Proteases and Tissue Remodeling Unit ? Z01DE0699-05? Thomas H. Bugge? ? ? Novel cell surface serine proteases in epidermal development, repair, and malignancy (33% effort) ? A newly defined syndrome, autosomal recessive ichthyosis with hypotrichosis (ARIH), recently was linked to mutation in the ST14 gene leading to a G827R substitution in the protease matriptase. The clinical manifestations of ARIH include tooth defects, corneal opacity with photophobia, epidermal abnormalities, and hypotrichosis. In collaboration with Benjamin Cravatt, The Scripps Institute, and Julia Segre NHGRI, we characterized the activity of G827R matriptase. The mutated protease retained the ability to incorporate fluorophosphonate probes, showing that some serine hydrolytic activity was retained in the mutant protein. However, activity of G827R towards a variety of peptide and natural substrates was undetectable. To further investigate the link between ARIH and reduced matriptase activity, we generated a mouse strain that displays a one-hundred-fold reduction in epidermal matriptase mRNA. Interestingly, the phenotype of these ST14 hypomorphic mice recapitulates key features of human ARIH. These include: a) Abnormal enamel surfaces of incisors and molars. b) Hypotrichosis with absence of erupted whiskers in neonates, sparse fur, and curly and thin whiskers. c) Impaired epidermal barrier function and desquamation, leading to marked retention and hyperproliferative ichthyosis, and outwardly wrinkled and scaly epidermis. d) Impaired proteolytic processing of the downstream matriptase targets, pro-prostasin and profilaggrin. These studies strongly indicate that reduced activity of a matriptase-prostasin proteolytic cascade is the etiological origin of human ARIH and provides an important mouse model for the exploration of matriptase function in ARIH, as well as other physiological and pathological processes. ? The transmembrane Kunitz-type serine protease inhibitor, hepatocyte growth factor activator (HAI)-1 is near-ubiquitously co-expressed with matriptase in both embryonic and adult epithelia, and that matriptase inhibition by HAI-1 is essential for preserving epithelial integrity and to complete mouse embryogenesis. To gain insights into the biochemistry of this critical transmembrane serine protease inhibitor, in a collaboration with Lotte Vogel, University of Copenhagen, we determined the biosynthesis and cellular localization of HAI-1 in polarized epithelial cells in culture. The study revealed that HAI-1 has a complex subcellular itinerary that differs from all serine protease inhibitors studied to date. Immediately after synthesis, HAI-1 is exocytosed to the basolateral plasma membrane, where a small fraction of the inhibitor undergoes ectodomain shedding. Residual HAI-1 subsequently recycles between the endosomal compartment and the basolateral plasma membrane, before the inhibitor is transcytosed to the apical membrane, from which it is finally cleared by endocytosis and lysosomal degradation. This shows that HAI-1 is exposed to many cellular compartments, where the inhibitor may functionally interact with matriptase and several other potential target serine proteases within and an outside of the matriptase proteolytic cascade. ? ? ? Urokinase receptor-associated protein (uPARAP) and intracellular collagen turnover (33% effort) ? We have worked with in collaboration with Kenn Holmbeck, MMPU, CSDB, Roberto Weigert, IMTU, OPCB, and Niels Behrendt, University of Copenhagen, Denmark, to gain mechanistic insights into the uPARAP-dependent intracellular pathway of collagen degradation, with the ultimate aim of integrating this novel collagen degradation pathway into known collagen catabolic pathways. These studies have revealed that the growth of explanted primary uPARAP-deficient fibroblasts on a native fibrillar collagen matrix leads to a dramatic accumulation of large collagen fragments in the culture supernatant, compatible with the hypothesis of uPARAP being engaged in the endocytic turnover of collagen fragments that have been proteolytically liberated from collagen-rich matrices by soluble or membrane-associated collagenases. In agreement with this notion, acid extracted soluble collagen that has been pre-cleaved by mammalian collagenases is taken up considerably faster by uPARAP-expressing cells than intact, native collagen. Unexpectedly, however, single or combined genetic deficiencies in several mesenchymal collagenases (membrane type (MT)-1 matrix metalloproteinase (MMP), MT3-MMP, and MMP-2) or even the pretreatment of cells with a broad-spectrum metalloproteinase inhibitor failed to markedly affect uPARAP-dependent collagen internalization. These data suggest that uPARAP either internalizes collagen independently of pericellular/extracellular collagen pre-cleavage, or more likely, that uPARAP-dependent collagen degradation is preceded by initial collagen cleavage via a hitherto uncharacterized collagen catabolic pathway. ? ? ? Reengineered bacterial cytotoxins as antitumor and protease imaging agents (33% effort)? We have continued our long-standing collaborations with Stephen Leppla, MPS, NIAID, and Art Frankel, Scott & White, Temple, Texas, on the development of cell surface protease-activated bacterial cytotoxins as therapeutic agents for cancer. Our previous work provided proof of concept by showing that toxins activated by proteolytic enzymes overexpressed in tumors display high tumoricidal activity and limited tissue toxicity in mouse models. We have also demonstrated that the therapeutic index of protease-activated toxins can be further increased by making toxin variants that require sequential activation by more than one tumor-associated protease, or by combining the requirement for tumor protease activation with retargeting of toxin receptor specificity. In an independent effort, we have recently shown that protease-activated cytotoxins can be retooled for use in non-invasive imaging of specific proteolytic activity in single living cells in culture.? We have worked with Silvio Gutkind, MCU, OPCB, on the development of a novel modified anthrax toxin that targets tumor vasculature. To this end, we have taken advantage of previous studies that have established that angiogenic endothelium, but not quiescent endothelium, expresses cell surface MMP activity, that the response of endothelial cells to VEGF is dependent on mitogen-activated protein kinase kinases (MEKs), and that anthrax lethal factor (LF) efficiently cleaves and inactivates MEKs when translocated to the cytoplasm during anthrax toxin intoxication. We thus generated a reengineered anthrax toxin in which the cytoplasmic translocation of LF was mediated by a MMP-activated variant of anthrax protective antigen (PA-L1; 164RRKR167 → 164GPLGMLSQ171). As predicted, when combined with LF, PA-L1 displayed very limited toxicity to normal tissues, due to the restricted expression of cell surface MMP activity and MEK-dependence of homeostatic tissues. Wound healing times increased by about 50% in toxin-treated mice, compatible with anti-angiogenic activity and impairment of angiogenesis-dependent processes. Unexpectedly, however, the novel toxin displayed greater bioavailability and lower antigenicity than wildtype toxin, as a consequence of the limited uptake and clearance of the modified toxin by normal cells. Furthermore, PA-L1 and LF displayed potent anti-angiogenic activity, as assessed in both in vitro and in vivo assays, which translated into strong anti-tumor activity towards a diverse array of xenografted and syngrafted human and murine tumors. All these properties taken together suggest that the novel MMP-activated anti-tumor toxin holds promise for use in cancer therapy.
Showing the most recent 10 out of 56 publications
