By bioisosteric transformations and successive optimization of known GABA uptake inhibitors, several series of novel GABA uptake inhibitors have been prepared by different synthetic approaches. These compounds are derivatives of nipecotic acid and guvacine, substituted at the nitrogen of these amino acids by various lipophilic moieties such as diarylaminoalkoxyalkyl or diarylalkoxyalkyl. The in vitro values for inhibition of [(3)H]GABA uptake in rat synaptosomes was determined for each compound, and it was found that the most potent compound from this series, (R)-1-(2-(3,3-diphenyl-1-propyloxy)ethyl)-3-piperidinecarboxyli c acid hydrochloride (29), is so far the most potent parent compound inhibiting GABA uptake into synaptosomes. Structure-activity results confirm our earlier observations, that an electronegative center in the chain connecting the amino acid and diaryl moiety is very critical in order to obtain high in vitro potency. Several of the novel compounds were also evaluated for their ability in vivo to inhibit clonic seizures induced by a 15 mg/kg (ip) dose of methyl 6, 7-dimethoxy-4-ethyl-beta-carboline-3-carboxylate (DMCM). Some of the compounds tested show a high in vivo potency comparable with that of the recently launched anticonvulsant product 6 ((R)-1-(4, 4-bis(3-methyl-2-thienyl)-3-butenyl)-3-piperidinecarboxylic acid).

On the basis of the SAR of a series of known gamma-aminobutyric acid (GABA) uptake inhibitors, including 4 (SKF 89976), new tricyclic analogues have been prepared. These novel compounds are derivatives of nipecotic acid, guvacine, and homo-beta-proline, substituted at the nitrogen of these amino acids by various lipophilic moieties such as (10,11-dihydro-5H-dibenz[b,f]azepin-5-yl)alkoxyalkyl or (10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5-ylidene)alkoxyalkyl. The in vitro values for inhibition of [(3)H]-GABA uptake in rat synaptosomes was determined for each compound in this new series, and it was found that several of the novel compounds showed a high potency comparable with that of the reference compounds 4, 5 (tiagabine), and 6 (CI-966). Several of the novel compounds were also evaluated for their ability in vivo to inhibit clonic seizures induced by a 15 mg/kg (ip) dose of methyl 6,7-dimethoxy-4-ethyl-beta-carboline-3-carboxylate (DMCM). One compound, (R)-1-(2-(2-(10,11-dihydro-5H-dibenz[b,f]azepin-5-yl)ethoxy)ethyl)-3-piperidinecarboxylic acid (23), was selected for further biological investigations and showed a protective index comparable to or slightly better than that of the recently launched anticonvulsant product 5 ((R)-1-(4,4-bis(3-methyl-2-thienyl)-3-butenyl)-3-piperidinecarboxylic acid).

Covalent cross-linking of biomolecules can be useful in pursuit of tissue targeting or dual targeting of two receptors on cell surfaces for avidity effects. Long linkers (>10 kDa) can be advantageous for such purposes, and poly(ethylene glycol) (PEG) linkers are most commonly used due to the high aqueous solubility of PEG and its relative inertness toward biological targets. However, PEG is non-biodegradable, and available PEG linkers longer than 5 kDa are heterogeneous (polydisperse), which means that conjugates based on such materials will be mixtures. We describe here recombinant linkers of distinct lengths, which can be expressed in yeast, which are polar, and which carry orthogonal reactivity at each end of the linker, thus allowing chemoselective cross-linking of proteins. A conjugate between insulin and either of the two trypsin inhibitor peptides/proteins exemplifies the technology, using a GQAP-based linker of molecular weight of 17 848, having one amine at the N-terminal, and one Cys, at the C-terminal. Notably, yeast-based expression systems typically give products with mixed disulfides when expressing proteins that are equipped with one unpaired Cys, namely, mixed disulfides with glutathione, free Cys amino acid, and/or a protein homodimer. To obtain a homogeneous linker, we worked out conditions for transforming the linker with mixed disulfides into a linker with a homogeneous disulfide, using excess 4-mercaptophenylacetic acid. Subsequently, the N-terminal amine of the linker was transformed into an azide, and the C-terminal Cys disulfide was reduced to a free thiol and reacted with halo-acetyl insulin. The N-terminal azide was finally conjugated to either of the two types of alkyne-containing trypsin inhibitor peptides/proteins. This reaction sequence allowed the cross-linked proteins to carry internal disulfides, as no reduction step was needed after protein conjugations. The insulin–trypsin inhibitor conjugates were shown to be stabilized toward enzymatic digestions and to have partially retained binding to the insulin receptor.