INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Galanin, a 29-amino-acid peptide, was first isolated from pig small intestine (Tatemoto et al. 1983) and is widely distributed in the peripheral and central nervous systems (Melander et al. 1986; Skofitsch and Jacobowitz 1985). The presence of galanin within dense core vesicles of nerve terminals and its stimulus-dependent release has suggested a neurotransmitter/neuromodulatory role for this peptide (Bartfai et al. 1993; Consolo et al. 1994). Recently, three distinct receptors for galanin have been cloned and identified to belong to seven-transmembrane G-protein-coupled family of receptors (for review, see Branchek et al. 2000). However, selective antagonists have yet to be found for any of the three cloned receptors.

In general, galanin has been identified to be an inhibitory peptide on the basis of its hyperpolarizing actions on neurons of the cardiac ganglia, locus coeruleus, dorsal raphe, hypothalamic supraoptic nucleus, and the CA3 region of the hippocampus (Konopka et al. 1989; Papas and Bourque 1997; Pieribone et al. 1995; X. J. Xu et al. 1995; Z. Q. Xu et al. 1999). In the hypothalamus, galanin also inhibits the depolarizing afterpotential that sustains bursting in magnocellular neurosecretory cells (Papas and Bourque 1997) and presynaptically depresses glutamate-mediated excitation of arcuate neurons (Kinney et al. 1998). Pharmacological experiments that examine transmitter release are less clear on the predominantly inhibitory effects observed at a cellular level. For example, galanin inhibits acetylcholine (ACh) release in the rat ventral hippocampus and from slices of the cerebral cortex (Fisone et al. 1987; Wang et al. 1999), whereas it is stimulates ACh release in the rat striatum (Amorso et al. 1992; Pramanik and Ogren 1993).

Galanin-immunoreactive cell bodies, fibers, and terminals and galanin receptors are located in the basal forebrain nuclei of rats, monkeys, and humans (Kohler and Chan-Palay 1990; Melander et al. 1986; Merchantaler et al. 1993). In particular, galanin-positive synapses have been identified on cholinergic neurons in the lateral part of the nucleus of the diagonal band of Broca (DBB), a basal forebrain nucleus (Henderson and Morris 1997). However, the source for the galanin input to cholinergic neurons is unknown at present. The cholinergic neurons in DBB and other forebrain nuclei are selectively lost in neurodegenerative conditions such as Alzheimer's disease (AD) (Price 1986). The degree of loss of cholinergic neurons in the basal forebrain and cholinergic markers in the cortex strongly correlates with the behavioral and cognitive deficits seen in AD (Price 1986). The role of galanin in AD is controversial. Galaninergic innervation of remaining basal forebrain cholinergic neurons is augmented in AD (Bowser et al. 1997). Given its inhibitory effects on ACh release in the hippocampus, this galanin hyperinnervation of cholinergic basal forebrain neurons is postulated to depress septo-hippocampal circuits involve in learning and memory (Mufson et al. 1998). On the other hand, administration of nerve growth factor (NGF), a target-derived growth factor that is important for growth and survival of cholinergic basal forebrain neurons, increases the gene expression of galanin (Planas et al. 1997) and raises the possibility that induction of galanin by NGF may have a neuroprotective role. Investigation of the cellular mechanism of action of this peptide in the cholinergic basal forebrain may thus provide a better understanding of its role in the pathology in AD.

In this study, we investigated the actions of galanin on acutely dissociated rat cholinergic basal forebrain neurons from the nucleus of the DBB using a combination of whole cell patch-clamp and single-cell reverse transcription-polymerase chain reaction (RT-PCR) analysis. Our data show that the blockade of specific potassium conductances is a potential underlying mechanism for the action of galanin on DBB neurons and may explain its effects in regulating their excitability.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Dissociation procedures

Details of the procedure for acute dissociation of neurons from the DBB are described in Jassar et al. (1999). Briefly, brains were quickly removed from decapitated male Sprague-Dawley rats (15-25 day postnatal) and placed in cold artificial cerebrospinal fluid (ACSF) that contained (in mM) 140 NaCl, 2.5 KCl, 1.5 CaCl₂, 5 MgCl₂, 10 HEPES, and D-glucose 33 (pH 7.4). Brain slices (350-µm thick) were cut on a vibratome, and the area containing the DBB was dissected out. Although most of the tissue contained the horizontal limb of the DBB, some slices may have included a component of the vertical limb of the DBB. Acutely dissociated neurons were prepared by enzymatic treatment of slice with trypsin (0.65 mg/ml) at 30°C, followed by mechanical trituration for dispersion of individual cells. Cell were then plated on poly-L-lysine (0.005% wt/vol)-coated cover slips and viewed under an inverted microscope (Zeiss Axiovert 35). All solutions were kept oxygenated by continuous bubbling with pure oxygen.

