β1 Adrenergic Receptor-Mediated Enhancement of Hippocampal CA3 Network Activity

Although the existence of adrenergic neurons within the central nervous system (CNS) has been well established, many of their functions remain unexplored. It is known that most norepinephrine (NE)-containing neurons arise from brain stem nuclei and project multiple branched axons throughout the CNS (Moore and Bloom, 1979). Due to this rich and widespread projection of adrenergic axons, it is thought that the neurotransmitter NE plays a modulatory role in some global aspects of brain function. Indeed, the adrenergic system is shown to be involved in a variety of CNS functions, including regulating the sleep-wake cycle (Berridge and Waterhouse, 2003), modulating electroencephalographic activity (Foote et al., 1983), promoting a state of vigilance (Aston-Jones and Bloom, 1981), thermoregulation (Philipson, 2002), and enhancing learning and memory (Kobayashi and Kobayashi, 2001).

Among the many targets of NE-containing axons in the brain is the hippocampus (Jones and Moore, 1977; Loy et al., 1980), which receives one of the highest densities of adrenergic terminals in the CNS (Schroeter et al., 2000). The hippocampus occupies a central position in the neural circuits that govern emotions, motivation, attention, and certain types of memory processes (Milner et al., 1998; Eichenbaum, 2000). The hippocampus also has a pathological role in Alzheimer's disease (Hyman et al., 1984; Braak and Braak, 1997) and schizophrenia (Luchins, 1990), has been shown to be exceptionally vulnerable to ischemia and anoxia (Tekkok and Ransom, 2004), and is of particular interest in epilepsy due to its low seizure threshold and frequent involvement in hyperexcitable episodes (Johnston and Amaral, 2004). The hippocampus is also the site where a form of activity-dependent synaptic enhancement, called long-term potentiation (LTP), was first shown. This potentiation of excitatory activity may be the cellular basis of learning and memory (Bliss and Collingridge, 1993).

Specificity in the adrenergic system is believed to be achieved primarily through a distinct and diverse expression pattern of postsynaptic adrenergic receptors (ARs). Activation of these ARs is often seen to produce state-dependent or enabling effects, with different receptor subtypes often mediating opposing effects within a cell. Pharmacological and molecular cloning studies have revealed the existence of three βAR subtypes (β1, β2, and β3), based upon both sequence homologies and affinity values for subtype-selective AR antagonists. Although all βARs are positively coupled to adenylyl cyclase via activation of the G protein G_s, each subtype has its own unique pharmacological characteristics, particularly for catecholamine AR agonists (Pupo and Minneman, 2001). For example, the relative potencies of isoproterenol (ISO), NE, and epinephrine (EPI) differ at each receptor subtype: ISO > NE ≥ EPI for β1; ISO > EPI > NE for β2; and ISO > NE > EPI for β3AR (Lands et al., 1967; Emorine et al., 1989; Bylund et al., 1994). In addition, ISO is essentially equipotent at β1 and β2AR subtypes, whereas it is considerably less potent for the β3AR (Liggett, 1992; Galitzky et al., 1995; Shimizu et al., 1996). Although much is known about the role of βAR subtypes in the periphery, far less is known about their function in the CNS.

All three βAR subtypes are found in the brain, albeit the β3AR subtype seems to have a very limited expression pattern (Pupo and Minneman, 2001). In the hippocampus, βAR activation has been shown to facilitate LTP (Hopkins and Johnston, 1984; Gelinas and Nguyen, 2005) as well as to enhance certain memory processes (Murchison et al., 2004). Delineating which subtype of βAR mediates specific physiological functions in the hippocampus would help clarify a role for the AR system in cognitive function.

The goal of this study was to characterize the effect of βAR subtype activation on hippocampal cornu ammonis 3 (CA3) network synchronization in the rat hippocampus. Immunohistochemical staining and rank order of potencies for catecholamine-mediated responses suggested that activation of a β1AR population was increasing synchronous burst activity in the CA3 region. Functional determination of subtype-selective βAR antagonist affinity values were used to support this assumption. K_b values calculated using Schild regression analysis confirmed our hypothesis that β1AR activation is responsible for the enhanced hippocampal CA3 network activity observed in this model of adrenergic transmission.

Materials and Methods

Reagents. The pharmacological agents used were obtained from the following sources: atenolol, butoxamine hydrochloride, (–)-epinephrine bitartrate salt, (–)-isoproterenol (+)-bitartrate salt dehydrate, (±)-metoprolol (+)-tartrate salt, (–)-norepinephrine (+)-bitartrate salt hydrate, phentolamine methanesulfonate salt, and picrotoxin (Sigma-Aldrich, St. Louis, MO); ICI-118,551 (Tocris Cookston Inc., Ellisville, MO); sodium pentobarbital (Ampro Pharmaceutical, Arcadia, CA); and isoflurane (Abbott Diagnostics, Chicago, IL). The immunohistological agents used were obtained from Sigma-Aldrich, including sodium nitrite, paraformaldehyde, para-picric acid, sodium phosphate buffer, sucrose, glycerol, and Triton X-100. All reagents used to make the artificial cerebrospinal fluid (ACSF) were from J. T. Baker (Phillipsburg, NJ).

