5.1 Effect of arachidonic acid on the transient potassium current


In the present study, an ultra low patch pipette concentration of AA (1 pM) consistently inhibited the maximal conductance of the transient K+ current IA in all studied preparations.

In CA1 pyramidal neurons intracellular arachidonic acid reduced the maximal conductance of the transient potassium current IA by about 45 %, while not affecting the delayed rectifier current Ik(v). The effect of arachidonic acid develops over 15 min of dialysis in agreement with diffusional exchange between the patch pipette and soma ⁄ proximal dendrites of a CA1 pyramidal neuron [Pusch, 88]. ETYA, a non-metabolizable analogue of AA requires higher concentration (100 pM) to mimic the effect of 1 pM arachidonic acid. This may probably be due to substitution of the four highly reactive double carbon bonds of AA with the much more stable triple bonds of ETYA. Higher concentration is also probably the reason for the effect of ETYA to develop faster in comparison to those of AA. As ETYA inhibits the AA-metabolizing enzymes cyclooxygenase, lipoxygenase and cytochrome P450, these results suggest a direct specific effect of AA (1 pM) or ETYA (100 pM) on IA. “Direct” should be understood as effect that is not exerted by AA metabolites and should not exclude the possibility that AA effects could be mediated by other processes or intermediate compounds, e.g. ROS production, altered redox potential or membrane fluidity.

The effect of 1 pM AA on the maximal conductance of IA is stronger in neurons from the entorhinal cortex, where IA was suppressed by 60-70 %.


1 pM AA affects clearly the inactivation of the A-channels. In addition to the reduction of the maximal conduction of IA, AA causes a significant shift (~12 mV) of steady-state inactivation to more negative potentials in hippocampal, as well as entorhinal cortex neurons. This will correspond to a decrease in the probability of the channels to be found in open state at certain test voltage.

Thus, AA shift will cause further reduction of IA by 10–40% when IA is activated from a membrane potential of -60 to -50 mV. Actually, in neurons the transient potassium current IA is comprised by Kv1.4 and Kv4.2 that belong to different gene subfamilies and the channels they form exhibit different types of inactivation. HEK-293 cell lines were transfected separately with cDNAs, coding only the pore-forming subunits of those channels. Interestingly, the effect of 1 pM AA on the maximal conductance was still present. IA was reduced by ~ 70% in Kv1.4- and by ~ 55 % in Kv4.2-transfected HEK cells. There is, at first sight, a contradiction of the findings that Kv1.4 (Shal family) channel shows insensibility to 25 µM AA in Xenopus oocytes [Villarroel, 96] and above presented results. In fact, it is already known, that the same channels in different expression systems may exhibit different values for time- and voltage-dependence or drug effects. It has been shown, for example, that kinetic properties of Kv 4.2- and Kv1.4-evoked currents depend strongly on the expression environment in experiment comparing ventricular myocytes, oocytes and HEK-293 cell lines [Petersen, 99]. Additionally, some members of Shaw family also encode A-type currents albeit with lower 4-AP sensitivity. Products of Kv3.1 and Kv3.2 genes express delayed-rectifier type currents in heterologous expression systems, while Kv 3.3 and Kv3.4 proteins express A-type currents. However, Kv3.3 currents, which are also transient when expressed in oocytes, resemble Kv3.1 and Kv3.2 currents when expressed in CHO or HEK-293 cells [Rudy, 99]. 

It must be noted that Kv1.4- and Kv4.2-elicited currents have some similarities with native currents, but in general possess quantitatively distinct properties, concerning time- and voltage-dependence. As observed by the present and other studies (Villarroel and Schwarz, 1996; Holmqvist et al., 2001), in different non-neuronal expression systems the reduction of the Kv-α-currents maximal amplitude by arachidonic acid isn’t accompanied by kinetic changes. 


One can assume that this is due to absence of important auxiliary subunits and/or cytosol factors in the expression model system. In agreement, Kv1.4-transfetced HEK-293 cells exhibited A-type current with distorted inactivation in control conditions.

