Several types of drug-resistant epilepsies of infancy have been associated with mutations in the KCNT1 gene, which encodes the sodium-activated potassium channel subunit KNa1.1. Approximately forty different amino acid substitutions have been identified in patients, involving both transmembrane and intracellular domains of the channel subunit, and with one exception all cause gain-of-function in channel gating. Channel inhibition, therefore, is potentially a stratified approach to treat the disorder. To date, quinidine therapy has been trialled with several patients, but mostly with unsuccessful outcomes, which have been linked to its low potency and lack of specificity; effects on cardiac ion channels being a major concern. Here we describe the use of high-resolution structures of KNa1.1 channels, generated through cryo-electron microscopy, to identify novel inhibitors using computational methods. From an initial analysis of seventeen compounds identified using this approach, we found six that inhibited KNa1.1 channels. Their IC50 values ranged between 0.6 and 7.4 μM, compared to 100 μM with quinidine, and likely inhibited through pore-blocking mechanisms. The compounds delivered varying results when evaluating state-dependence and in preliminary toxicity assays, with some demonstrating negligible cellular toxicity and low levels of hERG block. These compounds may provide starting points for the development of novel pharmacophores for KNa1.1 inhibitors, with the view to treating KCNT1-associated epilepsies. Furthermore, this illustrates the potential for utilising ion channel structures generated through cryo-electron microscopy in drug discovery.

The ongoing pain associated with arthritis reduces the quality of human life. Injection of complete Freund’s adjuvant (CFA) into the knee of mice induced inflammation and a concomitant decrease in their natural digging behaviour within 24-hours. After inflammation, retrogradely labelled knee-innervating dorsal root ganglion neurons showed a decreased action potential threshold and increased transient receptor potential vanilloid-1 (TRPV1) expression (validated with immunohistochemistry). Subsequently, systemic administration of a TRPV1 antagonist normalized mouse digging behaviour within 30-minutes, likely due to inhibition of pain signalling. To test the translational potential of these results, we recorded from mouse sensory neurons incubated overnight with synovial fluid from humans with osteoarthritis experiencing chronic pain (OA-SF). We find that OA-SF directly sensitizes mouse knee neurons by decreasing the action potential threshold and increasing the resting membrane potential implicating voltage-gated sodium (Na_v) and potassium (K_v) channels. Using Ca²⁺-imaging we also find that OA-SF increases TRPV1 and TRPM8 activity compared to synovial fluid from healthy individuals, highlighting the importance of TRP channels in OA pain. Taken together, our data suggest that inflammatory mediators in OA-SF excite sensory neurons and hence drive knee pain. This in vitro, translational model can be incorporated into clinical trials to unbiasedly assess the efficacy of drugs on sensory neurons and hence identify drivers of OA pain.

Growing functional and genetic evidence points to an emerging role for two-pore domain potassium channels (K2Ps) in numerous pathophysiologies. K2Ps carry background (or leak) potassium current and primarily function to maintain cellular resting membrane potential. However, K2Ps have proved a challenging superfamily to modulate with small molecules and there is currently a lack of pharmacological tools which target K2Ps. This has limited the investigation of K2Ps precise physiological function and efforts to generate therapeutics which act upon K2Ps. Our initial aim was to develop a screening system which would allow the identification of small molecule K2P activators. We initially showed that the identification of channel activators can be compromised in systems in which the target (K2P) is over-expressed at high levels. To avoid this, we used baculovirus (‘BacMam’) to transduce K2P expression in mammalian cells. ‘BacMam’ allows the precise titration of expression of the gene of interest. This enabled us to develop cell systems in which we were able to intricately and robustly select a level of K2P expression (via functional studies) optimized for the identification of channel activators. An example assay development using the K2P TRAAK is described. TRAAK is known to be expressed in nociceptors and has been shown to be a genetic predictor of post-surgical neuropathic pain. An appropriate level of TRAAK channel function was chosen after ‘BacMam’ optimization and channel function measured via thallium flux. The pharmacology of previously described TRAAK activators is also shown. A representative sub-group of K2Ps (TWIK-1, THIK-1, TASK-3, TASK-2, TREK-2) was screened using the BacMam/ Thallium flux assay against the LifeArc Index set (11k compounds). This allowed us to assess which K2Ps were ‘ligandable’ or ‘activatable’ by small molecules. Multiple novel activators were identified, and development is on-going. We also screened a library of drug-like molecules against the same group of K2Ps and identified Terbinafine as a novel activator of the TASK-3 channel. The selectivity of this molecule across the K2P superfamily was assessed and structure-activity relationship studies showed that relatively small changes to the molecule had significant effects on activity and selectivity.

