Loss of Functional A-Type Potassium Channels in the Dendrites of CA1 Pyramidal Neurons from a Mouse Model of Fragile X Syndrome

Abstract

Despite the critical importance of voltage-gated ion channels in neurons, very little is known about their functional properties in Fragile X syndrome: the most common form of inherited cognitive impairment. Using three complementary approaches, we investigated the physiological role of A-type K⁺ currents (I_KA) in hippocampal CA1 pyramidal neurons from fmr1-/y mice. Direct measurement of I_KA using cell-attached patch-clamp recordings revealed that there was significantly less I_KA in the dendrites of CA1 neurons from fmr1-/y mice. Interestingly, the midpoint of activation for A-type K⁺ channels was hyperpolarized for fmr1-/y neurons compared with wild-type, which might partially compensate for the lower current density. Because of the rapid time course for recovery from steady-state inactivation, the dendritic A-type K⁺ current in CA1 neurons from both wild-type and fmr1-/y mice is likely mediated by K_V4 containing channels. The net effect of the differences in I_KA was that back-propagating action potentials had larger amplitudes producing greater calcium influx in the distal dendrites of fmr1-/y neurons. Furthermore, CA1 pyramidal neurons from fmr1-/y mice had a lower threshold for LTP induction. These data suggest that loss of I_KA in hippocampal neurons may contribute to dendritic pathophysiology in Fragile X syndrome.

Introduction

Fragile X syndrome (FXS) is the most common form of inherited mental delay, affecting 1 in 4000 males and 1 in 6000 females. Patients with FXS display a variety of phenotypes related to the CNS, including lower cognitive function, impairments of both working and episodic memory, and placement along the autistic spectrum (Van Dam et al., 2000; Jin and Warren, 2003; Bernardet and Crusio, 2006; Dölen et al., 2007; Bassell and Warren, 2008). At the neuronal level, FXS has generally been considered a synaptic disorder. This is due to the large body of evidence showing alterations in synaptic structures and protein expression (Comery et al., 1997; Nimchinsky et al., 2001; Hou et al., 2006). However, synaptic elements are only part of the process that is involved in the regulation of neuronal signaling. Once a signal crosses the synapse, it is processed and integrated by neuronal dendrites, in particular, by the myriad of voltage-gated channels expressed in dendrites and spines (Magee, 2000; London and Häusser, 2005; Johnston and Narayanan, 2008; Sjöström et al., 2008). Investigations into the impact of FXS on voltage-gated channels and the processing ability of dendrites are surprisingly lacking. This is particularly noteworthy because FMRP, the mRNA-binding protein that is lacking in FXS, binds to the mRNAs encoding a number of ion channel proteins (Bell et al., 1991; Darnell et al., 2011).

One channel that is highly enriched in the distal dendrites of hippocampal CA1 pyramidal neurons and influences both synaptic inputs and back-propagating action potentials (b-APs) is the A-type K⁺ channel (I_KA) (Hoffman et al., 1997). This channel is thought to be from the shal or K_V4 family of voltage-gated K⁺ channels (Serôdio et al., 1994), and is critically important for normal dendritic function and excitability (Hoffman et al., 1997; Ramakers and Storm, 2002; Johnston et al., 2003). The rapid kinetics and quick recovery from inactivation allow I_KA to influence a variety of neuronal functions. Under normal conditions, I_KA regulates dendritic excitability by limiting the amplitude of both locally generated dendritic spikes and b-APs. The high density of I_KA can also reduce the magnitude of EPSPs in the distal dendrites and is subject to activity-dependent regulation with certain forms of synaptic plasticity (Frick et al., 2004; Kim et al., 2007; Lin et al., 2011). Two recently published reports using biochemical approaches arrived at conflicting conclusions as to whether expression of K_V4.2, the putative pore-forming subunit of hippocampal A-type K⁺ channels, is elevated (Lee et al., 2011) or reduced (Gross et al., 2011) in hippocampal neurons from the fmr1-/y mouse model of FXS.

Here we demonstrate that dendritic, but not somatic, I_KA is lower in CA1 pyramidal neurons from fmr1-/y mice. In addition to the reduced maximum dendritic I_KA, the voltage dependence of activation of A-type K⁺ channels is hyperpolarized in fmr1-/y neurons. Based upon the time course of recovery from inactivation, the I_KA present in fmr1-/y dendrites is likely mediated by K_V4 containing channels. This overall reduction in I_KA results in larger b-APs, more Ca²⁺ influx in the distal dendrites, and a lowered threshold for LTP induction in fmr1-/y neurons. We thus provide crucial evidence that demonstrates that reduced A-type K⁺ channel function alters hippocampal dendritic physiology in fmr1-/y mice. These deficits in hippocampal neuron function may, in part, be responsible for the cognitive impairments associated with FXS.

Materials and Methods

Preparation of acute hippocampal slices.

