Journal of Neurophysiology

Ts65Dn, a Mouse Model of Down Syndrome, Exhibits Increased GABAB-Induced Potassium Current

Tyler K. Best, Richard J. Siarey, Zygmunt Galdzicki


Down syndrome (DS) is the most common nonheritable cause of mental retardation. DS is the result of the presence of an extra chromosome 21 and its phenotype may be a consequence of overexpressed genes from that chromosome. One such gene is Kcnj6/Girk2, which encodes the G-protein-coupled inward rectifying potassium channel subunit 2 (GIRK2). We have recently shown that the DS mouse model, Ts65Dn, overexpresses GIRK2 throughout the brain and in particular the hippocampus. Here we report that this overexpression leads to a significant increase (∼2-fold) in GABAB-mediated GIRK current in primary cultured hippocampal neurons. The dose response curves for peak and steady-state GIRK current density is significantly shifted left toward lower concentrations of baclofen in Ts65Dn neurons compared with diploid controls, consistent with increased functional expression of GIRK channels. Stationary fluctuation analysis of baclofen-induced GIRK current from Ts65Dn neurons indicated no significant change in single-channel conductance compared with diploid. However, significant increases in GIRK channel density was found in Ts65Dn neurons. In normalized baclofen-induced GIRK current and GIRK current kinetics no difference was found between diploid and Ts65Dn neurons, which suggests unimpaired mechanisms of interaction between GIRK channel and GABAB receptor. These results indicate that increased expression of GIRK2 containing channels have functional consequences that likely affect the balance between excitatory and inhibitory neuronal transmission.


Mental retardation can be attributed to abnormal neural functioning caused by a spectrum of abnormalities that disrupt CNS function. The cognitive deficits associated with Down syndrome (DS) are the result of expression of genes from the third copy of chromosome (Chr.) 21 and their effects on CNS function. The major neurological DS phenotypes include mental retardation, muscle hypotonia, and appearance of Alzheimer disease neuropathology after age 35 (Antonarakis et al. 2004). Among the genes on Chr. 21 (Hattori et al. 2000) that might contribute to mental retardation in DS are genes located either within or in proximity to the DS critical region (DSCR) (Galdzicki et al. 2001; Toyoda et al. 2002). However, genes outside the DSCR also can be involved in the DS phenotype (Korenberg et al. 1994), and recently the concept and existence of the DSCR has been challenged (Olson et al. 2004). Although much investigation has focused on anatomic, chemical, and metabolic deficiencies in DS and DS mouse models, neurophysiological abnormalities in single neurons derived from segmental trisomy mouse models (Ts65Dn, Ts16Cje and Ts1Rhr) has not been widely investigated. In this study, we investigated electrophysiological properties of cultured Ts65Dn neurons to evaluate the impact that an extra gene copy of a potassium channel subunit (Kcnj6/Girk2) located within common segment of extra Chrs. in Ts65Dn, Ts1Cje, and Ts1Rhr mice has on intrinsic neuronal properties.

The distal segment of mouse Chr. 16 is homologous to nearly the entire long arm of human Chr. 21, thus trisomic mouse models have been generated that genetically model the human condition. Full trisomy 16 and segmental trisomy 16 (Ts65Dn, Ts1Cje, and Ts1Rhr) mice have been developed. Ts65Dn, and Ts1Cje mice mimic many of the behavioral, learning, and developmental deficits characteristics in DS individuals (Galdzicki and Siarey 2003; Galdzicki et al. 2001; Holtzman et al. 1996; Reeves et al. 1995; Sago et al. 1998). Protocols that induce long-term potentiation and depression (LTP and LTD) show abnormal synaptic plasticity in these mice. LTP is decreased, whereas LTD is elevated in Ts65Dn and Ts1Cje hippocampi (Costa and Grybko 2005; Kleschevnikov et al. 2004; Siarey et al. 1997, 1999, 2005). In Girk2 knockout mice, LTP and LTD are also abnormal (Adeniji-Adele 2004) but in the opposite direction, suggesting that the level of GIRK2 expression correlates with the level of potentiation or depression of hippocampal synapses.