Electrophysiological recordings

Whole cell patch-clamp recordings were performed at room temperature (20-22°C) using an Axopatch-1D amplifier. Patch electrodes (World Precision Instruments, thin wall with filament, 1.5-mm diam) were flame polished to yield resistances of 3-6 M. Internal patch pipette solution contained (in mM) 140 K-methylsulfate, 10 EGTA, 5 MgCl₂, 1 CaCl₂, 10 HEPES, 2.2 Na₂-ATP, and 0.3 Na-GTP (pH 7.2). Junction potential was nulled with the pipette tip immersed in the bath. Putative acutely dissociated DBB neurons were initially identified for recording by visual inspection. Current-voltage relationships and excitability characteristics were used to distinguish neurons from glial or other cell types (Jassar et al. 1999). Whole cell recordings were also done in the bridge current-clamp mode using an Axoclamp-2B amplifier to examine the effects of galanin on current-evoked changes in excitability of the acutely dissociated DBB neurons. Action potentials were evoked by brief current injection (0.6-1.5 nA, 600-ms duration) through the patch pipette. The resting membrane potential (RMP), number of spikes elicited, and interspike intervals were recorded for comparison under different experimental conditions. The membrane currents (voltage-clamp experiments) or the membrane voltages (current-clamp experiments) were recorded and analyzed on computer using pCLAMP software (version 6.0.3).

After whole cell configuration was established, we waited 5 min for steady-state currents to stabilize. The filter was set at 20 kHz during data acquisition. Cells were held in voltage clamp at 80 mV, which was close to the RMP observed in earlier studies on neurons from basal forebrain slices (Alonso et al. 1994; Easaw et al. 1997). A 1-s-long hyperpolarizing command to 110 mV was applied to remove inactivation of K⁺ channels so that the maximum current could be activated during the subsequent slow voltage ramp to +30 mV (20 mV/s) that followed it. No obvious tail currents were observed at the end of the ramp when the command potential was returned to 80 mV, suggesting that the ramp elicited mainly steady-state currents.

Cell size was estimated electronically using the whole cell capacitance compensation circuit on the Axopatch-1D amplifier. Series resistance compensation was continuously adjusted to >80% and monitored and readjusted as necessary during the course of each experiment. The average series resistance (electrode plus access resistance) was 7.3 ± 0.5 M (n = 51). Maximum voltage-clamp error in recording a current of 10 nA using a patch electrode with an electrode resistance of 8 M was 16 mV. This reflects the average maximum error because the currents recorded were usually <10 nA. In figures displaying difference currents, capacitance transients have been truncated.

To examine the effects of galanin on the contribution of Ca²⁺ to voltage-dependent ionic currents, we utilized an external solution which was nominally Ca²⁺-free and contained 50 µM Cd²⁺. In this solution, CaCl₂ was replaced with an equimolar concentration of MgCl₂. To record currents through calcium channels, we used Ba²⁺ as a charge carrier as previously described (Easaw et al. 1999). The external solution contained (in mM) 150 tetraethylammonium chloride, 2 BaCl₂, 10 HEPES, and 30 glucose (pH to 7.4 with TEA-OH). The internal patch pipette solution consisted of (in mM) 130 Cs-methanesulfonate, 2 MgCl₂, 10 HEPES, 10 BAPTA, 4 Mg-ATP, 0.2 Na-GTP, and 0.1 leupeptin (pH to 7.2 with CsOH). Depolarizing voltage steps from 80 to +70 mV (increment, 10 mV/step; 20-ms duration) were applied to voltage-clamped DBB neurons under control conditions and in the presence of galanin. Leak currents were minimal under our recording conditions. They did not change during the recordings and were not affected by application of galanin. Therefore we did not subtract these in subsequent measurements of steady-state barium or calcium currents.

We also studied the effects of galanin on I_Na. To isolate I_Na, the external solution contained (in mM) 125 NaCl, 20 TEA-Br (to block K⁺ currents), 2 MgCl₂, 5 MnCl₂ (to block Ca²⁺ currents), 10 HEPES, and 20 glucose (pH 7.4 with Tris-OH), and the internal solution consisted of (in mM) 130 cesium-methanesulfonate, 10 HEPES, 10 BAPTA, 5 Mg-ATP, and 0.3 Na₂-GTP (pH 7.2 with Cs-OH). The currents were evoked by 10-ms voltage steps from 80 mV to a maximum of +60 mV (increment, 10 mV/step; 10-ms duration) in the presence and absence of galanin.