All experiments were conducted in accordance with the University's Institutional Animal Care and Use Committee. Hippocampal slices (300 μm) were prepared from 2- to 4-month-old male wild-type (C57BL/6) and fmr1-/y mice as described in Brager et al. (2012). Briefly, animals were deeply anesthetized with a lethal dose of ketamine and xylazine and intracardially perfused with ice-cold modified ACSF containing the following (in mm): 210 sucrose, 2.5 KCl, 1.2 NaH₂PO₄, 25 NaHCO₃, 0.5 CaCl₂, 7.0 MgCl₂, and 7.0 dextrose bubbled with 95% O₂/5% CO₂. The brain was removed and bisected along the midline, an oblique cut was made to promote the planar orientation of the dendrites, the brain was mounted to the stage of a Vibratome, and sections were made from the middle hippocampus. Slices were placed in a holding chamber with ACSF containing the following (in mm): 125 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 25 NaHCO₃, 2.0 CaCl₂, 2.0 MgCl₂, and 21 dextrose, pH 7.4, bubbled with 95% O₂/5% CO₂ at 35°C for 45–60 min and then kept at room temperature.

Electrophysiology.

Slices were placed individually, as needed, into a submerged recording chamber and continuously perfused with oxygenated extracellular saline containing the following (in mm): 125 NaCl, 3 KCl, 1.25 NaH₂PO₄, 25 NaHCO₃, 2.0 CaCl₂, 1.0 MgCl₂, and 21 dextrose, pH 7.4, at 32–34°C. Slices were viewed with a Zeiss AxioExaminer D microscope fitted with a 60× water-immersion objective and Dodt contrast optics. Patch pipettes were pulled from borosilicate glass and wrapped with Parafilm to reduce capacitance.

Cell-attached patch recordings.

For cell-attached recordings, pipettes contained the following (in mm): 125 NaCl, 3 KCl, 10 HEPES, 2.0 CaCl₂, 1.0 MgCl₂, and 0.001 TTX, pH 7.3 with KOH. Membrane currents were recorded using an Axopatch 200B amplifier (Molecular Devices), sampled at 10 kHz, analog filtered at 2 kHz, and digitized by an ITC-18 interface connected to a computer running Axograph X. To separate I_KA from total I_K, the sustained component was isolated using a depolarized holding potential (−30 mV) in which I_KA is inactivated and subtracting that from the total I_K obtained from a hyperpolarized holding potential (−100 mV). Activation curves were constructed by using depolarizing voltage commands (−90 to 50 mV in 10 mV steps) to activate I_KA from a holding potential of −100 mV. Steady-state inactivation was studied using depolarizing test pulses to a fixed voltage (50 mV) preceded by a series of prepulse conditioning potentials ranging from −100 to −30 mV. Activation and inactivation data were fit to single Boltzmann functions using a least-squares program. Recovery from inactivation was determined by measuring the I_KA in response to two steps from −100 to 50 mV. The time between the two steps was varied from 5 to 300 ms. A single exponential fit the recovery from inactivation of I_KA. Linear leakage and capacitive currents were digitally subtracted by scaling traces at smaller command voltages in which no voltage-dependent current was activated.

Whole-cell dendritic recording.

The ACSF was the same as for cell-attached experiments except 20 μm DNQX, 50 μm d-AP5, 5 μm gabazine, and 5 μm CGP55845 were included to block synaptic transmission. For whole-cell dendritic recording, pipettes were filled with the following (in mm): 120 K-gluconate, 20 KCl, 10 HEPES, 4 NaCl, 4.0 Mg-ATP, 0.3 Na-GTP, and 14 K₂-phosphocreatine, pH 7.3 with KOH. Whole-cell recordings were made using a Dagan BVC-700A in current-clamp mode. Data were sampled at 40 kHz, analog filtered at 5 kHz, and digitized by an ITC-18 interface connected to a computer running Axograph X. Series resistance (R_S) was monitored throughout the recording and the experiment was discarded if R_S exceeded 30 MΩ or varied by >20%. Back-propagating action potentials were elicited by antidromic extracellular stimulation in the stratum oriens or the alveus. The liquid junction potential was 13 mV and was not corrected.

Ca²⁺ imaging.

For Ca²⁺-imaging experiments, slices were viewed using a Zeiss Axioskop FS2 fitted with a 60× objective and differential interference contrast optics. The extracellular saline was the same as for whole-cell dendritic recording except that 20 μm BaCl₂ was included to block inwardly rectifying K⁺ channels. For Ca²⁺-imaging experiments, the pipette solution was the same as for whole-cell recording except for the inclusion of the calcium indicator bis-fura-2 (100 μm). Neurons were allowed to fill with the Ca²⁺ indicator for 20 min before imaging. Voltage signals were recorded using a Dagan BVC-700A amplifier, sampled at 20 kHz, analog filtered at 3 kHz, and digitized by an ITC-18 interface connected to a computer running custom written software in IGOR Pro. A CCD camera (Redshirt Imaging) with an 80 × 80 pixel array, 2 kHz frame rate, and single wavelength (380 nm) excitation was used to measure calcium signals in response to back-propagating action potentials (2 nA for 2 ms into the soma). Changes in [Ca²⁺]_i were quantified by calculating ΔF/F₀, where F₀ is the fluorescence intensity before stimulation (after subtracting autofluorescence) and ΔF is the change in fluorescence during neuronal activity (corrected for bleaching). The autofluorescence of the tissue was measured in a region of equal size but adjacent to the dye-filled neuron in the dendritic field and bleaching was determined by measuring the change in fluorescence at rest (without stimulation). The ΔF/F₀ measurements were repeated five times and averaged.

