Decreased A-currents in hippocampal dentate granule cells after seizure-inducing hypoxia in the immature rat


  • Bi-Wen Peng,

    Corresponding author
    1. Department of Physiology, Wuhan University School of Basic Medical Sciences, Wuhan, Hubei, China
    • Address correspondence to Bi-Wen Peng, Department of Physiology, School of Basic Medical Sciences, Wuhan University, Donghu Rd185#, Wuhan 430071, Hubei, China. E-mail: and Russell M. Sanchez, Department of Surgery, College of Medicine, Texas A&M Health Science Center, 1901 S. 1st St., Bldg. 205, Temple, TX 76504, U.S.A. E-mail:

    Search for more papers by this author
  • Jason A. Justice,

    1. Department of Surgery, College of Medicine, Texas A&M Health Science Center & Central Texas Veterans Health Care System, Temple, Texas, U.S.A
    Search for more papers by this author
  • Xiao-Hua He,

    1. Department of Physiology, Wuhan University School of Basic Medical Sciences, Wuhan, Hubei, China
    Search for more papers by this author
  • Russell M. Sanchez

    Corresponding author
    1. Department of Surgery, College of Medicine, Texas A&M Health Science Center & Central Texas Veterans Health Care System, Temple, Texas, U.S.A
    • Address correspondence to Bi-Wen Peng, Department of Physiology, School of Basic Medical Sciences, Wuhan University, Donghu Rd185#, Wuhan 430071, Hubei, China. E-mail: and Russell M. Sanchez, Department of Surgery, College of Medicine, Texas A&M Health Science Center, 1901 S. 1st St., Bldg. 205, Temple, TX 76504, U.S.A. E-mail:

    Search for more papers by this author



Cerebral hypoxia is a major cause of neonatal seizures, and can lead to epilepsy. Pathologic anatomic and physiologic changes in the dentate gyrus have been associated with epileptogenesis in many experimental models, as this region is widely believed to gate the propagation of limbic seizures. However, the consequences of hypoxia-induced seizures for the immature dentate gyrus have not been extensively examined.


Seizures were induced by global hypoxia (5–7% O2 for 15 min) in rat pups on postnatal day 10. Whole-cell voltage-clamp recordings were used to examine A-type potassium currents (IA) in dentate granule cells in hippocampal slices obtained 1–17 days after hypoxia treatment.

Key Findings

Seizure-inducing hypoxia resulted in decreased maximum IA amplitude in dentate granule cells recorded within the first week but not at later times after hypoxia treatment. The decreased IA amplitude was not associated with changes in the voltage-dependence of activation or inactivation removal, or in sensitivity to inhibition by 4-aminopyridine (4-AP). However, consistent with the role of IA in shaping firing patterns, we observed in the hypoxia group a significantly decreased latency to first spike with depolarizing current injection from hyperpolarized potentials. These differences were not associated with changes in resting membrane potential or input resistance, and were eliminated by application of 10 m 4-AP.


Given the role of IA to slow action potential firing, decreased IA could contribute to long-term hippocampal pathology after neonatal seizure-inducing hypoxia by increasing dentate granule cell excitability during a critical window of activity-dependent hippocampal maturation.

Cerebral hypoxia is a major cause of neonatal seizures, and such seizures increase the risk of later epilepsy (Aicardi & Chevrie, 1970; Volpe, 1981; Hauser et al., 1993). Experimentally, hypoxia-induced seizures in neonatal rats are associated with immediate and subacute dysregulation of a number of ion channels that may render the hippocampus hyperexcitable during this critical maturational period (Jensen et al., 1991; Sanchez et al., 2001, 2005; Zhang et al., 2006; Sanchez et al., 2007; Rakhade et al., 2008). To date, most of these channelopathies have been identified in hippocampal CA1 pyramidal neurons, as these represent the major output of the hippocampus.