In addition, both Kv1.4 and Kv4.2 current inactivation is significantly slower compared to neuronal A-current. Moreover, in my study, the Kv-α-evoked currents inactivation in HEK 293 cells is in general insensitive to arachidonic acid. Taken together with the strong reduction of the maximal conductance of Kv-α-currents by 1 pM AA, one can hypothesize that there are at least two action sites for arachidonic acid to modulate A-channels and at least one of them is situated on or in the vicinity of the α-subunits. It can be also concluded that neuronal regulatory subunits are necessary for AA effect. Holmqvist et al, 2001, found out that KChIP1 is required for Kv4 current modulation by 10 µM arachidonic acid in Chinese hamster ovary (CHO) cells and Xenopus oocytes.

5.2 Oxidation is involved in arachidonic acid-mediated modulation of the A-current

Rationale for studying the oxidation as a mediating process for arachidonic acid modulation of potassium channels was its photo- and air-sensitivity. Unsaturated fatty acids are highly susceptible to oxidative breakdown (Frankel, 1984). PUFAs can autooxidize to form hydroperoxide, carbon or cyclic peroxy radicals. Therefore some researchers add ascorbic acid to protect arachidonic acid in solution from autooxidation. Further motivation was the fact that the concentration of arachidonic acid in effectively suppressing transient potassium current in my experiments is 1 pM. This extremely low concentration supposes very specific arachidonic acid-channel protein interaction or alternatively, an amplifying reaction, which would be in this case a radical chain reaction.


Moreover, oxidative stress is emerging as a cause and consequence of several neurodegenerative diseases, including Alzheimer’s disease, temporal lobe epilepsy, Parkinson’s disease, etc. Entorhinal cortex shows neuronal loss early in AD (layer II) and TLE (layer III). CA1 neurons from the hippocampus are damaged late in both diseases. Therefore, in this study the susceptibility of A-type potassium current to oxidative modulation in ECLII stellate cells, ECLIII and CA1 pyramidal neurons was compared. A-currents from the entorhinal cortex neurons are more sensitive to 1 pM AA.

Moreover, studies with antioxidants conclude involvement of oxidation in arachidonic acid effects. Both of the described AA effects, reduction of maximal conductance and shift of steady-state inactivation, are independently affected by GSH, ascorbic acid, Trolox and H2O2. These results further support the hypothesis that the effects of AA are mediated by at least two oxidation sites of the channel protein. The oxidation sites are most likely intracellular; as intracellular AA and membrane-impermeable antioxidants (Trolox, GSH) applied from the inside have strong effects. Surprisingly, intracellular application of 20 mM GSH did not block the effect of AA in CA1, ECLIII and ECLII neurons from brain slice, as reported for hippocampal neurons in primary culture [Bittner, 99; Muller, 02], but strongly enhanced it, particularly at 3 min of whole-cell recording. In contrast, combination of GSHi with extracellular application of ascorbic acid (0.4 mM) significantly inhibited the reduction of IA by AA in CA1 pyramidal neurons, but not in neurons from the entorhinal cortex. Shift of steady-state inactivation to more negative potentials was not affected by GSHi or by the combination of GSHi and ascorbic acid in CA1 or ECLII neurons, but was blocked in ECLIII pyramidal neurons. Additionally, GSH and ascorbic acid affected voltage dependence of activation in neurons from acute slice. Enhancement of the AA effect by some antioxidants may be mediated by catalysing the underlying reactions between AA and other radicals in the slice tissue and channel protein.

Antioxidants have some capacity to buffer oxidants but in this process they themselves become oxidants that may, in turn, more effectively oxidize certain targets. For example, GSH reductase becomes oxidized by the reduction of GSSG to GSH and, in turn, is reduced by oxidizing NADP+ to NADPH, thereby acting as a catalyst of the reaction between NADP+ and GSSG.


Very interesting and important finding of this study is the reduction of the transient and the delayed rectifier currents by antioxidants. In CA1 pyramidal neurons, GSHin did not affected the maximal amplitudes of these currents, but shifted the voltage-conductance dependence in AA-similar way. Ascorbic acidout in turn dramatically suppressed IA and IK(V) amplitudes, without affecting their voltage dependences of activation and inactivation. This suggests that potassium channels can sense redox changes.