Drugs and solutions

Galanin and galantide (M15) were obtained from Bachem (Torrance, CA) and iberiotoxin from Sigma Chemical (St. Louis, MO). All the agents were dissolved in distilled water to make 1,000× stock solution (stored at 70°C) and diluted in external perfusing medium just before the time of application. All drugs and chemicals were applied via bath perfusion at the rate of 3-5 ml/min, which allowed complete exchange in <0.5 min. Data are presented as means ± SE.

Student's two-tailed t-test was utilized for determining significance of effect.

Single-cell RT-PCR for chemical phenotyping

Neurons were harvested after electrophysiological recordings were completed and readied for RT-PCR according to a previously described protocol (Surmeier et al. 1996). In brief, contents of the electrode containing the cell and 5 µl of internal solution were expelled into a 0.2 ml PCR tube containing 5 µl sterile water (Sigma water W-4502), 0.5 µl dithiothreitol 0.1 M (DTT), 0.5 µl RNasin (10 U/µl), and 1 µl oligo-dT (0.5 µg/µl). The tube was then placed on ice. Single-stranded cDNA was then synthesized from mRNA by adding a solution containing 1 µl SuperScript II RT (200 U/µl), 2 µl 10 × PCR buffer, 2 µl 25 mM MgCl₂, 1.5 µl 0.1 M DTT, 1 µl 10 mM dNTPs, and 0.5 µl RNasin (10 U/µl). The PCR tube was gently mixed and incubated in a Techne Progene thermal cycler at 42°C for 50 min. The process was then terminated by heating to 72°C for 15 min, and the tube cooled to 4°C. Subsequently 2 µl of the RT product was taken and combined with 5 µl 10 × PCR buffer, 5 µl 25 mM MgCl₂, 0.5 µl Taq polymerase (5 U/µl), 31.5 µl sterile water (sigma water W-4502), 1 µl 25 mM dNTP mixture, and 1.5 µl of a specific set of primers (15 µM). All reagents were purchased from GibcoBRL. Primer sequences for choline acetyltransferase (ChAT) and for glutamate decarboxylase (GAD) have been previously described (Surmeier et al. 1996; Tkatch et al. 1998) and that for -actin was obtained from GenBank (the lower primer 5'-GAT AGA GCC ACC AAT CCA C, the upper primer 5'-CCA TGT ACG TAG CCA TCC A). All primers were synthesized at the University of Alberta Department of Biochemistry. The contents were mixed together and placed in the thermal cycler. The PCR amplification protocol was as follows: step 1: 94°C 4 min; step 2: 94°C 1 min, 53°C 1 min, 72°C 45 s (step 2 was repeated 35 times); step 3: 72°C 15 min; step 4: held at 4°C. A portion of the product was then run on a 2% TEA agarose gel, and the gel was then placed in a bath containing 2 µg/ml of ethidium bromide, after 10 min DNA bands were visualized with UV light box and photographed with a Polaroid camera.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Most of the acutely dissociated neurons from the DBB had neuron-like morphology (i.e., large cells with a conspicuous nucleus, nucleolus and a few blunt processes which were truncated axon/dendrites). The average membrane capacitance estimated electronically on the Axopatch-1D amplifier was 17.8 ± 0.3 pF (n = 134). Under our recording conditions, the average input conductance measured from the slope of the I-V relationships between 60 and 110 mV was 0.82 ± 0.18 nS (n = 51).

Based on the previous observations (Jassar et al. 1999), we utilized a voltage-ramp protocol where the cells were held at 80 mV and subjected to voltage ramps from 110 to +30 mV at the rate of 20 mV/s after conditioning at 110 mV for 1 s.

Effects of galanin on whole cell potassium currents in DBB neurons

The concentration of galanin (300 nM) used in the present experiment was based on previous electrophysiological studies of this peptide in acutely dissociated neurons (300 nM) (Puttick et al. 1994) or in brain slices (range, 100 nM-1 µM) (Kinney et al. 1998; Papas and Bourque 1997; Xu et al. 1999). Figure 1A shows the voltage-activated currents recorded from a DBB neuron under control conditions, in the presence of galanin (300 nM), and recovery after washout. Application of galanin resulted in a decrease in whole cell currents in the 30 to +30 mV range in 51 of 75 cells. In these cells, the amplitude of the currents at +30 mV was significantly reduced from 6.5 ± 0.3 nA under control conditions to 5.4 ± 0.2 nA (n = 51; P < 0.001) in the presence of galanin; this represents a decrease of 16.1 ± 1.2% at this potential. In 24 cells, galanin application either did not evoke any a change in whole cell current or caused an increase or decrease in the whole cell current that was <5% from control values at +30 mV.

