Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Nat Neurosci. Author manuscript; available in PMC 2009 November 1.
Published in final edited form as:
Published online 2009 April 12. doi:  10.1038/nn.2307
PMCID: PMC2674138

Hyperpolarization-activated cation channels inhibit EPSPs by interactions with M-type K+ channels


The processing of synaptic potentials by neuronal dendrites depends on both their passive cable properties and active voltage-gated channels, which can generate complex effects due to their nonlinear properties. In this study, we characterized the actions of the hyperpolarization-activated cation current (Ih) on dendritic processing of subthreshold excitatory postsynaptic potentials (EPSPs) in mouse CA1 hippocampal neurons. Although Ih generates an excitatory inward current that exerted a direct depolarizing effect on the peak voltage of weak EPSPs, it produced a paradoxical hyperpolarizing effect on the peak voltage of stronger but still subthreshold EPSPs. Using a combined modeling and experimental approach, we found that the inhibitory action of Ih is caused by its interaction with the delayed rectifier M-type K+ current. In this manner, Ih can enhance spike firing in response to an EPSP when spike threshold is low but inhibit firing when spike threshold is high.


Neurons actively process and integrate synaptic potentials through the actions of a wide array of voltage-gated ion channels that are often differentially expressed throughout a neuron's dendritic tree1. In some instances, the effects of voltage-gated channels on dendritic processing are relatively straightforward and well understood. For example, dendritic voltage-gated sodium and calcium channels can amplify synaptic potentials2 through the generation of local or propagated dendritic action potentials3, 4. In contrast, dendritic voltage-gated or calcium-activated K+ channels can reduce EPSP amplitude and dampen dendritic excitability5-7. However, in other cases, nonlinear interactions between dendritic voltage-gated channels give rise to complex effects that are less easily understood. In this study we focus on the paradoxical effects of the hyperpolarization-activated HCN cation channels on the processing of EPSPs in the apical dendrites of CA1 pyramidal neurons, where these channels are expressed in a gradient of increasing density with increasing distance from the soma8-11.

Unlike most voltage-gated channels, HCN channels activate with hyperpolarization and deactivate with depolarization. Their mixed permeability to K+ and Na+ ions results in a reversal potential (Eh) of approximately −30 mV, causing these channels to generate an excitatory inward current (Ih) at subthreshold potentials. These biophysical properties underlie the role of Ih as a pacemaker current in cardiac myocytes and thalamocortical relay neurons, where activation of Ih following action potential repolarization generates a depolarizing current that drives spontaneous, rhythmic firing12, 13. In neurons that are not spontaneously active, Ih contributes a 5–10 mV depolarizing influence on the resting membrane potential and increases the resting membrane conductance (that is, it lowers the input resistance), thereby regulating the spatial and temporal integration of synaptic inputs10, 14-16.

Despite the fact that Ih provides a depolarizing current at subthreshold potentials, several studies indicate that it has a paradoxical inhibitory effect on the ability of an EPSP to trigger an action potential. Thus, enhancement of Ih – by the anticonvulsant lamotrigine17, application of dopamine18, or induction of long-term potentiation19, 20 – decreases excitability and spike firing. Conversely, downregulation of Ih – through genetic deletion of HCN121, pharmacological blockade using cesium15, 22 or the organic antagonist ZD728815, 16, or following induction of long-term depression23 or seizures24 – increases EPSP amplitude, temporal summation, and spike firing.

The inhibitory effects of Ih, by which we mean the inhibition seen when Ih is enhanced, have generally been attributed to its action to increase the resting membrane conductance. This so-called ‘shunting effect’ on the excitatory postsynaptic current decreases the amplitude of an EPSP10, 22, where EPSP amplitude (ΔVEPSP) is defined as the difference between the peak voltage of an EPSP (Vpeak) and the resting potential. However, the impact of an EPSP depends not on its amplitude but on the voltage reached at its peak, which determines whether an EPSP is suprathreshold25. Importantly, Ih exerts two opposing influences on Vpeak: its shunting effect decreases EPSP peak voltage whereas its direct depolarizing effect increases Vpeak (see Fig. 1a).

Figure 1
Experimental paradigm and effects of pharmacological blockade of Ih

In this study, we first show in a simple computational model that, in the absence of other voltage-gated conductances, Ih should be always excitatory for EPSPs negative to the Ih reversal potential—that is, the depolarizing action of Ih on Vpeak is always greater than its shunting effect. This implies that any inhibitory effect of Ih on Vpeak must be caused by its interactions with other voltage-gated conductances.

One such interaction results in an inhibitory effect of Ih on the duration of Ca2+ action potentials in the distal dendrites of CA1 pyramidal neurons26. In this instance, the depolarizing effect of Ih on the resting membrane increases the resting inactivation of N- and T-type voltage-gated Ca2+ channels, thus inhibiting the Ca2+ spikes. In principle, this effect of Ih on resting potential and resting inactivation could also explain how Ih suppresses the firing of Na+ action potentials. However, it remains unclear whether Ih can exert an inhibitory effect on Vpeak for subthreshold EPSPs.

Here, we found that Ih, through interactions with voltage-gated K+ channels, could indeed produce an inhibitory effect on peak voltage of subthreshold EPSPs. Interestingly, the influence of Ih on Vpeak depended on synaptic strength. Thus, whereas Ih shifted Vpeak to more positive potentials for weak EPSPs, Ih inhibited Vpeak for stronger, but still subthreshold, EPSPs. In other words, the effects of Ih on Vpeak crossed over from depolarizing to hyperpolarizing as a function of increasing synaptic strength, with the “crossover” potential occurring below both the reversal potential for Ih and the action potential threshold. This indicated that the net effect of Ih is essentially inhibitory as it made it more difficult for an EPSP to reach threshold. Both our computational and experimental results demonstrated that the inhibitory effect of Ih was caused by its action to depolarize the resting membrane, which enhanced the resting activation of the delayed-rectifier M-type K+ channels. Because the M-channels are under neuromodulatory control27, the influence of Ih on dendritic integration may switch from inhibitory to excitatory depending on the state of M-channel regulation. Such modulation may have important implications for regulation of long-term synaptic plasticity that contributes to learning and memory21, 28, and for the treatment of epileptic disorders in which both Ih and M-channels may play a role29-32.


Dual influence of Ih on EPSPs in CA1 pyramidal neurons

We first examined the influence of Ih on neuronal activity in mouse hippocampal CA1 pyramidal neurons by performing whole-cell current-clamp recordings of both resting membrane properties and somatic EPSPs evoked by stimulation of the Schaffer collateral pathway (Fig. 1b). In response to hyperpolarizing current steps injected in the CA1 neuron soma, the membrane voltage exhibited a depolarizing sag that is characteristic of Ih activation (Fig. 1c). We then applied focal synaptic stimulation of increasing strength to elicit EPSPs of increasing amplitude to determine the relationship between Vpeak and stimulus strength (Fig. 1d). In all experiments, inhibitory synaptic transmission was blocked using GABAA and GABAB receptor antagonists.

Next, we applied the organic antagonist ZD7288 to block Ih and repeated the above measurements of resting membrane properties, voltage sag and EPSP input-output curve. Relatively low concentrations of ZD7288 (10 μM) and short exposure times (10–15 min) were used to minimize nonspecific effects of this drug on synaptic transmission33. These conditions were sufficient to eliminate the voltage sag in response to hyperpolarizing currents, indicating effective block of Ih (Fig. 1c). The average sag ratio decreased from 10.2 ± 1.0% under control conditions to −3.1 ± 0.5% in the presence of ZD7288 (n=7, p < 0.001, paired t-test). Application of ZD7288 also shifted the resting membrane potential (RMP) by ~ 5 mV to more negative voltages (RMP under control conditions equaled −68.9 ± 1.5 mV; RMP in ZD7288 equaled −74.0 ± 1.4 mV; n=7; p < 0.001, paired t-test), and increased the input resistance by > 2-fold (control: 138.6 ± 8.2 MΩ; ZD7288: 287.8 ± 25.1 MΩ, n=7; p < 0.001, paired t-test), consistent with previous findings19, 23, 24, 26, 34.

A comparison of EPSP input-output curves in the presence and absence of ZD7288 showed that the effects of Ih on peak EPSP voltage depended on EPSP size (Fig. 2). For small EPSPs, the presence of Ih increased Vpeak, shifting it to more positive potentials as expected for an inward, excitatory current (Fig. 2a, 30 μA stimulus). However, as the stimulus strength was increased to evoke larger EPSPs, Vpeak in the absence of Ih approached its value in the presence of Ih (Fig. 2a, 45 μA stimulus). Eventually with even stronger stimuli, a crossover occurred, where the presence of Ih decreased the peak EPSP voltage to values more negative than those reached in the absence of Ih (Fig. 2a, 60 μA stimulus). This depolarizing/hyperpolarizing crossover effect was clearly seen when Vpeak was plotted as a function of stimulus strength in the presence and absence of Ih (Fig. 2b). Surprisingly, the crossover occurred for subthreshold EPSPs whose peak voltages were well below the Ih reversal potential of −30 mV, that is, at voltages where Ih provided an inward, depolarizing current.

Figure 2
Dual effects of Ih on peak voltage of subthreshold EPSPs

Subthreshold hyperpolarizing effects of Ih on Vpeak were seen in 6 out of 7 cells that we examined. The one exception occurred in a cell whose resting potential in the presence of Ih was unusually positive (−64 mV) so that spikes were evoked with small current stimuli (35 μA). In some cells, crossover occurred well below the action potential threshold, providing clear evidence that Ih can exert an unambiguously inhibitory influence on subthreshold EPSP peak voltage (see Fig 2b). In other cells, the crossover from depolarizing to hyperpolarizing effects occurred near threshold (Fig. 2c). In such cells Vpeak in the presence of ZD7288 approached or overlapped with Vpeak in the absence of drug up to potentials very near spike threshold. A slight increase in stimulus intensity could then evoke spikes in the presence but not absence of ZD7288. Such results are consistent with previous findings that Ih has a paradoxical effect to inhibit spiking17-19, 23.

Ih is purely excitatory when the sole active conductance

How can we explain the inhibitory effect of Ih to reduce the peak EPSP voltage at potentials negative to the Ih reversal potential? As discussed in the Introduction, whereas the depolarizing current carried by Ih should make Vpeak more positive, the shunting effect of the Ih conductance is expected to decrease Vpeak. To determine whether these two opposing effects of Ih could yield a net inhibitory influence on Vpeak, we first examined a single-compartment computational model containing only a passive leak conductance, a physiologically-realistic model of Ih10, 17, and a linear excitatory synaptic conductance modeled as an alpha function35 (Fig. 3a). We determined the Vpeak attained for different strengths of synaptic input when either the maximal conductance of Ih or the voltage at which Ih was half-maximally activated (V1/2) was varied across a range of physiologically and experimentally relevant values10, 17.

Figure 3
Computational model predicts excitatory role for Ih in absence of other voltage-gated channels

When we increased Ih by increasing its maximal conductance, the resting membrane potential was shifted to more positive values as expected. Similarly, depolarizing shifts of V1/2, which increased resting Ih, also depolarized the membrane (Fig. 3b). Enhancing Ih by either method diminished the EPSP amplitude (ΔVEPSP) for all synaptic strengths examined (Fig. 3c). These results confirmed previously reported findings for Ih10, 17, 21.

Despite the effect of Ih to decrease ΔVEPSP, the increase in Ih always had a depolarizing effect on the peak EPSP voltage, as long as the membrane potential was negative to the Ih reversal potential of −30 mV (Fig. 3d,e). Ih did shift Vpeak to more negative potentials for very large synaptic conductances that drove Vpeak positive to Eh, where Ih provided an outward current that hyperpolarized the membrane. Thus, because the threshold for firing a spike is negative to −30 mV (typically around −50 mV), our results show that Ih, in the absence of other voltage-gated conductances, was always excitatory for subthreshold EPSPs. This excitatory effect of Ih persisted even when its maximal conductance was varied by 100-fold (Fig. 3e), mimicking the range of Ih conductances reported along the somatodendritic gradient in CA1 and layer V neocortical pyramidal neurons8-11, 36. We also observed an excitatory effect on Vpeak when the magnitude of Ih was increased by a shift in its voltage-dependence of activation to more positive potentials (Supplementary Fig. 1a).

One well-characterized effect of Ih is to reduce temporal summation during a burst of EPSPs, an effect that has been attributed to the action of the Ih conductance to decrease the membrane time constant and to the hyperpolarization caused by the deactivation of Ih during the burst of EPSPs14-15. Our simulations confirmed that Ih did decrease the extent of temporal summation during a 100 Hz burst of 5 EPSPs. However, even during the burst, the net effect of Ih on membrane voltage was still excitatory, with the peak EPSP voltage during the burst reaching more positive potentials in the presence of Ih than in its absence (Supplementary Figs. 2a, 3a).

Previous studies reported a particularly strong inhibitory effect of Ih when its action to depolarize the resting membrane is compensated by injection of hyperpolarizing current or a reduction in external K+ concentration17-19, 23, 34. Such results are not surprising because when changes in resting potential are prevented the inhibitory effect of Ih to enhance the membrane conductance should predominate. Indeed, when we simulated this protocol by adjusting the leak conductance reversal potential to keep the resting potential constant, an increase in Ih had a marked inhibitory effect on Vpeak (Supplementary Fig. 1b). However, the more important question is how does Ih exert its inhibitory effect on Vpeak when the resting potential is free to adopt its intrinsic value, as occurs under physiological conditions.

Ih interacts with a K+ current to inhibit EPSPs

Our computational results above demonstrated a purely excitatory effect of Ih on subthreshold EPSPs and thus suggested that the inhibitory effect of Ih on large EPSPs observed in our experiments must involve an interaction with other voltage-gated conductances. One possibility we considered is that the resting membrane depolarization caused by the presence of Ih may increase the activation of a low-threshold voltage-gated K+ conductance, and that this interaction may have a net inhibitory impact on the peak voltage of the EPSP. To test this, we extended our computational model by adding a Hodgkin-Huxley delayed-rectifier K+ conductance35, 37 and repeated the above simulations.

In the presence of the delayed rectifier K+ conductance, an increase in Ih always depolarized the resting membrane and diminished EPSP amplitude (ΔVEPSP), as seen in the single-conductance model (Supplementary Fig. 4a,b). However, inclusion of the voltage-gated K+ conductance now revealed a clear inhibitory effect of Ih on EPSP peak voltage. For small EPSPs, Ih still exerted a net depolarizing influence on Vpeak (Fig. 4a, left), as observed in the model with Ih alone. However, for larger EPSPs, Ih now exerted a hyperpolarizing influence on Vpeak (Fig. 4a,right). This inhibitory effect was observed even when the peak EPSP voltage was negative to both the reversal potential of Ih (−30 mV) and typical values for action potential threshold (−50 mV) (Fig. 4b). Thus, these computational results reproduced the key findings of our synaptic stimulation experiments, including the presence of a subthreshold crossover voltage at which Ih shifted from having a depolarizing effect on Vpeak to having a hyperpolarizing influence (Fig. 4b). As observed in the model with Ih alone, when the resting potential was held fixed (by altering the leak conductance reversal potential), Ih had a purely inhibitory effect, reducing Vpeak at all synaptic strengths (Supplementary Fig. 4c).

Figure 4
Computational model including a Hodgkin-Huxley voltage-gated K+ conductance predicts subthreshold inhibitory effects of Ih on peak voltage of strong EPSPs

Which biophysical properties of the voltage-gated K+ conductance are required to enable the subthreshold inhibitory effects of Ih? We first examined the importance of the K+ conductance kinetics by making the rate of activation infinitely slow, such that the K+ conductance remained at its initial equilibrium value set by the resting potential during the entire time course of the EPSP. Under these conditions Ih still exerted a dual depolarizing/hyperpolarizing influence (Fig. 4c), increasing Vpeak for small EPSPS but reducing Vpeak for larger subthreshold EPSPs. In contrast, when we made the activation rate infinitely fast, so that the K+ conductance attained its steady-state level of activation instantaneously throughout the EPSP, Ih now exerted a purely excitatory effect (Fig. 4d), shifting Vpeak to more depolarized values for all EPSP sizes. Thus, the ability of Ih to inhibit Vpeak requires that the Ih-dependent enhancement in steady-state resting K+ conductance persists throughout the EPSP.

Next we examined how shifting the steady-state voltage-dependence of K+ current activation affects the ability of Ih to influence the EPSP. We found that the crossover voltage at which Ih changes from having a depolarizing influence on Vpeak to having a hyperpolarizing influence became more negative as the voltage-dependence of K+ conductance activation was shifted to more hyperpolarized potentials. Conversely, depolarizing shifts in K+ current activation properties moved the crossover voltage to more positive potentials (data not shown). Importantly, a subthreshold crossover voltage, and hence an inhibitory effect of Ih, was observed over a wide voltage range of K+ current activation parameters, indicating a robust effect.

Ih interacts with the M-type K+ current to inhibit EPSPs

Because the above computational results relied on the squid axon K+ conductance, it was important to explore whether a model incorporating a mammalian voltage-gated K+ conductance could also interact appropriately with Ih to yield subthreshold inhibitory effects. We reasoned that the KV7 M-type K+ current was a good candidate to mediate the inhibitory effects of Ih as the M-current is present in CA1 pyramidal neurons, activates at subthreshold voltages, exhibits a slow time course of activation, and displays non-inactivating gating properties38-41. We therefore examined a model containing Ih, a passive leak conductance, an excitatory synaptic input, and a model of the M-current based on previously published studies in mammalian pyramidal neurons42, 43.

We found that the M-current also enabled Ih to exert a dual depolarizing/hyperpolarizing effect on the peak voltage of subthreshold EPSPs. As observed with the Hodgkin-Huxley K+ conductance model, Ih interacted with M-current to depolarize the peak voltage of weak EPSPs but to hyperpolarize Vpeak for stronger but still subthreshold EPSPs (Fig. 4e). Also similar to the Hodgkin-Huxley K+ conductance, shifts in the V1/2 of M-current activation (Fig. 5a,b) or changes in M-current maximal conductance (Fig. 5c,d) altered the crossover voltage at which Ih began to exert an inhibitory influence. In the presence of M-current, Ih was also able to exert a net inhibitory effect on peak voltage during a burst of strong EPSPs, as observed previously14-15, although the influence of Ih by the final EPSP was minimal due to its deactivation during the burst (Supplementary Figs. 2b,c and 3b,c).

Figure 5
Computational results demonstrating that changes in M-current properties alters crossover voltage at which Ih first exerts an inhibitory influence on EPSP Vpeak

In contrast to the crossover in the EPSP input-output curves in response to changes in Ih, a comparison of EPSP input-output curves with or without M-current revealed that this K+ current exerted a purely inhibitory effect, shifting peak EPSP voltage to more negative potentials at all synaptic strengths. This inhibitory action of M-current was seen either in the absence or presence of a fixed level of Ih (data not shown). Such an effect is consistent with previous results that the M-current inhibits neural activity40-41.

So far we have considered the interaction of Ih and M-current in the context of a single compartment model. However, Ih is present in a gradient of increasing density along the apical dendritic tree of both CA1 and layer 5 pyramidal neurons, where Ih density at the distal tips of the dendrites is up to 50-fold larger than that in the soma8, 9, 11. In contrast, the precise subcellular localization of the M-type K+ channels is less clear, with some studies reporting dendritic M-currents38, 41 versus others claiming only somatic and/or axonal localization40, 44. To examine the importance of the subcellular localization of these channels, we incorporated the Ih dendritic gradient into a multicompartment model of a CA1 neuron in which excitatory inputs were targeted to the apical dendrites. Importantly, an inhibitory effect of Ih on the somatic peak EPSP voltage was still observed, regardless of whether M-current was present in dendrites or restricted to the soma (Supplementary Fig. 5b,c). When M-current was restricted to the soma, dendritic Ih exerted a purely depolarizing effect on the local dendritic EPSPs recorded at the site of synaptic input, 250 μm from the soma. However, the dendritic Ih still was able to inhibit the peak somatic voltage for large, subthreshold EPSPs. Conversely, in multicompartment models lacking an M-current, Ih produced a purely depolarizing effect on both local dendritic EPSPs and somatic EPSPs (Supplementary Fig. 5a). These results clearly show that the inhibitory effects of Ih on somatic EPSPs require the presence of M-type K+ channels but are not very sensitive to the subcellular distribution of either Ih or M-current.

Inhibitory effects of Ih prevented by M-current blockade

We next took an experimental approach to examine whether the inhibitory effects of Ih in CA1 neurons do indeed arise from its interaction with the M-current by applying the specific M-current inhibitor XE99130, 39, 40, using a drug concentration (10 μM) that does not alter synaptic transmission under the conditions of our experiments44. Application of XE991 produced a variable depolarization of the resting membrane that was large enough to lead to spontaneous firing in some CA1 neurons, consistent with the inhibitory role of the M-current. Such cells were not studied further because the spiking interfered with EPSP measurements. In those cells that did not fire spontaneously, XE991 produced a relatively small 3.4 mV (p = 0.10; n=7) depolarization of the resting membrane (Fig. 6d).

Figure 6
Pharmacological blockade of M-current caused Ih to have a purely excitatory influence

Next we examined the effects of Ih on somatic EPSPs with M-current blocked by measuring Vpeak as a function of stimulus strength, first in the presence of Ih and then following Ih blockade with ZD7288. Addition of ZD7288 in the presence of XE991 hyperpolarized the resting membrane by ~ 8 mV (RMP in XE991 alone = −65.5 ± 1.7 mV; RMP in XE991 + ZD7288 = −73.5 ± 1.8 mV; n = 7, p < 0.001, paired t-test), similar to its effect in the absence of XE991. Of particular interest, blockade of the M-current abolished the inhibitory effect of Ih, which now exerted a purely excitatory effect on the peak EPSP voltage for both weak and strong stimuli (Fig 6a-c; compare to Fig. 2 with normal M-current). This result was seen in all cells tested (n=7). In addition, when M-current was blocked, the presence of Ih now increased the excitability of the cell, as evidenced by spike firing at lower stimulus strengths in the absence of ZD7288 than in its presence (Fig 6a, right panel). These findings confirm the modeling results and further support the idea that Ih alone exerts an excitatory influence on neuronal activity and that the inhibitory effects of Ih require an interaction with voltage-gated K+ currents.


Here we found that Ih exerts dual depolarizing/hyperpolarizing effects on the peak voltage of subthreshold EPSPs as a function of synaptic strength. For weak EPSPs, Ih exerts a depolarizing effect on Vpeak. In contrast, for strong EPSPs, Ih exerts an inhibitory, hyperpolarizing effect. Whereas previous studies described an inhibitory influence of Ih on EPSP amplitude (ΔVEPSP) and firing of both Na+ and Ca2+ spikes10, 15, 17, 19, 23, 26, our results provide the first demonstration that Ih can also exert an inhibitory effect on subthreshold peak EPSP voltage. This is an important distinction because peak EPSP voltage, rather than EPSP amplitude, determines the impact of the EPSP on neuronal firing25.

Our results also provide insight into the mechanism of the paradoxical effect by which the depolarizing inward Ih can produce a net hyperpolarizing effect on peak EPSP voltage. First, using a simple computational model, we found that Ih acting in the absence of other voltage-gated channels exerts a purely depolarizing effect on the peak membrane potential achieved by subthreshold EPSPs. This indicates that the direct excitatory effect of Ih to depolarize the membrane predominates over its inhibitory effect to increase resting membrane conductance. This direct excitatory effect of Ih also underlies its classical contribution to the pacemaker depolarization that generates rhythmic firing in both cardiac myocytes and certain CNS neurons, such as thalamic relay neurons12, 13.

In contrast to the results of the simple model in which Ih is purely excitatory, we found that Ih produced an inhibitory effect on Vpeak of large, subthreshold EPSPs in models containing delayed-rectifier voltage-gated K+ channels. The excitatory to inhibitory crossover effect of Ih on peak EPSP voltage that occurred in such models provides an interesting example of the nonlinear interplay of voltage-dependent currents. The depolarization of the resting membrane by Ih enhances the resting voltage-gated K+ conductance beyond that attained in the absence of Ih. At the peak of a weak EPSP, the outward driving force on K+ is quite small compared to the large inward driving force on current through Ih channels, causing the direct depolarizing effect of Ih to be dominant. In contrast, at the peak of large EPSPs, the outward driving force on K+ is increased and the inward driving force on Ih is decreased. As a result, the inhibitory effect of the K+ current is dominant. Our computational results further suggest that the inhibitory effect of Ih requires that the K+ current kinetics be relatively slow compared to the time course of the EPSP. Under these conditions, the inhibitory effect of Ih to enhance the resting K+ conductance can persist throughout the EPSP.

Both our experimental and computational findings implicate the M-type K+ current as the likely mediator of the inhibitory effects of Ih in CA1 pyramidal neurons. The relatively negative voltage range of activation of the M-current allows it to respond to the small changes in resting potential mediated by Ih; the slow activation kinetics of the M-current ensure that such changes in resting activation influence the peak EPSP voltage30, 38, 40, 41. Moreover, in experiments where we blocked the M-current with XE991, Ih exerted a purely excitatory effect, confirming that M-current is necessary for the inhibitory effect of Ih. One other important computational result is that the inhibitory effect on somatic peak EPSP voltage caused by the interaction of Ih and M-current does not depend on channel distributions within the somatodendritic compartments. This is important as Ih is known to be present in a gradient of increasing density in apical dendrites whereas it is unclear whether M-current is restricted to axo-somatic compartments or is also present in dendrites38, 40, 41.

In contrast to the dual depolarizing/hyperpolarizing effects of Ih on Vpeak, all manipulations that increased the M-current, whether in the absence or presence of Ih, hyperpolarized Vpeak, with no crossover effect. This is consistent with a large number of previous studies showing an inhibitory influence of the M-current40, 41. The purely inhibitory nature of the M-current arises because its two actions to enhance membrane conductance and to generate a hyperpolarizing outward current both act in the same direction to inhibit peak EPSP voltage and neuronal firing. This is in contrast to Ih, whose direct depolarizing and shunting effects have opposing influences.

Interestingly, both Ih and M-current can be modulated by neurotransmitters and second messenger cascades, raising the possibility that the mode of action of Ih on dendritic integration can be tuned from inhibition to excitation. For example, both cyclic nucleotides45 and phosphatidylinositol (4,5)-bisphosphate (PIP2)46, 47 shift the voltage dependence of Ih activation to more positive potentials. As we reported above, upregulating Ih has a depolarizing effect on small EPSPs but an inhibitory effect on large EPSPs; downregulation of Ih leads to the opposite outcomes. Similarly, a loss of M-current, as occurs with muscarinic receptor stimulation48 and PIP2 depletion49, will drive Ih into a purely excitatory mode of action on Vpeak. In contrast, a large increase in M-current, as occurs in response to an increase in cAMP32, will cause Ih to exert a predominantly inhibitory effect on Vpeak (see Fig. 5). Changes in both the M-current and Ih have been implicated in epileptic diseases24, 29, 31, 32, and understanding how these two currents interact to regulate excitability may ultimately be important for developing new therapeutic approaches.

The dual depolarizing/hyperpolarizing effects of Ih on peak EPSP voltage have interesting implications for how this current may differentially regulate neuronal firing depending on the state of excitability of a neuron. Under conditions where spike threshold is low and negative to the Ih crossover voltage for inhibition, manipulations that enhance Ih will increase Vpeak and, thus, have an excitatory effect on the ability of an EPSP to trigger an action potential. In contrast, when spike threshold is high and positive to the crossover voltage, manipulations that enhance Ih will decrease Vpeak and, thus, inhibit the ability of an EPSP to elicit an output. Thus, the polarity of the effect that a change in Ih exerts on neuronal firing will depend on the overall excitability of the cell. Even if Ih and Im remain constant, the effect of their interaction on neuronal output can shift from excitatory to inhibitory as a result of modulatory changes in other voltage-gated conductances that alter spike threshold. Such nonlinear subthreshold interactions among voltage-gated channels provide a rich variety of mechanisms to fine-tune the relationship between excitatory synaptic input and neuronal output.


Slice Electrophysiology

Whole-cell recordings were obtained from hippocampal CA1 pyramidal cells in submerged horizontal brain slices from P28–P40 mice. Recordings were performed at 31°C–33°C with inhibitory transmission blocked by GABAA (2 μM gabazine) and GABAB receptor antagonists (1 μM CGP-55845). Stimulating current pulses (0.1–0.2 ms) were applied through focal extracellular electrodes with a constant current generator once every 15 s. For graded stimulation, current amplitude was adjusted to evoke an EPSP in control conditions and then incremented until spike threshold was reached. Identical current pulses were reapplied after addition of 10 μM ZD7288 to block Ih. All procedures conformed to US National Institutes of Health regulations and were approved by the Institutional Animal Care and Use Committees of Columbia University and the New York State Psychiatric Institute. See Supplementary Information for full details.

Statistical Analysis

Average sag ratio expressed as [(1 − ΔVss / ΔVmin) × 100%] where ΔVss = RMP−Vss and ΔVmin = RMP−Vmin. Comparisons were made using paired t-tests where appropriate. An unpaired t-test was used to compare control RMP to XE RMP. p values less than 0.05 were considered statistically significant. Results expressed as mean ± S.E.

Computational Modeling

Computational models were implemented and run in NEURON50 (version 5.9; The Ih10, 17 and M-type K+ conductance40,42, 43 models were based on experimental studies. See Supplementary Information for full details.