Kcnj6, which encodes GIRK subunit 2 of the G-protein-coupled inward rectifying potassium channel (GIRK), is located within the DSCR on human Chr. 21 (Hattori et al. 2000) and likewise within triplicated segments of mouse Chr. 16 from DS mouse models. Mammalian GIRK channels are homo- and hetero-tetramers formed from GIRK1–GIRK4 subunits and are activated by neuromodulators acting on G-protein-coupled receptors (GPCRs). Within the CNS, GIRK1–GIRK3 subunits contribute to the formation of functional GIRK channels with GIRK3 playing a minor role in neuronal membrane potassium conductance (Koyrakh et al. 2005). In the hippocampus, GIRK1 and GIRK2 subunits are primarily localized to postsynaptic compartments and particularly found in peri- and extrasynaptic regions of dendritic spines of CA1 pyramidal neurons (Drake et al. 1997; Koyrakh et al. 2005). Furthermore, tonic GIRK open channel probability increases with dendritic distance from the soma (Chen and Johnston 2005). The postsynaptic location and function of GIRK containing channels signify that these channels play an important role in synaptic function and modulation and that the expression levels would impact intrinsic dendritic properties and the ability of neurons to properly integrate, modulate, and encode excitatory and inhibitory input. Through their contribution to the potassium conductance they can influence resting membrane potentials (Koyrakh et al. 2005; Luscher et al. 1997) and can impede neuronal excitability via shunting and slowing frequency of spike trains (Ehrengruber et al. 1997; Hille 2001).

High hippocampal GIRK expression indicates the importance that GIRK channels may have in modulating learning and memory. Indeed, reports have established that rodents lacking GIRK1 and GIRK4 subunits demonstrate defective learning and memory (Kourrich et al. 2003; Wickman et al. 2000). Our previous study (Harashima et al. 2006) showed that the presence of an extra GIRK2 gene copy in Ts65Dn hippocampus leads to elevation in GIRK2 mRNA. As a consequence, GIRK2 subunit protein was overexpressed in Ts65Dn neurons. Interestingly, GIRK1 subunit protein was also overexpressed, whereas GIRK1 mRNA was normal (Girk1 gene is not localized to the Ts65Dn chromosome). These data strongly suggest that GIRK current should be greater in Ts65Dn hippocampus in comparison to diploid neurons because heterotetramers of GIRK1 and GIRK2 are the most predominant form of GIRK channel in CNS neurons. The physiological impact of GIRK1-2 channel subunit overexpression in DS has not been demonstrated. In this study, we sought to reveal that GIRK2 (and consequential GIRK1) subunit overexpression leads to an increase in GIRK current that would likely impact tonic inhibitory tone. We demonstrate herein that the reported overexpression levels of GIRK channel subunits in Ts65Dn mice correlates with increases in GIRK currents and changes in intrinsic neuronal properties from hippocampal neurons cultured from the Ts65Dn neonatal mouse.


Cell culture

Mouse hippocampal neurons were cultured in a similar manner to that described previously (Galdzicki et al. 1998). Litters from Ts65Dn mothers and diploid fathers were taken at postnatal days one and two (P1-2). Pups were decapitated, and hippocampi dissected out, after which free hippocampi were cut into pieces and incubated in 0.05% trypsin (wt/vol) for 10 min at 37°C. Tissue was washed twice with plating media after which hippocampal cells were mechanically separated by trituration through a polished glass Pasteur pipette. Hippocampal cells were then plated on 35-mm Nunc brand dishes previously coated with 25 μg/ml poly-d-lysine and maintained at 37°C and 5% CO2 in the presence of plating media. One day after plating, the media was replaced with maintenance media. Plating media was composed of Neurobasal-A media with B27 supplement (Invitrogen/Gibco BRL, Carlsbad, CA), fetal bovine serum (FBS, 10%), horse serum (HS, 5%), and glutamine (1%; Biosource, Camarillo, CA), whereas maintenance media was of the same composition but without FBS or HS. Throughout the dissection and plating procedure tissue from each pup was processed separately, and ploidy was determined afterward. Mice were karyotyped with blood as described previously (Harashima et al. 2006). All recordings and the initial analysis were done without prior knowledge of the genotype.