Trolox applied from the inside blocks the reduction of the maximal amplitude of IA by AA only in ECLIII pyramidal neurons, but abolishes AA effect on the voltage-conductance dependence inactivation in CA1 and ECLII neurons. In these two types of neurons Troloxin had the same enhancing influence on AA effect as GSH. Interestingly, intracellular Trolox by itself reduced IA in CA1 pyramidal neurons and produced AA-similar shift of inactivation voltage-dependence. On the other hand, Trolox applied both inside and outside of the membrane significantly recover the AA-mediated reduction of IA with prolonged recording time only in CA1 pyramidal neurons. Additionally, Trolox in/ out completely block the effect of arachidonic acid on the maximal conductance of Kv1.4 and Kv4.2 channels. This further supports the reduction of maximal IA conductance and shift of steady-state inactivation being independent modifications of the channel protein. Reduction of superoxide anions to H2O2 by superoxide dismutase gives an electroneutral molecule that much better permeates into lipid membranes (Nelson and Cox, 2005). Oxidation of channel proteins may be facilitated by antioxidants in a similar way. Such a view is further supported by the effects of H2O2 on IA.

5.3 H2O2 as a model oxidative modulation of ion channels

Intracellular application of H2O2 reduced IA more effectively than AA in CA1 pyramidal neurons, but to the same extent as AA in neurons from the entorhinal cortex. 


Furthermore, H2O2 strongly reduced IK(V) in CA1 neurons. It has been shown that H2O2 dose-dependently enhances the cell membrane associated PLA2 activity and stimulates AA-release [Chakraborti, 93; Yasuda, 99]. Therefore, it is possible that the stronger effect of H2O2 is due to stimulating of AA release that in turn additionally reduces the A-current. The effect of H2O2  on the maximal conductance was abolished by antioxidants only in CA1 pyramidal neurons, whereas entorhinal cortex neurons showed no recovery.

On the other hand H2O2 shifted voltage-dependence of inactivation to the same extent as AA.

Unlike AA, H2O2 also reduced the maximal conductance of the delayed rectifier current IK(V) in a GSHi-sensitive way, further supporting that the interaction of AA with IA is highly specific. 

5.3.1 Culture vs. Slices 


The reduction of IA by extremely low concentrations of intracellular arachidonic acid or by H2O2 is in agreement with experiments, conducted with cultured hippocampal neurons [Bittner, 99; Muller, 02]. Therefore, it was expected antioxidants to block arachidonic acid effects in acute slices, as it was reported for neurons from primary hippocampal culture [Bittner, 99; Muller, 02]. Unexpectedly, enhancement and acceleration of arachidonic acid effect by several antioxidative agents in brain slice were observed. This suggests fundamental differences of oxidative regulation of the A-current and Ikv in cultured hippocampal neurons, compared to neurons from brain slice.

It is possible that GSH is oxidized to GSSG that, in turn, more effectively oxidizes cystein, methionine or histidine residues in the channel protein that are not accessible to GSSG in cultured neurons because of subunit composition or differential expression of other interacting proteins. This may have implications for the relevance of excitotoxicity and oxidative stress research using neuronal cultures.

Neuronal cultures are criticized for abnormal neuronal behaviour that may result from developmental maturation under artificial conditions, leading many to favour acute brain slice preparations or in vivo recording. For example, inhibitory synapses in dissociated neuronal cultures exhibit abnormal physiology from postnatal day 6–15, continuing to show properties of early (postnatal day 1–5) postnatal synapses even after 13–21 days in vitro [Henneberger, 05]. Cortical oscillatory behaviour in vivo does not require GABA A-ergic transmission and disappears at postnatal day 6–7 [Garaschuk, 00], but develops much more slowly in cortical cultures only after days 9–15 in vitro and requires GABAA receptor activation [Opitz, 02].


In one hand, brain slices are usually maintained in a 95% O2 atmosphere in order to avoid O2-dependent cell damage at a depth of 90–150 μm below both surfaces of the slice, as observed for an interface slice configuration at 20% O2 with histology [Bingmann, 82]. On the other hand, in experiments concerning oxidative modulation of channel in brain slice neurons, 95% O2 is an important issue. Actually, with an O2 partial pressure of 600 mmHg (79%) in the perfusion, microelectrode measurements have shown steep gradients of O2 partial pressure in the bath adjacent to the slice surface, with an O2 pressure gradient inside brain tissue of 400-µm slices at 24°C from 60 to 15 mmHg, i.e. in a normal physiological range for neurons only 50 µm below the surface and a hypoxic range in the centre of the slice, with observations of vacuolization of cytoplasm even in 300-μm-thick slices [Bingmann, 82].

