Development of supramolecular anticoagulants with on-demand reversibility.

Drugs are administered at a dosing schedule set by their therapeutic index, and termination of action is achieved by clearance and metabolism of the drug. In some cases, such as anticoagulant drugs or immunotherapeutics, it is important to be able to quickly reverse the drug's action. Here, we report a general strategy to achieve on-demand reversibility by designing a supramolecular drug (a noncovalent assembly of two cooperatively interacting drug fragments held together by transient hybridization of peptide nucleic acid (PNA)) that can be reversed with a PNA antidote that outcompetes the hybridization between the fragments. We demonstrate the approach with thrombin-inhibiting anticoagulants, creating very potent and reversible bivalent direct thrombin inhibitors (Ki = 74 pM). The supramolecular inhibitor effectively inhibited thrombus formation in mice in a needle injury thrombosis model, and this activity could be reversed by administration of the PNA antidote. This design is applicable to therapeutic targets where two binding sites can be identified.

The N-terminus of GPR37L1 is proteolytically processed by matrix metalloproteases.

GPR37L1 is an orphan G protein-coupled receptor expressed exclusively in the brain and linked to seizures, neuroprotection and cardiovascular disease. Based upon the observation that fragments of the GPR37L1 N-terminus are found in human cerebrospinal fluid, we hypothesized that GPR37L1 was subject to post-translational modification. Heterologous expression of GPR37L1-eYFP in either HEK293 or U87 glioblastoma cells yielded two cell surface species of approximately equivalent abundance, the larger of which is N-glycosylated at Asn105. The smaller species is produced by matrix metalloprotease/ADAM-mediated proteolysis (shown by the use of pharmacological inhibitors) and has a molecular weight identical to that of a mutant lacking the entire N-terminus, Δ122 GPR37L1. Serial truncation of the N-terminus prevented GPR37L1 expression except when the entire N-terminus was removed, narrowing the predicted site of N-terminal proteolysis to residues 105-122. Using yeast expressing different G protein chimeras, we found that wild type GPR37L1, but not Δ122 GPR37L1, coupled constitutively to Gpa1/Gαs and Gpa1/Gα16 chimeras, in contrast to previous studies. We tested the peptides identified in cerebrospinal fluid as well as their putative newly-generated N-terminal 'tethered' counterparts in both wild type and Δ122 GPR37L1 Gpa1/Gαs strains but saw no effect, suggesting that GPR37L1 does not signal in a manner akin to the protease-activated receptor family. We also saw no evidence of receptor activation or regulation by the reported GPR37L1 ligand, prosaptide/TX14A. Finally, the proteolytically processed species predominated both in vivo and ex vivo in organotypic cerebellar slice preparations, suggesting that GPR37L1 is rapidly processed to a signaling-inactive form. Our data indicate that the function of GPR37L1 in vivo is tightly regulated by metalloprotease-dependent N-terminal cleavage.