Potassium Channels and Their Potential Roles in Substance Use Disorders

Substance use disorders (SUDs) are ubiquitous throughout the world. However, much remains to be done to develop pharmacotherapies that are very efficacious because the focus has been mostly on using dopaminergic agents or opioid agonists. Herein we discuss the potential of using potassium channel activators in SUD treatment because evidence has accumulated to support a role of these channels in the effects of rewarding drugs. Potassium channels regulate neuronal action potential via effects on threshold, burst firing, and firing frequency. They are located in brain regions identified as important for the behavioral responses to rewarding drugs. In addition, their expression profiles are influenced by administration of rewarding substances. Genetic studies have also implicated variants in genes that encode potassium channels. Importantly, administration of potassium agonists have been shown to reduce alcohol intake and to augment the behavioral effects of opioid drugs. Potassium channel expression is also increased in animals with reduced intake of methamphetamine. Together, these results support the idea of further investing in studies that focus on elucidating the role of potassium channels as targets for therapeutic interventions against SUDs.


Introduction
Substance use disorders (SUDs) are biopsychosocial disorders include neuropsychiatric symptoms such as loss of control of drug taking, repeated relapses to drug taking after intervals of forced or voluntary abstinence, and continued drug use in the presence of adverse consequences as described in the Diagnostic Statistical Manual of the American Psychiatric Association [1]. Individuals who use licit and illicit drugs, including alcohol, cocaine, methamphetamine, or opioids, do not all develop SUDs because of genetic and environment factors that enable them to be resilient [2]. On the other hand, several approaches have been taken to develop treatment approaches to alter the clinical course of SUDs in those patients who meet diagnostic criteria and subsequently seek medical and psychological treatment [3][4][5]. These approaches have met with variable degrees of success. However, there still remains a substantial need to develop pharmacotherapies that can provide important relief for individuals who have progressed from the use of these drugs for their reward properties to their compulsive abuse even in the presence of adverse medical and social consequences. Thus, the goal of the present review is to discuss the possibility of targeting potassium channels as potential pharmacological treatment for SUDs.

Classification of Potassium Channels
Ion channels are integral pore-forming transmembrane proteins that selectively control the influx and efflux of important physiological ions including Na + , K + , Ca2 + , and Cl − into and from cells or intracellular organelles. They serve to control cytoplasmic and intraorganellar concentrations of these ions as well as regulate membrane potential and The inwardly rectifying (K IR ) subgroup, named after its inward flow of K + ions, was first cloned in 1993 [31]. These channels are arranged into homo-or hetero-tetramers that react to magnesium and polyamines [32]. This group contains 16 genes (K IR 1-K IR 7) that are divided into 4 families: K + transport channels (KCNJ1, KCNJ10, KCNJ13, KCNJ15);  classical K IR channels (KCNJ2, KCNJ4, KCNJ12, KCNJ14, KCNJ16, KCNJ18); G-protein K + channels (KCNJ3, KCNJ5, KCNJ6, KCNJ9); ATP-sensitive K + channels (KCNJ8, KCNJ11).

K + Channels and Neuronal Function
Potassium channels perform multiple functions in the neuron. These include maintaining resting cell membrane potential, modulating neuronal excitability, neurotransmitter release, and maintaining homeostasis [11,33,34]. The sub-cellular localization of K + channels in axon terminals [35] provides a mechanism via which they can control cellular communication via dopamine (DA) neurotransmission in reward circuitries [36][37][38].
In fact, KCNQ2 and KCNQ3 channels that are expressed in the brain as heterotetramers (KCNQ2/3) [39] are localized on dopamine neurons in the ventral tegmental area (VTA) [37].
Activation of KCNQ2/3 has been shown to reduce midbrain dopamine neuronal excitability and to also attenuate psychostimulant-induced increases in extracellular dopamine in the nucleus accumbens [40][41][42]. Activation of KCNQ2/3 was shown to reduce the firing of midbrain dopaminergic neurons, to inhibit striatal dopamine synthesis [40]. Retigabine, another KCNQ activator, was also reported to prevent d-amphetamine-induced DA efflux in the nucleus accumbens and d-amphetamine-induced locomotor hyperactivity [42]. Similar results have been obtained using striatal slices whereby retigabine was able to inhibit KCl-dependent release of DA [43]. Together these results implicate potassium in the regulation of DA release and stimulant-induced behavioral activation.
The accumulated evidence suggests that K + channels can regulate and/or facilitate the development of neuroadaptations that may be important in drug-induced progression of behaviors that are associated with the development of a SUD diagnosis [48]. Indeed, K + channel signaling appears to be impacted by exposure to various rewarding drugs that include alcohol [49][50][51], cocaine [52], methamphetamine [53][54][55], and opioids [56,57], morphine [58,59], oxycodone [60], and fentanyl [61]. Detailed effects of these drugs on potassium channels are described below.