Weak theta-burst pairing LTP experiments.

The methods for the LTP experiments (see Fig. 7) were similar to those described previously (Watanabe et al., 2002; Chen et al., 2006). GABA_A- and GABA_B-mediated IPSPs were blocked by 2 μm SR95531 and 5 μm CGP55845, respectively. To prevent epileptiform activity, a cut was made between area CA3 and area CA1. Schaffer collateral EPSPs of 2–3 mV were elicited using tungsten-stimulating electrodes placed <20 μm from the apical dendrite ∼180–225 μm from the soma. Back-propagating action potentials were elicited by somatic current injection (1–2 nA from 2 ms). The weak theta-burst pairing (TBP) protocol consisted of a burst of 5 EPSPs at 100 Hz paired with a single somatic current injection (paired with the last EPSP in the burst) and each burst was delivered 10 times at 5 Hz. This train was repeated three times at 10 s intervals. Although there was only a single current injection paired with the burst of EPSPs, the number of elicited action potentials varied from cell to cell. Therefore, the actual number of elicited action potentials was reported for each burst.

Figure 7.

wTBP is more effective at producing LTP in fmr1-/y neurons. A, Representative traces showing the wTBP protocol. B, The mean number of APs per burst is not significantly different between wild-type and fmr1-/y neurons. C, Graph showing the time course of EPSP slope change after a wTBP protocol. Note that the LTP in fmr1-/y neurons is significantly greater than wild-type neurons 30 min after wTBP. Inset, Representative EPSPs during baseline (a) and 30–35 min post wTBP (b). For wild-type: n = 7 cells, 4 mice; fmr1-/y: n = 6 cells, 4 mice **p < 0.01 vs wild-type.

Statistical analyses.

All data were compared using Student's t test, two-way ANOVA, and linear regression where appropriate (see figure legends for specific instances). All fits were performed using Prism software (GraphPad). Data were considered significantly different if p < 0.05. All data are plotted as mean ± SEM.

Results

Fmr1-/y mice lack the distal dendritic enrichment of A-type K⁺ channels

We made cell-attached patch-clamp recordings from the soma or the distal apical dendrite (range: 175–250 μm from the soma; median: 200 μm) of CA1 pyramidal neurons from wild-type and fmr1-/y mice to measure the current through A-type K⁺ channels directly. Depolarization to 50 mV from a holding potential of −100 mV resulted in an outward current (total I_K) that had both a transient and a sustained component (Fig. 1A). Subtracting the sustained current measured when the patch was held at −30 mV from the total I_K isolated the inactivating A-type K⁺ current (I_KA, Fig. 1A,B). Consistent with the previously described increasing gradient of I_KA from soma to dendrite, we found that in wild-type neurons, the peak dendritic I_KA was significantly larger than the peak somatic I_KA (dendritic: 29.0 ± 4.72 pA, n = 10; soma: 14.0 ± 2.55 pA, n = 10; Fig. 1C, black symbols). In contrast, there was no significant difference between the peak somatic and dendritic I_KA in fmr1-/y neurons (dendritic: 13.1 ± 2.29 pA, n = 8; soma: 14.1 ± 2.07 pA, n = 7, p > 0.05; Fig. 1C, red symbols). These data suggest that the distal dendritic enrichment of I_KA present in wild-type neurons is absent in fmr1-/y mice.

Figure 1.

Maximum dendritic I_KA is smaller in fmr1-/y mice. A, Isolation of I_KA by subtraction of sustained I_K from total I_K. B, Representative A-type K⁺ currents measured at 50 mV in dendritic patches (200 μm from the soma) on wild-type (black) and fmr1-/y (red) neurons. C, Graph of the summary data showing that maximum dendritic I_KA is significantly larger than somatic I_KA in wild-type neurons, but not in fmr1-/y neurons. For wild-type: dendritic patches, n = 10 from 8 mice; somatic patches, n = 10 from 5 mice. For fmr1-/y: dendritic patches, n = 8 from 6 mice, somatic patches, n = 7 from 4 mice. *p < 0.05 vs soma; p̂ < 0.05 vs wild-type, two-way ANOVA.