View larger version (41K): [in this window] [in a new window]   Fig. 1. Effects of galanin on whole cell currents in chemically identified diagonal band of Broca (DBB) neurons. A: current-voltage plots of whole cell currents evoked in a DBB neuron under control conditions (before application of galanin), in the presence of 300 nM galanin and 9 min washout after galanin. B: photograph of a gel showing the results of single-cell RT-PCR on 2 cells, 1 that is galanin responsive and shows a band corresponding to molecular weight (MW) of the choline acetyltransferase (ChAT) primer and another that was galanin nonresponsive and shows a band corresponding to the MW of the glutamate decarboxylase (GAD) primer. The -actin lane on the gels serves as a control. MW for -actin is ~515, ChAT is ~308, and GAD is ~400.

Under our recording conditions, currents that are activated in the voltage range in which galanin exerts its actions include the classical Hodgkin-Huxley type delayed rectifier (I_K) and the calcium-activated potassium (I_K,Ca) currents. The fast inactivation kinetics of the transient outward potassium current (I_A) (Belluzzi et al. 1985; Cooper and Shrier 1985; Griffith and Sim 1990) makes it unlikely that these channels are activated during the relatively slow voltage ramps utilized in our experiments. This suggests that under these conditions the conductances affected by galanin include I_K and/or I_K,Ca.

Chemical identity of galanin responsive DBB neurons

Within the DBB, there are two main chemical neurotransmitter phenotypes of neuronsGABAergic and cholinergic. Definitive determination of the chemical phenotype was done by RT-PCR analysis. ChAT was used as a specific marker for cholinergic neurons, and GAD was used as a specific marker for GABAergic neurons. Figure 1B shows the photograph of a gel indicating RT-PCR product from a galanin responsive cell on the left (band corresponding to the molecular weight of the ChAT primer) and a galanin unresponsive neuron on the right (band corresponding to the molecular weight of GAD primer). Results from 21 DBB neurons recorded under voltage- or current-clamp modes and in which the PT-PCR reaction was unequivocal indicate that 17 galanin-responsive cells were ChAT positive and GAD negative. Four galanin unresponsive neurons were GAD positive and ChAT negative.

Effects of the blockade of calcium influx

Because the I_K,Ca currents are activated by Ca²⁺, which flows into the cell during the depolarizing phase of the action potential, their contribution to the voltage-activated conductances can be assessed by blocking Ca²⁺ influx (Lancaster et al. 1991). This was achieved by replacing the extracellular Ca²⁺ with Mg²⁺. Cd²⁺ (50 µM) was included in the external perfusing solution to ensure complete blockade of the Ca²⁺ influx. Figure 2A shows the average of current-voltage relationships obtained from eight neurons under control conditions, with 0 mM Ca²⁺ and 50 µM Cd²⁺ in the external medium, galanin in the presence of 0 mM Ca²⁺ and 50 µM Cd²⁺, and recovery. Replacing the external solution with 0 mM Ca²⁺ and 50 µM Cd²⁺ decreased the currents measured at +30 mV by 19.4 ± 3.4% (control = 5.4 ± 0.5 nA; 0 mM Ca²⁺ and 50 µM Cd²⁺ = 4.4 ± 0.5 nA). Application of galanin in the presence of 0 mM Ca²⁺ and 50 µM Cd²⁺ further reduced the currents by 4.9 ± 1.2% (0 mM Ca²⁺ and 50 µM Cd²⁺ with galanin = 4.1 ± 0.5 nA, P < 0.01). Thus although a significant amount of the galanin response was blocked in the presence of 0 mM Ca²⁺ and 50 µM Cd²⁺, a small but persistent reduction in currents remained under these conditions. This suggests that the effects of galanin are in part mediated via a Ca²⁺-dependent K⁺ current.