Supplementary Material


We thank Joshua Dudman for helpful advice and for providing custom data acquisition routines written in Igor. This work was partially supported by grant MH80745 from NIH to S.A.S. L.A and M.S.G. supported in part by an NIH Director's Pioneer Award, part of the NIH Roadmap for Medical Research, through grant number 5-DP1-OD114-02. M.S.G. was supported in part by Columbia University's Medical Scientist Training Program.


1. Migliore M, Shepherd GM. Emerging rules for the distributions of active dendritic conductances. Nat Rev Neurosci. 2002:362–370. [PubMed]
2. Magee JC, Johnston D. Synaptic activation of voltage-gated channels in the dendrites of hippocampal pyramidal neurons. Science. 1995:301–304. [PubMed]
3. Golding NL, Staff NP, Spruston N. Dendritic spikes as a mechanism for cooperative long-term potentiation. Nature. 2002:326–331. [PubMed]
4. Losonczy A, Magee JC. Integrative properties of radial oblique dendrites in hippocampal CA1 pyramidal neurons. Neuron. 2006:291–307. [PubMed]
5. Frick A, Magee J, Johnston D. LTP is accompanied by an enhanced local excitability of pyramidal neuron dendrites. Nat Neurosci. 2004:126–135. [PubMed]
6. Losonczy A, Makara JK, Magee JC. Compartmentalized dendritic plasticity and input feature storage in neurons. Nature. 2008:436–441. [PubMed]
7. Ngo-Anh TJ, et al. SK channels and NMDA receptors form a Ca2+-mediated feedback loop in dendritic spines. Nat Neurosci. 2005:642–649. [PubMed]
8. Kole MH, Hallermann S, Stuart GJ. Single Ih channels in pyramidal neuron dendrites: properties, distribution, and impact on action potential output. J Neurosci. 2006:1677–1687. [PubMed]
9. Lorincz A, Notomi T, Tamas G, Shigemoto R, Nusser Z. Polarized and compartment-dependent distribution of HCN1 in pyramidal cell dendrites. Nat Neurosci. 2002:1185–1193. [PubMed]
10. Magee JC. Dendritic hyperpolarization-activated currents modify the integrative properties of hippocampal CA1 pyramidal neurons. J Neurosci. 1998:7613–7624. [PubMed]
11. Notomi T, Shigemoto R. Immunohistochemical localization of Ih channel subunits, HCN1-4, in the rat brain. J Comp Neurol. 2004:241–276. [PubMed]
12. Luthi A, McCormick DA. H-current: properties of a neuronal and network pacemaker. Neuron. 1998:9–12. [PubMed]
13. Robinson RB, Siegelbaum SA. Hyperpolarization-activated cation currents: from molecules to physiological function. Annu Rev Physiol. 2003:453–480. [PubMed]
14. Angelo K, London M, Christensen SR, Häusser M. Local and global effects of I(h) distribution in dendrites of mammalian neurons. J Neurosci. 2007:8643–8653. [PubMed]
15. Magee JC. Dendritic lh normalizes temporal summation in hippocampal CA1 neurons. Nat Neurosci. 1999:508–514. [PubMed]
16. Williams SR, Stuart GJ. Site independence of EPSP time course is mediated by dendritic I(h) in neocortical pyramidal neurons. J Neurophysiol. 2000:3177–3182. [PubMed]
17. Poolos NP, Migliore M, Johnston D. Pharmacological upregulation of h-channels reduces the excitability of pyramidal neuron dendrites. Nat Neurosci. 2002:767–774. [PubMed]
18. Rosenkranz JA, Johnston D. Dopaminergic regulation of neuronal excitability through modulation of Ih in layer V entorhinal cortex. J Neurosci. 2006:3229–3244. [PubMed]
19. Fan Y, et al. Activity-dependent decrease of excitability in rat hippocampal neurons through increases in I(h) Nat Neurosci. 2005:1542–1551. [PubMed]
20. Narayanan R, Johnston D. Long-term potentiation in rat hippocampal neurons is accompanied by spatially widespread changes in intrinsic oscillatory dynamics and excitability. Neuron. 2007:1061–1075. [PMC free article] [PubMed]
21. Nolan MF, et al. A behavioral role for dendritic integration: HCN1 channels constrain spatial memory and plasticity at inputs to distal dendrites of CA1 pyramidal neurons. Cell. 2004:719–732. [PubMed]
22. Stuart G, Spruston N. Determinants of voltage attenuation in neocortical pyramidal neuron dendrites. J Neurosci. 1998:3501–3510. [PubMed]
23. Brager DH, Johnston D. Plasticity of intrinsic excitability during long-term depression is mediated through mGluR-dependent changes in I(h) in hippocampal CA1 pyramidal neurons. J Neurosci. 2007:13926–13937. [PubMed]
24. Shah MM, Anderson AE, Leung V, Lin X, Johnston D. Seizure-induced plasticity of h channels in entorhinal cortical layer III pyramidal neurons. Neuron. 2004:495–508. [PMC free article] [PubMed]