Electrophysiological recordings

Whole cell patch-clamp recordings of isolated hippocampal neurons were performed 10–20 days after plating. This age complies with functional GABAB receptors and GIRK channels in cultured neurons (Correa et al. 2004). Pipettes were backfilled with a solution containing (in mM) 100 K-gluconate, 20 KCl, 10 HEPES, 10 EGTA, 1 CaCl2, 4 Mg-ATP, 0.3 Na-GTP, and 7 Na-phosphocreatine, pH 7.35. In some instances, the 0.3 mM Na-GTP was replaced with 0.3 mM Li3-GTPγS, a nonhydrolyzable GTP analogue. At the time of recording, maintenance media was removed and replaced with room temperature bathing solution containing (in mM) 150 NaCl, 4 KCl, 2 CaCl2, 2 MgCl2, 10 Na-HEPES, and 10 glucose, pH 7.35. tetrodotoxin (1 μM) and MK-801 or d-2-amino-5-phosphonovaleric acid (d-APV; 50 μM) were added to block spontaneous firing, and N-methyl-d-aspartate (NMDA)-mediated currents, respectively. To elicit change in the potassium driving force, the bathing solution NaCl was replaced by equivalent KCl (high K+: 60 mM). The GABAB receptor agonist baclofen was applied in both basal and high potassium solutions via a 9-barrel rapid solution changer (RSC-200, Biologic, Calix, France/Molecular Kinetics, Pullman, WA) positioned ∼200 μm from the patched neuron. Recordings were performed in voltage-clamp configuration at a holding potential of −70 mV, and data were acquired by way of an EPC-7 amplifier (HEKA), filtered at 5 kHz (8-pole Bessel filter, NPI, ALA Scientific Instruments, Westbury, NY), and recorded on a personal computer as per Klein et al. (2001). A 5 mV hyperpolarizing step from a holding potential of 8–70 mV was applied to estimate membrane capacitance and resistance at the initiation of whole cell access and at intervals throughout the recording. Similarly, resting membrane potential was measured at the beginning of each recording and throughout the experiment to assess and monitor cell viability.

Data analysis and statistics

High potassium (60 mM) currents were determined as steady-state values subtracted from basal currents measured under 4 mM extracellular potassium. When delivered under high potassium, baclofen-evoked peak and steady-state currents were determined by subtraction from the value of the high extracellular potassium steady-state currents. Under basal potassium, baclofen currents were subtracted from the steady-state basal current. Dose response relationships are the best least-squares fit to Eq. 1/(1+ EC50/[baclofen]), and EC50 is concentration of baclofen, which produces 50% of the maximum possible response.

To assess the conductance of individual channels from mice of each genotype, stationary noise analysis of current recordings were performed similar to previous reports (Sciancalepore et al. 1990; Takigawa and Alzheimer 1999; Traynelis and Jaramillo 1998). Current variance from 200 ms stretches of GIRK current were plotted as a function of the mean GIRK current value for that given stretch. Data points from individual neurons were fit to the equation: σ2 = iII2/N where σ2 is the current variance, i is the unitary current amplitude, I is the whole cell current amplitude, and N is the number of available channels. The value of i determined from the preceding equation was then related to single-channel conductance by the formula: γ = i/(VhEK) where γ is the single-channel conductance, Vh is the holding potential (−83 mV after liquid junction potential correction), and EK is the Nernst potential for potassium at high potassium recording conditions (−18 mV). Channel density was then determined for each neuron by taking the ratio of estimated N and membrane capacitance. Values of i, γ, N, and channel density from each neuron were averaged and analyzed for statistical significance based on ploidy.

To evaluate impact of ploidy on kinetics of baclofen-induced current, tau of current activation (τa) and deactivation (τd) were calculated by fitting to a single-exponential function. GIRK current desensitization was determined by subtraction of peak from steady-state current values (see Fig. 1B, right arrow and dashed line). Changes in current from peak to steady state are likely due to desensitization rather than cell dialysis. This is because only cells that showed a return to baseline on washout were used in the analysis, and long-term agonist application showed rates of desensitization comparable to short-term application.

FIG. 1.

