Cell surface proteolysis in development, tissue repair, and malignancy

National Institute Of Dental & Craniofacial Research

Investigators

Linked publications, trials & patents

Abstract

The overall aim of this project is to understand the biochemistry, biology, and pathology of cell surface-associated proteolysis, with an emphasis on understanding the contribution of proteolytic pathways to the development, homeostasis, regeneration, and pathology of oral tissues. Overlapping functions of EpCAM and Trop2 in claudin stabilization and development of intestinal and extraintestinal epithelia EpCAM is a type I transmembrane protein expressed on the surface of epithelial cells that is reported to regulate the stability of the claudin family of tight junction proteins. We have previously shown that EpCAM is a candidate pathogenic substrate for matriptase in the autosomal recessive disorder, syndromic congenital sodium diarrhea, which is caused by homozygous or compound heterozygous loss-of-function mutations in the matriptase inhibitor, hepatocyte growth factor inhibitor (HAI)-2. Despite the widespread epithelial expression of EpCAM, loss of EpCAM function in humans (as seen in non-syndromic congenital sodium diarrhea) and in mice has been reported to selectively affect the development of the intestine, resulting in severe early onset enteropathy. To further explore EpCAM as a potential pathogenic substrate for matriptase, we generated and comprehensively phenotyped mice with single or combined deficiency in EpCAM and its closely related paralog, Trop2. Contrary to previous reports, we found that critical functions of EpCAM are not limited to intestinal tissues. Thus, loss of EpCAM resulted in hyperkeratosis of the epidermis and forestomach, alopecia, nephron hypoplasia, proteinuria, and the inability to secrete digestive enzymes by the exocrine pancreas. These epithelial deficiencies were associated with the loss of claudin-7 from tight junctions, demonstrating the critical role of EpCAM in claudin stabilization in vivo. Expression analysis revealed that Trop2 expression overlaps with EpCAM in a number of developing epithelia, suggesting a functional redundancy of the two proteins. Consistent with this hypothesis, single loss of Trop2 by itself had no overt impact on mouse development and survival, but the combined loss of EpCAM and Trop2 compounded the developmental defects observed in EpCAM single-deficient mice. This manifested as a 60% decrease in embryonic viability, and further shortened postnatal lifespan of born pups lacking both proteins. Importantly, in tissues that co-expressed both proteins, Trop2 was able to compensate for the loss of EpCAM in stabilizing claudin-7. These findings expand the spectrum of epitheliopathies linked to EpCAM deficiency, and it identifies overlapping functions of EpCAM and Trop2 in stabilizing claudin-7. Cellular orchestration of extracellular matrix degradation Fibrin is an ECM protein that is formed by the polymerization of thrombin-cleaved fibrinogen and cross-linking of the polymer by coagulation factor XIII. Fibrin is deposited into the extravascular space in response to tissue injury, where it serves to stem the loss of blood, immobilize bacteria, and provide a provisional matrix for tissue regeneration. Fibrin, however, is highly proinflammatory and causes chronic inflammation and tissue damage, unless removed in a timely manner. Importantly in this respect, humans with compromised fibrinolytic function display excess gingival fibrin deposition and severe periodontal disease, which in humans often results in complete tooth loss early in life. Fibrinolysis is achieved by the proteolytic conversion of the inactive zymogen, plasminogen (Plg), to the active protease, plasmin. In collaboration with Matt Flick, University of North Carolina, and Francis Castellino, University of Notre Dame, we found that mice with fibrinolytic defects (achieved through combined tissue plasminogen activator and urokinase plasminogen activator deficiency, Plg deficiency or plasmin inactivation) all displayed a severe age-related periodontal bone loss. Histological characterization and whole-transcriptome analysis of periodontal tissue revealed exuberant gingival fibrin deposition, and a signature neutrophil infiltrate. Interestingly, this Plg deficiency-associated periodontal bone loss was completely reversed in mice expressing no fibrinogen as well as in knockin mice that expressed a mutant fibrinogen that lacks the neutrophil M2 integrin binding site (fibrinogen-390-396A/390-396A mice). Surprisingly, although not developing periodontal bone loss, Plg-deficient mice expressing the mutant fibrinogen-390-396A presented with gingival fibrin deposition and neutrophil accumulation comparable to Plg-deficient mice expressing wildtype fibrinogen. This suggests that fibrin promotes periodontitis through local engagement, rather than recruitment, of neutrophils. Therefore, in collaboration with Andrew Doyle, NIC, we proceeded to develop assays to study the effect of fibrin on neutrophil effector functions in vitro. By applying these assays, we found that neutrophils plated on the mutant fibrin-390-396A showed decreased reactive oxygen species (ROS) production and neutrophil extracellular trap formation (NETosis) compared to neutrophils plated on wildtype fibrin, indicating that neutrophil M2 integrin engagement of fibrin leads to neutrophil activation. Accordingly, blunting neutrophil effector functions in vivo by genetic elimination of neutrophil elastase or by systemic treatment with DNAse I to remove neutrophil extracellular traps alleviated periodontal disease in Plg-deficient mice. Importantly, mice expressing the mutant fibrin-390-396A on a wildtype mouse background displayed reduced age-associated periodontal bone loss compared to their wildtype littermates. The tissue destructive fibrin-neutrophil axis was exacerbated by commensal or dysbiotic microbiota, as shown by the substantial diminution of periodontal bone loss in germ-free mice. These data demonstrate that the local engagement of gingival fibrin by neutrophils via fibrin-M2 ligation induces periodontal bone loss by triggering key neutrophil effector functions. Collectively, this study suggests a critical role of fibrin-neutrophil engagement in periodontal disease in individuals with compromised, as well as with normal, fibrinolytic function. MEK1/2-specific tumor endothelium-targeting anthrax toxin with high therapeutic index Previous studies indicate that our reengineered anthrax lethal toxins likely suppress the growth of solid tumors by targeting tumor endothelial cells via CMG-2. LF, the effector moiety of our tumor-targeting toxin, possesses potent proteolytic activity towards MEK1 and MEK2, thereby inactivating the RAS-RAF-MEK-ERK pathway required for endothelial cell proliferation and tumor angiogenesis. However, LF also cleaves other MEKs such as MEK3/6 and MEK4/7, thus affecting p38 and JUNK MAPK pathways, which may result in undesired toxicity. A previous study revealed that the substrate specificity of LF is determined by exosites different from its active site. This finding allowed us to generate a LF mutant (LF Trp271Ala) that cleaves MEK1/2, but not other MEKs. When this MEK1/2-specific LF was combined with the urokinase plasminogen activator and matrix metalloproteinase-activated PrAg previously generated by us, the mutant LF showed anti-tumor activity equivalent to wildtype LF in several tumor models, including Lewis lung carcinoma, B16 melanoma, and C32 human melanoma. However, LF Trp271Ala displayed much lower toxicity than wildtype LF to mice, thus achieving a therapeutic index of >15, more than an order of magnitude higher than clinically used small molecule MEK inhibitors.