The work of Bingmann and Kolde (1982), comprising data from 300–1000-μm thick slices at temperatures from 24 to 35 °C, shows somewhat lower O2 pressure at the submerged side of slices in comparison to the gas interface side, indicating even less efficient O2 supply in the submerged configuration. In order to avoid tissue hypoxia, recordings from submerged slices are usually performed at temperatures of 24–32°C, significantly below 37°C [Colbert, 99; Garaschuk, 00; Henneberger, 05]. This appears to be essential for healthy brain slice electrophysiology and histology, typically showing excellent agreement with in vivo findings.

5.3.2 Role of membrane fluidity

Lipid molecules surround an ion channel in its native environment of cellular membranes. However, the importance of the lipid bilayer and the role of lipid/protein interactions in ion channel structure and function are not well understood. Rod MacKinnon group the presence of negatively charged lipids is required for ion conduction through the potassium channel, suggesting that the lipid bound to the channel protein is important for ion channel function [Valiyaveetil, 02]. They also recently demonstrated that the bacterial potassium channel KcsA binds a negatively charged lipid molecule and that lipids are required for the in vitro refolding of the channel tetramer from the unfolded monomeric state.


Moreover, in my experiments Trolox, that is known to be a potent antioxidant against lipid peroxidation, had the strongest effects in reduction of AA-induced reduction of IA, particularly when applied on both sides of the membrane. Therefore, it is not excluded that arachidonic acid/ROS influence channel protein functions through changing annular and/or bulk fluidity of the membrane, although the hypothesis that arachidonic acid may affect Kv channels in oocytes by affecting the membrane fluidity was rejected by Villaroel and Schwarz in 1996 [Villarroel, 96].

5.3.3 Arachidonic acid as a retrograde messenger

The participation of arachidonic acid and its metabolites in retrograde signaling and in other forms of local modulation of neuronal activity has been proposed [Williams, 89; Kandel, 92]. It is well known that stimulation of glutamate receptors evokes arachidonic acid release in a variety of neural cell preparations. In agreement to that, nonselective PLA2 inhibitors prevent induction of LTP, while application of arachidonic acid (or other unsaturated fatty acids) to hippocampal slices causes a slow-onset enhancement of synaptic transmission that resembles LTP. It is known that the fatty acid may increase glutamate release from hippocampal nerve terminals, block glutamate uptake, or potentiate NMDA receptor current [Freeman, 90]. Alternatively, it may act by enabling presynaptic glutamate receptors to produce enhanced glutamate release. After the discovery of the prostaglandins (PGs) in the CNS, much attention has been given to the role the eicosanoids may play in modulating neurotransmission, by interacting with presynaptic or with postsynaptic PG receptors. The existence of such receptors is well-demonstrated, and, in the brain, high-affinity binding sites have been described for both PGE2 and PGD2 [Yumoto, 86].

On one hand, direct AA retrograde messaging is most likely not an issue, as the applied concentration of AA is extremely low (1 pM) and apparently not sufficient for building a gradient, that could facilitate AA diffusion through the neuronal membrane. On the other hand, various enzymes stay alert for immediate conversion of AA to one of its many metabolites, which had also been shown to play some role in induction of LTP.


1 pM AA most probably exert effects on synaptic transmission by attenuating the maximal conductance and diminishing the number of open A-type channels.

Kv4.2 and Kv1.4 are known to be substrates for all 4 enzymes, which participate in LTP: PKA, PKC, CaMKII and erk MAPK. On one hand, arachidonic acid can stimulate MAPK, probably through PKC [Hii, 95]. On the other hand it has been shown, that ROS could activate erk-s [Baas, 95] and PKC [Gopalakrishna, 89]. It is not excluded, that arachidonic acid modulates phosphorylation of channel proteins by these kinases via ROS production. However, the exact role of arachidonic acid in enhancing neurotransmission remains to be established. If AA is released in presynaptic terminals and suppresses IA then Ca2+ influx into presnypatic terminals may be augmented and transmitter release subsequently increased. Expression of IA has been described for mossy fiber terminals in CA3 and 4-AP has a strong effect on presynaptic Ca2+ uptake and on glutamate release.