Baclofen induces an increase in G-protein-coupled inward rectifying potassium channel (GIRK) current density in Ts65Dn in comparison to diploid neurons. A: current-voltage relationships of diploid and Ts65Dn neurons show baclofen current is inwardly rectifying and the reversal potential approximates the K+ Nernst potential. I-V curves were determined by subtraction of ramps with and without baclofen at basal potassium concentrations. B: example of the increased inward current density induced by 25 μM baclofen in a Ts65Dn neuron compared with diploid.

EC50 values were derived using the nonlinear regression–curve fit (sigmoidal dose response with variable slope) function of GraphPad Prism version 4.03 (GraphPad Software, San Diego, CA). Genotype and dose dependence was assessed using two-way repeated-measures ANOVA (GraphPad Prism version 4.03). In all other cases, unpaired t-test were employed. Significance has been assigned at P ≤ 0.05. Data are means ± SE unless otherwise indicated.


Preliminary experiments with agonists, such as baclofen, serotonin, acetylcholine, and 2-chloroadenosine, to receptors coupled to GIRK channels regularly elicited inward rectifying potassium channel in diploid and Ts65Dn neurons. However, the GABAB receptor agonist baclofen evoked the largest and most consistent current in both diploid and Ts65Dn neurons. Therefore we proceeded to focus on currents elicited by baclofen in our experiments. GABAB specificity was determined through use of CGP-55845 (2 μM), which completely blocked baclofen induced current (n = 3). Voltage ramps indicate that the reversal potential of this current approximates the Nernst potential for potassium ions as well as showing inward rectification consistent with Kir channels (Fig. 1A). All passive membrane properties were similar between genotypes (see Table 1). Resting membrane potential under current clamp was not significantly hyperpolarized in Ts65Dn neurons (although, P = 0.08). These results are similar to those obtained in full trisomy 16 cultured hippocampal neurons (Galdzicki et al. 1993).

View this table:

Passive neuronal properties of diploid and Ts65Dn neurons

Evaluation of steady-state current density evoked by application of high extracellular potassium (60 mM) found no significant differences between diploid and Ts65Dn neurons (diploid, −23.6 ± 3.0 pA/pF, n = 41; Ts65Dn, −22.3 ± 5.4 pA/pF, n = 19; P = 0.81). Likewise, membrane resistance changes on application of high potassium decreased at similar percentage between genotypes (diploid, 51 ± 7%, n = 6; Ts65Dn, 49 ± 7%, n = 8; P = 0.86). These data indicate that potassium-induced inward currents do not differentially effect membrane properties of Ts65Dn neurons. Similar results were previously reported for cultured murine “full” trisomy 16 hippocampal neurons and human DS dorsal root ganglion neurons (Galdzicki et al. 1993; Nieminen et al. 1988).

Baclofen was applied under both basal and high potassium solutions. Under basal potassium concentrations and at the holding potential (−70 mV, junction potential adjusted to −83 mV), the potassium driving force is outward. We measured this outward current after application of 25 μM baclofen and found current density responses was significantly increased in Ts65Dn (1.3 ± 0.3 pA/pF, n = 16) compared with diploid (0.6 ± 0.2 pA/pF, n = 11) neurons (P < 0.05). Under high potassium where the potassium driving force is inward, baclofen-induced GIRK current density of Ts65Dn neurons were significantly shifted left toward smaller concentrations of baclofen. Peak current density of baclofen-induced GIRK current shows dose dependence that is significantly different between diploid (n = 11–23) and Ts65Dn (n = 9–13) neurons [2-way repeated-measures (RM) ANOVA, P = 0.0002]. Peak EC50 shifts from 4.2 μM in diploid to 0.9 μM in Ts65Dn (Fig. 2A). The steady-state current density was also significantly different (2-way RM ANOVA, P < 0.0001). The EC50 shifted left from 3.0 μM in diploid to 1.0 μM in Ts65Dn (Fig. 2B). The increase in current density efficacy is consistent with an increase in channel number.

FIG. 2.