A-type K⁺ channels in fmr1-/y neurons are activated at more negative potentials compared with wild-type neurons

We measured the voltage dependence of A-type K⁺ channels in somatic and dendritic patches from wild-type and fmr1-/y mice. Outward K⁺ currents were elicited by depolarizing the patches in 10 mV increments from a holding potential of −100 mV (Fig. 2A). Consistent with previously published results (Hoffman et al., 1997), we found that in wild-type neurons, the activation of dendritic A-type K⁺ channels was significantly hyperpolarized compared with somatic A-type K⁺ channels (dendritic V_1/2: −0.1 ± 2.87 mV, n = 7; soma V_1/2: 9.6 ± 2.9 mV, n = 4, p < 0.05; Fig. 2B). In fmr1-/y neurons, the activation of dendritic A-type K⁺ channels was similarly hyperpolarized compared with the soma (dendritic V_1/2: −10.9 ± 2.43 mV, n = 8; soma V_1/2: −2.9 ± 3.6 mV, n = 4, p < 0.05; Fig. 2C). Interestingly, the activation of both somatic and dendritic A-type K⁺ channels in fmr1-/y neurons was significantly hyperpolarized compared with wild-type (Fig. 2D–F). There was no significant difference in slope factor for either dendritic (k_wt: 11.3 ± 0.9; k_fmr1-/y: 9.3 ± 1.2, p > 0.05) or somatic A-type K⁺ channels (k_wt: 13.0 ± 2.2 mV; k_fmr1-/y: 10.6 ± 1.1; p > 0.05) between wild-type and fmr1-/y neurons. These data suggest that although the I_KA current density is lower in the dendrites of fmr1-/y neurons, the channels that are present activate at more hyperpolarized voltages.

Figure 2.

Voltage dependence of I_KA activation is hyperpolarized in fmr1-/y neurons. A, Representative I_KA traces (top) in response to a family of depolarizing voltage commands from −100 to 50 mV in 10 mV steps (bottom). B, C, Activation curves of transient A-type K⁺ channels recorded from somatic and dendritic cell-attached patches in wild-type (B) and fmr1-/y (C) neurons. D, E, Activation curves of transient A-type K⁺ channels recorded from dendritic (D) and somatic (E) cell-attached patches in wild-type and fmr1-/y neurons. F, Group data showing that the V_1/2 of activation in dendritic patches is hyperpolarized compared with somatic patches for both wild-type and fmr1-/y neurons. For wild-type: dendritic patches, n = 7 from 7 mice; somatic patches, n = 4 from 3 mice. For fmr1-/y: dendritic patches, n = 8 from 7 mice, somatic patches, n = 4 from 3 mice. *p < 0.05 vs soma; p̂ < 0.05 vs wild-type, two-way ANOVA.

The I_KA present in fmr1-/y dendrites is mediated by K_V4-containing channels

The inactivation of A-type K⁺ channels is voltage dependent, with the time constant of inactivation (τ_inact) increasing with depolarization (Hoffman et al., 1997). In agreement with these results, we found that in wild-type neurons, τ_inact increased with depolarization (Fig. 3A₁,B, black symbols). In contrast, τ_inact in fmr1-/y neurons was less voltage dependent (Fig. 3A₂,B) and the slope of the line fit to the voltage dependence of τ_inact in fmr1-/y patches was significantly smaller compared with wild-type (wt: 1.0 ± 0.4 ms/10 mV, n = 8; ko: 0.3 ± 0.1 ms/10 mV, n = 6, p < 0.05; Fig. 3C). Due to the reduced voltage dependence, τ_inact for maximum I_KA was significantly slower in wild-type dendrites compared with fmr1-/y dendrites (wt: 19.1 ± 3.6 ms; fmr1-/y: 10.1 ± 2.3 ms, p < 0.05, t test; Fig. 3D).

Figure 3.

The kinetics of recovery from steady-state inactivation suggests that I_KA in both wild-type and fmr1-/y neurons is mediated by K_V4-containing channels. A, Representative traces showing the slower inactivation of I_KA at 50 mV (thick trace) compared with 0 mV (thin trace) in a dendritic patch from a wild-type neuron (A₁) and an fmr1-/y neuron (A₂). The traces are scaled to peak an on an expanded time scale below. B, τ_inact of I_KA is less voltage dependent in fmr1-/y dendrites compared with wild-type dendrites. C, The slope of the fit to the voltage dependence of τ_inact for fmr1-/y patches is significantly smaller than wild-type patches. D, The mean τ_inact for the maximum I_KA is significantly faster in fmr1-/y patches (wild-type: n = 8 patches, 8 mice; fmr1-/y: n = 6 patches, 5 mice; *p < 0.05 vs wild-type). E, Steady-state inactivation of dendritic A-type K⁺ channels is not significantly different between wild-type and fmr1-/y neurons. Inset, Representative traces and voltage protocol used to measure steady-state inactivation (wild-type: n = 4 patches, 3 mice; fmr1-/y: n = 4patches, 4 mice). F, Recovery from steady-state inactivation for dendritic A-type K⁺ channels from wild-type and fmr1-/y neurons. G, Group data showing the time constant of recovery from steady-state inactivation of dendritic A-type K⁺ channels for wild-type and fmr1-/y neurons is not significantly different. Note the τ_recovery for both wild-type and fmr1-/y neurons is consistent with K_V4 containing A-type K⁺ channels. For wild-type: n = 5 from 5 mice; fmr1-/y: n = 4 from 4 mice.