The dentate gyrus is in a unique position for information transfer from the entorhinal cortex to the hippocampus (Andersen et al., 1966), and can gate seizure propagation from the entorhinal cortex to the hippocampus proper (Walther et al., 1986; Collins et al., 1988; Heinemann et al., 1990; Dreier & Heinemann, 1991). A wealth of data from experimental models has suggested that compromise of this gating role of the dentate gyrus is critical to, or at least permissive of, long-term epileptogenesis (Dudek & Sutula, 2007). Therefore, aberrant channel function that renders principal dentate gyrus neurons hyperexcitable could contribute critically to epileptogenesis, but such changes consequent to neonatal seizure-inducing hypoxia remain largely unexplored.

Potassium channels critically regulate neuronal excitability (Pongs, 1999), and their dysfunction can result in epileptiform network activity in vitro (Traub et al., 2001; Gabriel et al., 2004) and in seizures and epilepsy in vivo (Pena & Tapia, 2000; Pena et al., 2002; Misonou et al., 2004; Binder et al., 2006). In particular, the “A-current” (IA) is a rapidly inactivating current that contributes to action potential repolarization and promotes single spike firing (i.e., inhibits burst firing) in many types of neurons and cardiac muscle (Pongs, 1999; Castro et al., 2001). IA is mediated by channels that are composed of molecular subunits from the Kv1 and Kv4 potassium channel families (Birnbaum et al., 2004). In the dentate gyrus, Kv1.1, Kv1.2, and Kv1.4 α-subunits are expressed mainly in the middle third of the molecular layer (Sheng et al., 1994; Rhodes et al., 1997), and Kv4.1, Kv4.2, and Kv4.3 are expressed in the granule cell layer (Sheng et al., 1994; Serodio & Rudy, 1998). A-type potassium channel function and regulation has been reported to be altered in animal models of status epilepticus, and may critically contribute to brain hyperexcitability and epileptogenesis (Bernard et al., 2004; Ruschenschmidt et al., 2006). Whether IA channel function is altered after perinatal seizure-inducing hypoxia and could contribute to consequent epileptogenesis has not been investigated.

In the current study, we asked if IA is pathologically altered in hippocampal dentate gyrus cells by neonatal seizure-inducing hypoxia, and may serve to alter their intrinsic firing properties, potentially contributing to limbic network hyperexcitability.

Materials and Methods


Male Long-Evans rat pups (Charles River, Wilmington, MA, U.S.A.), aged postnatal day 10–17, were used for these experiments. Litters were housed with their dam on a 12-h light/dark cycle. All procedures were in accordance with National Institutes of Health (NIH) guidelines on the ethical use of experimental animals.

Hypoxia treatment

Rat pups on postnatal day 10 were taken in pairs, weighed, their rectal temperatures recorded, and each was individually placed on a heating pad in a custom-made acrylic glass chamber with an O2 sensor mounted inside and three gas inlets for N2 infusion. One pup (littermate control) of each pair was placed in a chamber that remained open to room air for the duration of hypoxia treatment of the second pup. For the treated pup, the chamber O2 concentration was lowered by N2 infusion to 7% within 30–40 s, maintained at 6–7% for an additional 4 min, lowered to 5–6% for 8 min, and then lowered by 1% per minute until the pup became apneic for 30 s, at which time the chamber lid was removed for exposure to room air. The total time of hypoxia exposure was 14–16 min. Using this approach, hypoxia-treated pups typically exhibited spontaneous convulsive seizures lasting 10–60 s beginning 2–4 min after hypoxia onset, which recurred throughout hypoxia exposure and continued for several minutes after return to room air. Each pair of rat pups was earmarked, maintenance of core temperature verified, and immediately returned to their dam prior to beginning treatment of the next pair.

Slice preparation and in vitro whole-cell recordings

Rat pups were killed by decapitation under isoflurane anesthesia. The brains were removed and immediately placed into ice-cold oxygenated artificial cerebrospinal fluid (aCSF). A transverse razor cut was made anterior to the cerebellum and perpendicular to the midbrain, and the cut end was glued to the stage of a vibratome (Leica VT1200, Solms, Germany) for slicing. Three hundred seventy micrometer slices were cut in cold, continuously oxygenated aCSF, and then incubated for at least 1 h in a custom-made holding chamber filled with continuously oxygenated aCSF at room temperature.