5.3.4 Significance of potassium channel modulation by ROS

Oxidation-derived radical species were once considered to be metabolic by-products that have mostly negative function. In continuously growing body of evidence, ROS are recognized to have roles in cell signalling and to take part in number of cellular functions, such as O2 sensing, cell proliferation, and apoptosis [Schafer, 01]. Numerous K+-channels and other transport mechanisms are in fact redox-regulated [Kourie, 98]. Redox response of K+-channels may be a regulation mechanism to changes in the cellular metabolism.


Oxidizing and reducing agents produce functional changes, mostly of the channel protein itself. Typically, the gating characteristics are altered while the permeation properties remain largely unchanged.

Furthermore, each channel complex may contain multiple oxidation targets that determine the effect specificity. It is possible that different experimental reagents attack distinct target residues.

For example, cysteine oxidation decreases the overall channel activity by decreasing the number of channels available to open by driving the channels to a very long-lived closed state [Soto, 02]. Additional complexity in the oxidant sensitivity of K+-channel is conferred by oxidation of methionine residues. Reversible oxidation of methionine has been suggested to play a role both as a regulatory mechanism and also as an endogenous oxidant scavenger system [Hoshi, 01].


The pore-forming subunits Kv1.4 and Kv4.2 contain many different situated cysteine and methionine residues, and differential oxidation of these residues in HEK-293 cells may account for the diverse results reported here.

Oxidative regulation is also reported for many Kv channels both in native and heterologous systems [Rozanski, 02]. N-type inactivation of Kv1.4 is reversibly redox-modulated by means of a cysteine in the distal part of the N-terminal inactivating ball domain [Ruppersberg, 91]. Auxiliary subunits which specifically co-assemble with Kv-α, also provide cysteine-containing N-terminal ball domains to equip Kv-α with redox-dependent inactivation. Potentially, this may allow for greater precision in regulation, as thus metabolic and electric events are coupled. Still, experimental evidence to support this idea is not yet available. 

5.3.5 Arachidonic acid, oxidative stress and neurodegenerative diseases

Oxidative stress has met continuously growing interest for its roles in the physiology and pathophysiology of immune defense, ischemia, and neurodegenerative disease. Reactive oxygen species modify ion channels and other transport mechanisms [Kourie, 98].


In this study, entorhinal cortex (EC) layer II stellate neurons (that show neuronal loss early in AD) and layer III pyramidal neurons (are shown to be damaged early in TLE) are compared to hippocampal CA1 pyramidal neurons (that are damaged late in TLE and AD).

Furthermore, the present study supported the finding that antioxidants in particular cases could affect ion channel functions alone or by increasing the AA-induced reduction. Clearly, AA as well as H2O2, particularly in older age or in neurodegenerative disease [Auerbach, 97], will strongly affect K+-currents and, via depolarization and inactivation of Na+-channels, dendritic excitability and spike-timing-dependent synaptic plasticity / learning [Giese, 98]. Antioxidative therapy as a part of neurodegenerative condition treatment program may then have negative effects in this context.

5.3.6 Mechanism of arachidonic acid modulation of Kv channels still remain to be discovered

Members from the same Kv families can co-assemble to form functional channels. Functionally, this has crucial implications for both physiology and pharmacology of the final K+-channel expressed in the cell membrane. Therefore, one must exercise considerable caution when extrapolating data for a single channel subunit to a “real” channel in a “real” cell membrane. The complexity increases further, as K+-channels can associate with a large number of 'accessory subunits', which can alter the properties of the resulting oligomeric channel even more.


Indeed, the total number of different subunits is even larger since many of the subunit genes undergo RNA processing, such as alternative splicing resulting in multiple protein products from each gene. Additionally, the diversity of K+ channels is magnified by alternative splicing, RNA editing and posttranslational modifications [Coetzee, 99].

A difficult question that is yet to be addressed adequately is where and how arachidonic acid and/or ROS could modulate Kv channel function.

© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML-Version erstellt am: