Both protease- and reactive oxygen species (ROS)-mediated proteolysis are thought to be key effectors of tissue remodeling. We have previously shown that comparison of amino acid composition can predict the differential susceptibilities of proteins to photo-oxidation. However, predicting protein susceptibility to endogenous proteases remains challenging. Here, we aim to develop bioinformatics tools to (i) predict cleavage site locations (and hence putative protein susceptibilities) and (ii) compare the predicted vulnerabilities of skin proteins to protease- and ROS-mediated proteolysis. The first goal of this study was to experimentally evaluate the ability of existing protease cleavage site prediction models (PROSPER and DeepCleave) to identify experimentally determined MMP9 cleavage sites in two purified proteins and in a complex human dermal fibroblast-derived extracellular matrix (ECM) proteome. We subsequently developed deep bidirectional recurrent neural network (BRNN) models to predict cleavage sites for 14 tissue proteases. The predictions of the new models were tested against experimental datasets and combined with amino acid composition analysis (to predict ultraviolet radiation (UVR)/ROS susceptibility) in a new web app: the Manchester proteome susceptibility calculator (MPSC). The BRNN models performed better in predicting cleavage sites in native dermal ECM proteins than existing models (DeepCleave and PROSPER), and application of MPSC to the skin proteome suggests that: compared with the elastic fiber network, fibrillar collagens may be susceptible primarily to protease-mediated proteolysis. We also identify additional putative targets of oxidative damage (dermatopontin, fibulins and defensins) and protease action (laminins and nidogen). MPSC has the potential to identify potential targets of proteolysis in disparate tissues and disease states.

Proteases and protease inhibitors (P/PIs) are involved in many biological processes in human skin, yet often only specific families or related groups of P/PIs are investigated. Proteomics approaches, such as mass spectrometry, can define proteome signatures (including P/PIs) in tissues; however, they struggle to detect low-abundance proteins. To overcome these issues, we aimed to produce a comprehensive proteome of all P/PIs present in normal and diseased human skin, in vivo, by carrying out a modified systematic review using a list of P/PIs from MEROPS and combining this with key search terms in Web of Science. Resulting articles were manually reviewed against inclusion/exclusion criteria and a dataset constructed. This study identified 111 proteases and 77 protease inhibitors in human skin, comprising the serine, metallo-, cysteine and aspartic acid catalytic families of proteases. P/PIs showing no evidence of catalytic activity or protease inhibition, were designated non-peptidase homologs (NPH), and no reported protease inhibitory activity (NRPIA), respectively. MMP9 and TIMP1 were the most frequently published P/PIs and were reported in normal skin and most skin disease groups. Normal skin and diseased skin showed significant overlap with respect to P/PI profile; however, MMP23 was identified in several skin disease groups, but was absent in normal skin. The catalytic profile of P/PIs in wounds, scars and solar elastosis was distinct from normal skin, suggesting that a different group of P/PIs is responsible for disease progression. In conclusion, this study uses a novel approach to provide a comprehensive inventory of P/PIs in normal and diseased human skin reported in our database. The database may be used to determine either which P/PIs are present in specific diseases or which diseases individual P/PIs may influence.