For recording, slices were transferred to a 0.5 ml submersion chamber (Warner Instruments, Hamden, CT, U.S.A.). The chamber was continuously superfused with aCSF (1–2 ml/min by gravity flow) at room temperature. Patch pipettes were drawn on a Flaming-Brown micropipette puller (Model P-97; Sutter Instrument Co, Novato, CA, U.S.A.) to 1–2 μm tip diameter using borosilicate glass (inner diameter of 0.68 mm, outer diameter of 1.20 mm) (AM Systems, Carlsborg, WA, U.S.A.) and the tips polished with Microforge (MF-830; Narishige, Tokyo, Japan). Filled electrodes had resistances between 3 and 7 MΩ. Whole-cell patch-clamp recordings were obtained from granule cells of the dentate gyrus under visual guidance using infrared differential interference contrast microscopy (Zeiss Axioskop FS2 w/Dage-MTI camera, Oberkochen, Germany). Voltage-clamp recordings were obtained using a Multiclamp 700A amplifier and analog signals were digitized using a Digidata 1322A (Molecular Devices, Sunnyvale, CA, U.S.A.) for acquisition to a Windows-based computer. Analog data were low-pass filtered at 2 kHz and digitized at 10 kHz. Voltage-clamp protocols were generated and data acquired to computer using CLAMPEX (PClamp; Molecular Devices). Input and series resistances were monitored intermittently throughout experiments by applying 5 or 10 mV hyperpolarizing voltage steps from a holding potential of −65 mV. Initial series resistances were estimated to be <20 megohms, and data were discarded if series resistance changed by more than 20%. For all experiments, one cell was recorded per slice, and no more than two slices per animal were used.

IA was recorded with 1 μm tetrodotoxin (TTX) and 10 mm tetraethylammonium (TEA) activated under voltage-clamp by first stepping the command potential from a holding potential (VH) of −40 to −100 mV for 150 msec to remove steady-state inactivation, followed by a series of test voltage command steps from −40 to +30 mV in 10 mV increments. IA was isolated by inserting an inactivating 50 msec voltage step to −40 mV immediately before each test step, and subtracting the currents with this protocol from those using the activation protocol (see Fig. 1). To study the voltage-dependence of IA inactivation, 150 msec conditioning voltage steps from −160 to −40 mV (in 20 mV increments) were applied prior to an activation step to +30 mV. For IA isolation, a 50 msec step to −40 mV was applied between the hyperpolarizing conditioning step and the activation step (see Fig. 2), and IA was obtained by subtracting the currents in the second protocol from the currents in the first. Boltzmann fits to pooled data were generated using the equation I/Imax = 1/(1 + exp [(V − V1/2)/k]).

Figure 1.

Decreased IA in DGCs after seizure-inducing hypoxia. IA recorded in representative DGCs in slices from a control and hypoxia-treated rat. (A) Raw currents are shown in the left and middle panels, and subtracted currents are shown in the right panels (see Materials and Methods). For the cells shown, the peak subtracted amplitude activated at 30 mV was 963 pA for control compared to 608 pA for hypoxia. (B) Summary current-voltage relationships for IA in DGCs from the control and hypoxia-treated groups showed significantly decreased IA amplitudes in the hypoxia-treated group compared to controls (control, n = 26; hypoxia, n = 38, p < 0.0001, ANOVA).

Figure 2.

Voltage-dependence of IA was unchanged in DGCs after seizure-inducing hypoxia. (A, B) Traces show the voltage-protocols and example subtracted currents used to measure the voltage-dependence of IA activation (A) and removal of inactivation (B). (C, D) Summary data and Boltzmann fits (see Materials and Methods) showed no differences between control and hypoxia-treated groups in the voltage-dependence of activation (C) or removal of steady-state inactivation (D).

