released in response to mechanical stress, contributes to NETosis in a C5a-dependent manner, thereby exaggerating neutrophilic inflammation, leading to blistering and further NETosis. Histopathological analyses of EBA and PG cases also demonstrated NETs accumulation localized to the upper dermis, suggesting a conserved ATP-NET axis. Targeting this pathway may represent a promising therapeutic strategy for neutrophilic skin diseases.
Toward terminological clarity in digital biomarker research
Digital biomarker research has generated thousands of publications demonstrating associations between sensor-derived measures and clinical conditions, yet clinical adoption remains negligible. We identify a foundational