We next investigated whether there are any differences in steady-state inactivation for A-type K⁺ channels between dendritic wild-type and fmr1-/y neurons. The patch was depolarized to various prepotentials (−90 to −30 mV) for 500 ms to inactivate any A-type K⁺ channels present in the patch, which was followed by a test pulse (50 mV) to measure the maximum available I_KA (Fig. 3E). There were no significant differences in steady-state I_KA inactivation between wild-type and fmr1-/y neurons (V_{1/2 WT}: −63 ± 1.3 mV, n = 4; V_{1/2 fmr1-/y}: −66 ± 2.8 mV, n = 4, p > 0.05; k_WT: 7.0 ± 0.7; k_fmr1-/y: 7.7 ± 0.5; p > 0.05; Fig. 3E).

A-type K⁺ channels can be composed of either Kv4 or Kv1 subunits (Coetzee et al., 1999). One method of discriminating between A-type currents mediated by Kv1 and Kv4 potassium channels involves measuring the recovery from steady-state inactivation: K_V1 channels recover an order of magnitude slower than K_V4 channels (Castellino et al., 1995; Jerng et al., 2004). We found that the time constant of recovery from steady-state inactivation was not different between wild-type and fmr1-/y dendritic A-type channels and was consistent with K_V4-containing channels (wt: 12 ± 3 ms, n = 5; fmr1-/y: 18 ± 4 ms, n = 4,p > 0.05; Fig. 3F,G). These data suggest that, similar to wild-type neurons, the dendritic I_KA in fmr1-/y neurons is mediated by K_V4-containing channels.

Back-propagating action potentials are larger in fmr1-/y dendrites

One physiological role of A-type K⁺ channels in CA1 pyramidal neurons is to limit the amplitude of b-APs into the distal dendritic arbor (Fig. 4A). Based upon our cell-attached patch recordings and previous data from Kv4.2 knock-out mice (Chen et al., 2006), we predicted that b-AP amplitude would be greater in the distal dendrites of fmr1-/y neurons due to lower I_KA. We measured b-APs in distal dendrites (range: 180–225 μm from the soma, median 200 μm), proximal dendrites (range: 75–150 μm from soma, median 100 μm), and in the soma in wild-type and fmr1-/y neurons. In agreement with this prediction, we found that distal dendritic b-APs were significantly larger in fmr1-/y neurons compared with wild-type neurons (wt: 25.5 ± 3.2 mV, n = 5; fmr1-/y: 38.6 ± 2.1 mV, n = 7, p < 0.05; Fig. 4B–D). Consistent with our finding that I_KA was not significantly different at the soma between wild-type and fmr1-/y neurons, we found no significant difference in action potential amplitude measured at the soma (wt: 112 ± 1 mV, n = 5; fmr1-/y: 114 ± 1 mV, n = 5, p > 0.05) nor in the proximal dendrite (wt: 83 ± 3.4 mV, n = 6; fmr1-/y: 84 ± 2 mV, n = 6, p > 0.05; Fig. 4D).

Figure 4.

Distal dendritic b-APs are larger in fmr1-/y neurons. A, Representative voltage traces showing the attenuation of b-AP amplitude between the soma and distal dendrite of CA1 pyramidal neurons. B, Recordings of b-APs from the distal dendrites of wild-type (black) and fmr1-/y (red) neurons. Note the lower input resistance (arrow) and higher rebound amplitude (arrowhead) in fmr1-/y dendrites consistent with higher dendritic h-channels in fmr1-/y CA1 pyramidal neuron dendrites. C, The representative b-APs recorded from wild-type (black) and fmr1-/y (red) dendrites (∼ 220 μm from the soma) indicated in the box in B. D, Scatter plot showing the distance-dependent attenuation of bAP amplitude for wild-type and fmr1-/y neurons. Note the larger bAP amplitude in the distal dendrites of fmr1-/y neurons compared with wild-type. For wild-type recordings, n = 5 somatic, 3 mice; n = 6 proximal dendrite, 4 mice; and n = 5 distal dendrite, 4 mice. For fmr1-/y recordings, n = 5 somatic, 2 mice, n = 6 proximal dendrite, 4 mice; and n = 7 distal dendrite, 4 mice. *p < 0.05 vs wild-type, two-way ANOVA.

b-AP-mediated dendritic Ca²⁺-influx is greater in fmr1-/y neurons

The back-propagation of action potentials leads to increases in dendritic intracellular calcium via the opening of dendritic voltage-gated Ca²⁺ channels. The lower I_KA in fmr1-/y neurons should result in greater dendritic calcium influx in response to b-APs (Chen et al., 2006). We used fluorescent imaging to measure b-AP-mediated Ca²⁺ influx in the dendrites of CA1 pyramidal neurons (Fig. 5A). Under baseline conditions, and with inwardly rectifying K⁺ channels blocked by 20 μm Ba²⁺, b-AP-mediated Ca²⁺ influx significantly decreased with distance from the soma in wild-type neurons, consistent with an increasing density of A-type K⁺ channels along the apical dendrite (Fig. 5B₁,C, black symbols). Dendritic calcium signals in fmr1-/y neurons also decreased with distance (Fig. 5B₂,C, red symbols) from the soma, but were significantly larger than those observed in wild-type dendrites 200 and 250 μm from the soma.