For current-clamp recordings, resting membrane potential (RMP) was measured without current injection, and steady hyperpolarizing current was injected to bring the membrane potential to −80 mV. Latencies to first spike in response to depolarizing current steps were measured using CLAMPFIT (Molecular Devices) as the time from the current step onset to the onset (foot) of the rising phase of the first action potential. Action potential amplitudes were measured as the voltage difference from the foot to the peak of the rising phase, and half-widths were measured as the spike width at the voltage halfway between the foot and peak of the first action potential.


The aCSF consisted of (in mm): 126 NaCl, 3.3 KCl, 1.25 NaH2PO4, 1.3 MgSO4, 26 NaHCO3, 2 CaCl2, and 10 d-glucose, and was continuously bubbled with 95% O2/5% CO2. The internal patch solution consisted of (in mm): 145 K-gluconate, 10 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 1 ethylene glycol tetraacetic acid (EGTA), 10 KCl, 0.1 CaCl2, 0.2 NaATP and 2 MgATP (pH adjusted to 7.3; 280–290 mOsm). TTX, 4-aminopyridine (4-AP), and TEA were dissolved in distilled water at high concentration, diluted to final concentration in aCSF, and applied by bath superfusion. All drugs were obtained from Sigma-Aldrich (St. Louis, MO, U.S.A.).

Data analysis and statistics

Data were analyzed off-line using CLAMPFIT (Molecular Devices, Sunnyvale, CA, U.S.A.) on a Windows-based computer, and using AXOGRAPH 4.9 (Axon, Molecular Devices) and IGOR PRO (Wavemetrics, Lake Oswego, OR, U.S.A.) on a Macintosh computer. Data are presented as mean ± standard error of the mean (SEM). Significant differences between control and experimental treatments were determined using two-way analysis of variance (ANOVA) for comparison of IA amplitudes, and Student's t-test for other measured parameters.


Two types of dentate granule cells in the neonatal hippocampus

Dentate granule cell (DGC) properties were first characterized under current-clamp by examining voltage responses to injection of a family of 300-msec current pulses. According to the shape of the afterhyperpolarization following single action potentials (Dietrich et al., 1999), two distinct response patterns were observed, termed types I and II. Type I responses consisted of a delayed action potential followed by a distinctive large afterhyperpolarization (AHP), and were observed in the majority of cells in both the control (89/95, 93%) and hypoxia-treated (62/68, 91%) groups. Type II responses consisted of brief afterdepolarizations (ADPs) after single action potentials, and were observed in a small minority of neurons from both the control (6/95, 6.3%) and hypoxia-treated (6/68, 8.8%) groups. Given the low numbers of neurons encountered with type II responses, these were excluded from further analyses. No significant differences were observed between control and hypoxia-treated groups in resting membrane potential (p = 0.87), input resistance (p = 0.37), or action potential amplitudes (p = 0.79) (Table 1).

Table 1. Classification of granule cells by AHP/ADP
 AHP image_n/epi12150-gra-0001.pngADP image_n/epi12150-gra-0002.png
Control (n = 62)Hypoxia (n = 89)Control (n = 6)Hypoxia (n = 6)
  1. The top traces are current-clamp recordings of two types of dentate granule cells categorized according to their responses to 300 msec stepwise current injection (50 pA). Type I cells exhibited an afterhyperpolarization (AHP) after each action potential, whereas type II cells showed an apparent afterdepolarization potential (ADP). There was no difference in resting membrane potential (RMP) (p = 0.16), input resistance (IR) (p = 0.99), or action potential (AP) amplitudes (p > 0.05). (Values are means ± standard error [SE]; n is the number of cells.)