25. Koch C, Bernander O, Douglas RJ. Do neurons have a voltage or a current threshold for action potential initiation? Journal of computational neuroscience. 1995:63–82. [PubMed]
26. Tsay D, Dudman JT, Siegelbaum SA. HCN1 channels constrain synaptically evoked Ca2+ spikes in distal dendrites of CA1 pyramidal neurons. Neuron. 2007:1076–1089. [PMC free article] [PubMed]
27. Delmas P, Brown DA. Pathways modulating neural KCNQ/M (Kv7) potassium channels. Nat Rev Neurosci. 2005:850–862. [PubMed]
28. Wang M, et al. Alpha2A-adrenoceptors strengthen working memory networks by inhibiting cAMP-HCN channel signaling in prefrontal cortex. Cell. 2007:397–410. [PubMed]
29. Chen K, et al. Persistently modified h-channels after complex febrile seizures convert the seizure-induced enhancement of inhibition to hyperexcitability. Nat Med. 2001:331–337. [PMC free article] [PubMed]
30. Jentsch TJ. Neuronal KCNQ potassium channels: physiology and role in disease. Nat Rev Neurosci. 2000:21–30. [PubMed]
31. Peters HC, Hu H, Pongs O, Storm JF, Isbrandt D. Conditional transgenic suppression of M channels in mouse brain reveals functions in neuronal excitability, resonance and behavior. Nat Neurosci. 2005:51–60. [PubMed]
32. Schroeder BC, Kubisch C, Stein V, Jentsch TJ. Moderate loss of function of cyclic-AMP-modulated KCNQ2/KCNQ3 K+ channels causes epilepsy. Nature. 1998:687–690. [PubMed]
33. Chevaleyre V, Castillo PE. Assessing the role of Ih channels in synaptic transmission and mossy fiber LTP. Proc Natl Acad Sci U S A. 2002:9538–9543. [PubMed]
34. Gu N, Vervaeke K, Hu H, Storm JF. Kv7/KCNQ/M and HCN/h, but not KCa2/SK channels, contribute to the somatic medium after-hyperpolarization and excitability control in CA1 hippocampal pyramidal cells. J Physiol. 2005:689–715. [PubMed]
35. Dayan Peter, Abbott LF. Theoretical Neuroscience: Computational and Mathematical Modeling of Neural Systems. 2005:576.
36. Santoro B, et al. Molecular and functional heterogeneity of hyperpolarization-activated pacemaker channels in the mouse CNS. J Neurosci. 2000:5264–5275. [PubMed]
37. Hodgkin AL, Huxley AF. A quantitative description of membrane current and its application to conduction and excitation in nerve. The Journal of Physiology. 1952:500–544. [PubMed]
38. Chen X, Johnston D. Properties of single voltage-dependent K+ channels in dendrites of CA1 pyramidal neurones of rat hippocampus. J Physiol. 2004:187–203. [PubMed]
39. Hu H, Vervaeke K, Storm JF. Two forms of electrical resonance at theta frequencies, generated by M-current, h-current and persistent Na+ current in rat hippocampal pyramidal cells. J Physiol. 2002:783–805. [PubMed]
40. Hu H, Vervaeke K, Storm JF. M-channels (Kv7/KCNQ channels) that regulate synaptic integration, excitability, and spike pattern of CA1 pyramidal cells are located in the perisomatic region. J Neurosci. 2007:1853–1867. [PubMed]
41. Yue C, Yaari Y. Axo-somatic and apical dendritic Kv7/M channels differentially regulate the intrinsic excitability of adult rat CA1 pyramidal cells. J Neurophysiol. 2006:3480–3495. [PubMed]
42. Kole MH, et al. Action potential generation requires a high sodium channel density in the axon initial segment. Nat Neurosci. 2008:178–186. [PubMed]
43. Poirazi P, Brannon T, Mel BW. Arithmetic of subthreshold synaptic summation in a model CA1 pyramidal cell. Neuron. 2003:977–987. [PubMed]
44. Vervaeke K, Gu N, Agdestein C, Hu H, Storm JF. Kv7/KCNQ/M-channels in rat glutamatergic hippocampal axons and their role in regulation of excitability and transmitter release. J Physiol. 2006:235–256. [PubMed]
45. Wainger BJ, DeGennaro M, Santoro B, Siegelbaum SA, Tibbs GR. Molecular mechanism of cAMP modulation of HCN pacemaker channels. Nature. 2001:805–810. [PubMed]
46. Pian P, Bucchi A, Decostanzo A, Robinson RB, Siegelbaum SA. Modulation of cyclic nucleotide-regulated HCN channels by PIP(2) and receptors coupled to phospholipase C. Pflugers Arch. 2007:125–145. [PubMed]
47. Zolles G, et al. Pacemaking by HCN channels requires interaction with phosphoinositides. Neuron. 2006:1027–1036. [PubMed]
48. Brown DA, Adams PR. Muscarinic suppression of a novel voltage-sensitive K+ current in a vertebrate neurone. Nature. 1980:673–676. [PubMed]
49. Suh BC, Inoue T, Meyer T, Hille B. Rapid chemically induced changes of PtdIns(4,5)P2 gate KCNQ ion channels. Science. 2006;314:1454–1457. [PMC free article] [PubMed]
50. Carnevale NT, Hines ML. The NEURON book. 2006