Baclofen-induced normalized GIRK current in diploid and Ts65Dn neurons show similar dose dependence. A: normalized peak EC50 of 2.5 μM for diploid and 1.4 μM for Ts65Dn. B: normalized steady-state EC50 of 2.1 μM for diploid and 1.4 μM for Ts65Dn are not significantly different. Peak and steady-state current values were normalized to the maximum current in each cell. C: peak (→ in Fig. 1B); D: steady state (- - - in Fig. 1B), current density of baclofen-induced GIRK current show dose dependence that is significantly different between diploid and Ts65Dn neurons (P < 0.001). Peak EC50 shifts from 3.2 μM in diploid to 1.4 μM in Ts65Dn and steady-state EC50 shifts 2.1 μM in diploid to 0.9 μM in Ts65Dn. Current density is determined by dividing current by capacitance value (pA/pF).

Comparisons of normalized currents were also performed to determine whether GABAB receptor affinity for baclofen and/or GABAB receptor coupling with the GIRK channel was altered in Ts65Dn neurons. When the responses were normalized to the greatest current within each neuron, diploid and Ts65Dn dose responses were similar to each other (2-way RM ANOVA, P = 0.08). Normalized peak EC50 was 2.5 μM for diploid (n = 14–26) and 1.4 μM for Ts65Dn (n = 9–15; Fig. 2C). Normalized steady-state currents were also similar across genotype (2-way RM ANOVA, P = 0.19). Normalized steady-state EC50 was 2.1 μM for diploid and 1.4 μM for Ts65Dn (Fig. 2D). These EC50 values are comparable to those previously reported for baclofen of 2.7–3.0 μM (Sodickson and Bean 1996, 1998). Because normalized GIRK currents are similar between genotypes, it is expected that coupling mechanisms between the GABAB receptor and the GIRK channel are not abnormal in Ts65Dn neurons. Likewise, the affinity for baclofen by the GABAB receptor appears to be unaltered in Ts65Dn neurons.

GIRK channel blockers were used to block the baclofen-induced currents. Tertiapin-Q (50 nM), a specific GIRK channel blocker inhibited baclofen (25 μM) induced current by 46 ± 8% in diploid (n = 7) and 56 ± 5% in Ts65Dn (n = 12) neurons (Fig. 3A). No significant difference in the percentage of block by tertiapin-Q were found between genotypes (P = 0.25). Tertiapin-Q produced a biphasic response as shown in Fig. 3A. This unique response may be indicative of the use dependency of tertiapin-Q block on GIRK channels (Huang et al. 2005; Kanjhan et al. 2005). Application of barium (Ba2+) (200 μM), which blocks inward rectifying potassium channels, also inhibited baclofen (25 μM)-induced currents in both Ts65Dn and diploid neurons Fig. 3B. Interestingly, the inhibitory effect of Ba2+ on diploid neurons was dichotic. In one subset of neurons (7 of 11), the Ba2+ blocked 60 ± 14% of baclofen-induced current. In the other subset of diploid neurons (4 of 11), Ba2+ blocked 540 ± 23% of the baclofen-induced current (i.e., 100% of baclofen current and 55 ± 5% of the high potassium current). This dichotic effect was not seen in Ts65Dn neurons. Ba2+ block in Ts65Dn neurons was similar to the first subset of diploid neurons in that block was 56 ± 9% (n = 11) of baclofen current. This suggests that the overexpression of GIRK leads to a decrease in endogenously active Ba2+-sensitive potassium channels in a subset of hippocampal cultured neurons.

FIG. 3.

Block of baclofen-induced current by tertiapin-Q and Ba2+. A: example traces showing the action of tertiapin-Q (50 nM) on baclofen-induced (25 μM) current. Tertiapin-Q blocked 46 ± 8 and 56 ± 5% of the baclofen-induced current in diploid (n = 7) and Ts65Dn (n = 12) neurons, respectively. The biphasic block by tertiapin-Q in the Ts65Dn recording may be indicative of use dependence. B: traces demonstrating Ba2+ (200 μM) block of all baclofen (25 μM)-induced current. Diploid neurons responded in disparate ways to Ba2+. In 4 of 11 neurons Ba2+ blocked all of the baclofen-induced current and part (55 ± 5%) of the high K+ current. In the other 7 of 11 neurons, the response was similar to that in Ts65Dn neurons where block was 60 ± 14% compared with 56 ± 9% in Ts65Dn neurons (n = 11).

The increased GIRK current of Ts65Dn neurons may be explained by increased GABAB receptor number, receptor affinity for baclofen, coupling efficiency among receptor, G proteins, and channel and/or GIRK channel expression. We used a nonhydolyzable form of GTP, GTPγS, that can directly activate GIRK channels to bypass the receptor and G proteins and thus possibly discriminate between mechanisms for increased GIRK current. We replaced the GTP of the intracellular pipette solution with the equivalent amount of GTPγS. Current density responses to high potassium with GTPγS in the pipette were compared and no significant difference was found between genotypes (diploid, −28.7 ± 7.6 pA/pF, n = 5; Ts65Dn, −33.7 ± 4.4 pA/pF, n = 5; P = 0.59). The high potassium current density, although greater with GTPγS in the pipette as opposed to GTP, was not significantly different for each genotype (diploid, P = 0.58; Ts65Dn, P = 0.31). This may be the result of variable contribution of other potassium channels modulated by GTP and G proteins (Sanchez et al. 1998; Trapp et al. 1997).

Because the use of GTPγS was unable to substantially discriminate possible explanations for the increase in GIRK current density, we used stationary fluctuation analysis. An increase in channel number or density would likely be a principle cause for the increased GIRK current density and not changes in GABAB receptor or G-protein properties. The values of the single-channel conductance (γ) estimated for diploid (n = 12; 19.5 pS) and Ts65Dn neurons (n = 8; 25.1 pS) are within range of GIRK conductance obtained from acutely dissociated rat hippocampal neurons (Takigawa and Alzheimer 1999) but slightly smaller than that estimated from excised dendritic recordings (Chen and Johnston 2005). γ is not significantly different between diploid and Ts65Dn neurons (P = 0.46; Fig. 4, Table 2). However, when evaluating channel number per membrane capacitance, Ts65Dn neurons show a significant increase in channel density (80%, P < 0.05) that can explain the significant shift in EC50 of GIRK currents and corroborate our previous findings that GIRK channel number is increased in Ts65Dn neurons compared with diploid (Harashima et al. 2006).

FIG. 4.

Single-channel conductance (γ) and channel number (N) are similar between diploid and Ts65Dn neurons. A: sample recordings from increasing doses of baclofen in diploid and Ts65Dn neurons. B: current variance from 200 ms stretches of GIRK current were plotted as a function of the mean GIRK current value for that given stretch. Single-channel conductance (γ) and channel number (N) are not significantly different between genotypes (P = 0.46, P = 0.74, respectively); however, channel density is increased by 80% (P = 0.02; see Table 2).

View this table:

Fluctuation analysis of GIRK current from diploid and Ts65Dn neurons

We then examined the time course of activation (τa) and deactivation (τd) for the baclofen-induced GIRK current in an attempt to address the coupling between receptor and channel. We found a significant dose-dependent decrease in τa (speeding of activation) in both diploid (n = 4–20) and Ts65Dn (n = 5–12) neurons (Fig. 5A; 2-way RM ANOVA, P < 0.0001 for both diploid and Ts65Dn). A significant interaction of τa was found between the genotypes (2-way RM ANOVA, P < 0.05). Bonferroni posttests showed that no individual τa for a separate doses was significantly different. Deactivation kinetics show significant dose dependence for both diploid (n = 3–15) and Ts65Dn (n = 2–9; 2-way RM ANOVA, P < 0.05) neurons (Fig. 5B). No significant interaction or effect of genotype (2-way RM ANOVA, P = 0.07 and P = 0.42, respectively) was found. These data support the idea that coupling mechanisms among baclofen, GABAB receptor, and the GIRK channel is similar between genotypes.

FIG. 5.

Normal kinetics of baclofen induced GIRK current. A: kinetics of GIRK current activation are dependent on baclofen dose (P < 0.001) but are not significantly different between diploid and Ts65Dn neurons. B: decay kinetics show no dose dependence and are similar between ploidy. C: short-term (∼15 s) desensitization of GIRK current density evaluated by the difference between peak and steady-state values show dependence on baclofen dose for diploid and Ts65Dn neurons (2-way RM ANOVA, P < 0.05). A significant interaction was found (P < 0.05) and Bonferroni posttests indicate significant differences at 5 and 50 μM baclofen (*P < 0.05, **P < 0/01).