Figure 5.

b-AP-mediated Ca²⁺ influx is larger in the distal dendrites of fmr1-/y neurons. A, CCD images of bis-fura-2-filled CA1 pyramidal neurons from wild-type and fmr1-/y neurons. B, Representative traces showing Ca²⁺ influx (top) at the indicated regions in A in response to a single action potential elicited by somatic current injection (bottom) for wild-type (B₁) and fmr1-/y (B₂) neurons. C, Group data showing that distal dendritic b-AP-mediated Ca²⁺ influx is larger in fmr1-/y neurons (n = 7 cells in 5 mice) compared with wild-type neurons (n = 11 in 8 mice). *p < 0.05 vs wild-type; **p < 0.01 vs wild-type, two-way ANOVA.

To determine whether the difference in b-AP-mediated calcium signals was due to lower I_KA in fmr1-/y dendrites, we blocked I_KA by raising extracellular Ba²⁺ to 150 μm (a concentration sufficient to block I_KA (Gasparini et al., 2007). Extracellular Ba²⁺ reversibly increased b-AP-mediated calcium influx in wild-type dendrites (Fig. 6A,B), but not fmr1-/y dendrites (Fig. 6A,C). Furthermore, in the presence of 150 μm Ba²⁺, there was no difference in the distance-dependent decrease in b-AP-mediated calcium signaling between wild-type and fmr1-/y neurons (Fig. 6D). These data suggest that the lower expression of functional A-type K⁺ channels in fmr1-/y neurons contributes to the difference in the observed calcium signal.

Figure 6.

Block of I_KA reduces the distance-dependent decrease in b-AP-mediated Ca²⁺-influx in wild-type, but not fmr1-/y neurons. A, Representative traces showing Ca²⁺ influx 250 μm from the soma in response to a single action potential during baseline, 150 μm extracellular Ba²⁺, and washout measured from a wild-type and fmr1-/y neuron. B, 150 μm extracellular Ba²⁺ increases distal dendritic Ca²⁺ influx in a reversible manner in the distal dendrites of wild-type neurons. C, 150 μm extracellular Ba²⁺ has no effect on distal dendritic Ca²⁺ influx in fmr1-/y neurons. D, Group data showing that distal dendritic b-AP-mediated Ca²⁺-influx is not different between fmr1-/y and wild-type neurons in the presence of 150 μm extracellular Ba²⁺. *p < 0.05 vs baseline.

Induction threshold for theta-burst pairing LTP is lower in fmr1-/y neurons

Functional A-type K⁺ channels and their influence on dendritic b-APs and calcium influx play a role in determining the threshold for TBP-LTP in the hippocampus (Magee and Johnston, 1997; Watanabe et al., 2002). We previously demonstrated that after strong TBP (5 spikes per burst of 5 EPSPs), there was no significant difference in LTP between wild-type and fmr1-/y slices (Brager et al., 2012). Previous work with K_V4.2 knock-out mice demonstrated that fewer b-APs were required to induce TBP-LTP in absence of functional A-type K⁺ channels (Chen et al., 2006). We hypothesized that the larger bAPs and greater calcium influx in fmr1-/y dendrites would permit a weak TBP protocol (wTBP: 1–3 spikes per burst of 5 EPSPs) to induce greater LTP compared with wild-type neurons. We elicited EPSPs using a stimulating electrode placed 180–240 μm from the soma in stratum radiatum. This location corresponded to where we observed the greatest difference in calcium influx between wild-type and fmr1-/y dendrites (Fig. 5C). Although the number of action potentials did not always match the number of current injections, the average number of action potentials elicited per burst of EPSPs was not significantly different between wild-type and fmr1-/y neurons (Fig. 7A,B). Immediately after wTBP, both wild-type and fmr1-/y neurons showed short-term potentiation of EPSP slope (wt: 195 ± 30%, n = 7; ko: 200 ± 21%, n = 6, p > 0.05). In wild-type neurons, EPSP slope decreased over the next 20–30 min, whereas in fmr1-/y neurons, the EPSP slope remained significantly more potentiated (Fig. 7C). The amount of LTP 30 min post-wTBP was significantly greater in fmr1-/y neurons compared with wild-type neurons (wt: 161 ± 9%; ko: 223 ± 10%; p < 0.01).

Discussion

We found that there were significantly less functional A-type K⁺ channels in the distal dendrites of CA1 pyramidal neurons from the fmr1-/y mouse model of FXS. The maximum dendritic I_KA was reduced by 50% in neurons from fmr1-/y mice compared with wild-type. I_KA displayed mono-exponential, rapid recovery from inactivation for both wild-type and fmr1-/y dendrites, suggesting that I_KA was most likely carried predominantly by K_V4-containing channels. Despite smaller maximal I_KA, the A-type K⁺ channels present in fmr1-/y dendrites were more responsive than those in wild-type neurons due to a hyperpolarization of the midpoint of activation. The more negative voltage range for activation for the remaining channels may account for the smaller difference in b-AP amplitude than might be expected for a 50% reduction in dendritic I_KA. Nonetheless, the net effect of these changes in A-type K⁺ channel function was reduced attenuation of b-APs and greater b-AP-mediated calcium influx in fmr1-/y distal dendrites compared with wild-type.

Recently, there have been a few investigations into the pathophysiology of ion channels in FXS. In the auditory brainstem, FMRP interacts directly with and activates the sodium-activated K⁺ channel Slack-B (Brown et al., 2010), and the tonotopic gradient of K_V3.1 in the medial nucleus of the trapezoid body is flattened in fmr1-/y knock-out mice (Strumbos et al., 2010). In a previous study, we provided the first physiological evidence that dendritic function is altered in a model of FXS. We demonstrated that CA1 pyramidal neuron dendrites have elevated I_h due in part to higher expression of the HCN1 subunit of h-channels in neurons from the fmr1-/y mouse compared with wild-type (Brager et al., 2012). In a more recent study, a loss of BK channel function in CA3 pyramidal neurons was thought to underlie increased neurotransmitter release in fmr1-/y mice (Deng et al., 2013).

The K_V4.2-fmr1 controversy

Two recently published studies used biochemical and molecular approaches to examine changes in the expression of K_V4.2, the putative pore-forming subunit of A-type K⁺ channels in hippocampal pyramidal neurons, in fmr1-/y and wild-type mice (Gross et al., 2011; Lee et al., 2011). These two studies came to opposite conclusions. Gross et al. (2011) found that there was less K_V4.2 protein expressed in the hippocampus of fmr1-/y mice using immunohistochemistry and Western blotting. Furthermore, they demonstrated that there was reduced surface expression of K_V4.2, suggesting that there were fewer A-type K⁺ channels on the neuronal membrane. In contrast, Lee et al. (2011) found that there was more K_V4.2 protein and greater surface expression in fmr1-/y neurons. Our results are in agreement with those of Gross et al. However, there are several possible explanations for why our results differ from those of Lee et al.

First, we cannot exclude differences in mouse strains. In particular, it has been suggested that strain differences between 129 and C57BL/6 affect expression of the fmr1-/y phenotype (Paradee et al., 1999). It should be noted that Lee et al. used the 129 mouse, whereas our study and those of Gross et al. used the C57BL/6 mouse. Interestingly, much of the seminal research on the cellular neurobiology of FXS was performed on the C57BL/6 strain of fmr1-/y mice (Huber et al., 2002; Bear et al., 2004; Hou and Klann, 2004; Auerbach et al., 2011). Second, it is possible that hippocampal K_V4.2 protein expression is actually greater in fmr1-/y mice, but that there are fewer functional channels expressed on the apical dendrite. For example, there are cases where immunohistochemistry and physiology do not always agree (Hoffman et al., 1997; Kerti et al., 2012). Last, K_V4.2 protein and A-type K⁺ channels are located on individual dendritic spines on CA1 neurons (Kim et al., 2007). It is therefore possible that there is higher expression of K_V4.2 at the spines of hippocampal pyramidal neurons and lower expression on the apical dendritic trunk in fmr1-/y mice. If the increase on spines is greater than the decrease on the trunk, then it is possible that by immunohistochemistry and Western blotting, one would observe greater expression of K_V4.2 protein while observing less I_KA recorded from the apical dendrite. In addition, if there is a greater expression of A-type K⁺ channels on individual spines, then this may explain the reduced LTP expression observed in fmr1-/y that was rescued by heteropodotoxin (Lee et al., 2011).

Potential mechanisms

How does the loss of FMRP in fmr1-/y mice result in a loss of A-type K⁺ channel function? The most parsimonious possibility is that, in wild-type mice, FMRP binds to K_V4.2 mRNA and promotes translation. The absence of FMRP in fmr1-/y mice would result in a loss of K_V4.2 mRNA translation and a subsequent reduction I_KA. Alternatively, the loss of FMRP may alter the expression or function of a number of auxiliary A-type K⁺ channel subunits, including DPP6 and/or KChIP, which could result in lower I_KA. The association of K_V4 subunits with DPP6 and/or KChIP regulates both the surface expression and biophysical properties of dendritic A-type K⁺ channels (An et al., 2000; Covarrubias et al., 2008; Sun et al., 2011). We found no significant difference in the recovery from inactivation for I_KA from wild-type and fmr1-/y neurons (Fig. 2C,D). The absence of either DPP6 or KChIP should produce a longer recovery from inactivation, suggesting that A-type K⁺ channels in fmr1-/y dendrites are still in a complex with DPP6 and KChIP. We cannot rule out, however, that an overall reduction in either of these two auxiliary proteins occurs and leads to lower dendritic I_KA in fmr1-/y neurons. It is also possible that the loss of A-type K⁺ channel function is compensatory for other changes in neuronal activity. For example, inactivation of A-type K⁺ channels occurs in response to some LTP induction paradigms (Frick et al., 2004). It is therefore possible that changes in neuronal activity during development in FXS may lead to compensatory changes in A-type K⁺ channel function and/or expression.