RMP (mV)−70.3 ± 0.48−69.4 ± 0.41−71.6 ± 1.62−71.3 ± 1.11
Rin (MΩ)1268.5 ± 97.031217.9 ± 108.02412 ± 51227 ± 84
AP amplitude (mV)96 ± 3.8102.3 ± 1.82114.16 ± 5.43111.33 ± 8.4

Decreased IA in dentate granule cells after seizure-inducing hypoxia

We observed decreased IA in DGCs recorded 1–5 days posthypoxia without a change in voltage-dependence of activation. Mean IA amplitudes were significantly decreased in DGCs from the hypoxia-treated group compared to control (Fig. 1; p < 0.0001, two-way ANOVA; n = 26 cells control, 38 hypoxia). Boltzmann fits to the normalized currents showed no significant difference in the voltage-dependence of IA activation between the two groups (Fig. 2C; p = 0.85). For the control group, the mean voltage of half-activation (V-act1/2) was −4.6 ± 1.2 mV and average slope factor (K) was 13.7 ± 0.1 (n = 6). For the hypoxia group, V-act1/2 = −7.4 ± 2.1 mV and K = −15.8 ± 0.7 (n = 4).

The voltage-dependence of steady-state IA inactivation removal in DGCs was not significantly different between the two groups (Fig. 2). In control DGCs, the mean voltage of half-inactivation removal (Vinact1/2) was −63.81 ±3.31 mV and the mean slope factor was 16.32 ± 3.18 (n = 7). In the hypoxia-treated group, mean Vinact1/2 was −57.97 ± 2.79 mV and the mean slope factor was 12.09 ± 2.65 (n = 9). There was no significant difference in Vinact1/2 (p = 0.12) or in slope factor (p = 0.23) between the control and hypoxia-treated groups.

To examine the persistence of IA changes, we also examined IA in DG neurons in the second week posthypoxia (P17–P21) when the subacute seizure activity from the initial insult would be expected to have ceased and in the third week posthypoxia (P23–P27), when spontaneous seizures could be beginning (Rakhade et al., 2011). In both groups, no difference was found between control (P17–P21, n = 26; P23–P27, n = 9) and hypoxia-treated groups (P17–P21, n = 28; P23–P27, n = 8) (data not shown). This indicated that the decrease in IA in DGCs persists for only up to approximately a week posthypoxia.

Delayed but complete inhibition of IA in DGCs by 4-AP

Subtracted currents before and after 5 min of 10 mm 4-AP application initially showed an apparent decrease in the 4-AP-induced IA inhibition after seizure-inducing hypoxia (Fig. 3A). Pooled and normalized IA amplitudes confirmed that inhibition of IA was significantly greater in control DGCs compared to those from the hypoxia-treated group (Fig. 3B, C). In control DGCs, 4-AP inhibited IA by 33.72 ± 2% (n = 11), and in DGCs from the hypoxia-treated group, 4-AP inhibited IA by 16 ± 4% (n = 9; p = 0.03). However, continued application of 4-AP (at a constant fluid exchange rate of 2 ml/min) completely blocked IA in both control and hypoxia DGCs within 15 min (data not shown). Therefore, although IA blockade by 4-AP was delayed in the hypoxia-treated group, complete inhibition was achieved in both groups with prolonged application. This raised the possibility of potential group differences in the binding kinetics of 4-AP, but importantly confirmed conditions sufficient to block IA in both groups for subsequent experiments.

Figure 3.

4-Aminopyridine (4-AP) inhibition of IA in DGCs is slowed but intact after seizure-inducing hypoxia. (A) Representative raw traces using the IA activation protocol with test step to +30 mV are shown before and after 5-min application of 10 mm 4-AP. (B, C) Summary data illustrate greater apparent inhibition of IA by 4-AP after 5 min application in the control group compared to the hypoxia-treated group. However, prolonged 4-AP application (>15 min) resulted in complete inhibition of IA in both groups (see Results).

Increased DGC membrane excitability after seizure-inducing hypoxia

IA inhibits burst firing by prolonging membrane hyperpolarization after single action potentials, thereby delaying the onset of the next spike. To determine if decreased IA compromised this function in DGCs, we measured the latency to the first action potential relative to the onset of depolarizing current steps from either resting potential or −80 mV. DGCs in the hypoxia-treated group showed a shorter latency to the first action potential compared to controls when injected with depolarizing current from the more negative potential (Fig. 4A). This latency was 19.8 ± 4.69 msec for the control group (n = 20), and was significantly decreased to 3.49 ± 1.29 msec in hypoxia-treated group (n = 9; p < 0.01). Inhibition of IA by 15-min application of 10 mm 4-AP significantly decreased the latency in the control group to 7.33 ± 1.29 msec (p < 0.01), but the decreased latency in the hypoxia-treated group (8.25 ± 1.56 msec) was not significantly different (p = 0.15).

