The Lectin-Like Domain of Thrombomodulin Protects Against Diabetic Nephropathy by Inhibiting Complement Activation

Thrombomodulin (TM) is type 1 transmembrane glycoprotein predominately expressed on vascular endothelial cells. TM consists of distinct structural domains, which enable interaction with multiple ligands and thus modulation of coagulation, fibrinolysis, complement activation, inflammation, cell proliferation, and apoptosis. We have previously shown that TM protects against diabetic nephropathy (dNP) through activation of protein C (PC), which inhibits glomerular apoptosis. Recent studies showed that TM mediates cytoprotective effects independent of PC activation through it's lectin-like domain (LLD). The LLD directly interferes with complement activation and protect against arthritis. dNP is associated with an infiltration of inflammatory cells, enhanced cytokine production, and complement activation raising the question whether endothelial TM may protect against DN through a dual mechanism depending on (a) PC activation and (b) complement inhibition through the LLD.

Experimental dNP was evaluated in mice with targeted deletion of the LLD of TM (TM^Led/Led mice). Persistent hyperglycemia was induced using streptozotocin in TM^Led/Led mice and in control wild type mice (wt). Subgroups of mice were treated with low molecular heparin (LMWH) to inhibit complement activation and coagulation activation, fondaparinux to inhibit coagulation activation, or treated with minocycline to inhibit glomerular apoptosis. After 28 weeks albuminuria was determined and mice were sacrificed to isolated tissues and blood samples for analyses.

Albuminuria, histological indices of diabetic nephropathy and glomerular C3 deposition were increased in diabetic wt mice and further increased in diabetic TM^Led/Led mice. Only treatment with LMWH, but not with fondaparinux corrected the aggravated DN in TM^Led/Led mice, despite equal anticoagulation in both groups. Glomerular complement depositions (immunohistochemistry) were significantly reduced following in vivo LMWH treatment, but no in fondaparinux treated mice. In vitro, endothelial cells expressing a TM mutant lacking the LLD had diminished capacity to bind and inactivate C3b and were thus sensitive to complement mediated cell lyses. Of note, apoptosis (TUNEL) and expression levels of apoptosis regulators (p53, Bax/Blc-2 ratio, Western) did not differ between untreated or LMWH/fondaparinux treated diabetic TM^Led/Led mice, indicating that the protective effect of the LLD is independent of apoptosis inhibition. Consistently, inhibition of apoptosis with minocycline failed to normalize albuminuria or prevent histological indices of dNP in diabetic TM^Led/Led mice. Likewise, we did not observe difference in HMGB-1 levels within the kidney or in plasma sample, and inhibition of HMGB-1 using sodium Tanshinon II asilate failed to reduce dNP in diabetic TM^Led/Led mice.

We identify a novel pathway through which endothelial TM protects against dNP. TM protects against dNP by inhibiting complement activation through its LLD. This function is independent of the recently identified activated PC dependent inhibition of glomerular apoptosis. Thus, endothelial TM prevents dNP through two independent pathways. These results provide further experimental support for a causative role of endothelial dysfunction for DN.

No relevant conflicts of interest to declare.