Potassium Channels and Alcohol Use Disorder
Alcohol consumption is widespread in the general population in the world [62,63]. However, although relatively few individuals meet criteria for alcohol use disorder (AUD) [64], AUD is associated with significant medical, neurological, and psychiatric comorbidities with consequent increased morbidity and mortality [65]. Presently, there are few FDAapproved pharmacological treatments for AUD. These include disulfiram, acamprosate, and naltrexone and these medications have limited success at reducing the high rates of relapse and are not necessarily efficacious in some populations [66,67]. Thus, there is a need to develop other pharmacological agents based on the elucidation of neurobiological mechanisms that might be the substrates of AUD.
Herein, we have reported both the causal and non-causal role of K + channels in AUD. Studies that show changes in gene expression should be considered descriptive and of need for further validation. However, data from pharmacological interventions and knockout animal models should be considered as having shown potential causal relationships. High throughput microarray analyses using rodent models of AUD have found differential expression of genes that encode K + channels in the brain of animals with different levels of alcohol intake [68,69]. Significantly, FDA-approved potassium channel regulators reduce alcohol intake in rats [70].

Alcohol Use Disorder and K V Family
Rinker et al. (2017) identified significant changes in the expression of several genes that belong to K V families in the PFC and NAc of mice that showed different levels of alcohol intake [68]. In the PFC, they reported that voluntary ethanol consumption (VEC) results in changes in Kcna5, Kcnc3, Kcnd2, and Kcnq5 expression [68]. Chronic [68]. These observations are, in part, consistent with a previous report that chronic ethanol exposure caused downregulation of Kcnq2 mRNA expression was observed in the mouse amygdala [71]. Another K V channel, K V 4.2/KCND2, has been reported to show alcohol exposure-induced decreases in protein levels with increasing concentrations of alcohol in both in-vitro and in-vivo model [72,73]. These results are consistent with the report that administration of retigabine, an activator of K V 7.2/KCNQ2 and K V 7.3/KCNQ3, led to a reduction of voluntary alcohol intake in rodents [74,75]. It is to be noted that retigabine has been withdrawn from the market by its manufacturer GlaxoSmithKline.

AUD and K 2P Families
Members of K 2P subfamilies are also differentially regulated by alcohol intake [50,80]. Specifically, K 2P 7.1/Kcnk13/Thik1 mRNA expression in the mouse ventral tegmental area (VTA) was upregulated by acute ethanol exposure [50] whereas mRNA levels of K 2P 1.1/Kcnk1/Twik1, a weak inward rectifying channel, were down-regulated in the mouse cerebellum by chronic ethanol exposure [80]. Those results are interesting because Twik1 functions to set the resting membrane potential in cerebellar cells [87]. Interestingly, chronic nicotine exposure also down-regulated Twik1 expression in the amygdala [88], suggesting the possibility that Twik1 might participate in the long-term effects of rewarding drugs on the brain. Genome-wide studies have suggested the possibility that KCNK2 gene might be linked to AUD in the native Indians and the Euro-Americans [89].