View larger version (49K): [in this window] [in a new window]   Fig. 2. Galanin effects on calcium-dependent and potassium currents. A: current-voltage (I-V) relationship of mean whole cell currents evoked by applying voltage ramps under control conditions, in the presence of 0 mM Ca2+ (and 50 µM Cd2+), and galanin (300 nM) in the presence of 0 mM Ca2+ (and 50 µM Cd2+; n = 8). Blockade of the Ca2+ influx by changing the normal external solution to the one containing 0 mM Ca2+ and 50 µM Cd2+ resulted in reduction in the outward currents in the voltage range from 110 to +30 mV. Galanin, under such conditions, caused a smaller but significant further reduction in the outward currents in the same voltage range. B: I-V relationship of mean whole cell currents evoked by applying voltage ramps under control conditions, in the presence of iberiotoxin (IBTX, 25 nM) alone, and galanin (300 nM) with IBTX. IBTX application resulted in a reduction in the outward currents in the voltage range from 110 to +30 mV. Galanin, under such conditions, caused a smaller but significant reduction in the outward currents in the same voltage range. Histograms in the inset depict the response of 3 consecutive galanin applications 10 min apart. C: I-V relationship of mean whole cell currents indicating that in the presence of 5 mM tetraetylammonium (TEA), the galanin-evoked reduction in currents are blocked. D: effects of galanin (300 nM) on mean barium currents (IBa) in DBB neurons. I-V relationships of the averaged IBas evoked in 16 neurons under control conditions, in the presence of galanin and recovery on washout. Inset: IBa evoked in a DBB neuron by applying a step command to 10 mV under control conditions and in the presence of galanin.

Effects of blocking calcium-activated potassium currents (I_K,Ca) on the galanin response

The contribution of I_K,Ca or I_C (BK channels) to the voltage-activated currents can be determined by using the selective blockers of this family of voltage-sensitive calcium-activated potassium channels. The more sensitive and specific ones are: iberiotoxin (IBTX; IC₅₀ = 0.25 nM) (Galvez et al. 1990), a snail toxin from Buthus tamulus.

Figure 2B shows the an average of voltage-ramp relationship of currents from 13 cells evoked under control conditions, with galanin (300 nM) alone, with IBTX (25 nM) alone, and with a combination of galanin and IBTX. Galanin or IBTX applied individually and sequentially caused a reduction in outward currents in the same voltage range (40 to +30 mV). Control currents were decreased 15.1 ± 3.1% by galanin alone (control = 7.8 ± 0.4 nA; galanin = 6.6 ± 0.4 nA) and 21 ± 5.6% by IBTX alone (control = 7.4 ± 0.5 nA; IBTX = 5.7 ± 0.4 nA). When galanin was applied in the presence of IBTX, there was an additional significant reduction in current (9.0 ± 2.3%, P < 0.001). Thus the galanin-evoked decrease in whole cell currents are partly mediated through an IBTX-sensitive calcium-activated potassium conductance. The inset in Fig. 2B also illustrates another important point, namely that the galanin-evoked reduction of whole cell currents under control conditions is a time-independent response. Three sequential applications of galanin, 10 min apart, did not show a significantly different magnitude of response (n = 6, P > 0.6), indicating that the galanin response does not desensitize with repeated applications.

The IBTX-induced decrease in the currents occurred in the same voltage range and was of a similar magnitude as that obtained by omitting calcium in the external perfusing medium (see Fig. 2A). Collectively, these data support the specificity and efficacy of 25 nM IBTX in blocking voltage-sensitive calcium-activated K⁺ currents in our preparation.

Blockade of galanin response with TEA

TEA ions at a concentration of 5 mM block I_K and I_C (Jassar et al. 1999). Figure 2C shows the average current-voltage relationships obtained from six DBB neurons under control conditions, in the presence of 5 mM TEA, and galanin in the presence of TEA. TEA blocked 74.3 ± 0.9% of the outward current at +30 mV (control = 7.3 ± 0.8 nA, TEA = 1.7 ± 0.2 nA, n = 6). Galanin failed to produce any significant effect on the remaining currents in the presence of TEA (TEA + galanin = 1.7 ± 0.2 nA, n = 6, P = 0.64).

Galanin has no effect on barium currents

Because galanin blocks I_C, one possible target of action can be the voltage-gated calcium channels through which the calcium influx responsible for the activation of I_C occurs. Currents through calcium channels were recorded using Ba²⁺ (I_Ba) as the charge carrier. The use of Ba²⁺ as a charge carrier has advantages over Ca²⁺ as it avoids Ca²⁺-induced inactivation of Ca²⁺ channels and is a better charge carrier (Griffith et al. 1994). Figure 2D shows an average of barium currents flowing through calcium channels evoked in 16 DBB neurons under control conditions and subsequently in the presence of galanin. The inset shows the effects of galanin on I_Ba evoked by applying a 25-ms step command to 10 mV from a holding potential of 90 mV. The maximum current-density was 96.0 ± 10.2 pA/pF at 0 mV under control conditions (n = 16) in this group of cells. Galanin did not affect the whole cell I_Bas at any voltage between 80 and +70 mV. The whole cell I_Bas at 0 mV were not significantly different from those under control conditions (control: 2.0 ± 0.1 nA; galanin: 1.9 ± 0.1 nA; n = 16, P > 0.05). Although three different kinds of voltage-gated calcium channels have been shown to be present in the rat magnocellular cholinergic basal forebrain neurons (Allen et al. 1993; Jassar et al. 1999), because of the absence of effects of galanin, it was deemed unnecessary to carry out any further biophysical or pharmacological characterization of Ca²⁺ currents.