Impact of reduced I_KA on dendritic function

What are the physiological implications of reducing I_KA in hippocampal dendrites? As we have demonstrated, there will be greater calcium influx in response to b-APs. b-APs play a crucial role in some forms of associative long-term synaptic plasticity. Large EPSPs (e.g., coincident synaptic events) can inactivate I_KA and thereby lead to a temporally restricted window of increased dendritic excitability, during which the amplitude of a single b-AP may be boosted and the threshold for dendritic spike initiation lowered (Magee and Johnston, 1997; Watanabe et al., 2002). By setting up this temporal restriction, I_KA can control the narrow time window required for some forms of long-term synaptic plasticity. We previously reported that a strong TBP protocol (five well timed action potentials per burst of EPSPs) produced the same amount of LTP in both wild-type and fmr1-/y neurons (Brager et al., 2012). Furthermore, a train of 10 theta-bursts of EPSPs (TBS, four EPSPs per burst) produces similar increases in field EPSP slope between wild-type and fmr1-/y hippocampal slices (Lauterborn et al., 2007). In this study, we found that a weak-TBP protocol (1–3 action potentials per burst, which were not well timed to the EPSPs) produced greater LTP in fmr1-/y neurons than in wild-type. This suggests that the threshold for TBP-LTP induction, which normally requires precise action potential to EPSP timing, is lower in fmr1-/y neurons, presumably due in part to the greater calcium influx per b-AP (Watanabe et al., 2002; Chen et al., 2006). Previous work demonstrated that a weak TBS (five bursts) produced less LTP of field EPSPs in fmr1-/y mice (Lauterborn et al., 2007; Lee et al., 2011). One major difference between our wTBP-LTP experiments and the weak TBS experiments of Lauterborn et al. and Lee et al. is the requirement of bAPs for TBP-LTP induction (Magee and Johnston, 1997). It is possible that, without bAPs, dendritic EPSPs alone produce insufficient calcium influx to promote LTP induction in fmr1-/y neurons. This hypothesis is consistent with our previous results showing that there is elevated I_h in fmr1-/y dendrites, which could reduce the temporal summation of dendritic EPSPs during the TBS induction train (Brager et al., 2012). The pairing of bAPs with EPSPs during the TBP protocol could overcome this reduced temporal summation and promote calcium influx. In fmr1-/y dendrites, the calcium influx is enhanced due to the lack of function A-type K⁺ channels, leading to a reduced threshold for LTP induction.

The normalization of calcium signals between the apical trunk and the smaller oblique dendrites occurs in part due to the presence of A-type K⁺ channels (Frick et al., 2003; Cai et al., 2004). It is possible that, in addition to the greater calcium influx into the distal apical dendrite, the loss of dendritic I_KA in fmr1-/y mice will reduce b-AP attenuation into oblique dendrites, leading to greater calcium signals there as well. Because the majority of synaptic inputs are found on oblique dendrites, the elevated calcium levels here would have an even greater influence on synaptic plasticity mechanisms. Furthermore, A-type K⁺ channels themselves undergo activity-dependent plasticity in response to long-term potentiation—a downregulation of I_KA due to a hyperpolarizing shift in inactivation (Frick et al., 2004) and internalization of channels from spines (Kim et al., 2007). It is not known whether I_KA plasticity occurs in fmr1-/y mice. However, even if these forms of plasticity persist in fmr1-/y mice, the influence of this shift in A-type K⁺ channel function would be greatly reduced compared with wild-type neurons simply due to the fact that there is less I_KA to begin with.

In addition to the attenuation of b-AP amplitude, the high density of I_KA influences the initiation of local dendritic spikes and the magnitude of EPSPs in the distal dendrites (Kim et al., 2007; Losonczy et al., 2008). If our observation of less I_KA on the apical dendrite extends to dendritic spines, then miniature EPSPs would be larger in fmr1-/y due to the absence of the inhibitory influence of A-type K⁺ channels (Kim and Hoffman, 2008). It should be noted that, in addition to these direct consequences of lower I_KA, there could be additional compensatory changes (e.g., altered synaptic scaling; Andrásfalvy et al., 2008) that further alter dendritic signal processing in fmr1-/y dendrites.

In summary, we have shown how a loss of A-type K⁺ channel function in CA1 pyramidal neuron dendrites in the fmr1-/y hippocampus leads to an increase in dendritic b-AP amplitude and resulting calcium influx. We further show that these changes in I_KA lower the threshold for TBP-LTP induction. Additional possible ramifications of this loss of A-type K⁺ channel function include dendritic hyperexcitability, altered intrinsic plasticity, and loss of dendritic compartmentalization. These results could provide mechanistic insight into the cellular mechanisms that may contribute to the cognitive impairments associated with FXS.