Figure 4.

Decreased spike latency in DGCs upon rapid depolarization. (A) Raw voltage traces are shown to compare spikes evoked by stepwise current injection from rest (solid lines) and from −80 mV (dotted lines). In the control neuron (left), current injection from −80 mV resulted in increased latency to first spike compared to rest consistent with the removal of IA inactivation by membrane hyperpolarization and subsequent IA activation. In the neuron from the hypoxia-treated group (right), there was no significant increase in latency to first spike upon current injection from −80 mV compared to rest consistent with decreased IA activation. (B) Decreased IA was associated with increased action potential duration in DGCs after hypoxia. The bar graphs show summary action potential half-widths (left) and amplitudes (right) in each group with and without 4-AP. Action potential half-widths were significantly increased in the hypoxia-treated group compared to controls, consistent with slowed repolarization consequent to decreased IA. However, 4-AP comparably increased action potential half-widths in both groups, suggesting that other 4-AP-sensitive currents that underlie action potential repolarization were unaffected by hypoxia treatment. Action potential amplitudes were not significantly different between groups. (C) Summary bar graphs show no differences in resting membrane potential between control and hypoxia-treated groups. 4-AP comparably depolarized RMP in both groups (p < 0.001, paired t-test), and this effect also was not different between the control and hypoxia-treated groups. (D) Current-clamp recordings before and after 4-AP showed that the 4-AP-induced depolarization of the resting potential observed in both groups was sufficient to elicit spontaneous action potential firing in DGC neurons.

IA is strongly activated during action potentials, and promotes the rapid repolarization of membrane potential. Therefore, decreased IA would also be expected to slow repolarization and increase the half-width of action potentials. Accordingly, we found that action potentials recorded from DGCs in the hypoxia-treated group were broader than controls. Mean action potential half-widths in the control group were 2.1 ± 0.08 msec (n = 12) compared to 2.7 ± 0.35 msec in the hypoxia-treated group (n = 15), p < 0.05. Action potential half-widths were comparably increased after 15 min of 4-AP application in both groups (Fig. 4B), with a 75% increase (1.59 ± 0.4 msec) in the control group (p < 0.05), and an 82% increase (2.23 ±0.5 msec) in the hypoxia-treated group (p < 0.001).

We also recorded resting membrane potential and spontaneous action potentials in DGCs under current-clamp before and after 4-AP to investigate the resting contribution of IA to intrinsic excitability. 10 mm 4-AP depolarized the RMP from −69.91 ± 1.01 to −64.17 ± 0.73 mV (p < 0.05) in the control group, and from −66.2 ± 1.25 mV to −61.07 ± 1.88 mV (p < 0.05) in the hypoxia-treated group (Fig. 4C). Therefore, the effect of 4-AP on RMP was not different between the control and hypoxia-treated groups, suggesting that other 4-AP-sensitive currents predominate in setting the RMP. Despite the small change of RMP in DGC neurons, action potentials were induced after 4-AP treatment. The membrane potential of DGC neurons was set to −70 mV with positive or negative current injection, close to the reported average resting membrane potential. 10 mm 4-AP depolarized DGC neurons both in control and hypoxia group and most of the cells recorded (10/11) induced trains of action potential as shown in Fig. 4D.


The key findings of this study were that IA was significantly decreased in the majority of DGCs recorded 1–5 days after seizure-inducing hypoxia, and that this was associated with more rapid spiking in response to depolarizing current injection from hyperpolarized membrane potentials. Therefore, dysregulation of intrinsic firing properties secondary to altered IA could contribute to hyperexcitability of DGCs subacutely after neonatal seizure-inducing hypoxia, at a time of maturation when activity-dependent anatomic and synaptic patterning is highly labile. This conceivably could have proepileptogenic consequences or promote cognitive delay, but such functional consequences remain to be addressed. Our findings only establish altered IA in DGCs as a potential contributing factor to limbic pathophysiology consequent to seizure-inducing neonatal hypoxia.

A-type K+ channels are crucial determinants of neuronal firing patterns, and could be particularly important in controlling seizures. Reducing the amplitude of A-type currents (IA) increases seizure susceptibility (Juhng et al., 1999) and lack of A-type Kv4.2 potassium channels contributes the increased excitability and decreased seizure thresholds in methylazoxymethanol acetate–exposed rats (Castro et al., 2001). Decreased IA in hippocampal CA1 pyramidal neurons after kainate-induced status epilepticus has been observed to increase distal dendritic excitability by allowing increased back-propagation of action potentials to apical dendrites (Bernard et al., 2004). Notably, after pilocarpine-induced status epilepticus, IA in DGCs did not exhibit similar changes to that observed in other hippocampal subregions, and in fact, appeared resistant to seizure-associated changes (Ruschenschmidt et al., 2006). Nonetheless, these authors reported powerful regulation of IA recovery from inactivation in DGCs by the intracellular redox milieu, which could be profoundly altered by hypoxia or prolonged seizures. In the current study, we examined the voltage-dependence but not the time course of inactivation removal (and at room temperature), and thus did not address this mode of IA regulation. However, our finding of decreased channel availability in the hypoxia-treated group suggests that oxidation-promoted speeding of recovery from IA inactivation did not occur under our recording conditions. Under physiologic conditions, it is possible that redox changes could oppose the observed downregulation of IA in DGCs, but this has yet to be explored, as well as other mechanisms of posttranslational regulation that could be triggered by hypoxia, seizures, or both, at this early maturational stage.

Sensitivity to 4-AP is characteristic of each IA subtype, but can vary depending on the molecular composition of the channels (Bekkers, 2000; Korngreen & Sakmann, 2000; Shibata et al., 2000). The potassium channel blocker, 4-AP, significantly inhibited the IA-type channel recorded in DGCs in our study, further confirming that the transient potassium current is an IA-type current. In addition, 4-AP eliminated the difference between groups in latency to spike onset upon rapid depolarizing current injection from a hyperpolarized potential, consistent with decreased IA as underlying the more rapid spiking in the hypoxia-treated group.

IA has been shown to contribute to the regulation of action potential firing rates as well as action potential duration (Kocsis et al., 1982; Waddell & Lawson, 1990; Honmou et al., 1994). We observed that the duration of the first action potential was significantly increased by 4-AP, and consistently, AP duration was also prolonged in the hypoxia group compared to controls. Nonetheless, 4-AP had comparable effects on action potential duration and RMP in both the hypoxia and control groups, suggesting that additional 4-AP sensitive currents that contribute to these remained intact after hypoxia treatment.

Voltage-dependent potassium currents play important roles in regulating membrane excitability. IA has been implicated in determining the latency to first spike and the threshold and repolarization of action potentials (Rudy, 1988; Storm, 1990; Jagger & Housley, 2002). The first-spike latency in DGCs can be shaped by the expression, kinetic properties, and relative densities of IA. The presence of IA allows moderate hyperpolarizations (−80 mV) to evoke a well-recognized increase in the delay to first spike. In the current study, the latency to the first spike was dramatically shortened in hypoxia DGCs compared to that of P10–15 control neurons. The shortened latency would certainly increase the probability of action potential generation and change the repetitive firing pattern of DGCs. These changes could have significant impact on the maturation of limbic circuits. Therefore, the alterations in the excitability and firing pattern of DGCs in the neonatal brain could profoundly influence the activity of the whole hippocampus and vulnerability to epilepsy and other limbic pathologies.


This work were supported by NIH/NINDS R01 NS047385 (RMS); Epilepsy Foundation of America Postdoctoral Research Training Fellowship, National Natural Sciences Foundation of China No. 81100970 (BWP); Program for New Century Excellent Talents (No. NCET-07-0630), National Basic Research Program of China (No.2010CB529803).


None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.