Effects of galanin on transient outward (I_A) and the delayed rectifier (I_K) potassium currents

I_A and I_K are voltage-sensitive currents, and their activation and inactivation are strongly voltage dependent. I_A requires the holding potential to be relatively hyperpolarized (approximately 110 mV) for removal of its inactivation, whereas it is inactivated at 40 mV. On the other hand, I_K is not inactivated at 40 mV. These biophysical properties of I_A and I_K can, thus be utilized to isolate these currents. Therefore a conditioning pulse to 40 mV will activate I_K without any significant contamination by I_A (Connor and Stevens 1971; Easaw et al. 1999). A conditioning pulse to 120 mV will activate both I_A and I_K. The difference currents obtained by subtracting the currents evoked by depolarizing pulses following a conditioning pulse to 40 mV from those evoked following a conditioning pulse to 120 mV provide an accurate estimate of I_A. Figure 3A shows the currents recorded from a DBB neuron with a conditioning pulse to 40 mV for 150 ms, representing mainly I_K, under control conditions, in the presence of galanin and recovery on washout of galanin. Galanin reduced I_K by 9.2 ± 1.8% (control = 9.8 ± 0.5 nA, galanin = 9.1 ± 0.1.8 nA, n = 27, P < 0.001 at +30 mV). Figure 3B shows the difference currents recorded from the same neuron representing mainly I_A, under control conditions, in the presence of galanin and on washout. I_A was reduced significantly by galanin by 15.1 ± 4.1% in 18 of 27 cells (control = 5.0 ± 0.5 nA, galanin = 4.2 ± 0.4 nA, n = 18, P < 0.001). We have previously shown that the residual sustained current remaining at the end of the 100-ms test pulse (shown in Fig. 3B) consists mainly of I_K and I_C (Easaw et al. 1999), both of which are also reduced by galanin. Figure 3, C and D, shows the current-voltage relationships of averaged peak I_K (n = 27) and I_A (n = 18), respectively, using the voltage step protocols shown in Fig. 3, A and B. Galanin did not affect the activation characteristics of I_A (data not shown).

View larger version (27K): [in this window] [in a new window]   Fig. 3. Effects of galanin on delayed rectifier (IK) and transient outward (IA) potassium currents. A and B: the effects of galanin (300 nM) on IK and IA in a DBB neuron. In A, voltage protocol for recording IK is depicted on the left with holding potential of 80 mV and a 150-ms conditioning pulse to 40 mV. In this protocol, outward currents are mediated through IK (delayed rectifier) and IC (calcium-activated potassium conductance). B: in the same neuron, galanin causes a decrease in transient outward K+ currents (IA). IA was obtained as difference currents by subtracting the currents obtained by the voltage protocol shown in A from that obtained by applying the voltage protocol shown in B where cells were held at 80 mV and a 150-ms conditioning pulse to 120 mV was applied. C and D: the voltage-dependent (40 to +40 mV) changes in mean IK (n = 27) and IA (n = 18) currents, respectively, under control conditions and in the presence of galanin.

Effects of galanin on sodium currents

Sodium currents are involved in the fast depolarizing phase of the action potential. Enhancement of these currents can increase the excitability and vice versa. We recorded sodium currents in isolation to assess if galanin has any effects on these currents. The sodium currents were TTX-sensitive. The current-voltage relationships obtained from 7 cells reveal that galanin did not influence sodium currents in DBB neurons (control = 5.2 ± 0.4 nA, galanin = 5.2 ± 0.3 nA, P = 0.3 at +20 mV, not illustrated).

Effects of galanin on action potentials and excitability

We performed bridge current-clamp recordings in whole cell mode to examine the effects of galanin on DBB neurons. In addition, we also determined whether the galanin-induced changes in excitability that could be predicted from the observed decrease in currents through I_C channels that it produces under voltage-clamp conditions.