The amount to which baclofen-induced GIRK currents desensitize was evaluated in a dose-dependent manner at ∼15 s (short-term), whereas ∼1 min (long-term) desensitization was examined for 25 μM baclofen. Short-term desensitization of baclofen-induced GIRK current density was significantly dose-dependent for diploid (n = 12–27) and Ts65Dn neurons (n = 4–15; n = 4–15, 2-way RM ANOVA, P < 0.0001; Fig. 5C). This is similar to what has been previously reported in rat primary hippocampal cultures (Leaney 2003). Desensitization of GIRK current density also showed significant interaction (2-way RM ANOVA, P < 0.0001) between diploid and Ts65Dn neurons. Bonferroni posttests showed that 5 and 50 μM baclofen produced significantly different desensitization rates between genotypes (P < 0.05 and P < 0.01, respectively). Long-term baclofen (25 μM) exposure showed no significant difference in desensitization between genotypes (diploid, 27.8 ± 4.4%, n = 7; Ts65Dn, 23.6 ± 2.8%, n = 14; P = 0.41). The interpretation of these findings is not clear but may reflect complex relationships between the duration of baclofen exposure and the receptor-GIRK channel assembly.


The data presented in this report suggest that cultured hippocampal neurons from Ts65Dn neonates have elevated responses to GABAB receptor activation as evidenced by a leftward shift in the dose-response curves of GIRK current density. Variance analysis of GIRK current fluctuations suggests that the leftward shift of the dose response curve can be explained by a significant increase in channel density but not by a significant change in single-channel conductance.

The GIRK currents seen in Ts65Dn neurons were consistently larger than those measured in diploid neurons for all tested doses. We see no evidence for a change in the activation and deactivation kinetics of GIRK current in Ts65Dn neurons suggesting that coupling of receptor and channel remain similar to diploid neurons. Disparate desensitization responses to baclofen between diploid and Ts65Dn neurons, however, may be indicative of more complex changes in GIRK current induced by changes in channel subunit ratios.

Because hippocampal neurons from GIRK2 knockout animals show significant depolarization (Koyrakh et al. 2005; Luscher et al. 1997), we evaluated the impact of the extra Ts65Dn segment, which contains GIRK2, on resting membrane properties. Resting membrane potential from each genotype was not significantly different. In full trisomy 16 hippocampal culture, we also did not find significant changes in membrane resting potential (Galdzicki et al. 1993). Membrane resting potentials in Ts65Dn neurons may not be affected by level of GIRK function alone because the overexpression of other genes from the Ts65Dn Chr. may contribute to neuronal resting membrane potentials.

Chen and Johnston (2005) suggested that the low GIRK conductance at the soma and low surface to volume ratio would contribute little to somatic membrane properties. Yet at the dendrites where GIRK conductance is high and surface to volume ratio is also high, GIRK expression would have profound influence on dendritic membrane properties. Within the CA1, it is at these distal synapses where the GABAB-mediated potassium currents generated by perforant pathway stimulation within the stratum lacunosum-moleculare are greater than those within the more proximal stimulation of Schaffer collaterals within the stratum radiatum (Otmakhova and Lisman 2004). This poses the possibility that perforant path input at CA1 pyramidal synapse is more profoundly influenced in the Ts65Dn mouse than Schaeffer collateral input. A change in the integrative properties of pyramidal neurons is the likely result.

Plasma membrane associated GIRK1 and GIRK2 channel subunits have been identified primarily in extra- and perisynaptic regions of the postsynaptic membrane of mouse stratum radiatum. GIRK2 but not GIRK1 signal was also found within postsynaptic specializations (Koyrakh et al. 2005). More specifically, GIRK2 immunoreactivity colabeled with GABAB receptors on dendritic spines of adult rat hippocampal pyramidal neurons. This coexpression, however, was not found in dendritic shafts where GABAB receptor and GIRK2 protein were segregated (Kulik et al. 2006). Remarkably, in rat hippocampus, extrasynaptic GABAB receptors show heterogenous potassium currents and may not couple Ba2+-sensitive inward rectifying potassium channels, whereas synaptic GABAB receptors are homogenous and solely elicit characteristic GIRK currents (Pham et al. 1998). What an excess of GIRK1-2 expression would do to the balance between extrasynaptic and synaptic GABAB-potassium channel coupling remains to be seen, but the extra GIRK1-2 would increase the overall inhibitory tone. (Note: because our protocol involves bath application of baclofen, both synaptic and extrasynaptic receptors were activated).

