A Potassium Channel Activator Possesses Anticonvulsant Activity
By: Carol A. Rouzer, VICB Communications
Published: June 9, 2013
The first potent and specific G protein-activated inward-rectifying potassium channel activator suppresses seizures in two mouse models of epilepsy.
Inwardly rectifying potassium channels (IRK) conduct potassium ions across the cell membrane with a directionality that favors movement of K+ into the cell. These channels play a critical role in the restoration and stabilization of the cellular membrane potential. The activity of the G protein-activated inwardly rectifying potassium channel (GIRK) subclass of IRKs is modulated by G protein-coupled receptors that act through the pertussis toxin-sensitive G protein (Gi) (Figure 1). GIRKs are homo- or heterotetramers comprising four subunits, which are designated GIRK1 through GIRK4. The proteins are abundant in the brain, with the GIRK1/2 subtype (composed of the GIRK1 and GIRK2 subunits) predominating. GIRK1/4, GIRK2, and GIRK2/3 are also present in brain, albeit in lesser quantities. Studies of GIRK knockout mice have suggested that the proteins play roles relevant to multiple nervous system functions and/or disorders, including addiction, pain, anxiety, spatial learning, and seizures, while GIRK1/4 expressed in atrial cardiomyocytes has engendered considerable interest as a target for the treatment of atrial fibrillation. However, the lack of potent and specific modulators of GIRK activity has hampered studies of their function in both physiologic and pathophysiologic processes. To address this issue, a team of VICB researchers, including Dave Weaver, Jerod Denton, Craig Lindsley, and Gary Sulikowski, joined efforts to develop these critical tools. They now report the first potent and selective GIRK activator and show that it displays anticonvulsant activity in vivo [K. W. Kaufmann, et al. (2013) ACS Chem. Neurosci., published online June 3, doi:10.1021/cn400062a.]
Figure 1. Control of a G protein-coupled receptor-activated ion channel. Binding of a signaling molecule to the receptor leads to binding of GTP to the α subunit of the G protein. In pathway A, the alpha subunit dissociates from the β and γ subunits, and activates a separate effector molecule, which then activates the ion channel. In pathway B, the α subunit activates the ion channel directly. In pathway C, the combined βγ subunits activate the ion channel. GIRK channels are activated by pathway C. Figure reproduced from Wikipedia, and is in the public domain.
To identify activators of GIRK-dependent ion flux, the investigators developed a high-throughput screen exploiting the ability of GIRK channels to transport thallium ions. The passage of extracellular Tl+ through GIRK channels into cells preloaded with the thallium-sensitive dye Thallos-AM (developed by Dave Weaver, Gary Sulikowski, and Kwangho Kim) resulted in a fluorescent signal that served as a readout of channel activity. Initially, as part of a grant awarded to Colleen Niswender, the Molecular Libraries Small Molecule Repository (MLSMR) compound collection was screened using HEK-293 cells co-expressing the human, Gi/o-dependent metabotropic glutamate receptor 8 (MGlu8) and GIRK1/2. The investigators then counter-screened hit compounds from the first screen in cells expressing only MGlu8 to eliminate those whose activity was not dependent on GIRK1/2. From these screens, they identified a compound (CID 736191) that produced a GIRK1/2-dependent increase in fluorescence with an EC50 of ~1 μM. Synthesis of structural analogs of CID 736191 indicated that the N-phenyl pyrrazole and urea moieties were both required for activity. Modification of the aryl ring, however, was tolerated and led to the discovery of ML297 (ML297) (CID 56642816), with an EC50 of 160 nM, and an efficacy of 194% compared to the parent lead compound (CID 736191) (Figure 2).
Figure 2. Structure of the lead compound CID736191 (top). The only region of the lead molecule amenable to structural modification was the phenyl ring (center). Remaining portions of the model were required for GIRK stimulatory activity. Structure of ML297 (CID 56642816)(bottom).
Further characterization of ML297 using cells expressing GIRK1/2, GIRK1/4, GIRK2, and GIRK2/3 indicated that it could activate both GIRK1/2 and GIRK1/4, but was inactive for channels lacking GIRK1. The compound was also inactive against the closely related IRK, Kir2.1, and the voltage-gated potassium channel, Kv7.4. In contrast, ML297 inhibited the voltage-gated channel hERG, but only with low potency. Screening of ML297 against a panel of 68 potential target proteins indicated activity at the 5-HT2b receptor, the sigma σ1 receptor, and the GABAA receptor. However, in all three cases, the compound’s potency was so low compared to its potency for activation of GIRK1/2 that these activities were judged to be unlikely to have pharmacologic significance. The results suggested that ML297 interacts directly with GIRK1-containing GIRKs.
In GIRK1/2-expressing HEK293 cells, ML297 elicited a potassium flux that was twice as high as that obtained from activation of MGlu8 with glutamate. The inability of pertussis toxin to block the activity of ML297 provided further support for the hypothesis that the compound acts directly at the level of the GIRK. Measurements of potassium flux using the whole cell voltage clamp method yielded results consistent with those from the thallium-based screen. These experiments also demonstrated that ML297 produces a rapid GIRK activation, with a slower washout. ML297-induced GIRK activation was inhibited by barium ion, a well-recognized blocker of GIRKs; however the efficacy of Ba2+ appeared to be reduced in the presence of ML297.
In drug metabolism and pharmacokinetic studies, ML297 exhibited good solubility and modest plasma protein binding. Although the compound was rapidly converted by liver microsomes to a single inactive metabolite, administration to mice provided adequate plasma and brain levels to allow in vivo studies.
Following intraperitoneal injection of ML297, mice appeared normal and exhibited no obvious discomfort or distress. A clear decrease in locomotor activity could not be attributed to demonstrable motor deficits, so the investigators concluded it was likely due to general sedation. Since studies with GIRK knockout mice suggested that the channels are involved in regulating neuronal excitability, the investigators tested the ability of ML297 to suppress seizure activity in two mouse models of epilepsy. In the maximal electroshock model, ML297 caused a substantial increase in the time that elapsed prior to the onset of seizures with an efficacy similar to that of the well-known anticonvulsant valproate (Figure 3). In the second model, in which seizures were induced by administration of the GABAA antagonist pentylenetetrazol, ML297 decreased the percent of animals that experienced seizures and increased survival. In this model, ML297 was more efficacious than valproate (Figure 3).
Figure 3. Anticonvulsant activity of ML297. (A) In a mouse model that uses electric shock to induce convulsions, ML297 (60 mg/kg) caused a significant increase in seizure latency as compared to vehicle (VHL). ML297’s potency was similar to that of valproate (VAL, 150 mg/kg) in this model. (B) In a mouse model that uses a drug to induce seizures, ML297 was more effective than valproate at reducing the percent of treated mice exhibiting convulsions. (C) ML297 was also more effective than valproate in reducing mortality in this model. Image reproduced with permission from K. W. Kaufmann, et al. (2013) ACS Chem. Neurosci., published online June 3, doi:10.1021/cn400062a. Copyright 2013, American Chemical Society.
ML297 is the first potent and selective GIRK activator, with greatest potency and efficacy for GIRK1/2. This exciting discovery paves the way for the exploration of GIRK function and the evaluation of GIRK1/2 as a therapeutic target for epilepsy as well as other illnesses. Furthermore, the approach used by the VICB team can be applied to the discovery of other GIRK modulators with different specificities.