Figure 4A shows the trains of action potentials generated by injecting a 600-ms current pulse (1.0 nA) under control conditions, in the presence of galanin (300 nM), and on washout (recovery). The number of spikes generated by a 600-ms depolarizing current pulse was 8 ± 1.2 under control conditions and 13.5 ± 1.1 in the presence of galanin (n = 13, P < 0.001) and a recovery to near control values (6.7 ± 1.5). The ratio of interspike interval between the first two and the last two action potentials provides a measure of accommodation. A ratio approaching one indicates loss of accommodation. Under control conditions, this ratio was 0.56 ± 0.05 (n = 13), and it increased to 0.73 ± 0.04 (n = 13) in the presence of galanin (P < 0.01), indicating attenuation of accommodation in the presence of the peptide (Fig. 4B). On recovery, the ratio was 0.54 ± 0.04. Because I_C channels govern, in part, spike adaptation and excitability (Vergara et al. 1998), the depression of I_C currents by galanin is consistent with the loss of accommodation and increase in excitability caused by this peptide under current-clamp conditions.

View larger version (32K): [in this window] [in a new window]   Fig. 4. Galanin effects on excitability of DBB neurons. A: action potentials evoked by sustained current injections (600 ms: 1.0 nA) under control conditions, in the presence of 300 nM galanin, and on recovery, showing the increased excitability and loss of accommodation with the application of galanin. The number of action potentials generated by the same amount of current injection was increased in the presence of galanin. B: bar histograms that show the ratio between the 1st 2 (1st interspike interval) and the last 2 action potentials (last interspike interval) generated by a current injection as in A plotted for 13 cells under control conditions, in the presence of galanin and on washout (recovery). There is a significant increase in the ratio (P < 0.01) of the 1st and last intervals with galanin (compared with control), which indicates a loss of accommodation. C: chart recording showing the changes in resting membrane potential (RMP) induced by application of galanin (300 nM, top) and the recovery on washout. Galanin-evoked depolarization of the membrane is followed by increased excitability leading to spontaneous firing of action potentials (RMP of this cell = 67 mV). Middle: the changes in membrane voltage induced by application of galantide (10 nM) and the recovery on washout (RMP of this cell = 65 mV). IBTX (25 nM) application also evokes a membrane depolarization (RMP of this cell = 67 mV).

Figure 4C (top) shows the reversible, galanin-evoked depolarization of the membrane potential resulting in an increased firing rate (approximately 4 Hz). In a majority of cells, this depolarization was sufficient to bring the cell to threshold leading to spontaneous firing of action potentials. In 20 DBB neurons, application of 300 nM galanin resulted in depolarization of the RMP from 66.9 ± 1.4 to 57.6 ± 2.4 mV with recovery to 64.1 ± 7.7 mV. We also examined whether galantide, a putative antagonist for galanin, was able to block the excitatory effects of galanin. In eight cells, application of galantide even at doses of 1-100 nM resulted in a membrane depolarization similar to that observed for galanin (control RMP = 72.0 ± 2.2 mV, with galantide RMP = 63.9 ± 1.9 mV, Fig. 4C, middle). Iberiotoxin (25 nM), a blocker of I_C channels, applied to DBB neurons also evoked a similar degree of membrane depolarization as galanin and galantide (control RMP = 69.0 ± 1.4 mV, with iberitoxin RMP = 59.5 ± 1.7 mV, n = 8, Fig. 4C, bottom).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The findings reported here provide the first electrophysiological evidence that the neuropeptide galanin causes an excitation of mammalian central neurons. Activation of galanin receptors results in membrane depolarization and an increase in excitability of basal forebrain neurons, that is related in part, through its effects in reducing specific voltage- and calcium-dependent potassium conductances (I_C, I_A, and I_K). Furthermore, these actions of galanin are specific to cholinergic, and not GABAergic, basal forebrain neurons as identified by single-cell RT-PCR analysis.

Galanin modulation of ionic conductances and neuronal excitability

Several studies have examined the actions of galanin application on CNS neurons, and with one exception these studies have identified an inhibitory effect of this peptide. In the locus coeruleus, dorsal raphe, hypothalamic supraoptic nucleus, and CA3 hippocampal neurons, galanin caused a hyperpolarization of the postsynaptic membrane via a potassium conductance (Konopka et al. 1989; Papas and Bourque 1997; Pieribone et al. 1995; X. J. Xu et al. 1995; Z. Q. Xu et al. 1999) that was not specifically characterized but in locus coeruleus and dorsal raphe neurons was TEA sensitive. In dorsal root ganglion cells, galanin evoked an inward current that was accompanied by an increase in input conductance, but the specific ionic conductance underlying these effects was not identified (Puttick et al. 1994).

Our data show that galanin affects a suite of potassium conductances in basal forebrain DBB neurons. Among the potassium conductances affected, galanin decreases I_C currents without influencing calcium currents and leads us to conclude that galanin has a direct effect on I_C channels, similar to that we have previously observed for another peptide, vasopressin, in this region (Easaw et al. 1997). Since I_C provides a drive for the repolarization phase of the action potential and plays an important role in the process of spike frequency adaptation (accommodation) (Vergara et al. 1998), the galanin-induced blockade of I_C we observed here could explain the increase in excitability and loss of accommodation in DBB neurons with galanin. In addition, galanin consistently depolarized DBB neurons. Inhibition of I_C could contribute to membrane depolarization induced by galanin because application of iberiotoxin, a specific blocker I_C or BK channels also evoked a similar depolarizing response.

Fast-inactivating potassium channels (I_A) and the slower inactivating I_K are also important in modulating neuronal excitability, in part, through their effects on the early phase of spike repolarization (Storm 1990). Galanin-evoked inhibition of both I_A and I_K could lead to an overall increase in excitability of DBB neurons. Although galanin did not influence Ca²⁺ currents, prolonged depolarization resulting from a blockade of K⁺ currents could also lead to increased total Ca²⁺ influx and influence the excitability of the neuron.

Galanin receptor activation on cholinergic basal forebrain neurons

Our single-cell RT-PCR data indicate that the galanin-induced reduction of whole cell currents and the increase in excitability is specific to cholinergic and not GABA-synthesizing DBB neurons. Three G-protein-coupled receptor subtypes for galanin have been cloned (GALR1, GALR2, and GALR3) (for review, see Branchek et al. 2000). Although both GALR1 and GALR2 have been localized in the DBB (O'Donnell et al. 1999), co-expression of GALR1 and ChAT (a marker for cholinergic neurons) mRNAs was not detected in the basal forebrain nuclei (Miller 1998). This suggests that the effects of galanin on DBB neurons may be mediated via GALR2 or possibly GALR3 receptors. We attempted to utilize galantide (M15) as a putative galanin antagonist, but even at relatively low doses, this chimeric ligand acted as a galanin agonist causing membrane depolarization and increase in firing of DBB neurons. These galanin agonist-like actions have been reported in several previous studies using galantide and similar other chimeric galanin receptor ligands (Kask et al. 1995; Kinney et al. 1998; Papas and Bourque 1997). Development of pharmacologically selective antagonists for the cloned galanin receptors is necessary before a definitive characterization of the receptor on cholinergic basal forebrain neurons is possible.

Functional implications of excitatory effects of galanin

Although the origin of the endogenous galanin input to the DBB is unknown, the expression of galanin and its behavioral effects in the cholinergic basal forebrain have become increasingly important in the context of cognitive impairments seen in normal aging and diseases such as AD. In AD, degeneration of cholinergic basal forebrain neurons, which is linked to memory and spatial learning deficits, has been associated with a parallel upregulation of galanin synthesis (Mufson et al. 1998). There is thus a galininergic hyperinnervation of the remaining cholinergic basal forebrain neurons in the brains of AD patients (Chan-Palay 1988). On the basis of galanin-induced inhibition of ACh release (Fisone et al. 1987), it has been hypothesized that the increased galanin innervation of cholinergic neurons could play an important role in worsening cognitive function of AD patients. This postulate fits with some behavioral studies where central injections of galanin into the DBB and medial septum result in disruption of short-term memory tasks in rodents (Givens et al. 1992; Mastropaolo et al. 1988) and the inhibitory actions of galanin at a cellular level that have been observed at CNS sites other than the basal forebrain.

Our findings of an excitatory effect of galanin on cholinergic basal forebrain neurons, which are at the epicenter of pathology in AD, cannot be reconciled with the above hypothesis. An alternative interpretation is that galanin overexpression in AD may serve a compensatory role by augmenting the release of ACh from the remaining cholinergic basal forebrain neurons. This notion is supported by the observation that NGF, which rescues degenerating cholinergic neurons and increases ACh turnover by upregulating ChAT synthesis (Rylett et al. 1993; Williams et al. 1986), also induces galanin gene expression in the cholinergic basal forebrain (Planas et al. 1997). We suggest that the such an increase in galanin, through its excitatory effects on the cholinergic basal forebrain neurons, could cause increased release of ACh, which would be entirely consistent with the trophic effects of NGF on cholinergic function. Our findings would also explain the recent observation that adult mice carrying a targeted loss-of-function mutation in the galanin gene demonstrate deficits in cholinergic function and impairments in behavioral and electrophysiological tests of memory (O'Meara et al. 2000). An important consequence of the findings presented here relates to the development of galanin antagonists, which have been promoted as a possible treatment option for AD. Our findings should inject a note of caution in viewing such antagonists as "neuroprotective," a notion that is based on their potential to reverse the inhibitory effects of galanin, which may be receptor and region specific in the CNS.