An increase in GIRK current through overexpression should have profound impact on the physiology of the neuron. It could perhaps account for the abnormalities in synaptic plasticity of Ts65Dn and Ts1Cje mice (Costa and Grybko 2005; Kleschevnikov et al. 2004; Siarey et al. 1997, 1999, 2005) through shifts in the dynamic range of these synapses and a disruption of the balance between excitation and inhibition. Intrinsic interneuron/GABAergic tone would impinge an elevated level of GABAB receptor mediated inhibition in mice overexpressing GIRK1-2. In native systems, it is likely that lower concentrations of GABA would elicit larger GIRK currents. Thus minimal GABAergic neurotransmission should have greater shunting ability.

A heteromeric GABAB receptor consisting of subunits 1 and 2 is considered necessary for metabotropic GABA signaling. GABAB1 is responsible for binding GABA, whereas GABAB2 mediates surface trafficking and G-protein coupling (Calver et al. 2001; Galvez et al. 2001; Kaupmann et al. 1998; Margeta-Mitrovic et al. 2000, 2001; Pagano et al. 2001; Robbins et al. 2001). In hippocampal primary cultures from E18 rats, GABAB1 isoforms and GABAB2 receptor expression matches the time-dependent increases in baclofen induced potassium currents. In contrast to diffuse GABAB staining and minimal potassium current at 3 days in vitro (DIV), dendritic and punctate distribution of staining after 14 DIV paralleled the time at which maximum current density occurred. Interestingly, as the expression of the GABAB subunits developed according to DIV, the levels of GIRK1 protein remained constant suggesting that GABAB receptor expression was independent of GIRK (Correa et al. 2004). It is therefore unlikely that increases in baclofen-induced currents are mediated by changes in GABAB receptor expression. Likewise, our data derived from variance analysis indicating increases in channel density alone can account for the shift in dose response and argue against a change in GABAB receptor expression in these cultures.

For the most part any neurotransmitter that activates a Gi/o-coupled receptor can activate GIRK channels. Indeed, in native neurons and expression systems, acetylcholine, adenosine, endocannabinoids, dopamine, GABA, glutamate, histamine, melatonin, neuropeptide Y, norepinephrine, orexins (hypocretins), opioids, serotonin, and somatostatin, have all been shown to modulate GIRK channel activity (Bunemann et al. 2001; Hoang et al. 2003; Kobayashi et al. 1996; Kofuji et al. 1995; Kreienkamp et al. 1997; Kuzhikandathil et al. 1998; Luscher et al. 1997; McAllister et al. 1999; Nelson et al. 1996; Paredes et al. 2003; Saugstad et al. 1996; Spauschus et al. 1996; Takigawa and Alzheimer 1999; Ulens et al. 1999). Therefore it is likely that any of these systems with activity in the hippocampus and CNS that activate GIRK channels would result in abnormal DS function and may contribute to neurological and cognitive phenotypes found in DS individuals. In fact, Ts65Dn mice show an increase in GIRK2 mediated hypothermic responses to a serotonin (5-HT)1A/5-HT7 receptor agonist (Stasko et al. 2006), an effect that is most likely due to the presence of the extra Girk2 gene.

In summary, hippocampal neurons derived from Ts65Dn mice show increased sensitivity to GABAB signaling with significant shift to the left in the dose-dependence relationships. The kinetics of GIRK current activation, deactivation, and acute desensitization were unchanged in Ts65Dn neurons, suggesting that the increased GIRK current does not affect coupling mechanisms among the GABAB receptor, G protein, and GIRK channel. The increase in GIRK current can be attributed to a ∼80% increase in channel density and not significant changes in GIRK single-channel conductance.


This work was supported by National Institute of Child Health and Human Development Grant HD-38417, The Jerome Lejeune Foundation, and Uniformed Services University of the Health Sciences.


The authors thank M. Cho for assistance with the care and genotyping of the Ts65Dn mice. We are also grateful to A. Balbo for help with the neuronal cultures.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract