Ethanol (alcohol, ethyl alcohol) is the world's most commonly used recreational drug, yet, despite many years of research, ethanol's exact pharmacological mechanism remains somewhat elusive. Ethanol has been shown to affect a member of nearly every type of ion channel in the body, but often at concentrations far above those found in recreational users. In order to make sense of the huge amount of research into the effects of ethanol on neurons, one must gauge it against the threshold plasma concentration of ethanol needed to produce an effect and a rough estimation of the fatal concentration in humans, i.e. the range of ethanol's concentration found during recreational usage. It is estimated that the plasma concentration necessary to cause threshold effects in humans is 5 mM (40 mg/100 mL) (Rang et al., 2001) while in a review of 808 fatal alcohol poisonings, the mean plasma concentration was 72 mM (Koski et al., 2002). To put blood concentrations further in context, the blood-alcohol limit in most states in the US is 0.08% wt/vol or 17.4 mM.

Because of ethanol's depressive effects, there has been constant speculation that its site of action is the GABA-A receptor (the major inhibitory receptor in the brain, the same receptor enhanced by benzodiazepines and barbiturates). Many experiments have looked into this possibility, and they broadly fit into two categories: ones that found no effect and ones that found a relatively high potency effect, well within recreational concentrations. Exactly why there is this stark dichotomy is unclear, though it may have something to do with more recently discovered facts about the distribution of "subunits" of GABA-A receptors.

The GABA-A receptor is not homogenous throughout the brain, it is a protein made up of 5 subunits, selected from a possible 16 different subunits, named alpha 1 to 6, beta 1 to 3, gamma 1 to 3, delta, epsilon, theta and pi. Delta-subunit-containing GABA-A receptors are relatively rare, making up about 5% of GABA-A receptors throughout the brain, found primarily in the granule cells of the denate gyrus of the hippocampus and the granule cells of the cerebellum. It has been shown that ethanol potentiates the activity of delta-subunit-containing GABA-A receptors with a very high potency (0.1-3mM). It could be that experiments showing ethanol has a high potency effect somehow inadvertently targeted the delta-subunit-containing GABA-A receptors and caused them to activate more than in other experiments. The high potency of this effect, however, might discount delta-subunit containing GABA-A receptors as the target of ethanol's effects, because at plasma concentrations which cause a maximum effect at delta-subunit-containing GABA-A receptors, most people would be barely feeling any effect of ethanol. Also, the relatively low and restricted expression of delta-subunit containing GABA-A receptors throughout the brain makes it hard to explain how ethanol can have such a wide variety of effects, if this specific GABA-A receptor was its site of action.

Inhibition of NMDA receptors by ethanol is another theme that has received a lot of attention, because of the similar anaesthetic nature of ethanol and the classical NMDA receptor antagonists, the dissociative anaesthetics (ketamine and PCP). The NMDA receptor is a compelling site for explaining ethanol's actions, as the concentrations of ethanol that start to cause a significant inhibition of the NMDA receptor are the same concentrations at which the effects of ethanol are beginning to be registered in humans (~5 mM). NMDA receptor antagonists are classically known for their ability to block the formation of memories, an effect that larger amounts of ethanol are renowned for. However, one problem with this theory is the lack of similarity between alcohol's intoxication state and that induced by ketamine or PCP.

Besides the NMDA receptor, other glutamate receptors are affected by ethanol. The AMPA and kainate receptors are both inhibited by ethanol, but generally at concentrations above those needed to induce intoxication. On the other hand, the concentration needed to produce the 50% of maximal inhibition (IC50) of AMPA and kainate receptors is approximately the same as the mean fatal concentration of ethanol (~70 mM), indicating that the inhibition of AMPA or kainate receptors could be the cause of ethanol-induced fatalities. This makes sense, as large scale AMPA inhibition is almost certainly fatal. However, it has been demonstrated that there is a small population of kainate receptors expressed on interneurons (small, GABA-releasing neurons) in the hippocampus which are exquisitely sensitive to ethanol (IC50 = ~5 mM). Because of the crossover between the concentration needed to inhibit these receptors, they seem a possible target for ethanol. But again, because this effect is localized to a small area of the brain and is activated by a lower concentration of ethanol than is needed to induce intoxication, they might not be the cause of the major effects of ethanol.

The nicotinic acetylcholine receptor is another interesting target of ethanol. Ethanol potentiates the alpha7 subtype of the nicotinic acetylcholine receptor at concentrations that overlap the concentration needed to intoxicate. The alpha7 subtype is one of the most common nicotinic acetylcholine receptor subtypes in the brain, the other common subtype, the alpha4beta2 subtype, is not affected by ethanol. It is almost certain that ethanol's classical "depressant" effects are not mediated by this receptor, as the classical nicotinic acetylcholine receptor agonist nicotine is not depressant. However, one effect that nicotine and ethanol share is their addictive nature. It has been demonstrated that nicotinic acetylcholine receptor antagonists inhibit the increase in dopamine release in the brain caused by ethanol (Blomqvist et al., 1997) and the rewarding effects in humans (Blomqvist et al., 2002).

Ethanol has been shown to inhibit voltage-gated sodium channels (the same channel local anesthetics inhibit), but only at very high concentrations.

It has been demonstrated that ethanol modulates potassium channels. Potassium channels are very important in reducing neuronal excitability. The G-protein coupled inward rectifying (GIRK) channel is a channel modulated by CB1-cannabinoid and Mu-opioid receptors (as well as many others). Ethanol has been shown to directly open (or at least potentiate the opening of) this channel at concentrations starting at ~20-75 mM. While this puts the GIRK channel outside the range for mediating ethanol's recreational effects, it could be responsible for some of the effects of ethanol seen at higher doses, such as analgesia. In mice lacking a functional GIRK channel, ethanol had severely reduced potency as an analgesic (Kobayashi et al., 1999). Ethanol has also been shown to enhance the activity of calcium-activated potassium (BK) channels. These channels become active when a neuron fires repetitively, slowing and eventually stopping the neuron from firing. Ethanol activates these receptors in the high range of ethanol blood concentrations, and so may be responsible for some of the depressant effects of ethanol at high concentrations.

The glycine receptor is an inhibitory receptor distributed largely in the spinal cord and brainstem. Ethanol potentiates this receptor at concentrations directly overlaying its recreational plasma concentrations. This could cause a variety of effects like inhibition of signaling from the brain to the body, and might also be responsible for respiratory depression at high doses, but is unlikely to mediate the classical cognitive of ethanol.

The effect of ethanol on voltage-gated calcium channels is one effect that has received a lot of press. Ethanol inhibits these channels at concentrations very similar to its recreational plasma concentration and has an IC50 in the middle of the recreational range. The opening of voltage-gated calcium channels is responsible for the release of neurotransmitters throughout the body and, by inhibiting this, ethanol will decrease neurotransmitter release. This inhibition causes a decrease in both excitatory and inhibitory synaptic transmission.

Ethanol has also been shown to potentiate the serotonin 5-HT₃ receptor. This receptor is somewhat enigmatic, but is believed to modulate excitation in some parts of the brain, and to facilitate vomiting. The ethanol concentrations that are needed to potentiate the 5-HT₃ receptor are outside the range reached during recreational usage.

Despite ethyl alcohol being one of the most widely used psychoactives, the pharmacology of its altering effects is still not yet fully understood. There are many systems that it is known to affect, but many of the documented neuro-physiological effects only occur at plasma concentrations outside the range that ethanol is used in humans.

Its effects on GABA-A receptors seem tantalizing, but generally unlikely, due to either a far too high or too low potency, depending on the subunit makeup. Ethanol's inhibition of NMDA receptors seems likely to be a contributor to its effects. The inhibition of AMPA and kainate receptors is too weak to mediate recreational effects, but it may be responsible for ethanol-induced coma. Ethanol's potentiation of alpha7 nicotinic acetylcholine receptors seems a likely cause of some of its rewarding effects, but not a mediator of its general "depressant" character. Ethanol's effects at voltage-gated sodium channels are far too weak to have any effect during recreational usage. BK channel activation might cause some of the effects of ethanol at high concentrations and GIRK channel potentiation could be responsible for its analgesic effects. Potentiation of glycine receptors might cause spinal inhibition and lead to some effects of ethanol, but not the classical cognitive effects. 5-HT₃ receptor effects are too weak to mediate ethanol's effects. Inhibition of voltage-gated calcium channels is another likely mediator of ethanol's effects, as it would cause general inhibition of synaptic activity.

Figure 1. The concentration ranges of ethanol's effects at the relevant pharmacological targets relative to the threshold concentration needed to induce noticeable intoxication (1 drink) and death. The start of the bar represents an approximate average of the lowest concentration needed to get a statistically significant effect across several published papers. Vertical bars represent an approximate average of the IC50/EC50 for that particular effect across several published papers. Targets without vertical IC50/EC50 bars have no studies which have shown this value. A perfect pharmacological target would have effects starting at the low recreational plasma concentration threshold and an IC50/EC50 somewhere between the low and uppter threshold.

Collection of Evidence for Ethyl Alcohol's Activity on Various Neuropharmacological Systems
Target of Activity	Evidence of Activity
Mixed GABA receptor	Dissociated rat cortical cells, no effect up to 300 mM (Ming et al., 2001) Dissociated rat lateral septal nucleus, substantia nigra, thalamus, hippocampus, and cerebellum cells, no effect up to 300 mM (Criswell et al., 2003)) Dissociated rat hippocampal cells, no effect up to 200 mM, dissociated mouse hippocampal cells 35% of cells showed a potentiation at 0.5 mM and upwards (Aguayo et al., 1994) Dissociated rat medial septum neurons, no effect up to 300 mM (Frye et al., 1994) Rat hippocampal slice neurons 70% of neurons showed potentiation from ~20 mM, max potentiation 50% (Wan et al., 1996) Rat central amygdala slice neurons 70% of neurons showed a potentiation from 10mM with an EC50 = 20 mM, max potentiation 60% (Roberto et al., 2003)
Delta containing GABA receptor	Rat receptor expressed in frog eggs, potentiation with as little as 3 mM ethanol (Wallner et al., 2003) Rat/mouse/human receptor expressed in frog eggs, potentiation starting from 0.1mM with an EC50 ~2 mM, max potentiation 80% (Sundstrom-Poromaa et al., 2002)
NMDA receptor	Dissociated Rat cortical neurons. Inhibited NMDA receptor from 6mM with an apparent IC50 of ~20 mM, max inhibition 39.5% (Ming et al., 2001) Dissociated rat hippocampal neurons, inhibited NMDA receptor from 5mM with an IC50 of ~20 mM, max inhibition 61% (Lovinger et al., 1989) Cultured rat hippocampal and cortical neurons, inhibited NMDA receptor from 5mM with an IC50 of 120 mM, max inhibition 80% (Wright et al., 1996)
Kainate receptor	Inhibited kainate receptor increased rate of GABA release in rat hippocampus from 0.5mM with an IC50 of 4.6 mM, max inhibition 99% (Carta et al., 2003) Inhibited kainate receptor in cultured cerebellar neurons from 25mM with an IC50 ~50 mM, max inhibition 22% (Valenzuela et al., 1998) Inhibited rat kainate receptors expressed in frog eggs from 30mM with an IC50 ~100 mM, max inhibition 90%. (Dildy-Mayfield & Harris, 1995)
AMPA receptor	Inhibited AMPA receptor from 20 mM with an IC50 100-300 mM in dissociated rat medial septum cells, max inhibition ~50% (Frye & Fincher, 2000) Inhibited AMPA receptor from 20 mM with an IC50 ~100 mM in cultured rat neurons, max inhibition ~50% (Wirkner et al., 2000)
Nicotinic acetylcholine receptor	Inhibited rat alpha7 receptor expressed in frog eggs from 10 mM with an IC50 of 33 mM (Yu et al., 1996) Inhibited rat alpha7 receptors expressed in frog eggs from 3 mM with an IC50 of ~20 mM (Oz et al., 2005) Rat alpha4beta2 nicotonic receptors are generally insensitive to ethanol at concentrations <300mM (Covernton & Connolly, 1997)
Voltage gated sodium channels	Inhibited rat Nav1.2 expressed in frog eggs at 50-200 mM maximum inhibition ~40% (Shiraishi & Harris, 2004)
Potassium channels	Activates rat G-protein coupled inward rectifying potassium channels at concentrations >20 mM (Kobayashi et al., 1999) Enhances rat G-protein coupled inward rectifying potassium channels at concentrations >75 mM (Chiou et al., 2002) Activates rat Ca2+ activated Potassium channels (BK) from >20 mM (Chiou et al., 2002) Activates rat Ca2+ activated potassium channels (BK) expressed in frog eggs from >10 mM, with an EC50 of 24 mM (Dopico et al., 1998)
Glycine receptor	Potentiates glycine receptor in dissociated rat VTA neurons from 1mM, EC50 of 35 mM (Ye et al., 2001) Potentiates rat glycine receptors expressed in frog eggs from >10mM (Mascia et al., 1996)
Voltage gated calcium channels	Inhibits voltage gated calcium channels (N, P/Q type) in cultured PC12 cells with an IC50 of ~15 mM (Solem et al., 1997) Inhibits voltage gated caclium channels (L-type) in cultured PC12 cells starting at 10 mM (Mullikin-Kilpatrick & Treistman, 1995)
5-HT3 receptor	Potentiates 5-HT3 receptor current in rat vagal ganglion neurons at concentrations >25 mM (Lovinger & White, 1991)