AUD and K Ca Families
K Ca (Ca 2+ -activated) channels are targets for both acute and chronic alcohol exposure [70,[82][83][84][85]. Large (BK) and small conductance (SK) K Ca channels influence cell repolarization and impact dendritic Ca 2+ spikes [84]. Rinker et al. (2017), using their highthroughput assay following different levels of ethanol exposure, have reported significant changes in BK channel expression of Kcnma1 and Kcnmb2, in both PFC and NAc of mice [68]. In addition to changes in mRNA expression, acute ethanol administration induced significant decreased KCNMA1 protein levels in primary hippocampal neuronal cultures [82]. Marrero et al. (2015) reported that moderate drinkers did not exhibit changes in KCNMA1 expression whereas heavy drinkers showed decreased KCNMA1 expression [86].
Unlike, BK channels whose levels are changed with acute ethanol exposure, the expression of SK (KCNN) channels is influenced by chronic ethanol administration, with reduced expression of K Ca 2.3/Kcnn3 channels and KCNN3 protein levels being observed in the rodent NAc [79]. Interestingly, protracted withdrawal from chronic ethanol reduced the function and trafficking of KCNN3 in the rodent NAc [78].
Potassium channels might also play a role in the fetal alcohol syndrome. Specifically, Ramadoss et al. (2008) infused pregnant ewes with ethanol or saline using a "3 days/week binge" pattern during the third trimester [90]. They reported that chronic ethanol causes a 45% reduction in the total number of fetal cerebellar Purkinje cells. They also found that TASK1 channels were expressed in Purkinje cells and that the TASK3 isoform was expressed in granule cells of the ovine fetal cerebellum. Importantly, pharmacological blockade of both TASK1 and TASK3 channels simultaneous with ethanol was able to prevent the reduction in fetal cerebellar Purkinje cell number [90].
The study by Herman et al. (2015) exhibits the role of K IR 3.3/Kcnj9 as a critical gatekeeper of ethanol incentive salience and projects as a potential target for the treatment of excessive ethanol consumption [81]. Similar to Herman et al. (2015), knock-out (KO) of K IR 3.3/KCNJ9/protein enhanced ethanol conditioned place preference [94,95].
Rinker et al. (2017) also identified several K IR genes, in addition to K V , K Ca and K 2P genes in all the three paradigms of alcohol intake [68] (voluntary, CIE and heavy use-see Table 2). Interestingly, the major gene expression changes were identified in the NAc.

Potassium Channels and Cocaine Use Disorder
In recent years, overdose death related to cocaine use disorder (CUD) is on the rise and have almost tripled in the US [96]. Amongst the US population of cocaine users, about 5.5 million (2.0% of the population) were above the age of 12 and about 2.2% of high school seniors used cocaine [97]. Cocaine exerts its actions by blocking the dopamine (DA) transporter, with secondary increases in DA in the synaptic cleft [98]. These facts have led to therapeutic approaches that include the development of anti-DAergic drugs in attempts to treat CUD. These approaches have not met with a high degree of success [99,100], suggesting there are potential targets that might offer other avenues to fight cocaine addiction.
Recent studies have indeed suggested a causal role for potassium channels in CUD. Specifically, Mooney & Rawls, (2017) reported that flupirtine, an agonist of the voltagegated KCNQ2/3 channel, can reduce the development of cocaine place preference and locomotor activation [101]. Studies using knock-out mice models have reported that Kcnj6 and Kcnj9 knockout mice exhibited reduced cocaine seeking compared to WT mice [102]. However, only Kcnj6 knock-out mice displayed increased locomotion behavior [102]. In contrast, McCall et al., (2017) reports increased cocaine self-administration in mice with Kcnj6 deletion in VTA DA neurons [52]. The discrepancy observed between these two studies may be due to the way that the gene was deleted. The former uses whole animal KO whereas the latter use site-specific ablation. In addition, Kcnj6 ablation is known to play a key in the regulation of K IR 3/GIRK channel activity in midbrain DA neurons [103,104] (See Table 3).

Potassium Channels and Methamphetamine Use Disorder
METH use disorder (MUD) is highly prevalent throughout the world and characterized by loss of control over drug use despite adverse consequences as well as strong urges to seek and use the drug [1]. Cognitive and psychiatric deficits consequent to structural and functional pathologies have been well documented [105,106].
Researchers in the Cadet laboratory have recently published several papers in which they attempt to mimic human MUD conditions in rodent models [53,55,[107][108][109]. In those models, they have introduced an additional DSM5 criterion that is related to compulsive drug intake despite adverse consequences [107]. In that model, they use drug selfadministration and then introduced footshock punishment to represent adverse consequences that humans might encounter during their frequent experiences with drugs. In the case of METH, shocks are introduced after the animals had escalated their intake and reached a plateau of consistent daily consumption of large amount of METH. Our group has shown that the introduction of footshocks administered contingently with METH can help to dichotomize rats into punishment or shock-resistant (compulsive, vulnerable to addiction, addicted) and shock-sensitive (not vulnerable, non-addicted) animals [107,[109][110][111]. In addition to the behavioral manifestations, the resistant and sensitive rats have also been shown to exhibited interesting transcriptional and biochemical changes in certain brain regions including the dorsal striatum and nucleus accumbens [107,110,112]. It is to be noted here that unlike alcohol and cocaine, studies with METH have focused on non-causal association with potassium channels and requires further elucidation.

MUD and K V Family
Recently, we used that model to investigate potential alterations in global DNA hydroxymethylation in the nucleus accumbens (NAc). We used the NAc because behavioral phenomena that occur METH self-administration (METH SA) are thought to be regulated by interconnections between distinct brain regions that include that structure [113]. Neuroplastic changes in the NAc are thought to participate in the development and maintenance of drug-taking behaviors [114]. We found, for the first time, that rats that continued to take METH compulsively exhibited differential DNA hydroxymethylation in comparison with both control and nonaddicted rats [55]. Rats that suppress their intake of METH in the presence of footshocks also showed differences in DNA hydroxymethylation from control rats, suggesting that exposure to the drug can result in prolonged effects on this epigenetic marker. The changes in DNA hydroxymethylation in the NAc of non-addicted rats were observed mostly at intergenic sites located on long and short interspersed elements. Of significant relevance to the present review, we also observed differentially hydroxymethylated regions in genes that encoded voltage (K V 1.1, K V 1.2, K Vb 1 and K V 2.2) potassium channels [55]. In order to test if these changes in DNA hydroxymethylation were accompanied with changes in transcription, we used quantitative PCR and measured their mRNA levels in the experimental groups of rats. We found, indeed, that the mRNA levels of these potassium channels were increased in the non-addiction rats. These results suggest that increased expression of these channels might participate in suppressing METH self-administration by rats.
The potential role of potassium channels in METH taking behaviors was also investigated in a subsequent paper. In that paper, we tested the possibility that a single preexposure to METH could potentiate its intake during self-administration experiments [53]. Animals were therefore pre-treated with saline or METH prior to METH SA. The experiment consisted of three experimental groups: (1) a single saline injection followed by saline self-administration (SS); (2) a single saline injection followed by METH SA (SM); and (3) a single METH injection followed by METH SA (MM). METH-pretreated rats escalated METH SA earlier and took more METH than saline-pretreated animals. Because compulsive METH takers and METH-abstinent rats show differences in potassium (K + ) channel mRNA levels in their nucleus accumbens (NAc), we tested the possibility that the expression of K V potassium channels might also help to distinguish between rats that escalated METH earlier (MM group) than the other group (SM). Increased levels of mRNA and protein expression of voltage-gated K+ channels (Kv1: Kcna1, Kcna3, and Kcna6) were indeed found in the NAc of rats that escalated METH later and took less METH [53]. Rats with increased mRNA expression also showed decreased DNA methylation at the CpG-rich sites near the promoter region of these genes.
It is to be noted that potassium channels may also be involved in the toxic effects [115] that have been reported after various doses of METH [116]. Specifically, Zhu et al., (2018) treated primary cultured hippocampal neurons with METH and reported that the drug caused time-and dose-dependent increases in K V 2.1 protein expression which was accompanied by elevated cleaved-caspase 3 and declined bcl-2/bax ratio, markers of neuronal apoptosis as previously reported after METH injections into rats [117]. Blockage of K V 2.1 with the inhibitor, GxTx-1E, or its knockdown attenuated the toxic effects of the drug [116].

MUD and K Ca Family
The investigators from Cadet's laboratory also documented changes in DNA hydroxymethylation in genes that encode K Ca potassium channel genes in the NAc of animals exposed to METH and subsequently dichotomized by footshocks [55]. These genes included K Ca 2.1/KCNN1 and K Ca 2.2/KCNN2. Similar to the observations for K V channels, these changes were associated with increased expression of their mRNA levels. We also investigated the expression of calcium-activated K + channels in animals that had received saline or METH prior to being put through the METH SA experiments [53]. Interestingly, only Kcnn1 (K Ca 2.1) showed increased expression in the NAc of rats that received saline first (SM rats) in comparison to those that received an injection of METH first (MM rats). In contrast, Kcnn3 (KCa2.3/SKCa3) and Kcnma1 mRNA levels were increased in all rats that had self-administered METH [53], suggesting that METH SA is enough to alter the expression of these genes irrespective of the amount of METH taken. (See Table 4).

Potassium Channels and Opioid Use Disorders
Opioid use disorders (OUDs) have reached epidemic levels [117]. The number of opioid-related overdose deaths from both prescription and illicit opioids reached close to 450,000 between 1999-2018 [118]. In 2018 alone, 46,802 people died of opioid overdoses, representing 69.5% of all drug overdose deaths [119]. Some of these terrible consequences might be related to treatment availability which is quite limited. Presently, FDA-approved medications for OUD include the mu-opioid receptor full agonist methadone, mu-opioid receptor partial agonist buprenorphine, and mu-opioid receptor antagonist naltrexone [120]. There are therefore windows for the development of additional medications that may be derived from a basic understanding of the effects of opioid drugs.