Pathogenic T cell effector memory (TEM) cells drive many autoimmune disorders and are uniquely dependent on the Kv1.3 channel. A number of venom derived knottin (cysteine-rich mini-protein) inhibitors of Kv1.3 are being developed as potential drug candidates but can suffer from manufacturing difficulties, short half-lives and a lack of specificity. We have developed a novel molecular format wherein a peripheral CDR loop of an antibody has been replaced by a knottin. In this novel KnotBody^TM format, the knottin benefits from the improved therapeutic functionality of an antibody and the antibody gains additional diversity by the addition of a scaffold which is pre-disposed to blockade of ion channels.

A proof-of-concept fusion protein of one structural domain within another was initially achieved by inserting a trypsin inhibiting knottin (EETI-II) flanked by diverse repertoire of short linker sequences into the CDR2 position of naïve antibody light chain sequences. Functional KnotBody^TM molecules were selected from this library using phage display technology on the basis of retained trypsin binding, with the correct folding of both domains confirmed using X-ray crystallography.

To further demonstrate the benefits of this novel format, the modular nature of the KnotBody^TM binding surface was exploited to: (i) improve existing knottin binding by introducing additional V_H contacts; (ii) create a bispecific molecule by introducing a V_H chain that binds to a different target; (iii) substitute the proof-of-concept knottin (EETI-II, a trypsin inhibitor) with ShK, a Kv1.3 ion channel blocking toxin; (iv) develop a panel of low-nM Kv1.3 inhibitors with selectivity exceeding 3000-fold over the Kv1.1 channel, a closely related Kv family member.

Pathogenic T cell effector memory (TEM) cells drive many autoimmune disorders and are uniquely dependent on the Kv1.3 channel. A number of venom derived knottin (cysteine-rich mini-protein) inhibitors of Kv1.3 are being developed as potential drug candidates, but can suffer from manufacturing difficulties, short half-lives and a lack of specificity. We have developed a novel molecular format wherein a peripheral CDR loop of an antibody has been replaced by a knottin. In this novel KnotBody^TM format, the knottin benefits from the improved therapeutic functionality of an antibody and the antibody gains additional diversity by the addition of a scaffold which is pre-disposed to blockade of ion channels.

A proof-of-concept fusion protein of one structural domain within another was initially achieved by inserting a trypsin inhibiting knottin (EETI-II) flanked by diverse repertoire of short linker sequences into the CDR2 position of naïve antibody light chain sequences. Functional KnotBody^TM molecules were selected from this library using phage display technology on the basis of retained trypsin binding, with the correct folding of both domains confirmed using X-ray crystallography.

To further demonstrate the benefits of this novel format, the modular nature of the KnotBody^TM binding surface was exploited to: (i) improve existing knottin binding by introducing additional V_H contacts; (ii) create a bispecific molecule by introducing a V_H chain that binds to a different target; (iii) substitute the proof-of-concept knottin (EETI-II, a trypsin inhibitor) with ShK, a Kv1.3 ion channel blocking toxin; (iv) develop a panel of low-nM Kv1.3 inhibitors with selectivity exceeding 3000-fold over the Kv1.1 channel, a closely related Kv family member.

Voltage-gated sodium (Na_V) channels are studied extensively due to their potential as targets for several indications, such as pain, epilepsy, cardiac and muscle paralysis. Some Na_V channel modulators show state-dependence and bind preferentially to the inactivated state of the channel. The potency of state-dependent compounds is known to vary depending on the percentage inactivation of the channels.

To calculate accurate compound activity the precise value for the VHalf of inactivation should be used for each cell. The adaptive protocol block for the Sophion Qube 384-well automated patch clamp platform has made it possible to separately define the voltage applied to individual wells for both the activation and inactivation of the channels. This enables the generation of more precise data for voltage-gated ion channels.

Kv4.3 α-subunits associate with ancillary β-subunits of the KChIP2 family in heart muscular tissue to channel transient outward (Ito) currents. Decreased expression or dysfunction of Kv4.3 channels following myocardial infarction or during heart failure can contribute to abnormal repolarization, which may result in ventricular arrhythmias. Therefore, drug-induced inhibition of Kv4.3/KChIP2-mediated Ito exposes to potential cardio safety liabilities which are important to document early during the drug discovery process. In this study, we have characterized basic electrophysiological properties pertaining to Ito currents in CHO cells expressing recombinant Kv4.3 and KChIP2.2 subunits. The currents obtained were validated pharmacologically by assessing the efficacy of small molecule and peptide toxin inhibitors known from manual patch-clamp studies to block Ito. We then examined the inhibitory activities of 28 drugs with clinically documented high, medium or low pro-arrhythmic risk encompassing the test- and validation-sets of the Comprehensive in vitro Pro-Arrhythmia (CiPA) panel. While none of the CiPA drugs inhibited Kv4.3 at submicromolar concentrations (IC50 < 1 µM), two drugs (loratadine and nitrendipine) displayed intermediate potency (1 µM < IC50 < 10 µM), and four drugs (bepridil, quinidine, chlorpromazine and astemizole) revealed weak, but significant (10 µM < IC50 < 30 µM) inhibitors.