OUDs and K V Family
There is causal evidence that potassium channels might be involved in the effects of opioid drugs, suggesting the possibility that activators of potassium channels might be useful in the treatment of pain and/or opioid addiction [121]. For example, a genomewide association study using opioid-dependent humans has identified a risk variant in the voltage-gated K+ channel gene K V 6.2/Kcng2 [57]. In addition, the use of selective K V 7/KCNQ2 -3 K+ channel activator, flupirtine, increased the analgesic effects of morphine in animal pain models [122,123]. Moreover, another KCNQ K + channel activator, retigabine/ezogabine, elevated pain threshold and prolonged withdrawal latency after thermal stimulation [124]. Interestingly, the synergistic analgesic effect observed may eventually translate to reduced opioid prescription. The effects of retigabine could be blocked by the potassium channel antagonist, linopirdine [124]. In-vitro studies have shown the effects of opioid agonists, fentanyl, heroin, and methadone, on voltage-dependent potassium channels, K V 11.1/KCNH2/hERG [61,125,126]. It needs to be emphasized that the pharmacological activation of cardiac hERG potassium channels have adverse effects and warrants caution in clinical research. (see Table 5).

OUDs and K Ca Family
Knock-down of BK channel, Kcnb3 via small interference RNA was reported to ameliorate the hyperalgesia and anti-nociceptive effects of chronic morphine [128]. In addition to the big conductance K + channel, the small conductance SK channel also plays a specific role in reward circuits following morphine exposure. Specifically, Fakira et al. (2014) treated mice with escalating doses of morphine over 4 days followed by a challenge dose of morphine a week later. The dosing morphine schedule resulted in increased locomotor activity and increased SK2 activation [127]. The animals that received repeated morphine followed by a morphine challenge exhibited increased expression of SK2 protein levels in whole hippocampal homogenates but decreased SK2 expression in post-synaptic density [127]. A more recent study also provided evidence for the involvement of SK channels in the actions of opioid drugs [58]. The authors sought to determine if SK channels located in the nucleus accumbens played any role in morphine withdrawal. They found that firing of neurons in the shell of the nucleus accumbens was enhanced secondary to decreased expression of SK channels during morphine withdrawal, with SK2 and SK3 protein levels being significantly decreases after 3 weeks of withdrawal [58]. These observations support a role of potassium channels in the longterm effects of morphine in the brain.

OUDs and K IR Family
Opioids also exert significant effects on KIR family channels. These have been reported on the G-protein gated class that includes GIRK1-4/KCNJ3, 5, 6 and 9. The analgesic effects of morphine were potentiated in knock-out mice with Kcnj6/ GIRK2/ K IR 3.2 deletion [129]. Similarly, the antinociceptive effects of oxycodone were attenuated by knock-down of Kcnj3/ GIRK1/ K IR 3.1channels using short interfering RNA [60]. Studies by Kotecki et al. (2015) using Kcnj3 and Kcnj6 knock-out mice models have shown that enhancement of morphine-induced locomotion.
Blockers of ATP-sensitive class of K IR subfamily, glibenclamide and tolbutamide sulfonylureas, effectively reversed the peripheral antinociceptive effect of fentanyl [130]. Glibenclamide was also shown to play an active role in morphine reward [131] by facilitating morphine-induced conditioned place preference [132].

Conclusions
This review has provided descriptive and causal evidence for diverse roles of K + channels in the behavioral manifestation of substances of abuse including alcohol, cocaine, methamphetamine, and opioids. These substances can alter the expression of potassium channels at both mRNA and protein levels. These changes were found to occur in various brain regions including the nucleus accumbens and hippocampus that are involved in various aspects of addiction. In addition, we reviewed, evidence that activators of K + channels can suppress behaviors induced by some of these rewarding agents. These behaviors include drug acquisition, maintenance, and withdrawal-associated phenomena. The evidence discussed herein supports the view that there is a need to invest in the development of pharmacotherapeutic agents that target K + channels in the brain. These studies will help to develop non-dopaminergic agents against SUDs since DAergic drugs have not been shown to be very efficacious against these psychiatric disorders.

Conflicts of Interest:
The authors declare no conflict of interest.