Animals. Sprague-Dawley rats, postnatal day 12 to 29 (P12–29) were housed with their mothers in cages (16.5 × 8.5 inches) kept in rooms maintained at a temperature of ∼22°C with a relative humidity of ∼55%. Water and dried laboratory food (Teklad Global 18% Protein Rodent Diet; Harlan Teklad, Madison, WI) were provided ad libitum. Lighting was set to a 12-h light/dark cycle (lights on at 7:00 AM). Rats were allowed to acclimate for 4 days after arrival from Harlan (Indianapolis, IN) before their use. All protocols described were approved by the Institutional Animal Care and Use Committee of the University of North Dakota in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Immunocytochemistry. Five Sprague-Dawley rats (P20–21; 50–70 g) were anesthetized with pentobarbital and perfused transcardially with 100 ml of ice-cold (4°C) 0.9% saline containing 0.1% sodium nitrite followed by 400 ml of ice-cold, freshly prepared fixative consisting of 4% paraformaldehyde and 0.16% para-picric acid in 0.1 M sodium phosphate buffer, pH 7.4. The brain was removed immediately after perfusion and postfixed for 2 h in ice-cold fixative followed by cryoprotection for at least 48 h in ice-cold (4°C) 25% sucrose and 10% glycerol in 50 mM phosphate buffer. Sagittal sections of brain (20 μm in thickness) were cut on a freezing sliding microtome (Leica 3000R) and collected into 0.1 M sodium phosphate buffer, pH 7.4, containing 0.9% saline (phosphate-buffered saline) and processed for immunohistochemistry as described below.

Whole brain sections were incubated for 48 to 72 h at 4°C with anti-β1AR (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) or anti-calmodulin kinase-II (CaMK-II; Chemicon International, Temecula, CA) primary antibodies alone or simultaneously. CaMK-II is a marker of pyramidal neurons. All primary antibodies were diluted 1:200 in phosphate-buffered saline containing 0.3% Triton X-100 (PBS-T). After the primary incubation, the sections were washed in two 20-min washes in PBS-T and then incubated for 1.5 h at room temperature with CY3-conjugated donkey anti-rabbit or donkey anti-mouse IgG secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA) diluted 1:100 in PBS-T. Sections were then washed for 20 min in PBS-T, 20 min in 50 mM Tris-HCl, and mounted onto gel-coated slides using 50 mM Tris-HCl, pH 7.4. Slides were coverslipped with Vectashield (Vector Laboratories, Burlingame, CA) antifade mounting medium. For double-immunofluorescence labeling, anti-β1AR and anti-CaMK-II primary antibody labeling was visualized using CY3-conjugated donkey anti-rabbit IgG and fluorescein isothiocyanate-conjugated donkey anti-rabbit IgG secondary antibodies (both 1:100; Jackson ImmunoResearch Laboratories). Immunohistochemical controls consisted of primary omission for both single and double immunohistochemistry labeling. Analysis of immunoreactivity and determination of labeling colocalization were undertaken using standard fluorescence microscopy (BX-60; Olympus, Melville, NY) and image analysis (Spot RT slider; ImagePro, Sterling Heights, MI).

Slice Preparation. Hippocampal brain slices were prepared from Sprague-Dawley rats (P12–29; 25–90 g) as follows. Briefly, animals were deeply anesthetized with isoflurane, sacrificed by decapitation, and their brains were rapidly removed. Hippocampi were dissected from each hemisphere and placed into a beaker of ice-cold saline solution containing 110 mM choline chloride, 2.5 mM KCl, 7 mM MgSO₄, 0.5 mM CaCl₂, 1.25 mM NaH₂PO₄, 25 mM NaHCO₃, 25 mM glucose, 11.6 mM sodium ascorbate, and 3.1 mM sodium pyruvate. The hippocampi were sectioned transversely into 500-μm-thick slices using a conventional tissue sectioning apparatus (Stoelting, Wood Dale, IL). The hippocampal slices were incubated at 34 ± 1°C in ACSF containing in 119 mM NaCl, 5 mM KCl, 1.3 mM MgSO₄, 2.5 mM CaCl₂, 1 mM NaH₂PO₄, 26.2 mM NaHCO₃, and 11 mM glucose, for 30 min before being allowed to recover for at least an additional 30 min at room temperature (22 ± 1°C). All solutions were continually aerated with 95% O₂, 5% CO₂.

Electrophysiological Recordings. A single slice was transferred to the recording chamber, where it was submerged and superfused continuously at a rate of 2 to 4 ml/min with ACSF. Glass microelectrodes were made using a two-stage puller (PP-830; Narashige, Tokyo, Japan). Extracellular field potentials were recorded using microelectrodes filled with 3 M NaCl and placed in the stratum pyramidale of CA3 region of the hippocampus. Currents were detected using an Axoclamp 2B (Axon Instruments Inc., Union City, CA), amplified at 10 or 100× using a Brownlee Precision model 440 instrumentation amplifier (Brownlee Precision, San Jose, CA), digitized at 1 kHz with a Digidata 1322A analog-to-digital converter (Axon Instruments Inc.), and recorded using Axoscope 9.0 software (Axon Instruments Inc.).

It has been established that hippocampal CA3 pyramidal neurons will fire burst discharges due in part to extensive recurrent circuitry (Traub et al., 1991). This activity can be elicited by attenuating synaptic inhibition using a GABA_A receptor antagonist such as picrotoxin. Since this particular characteristic is present exclusively in CA3 pyramidal cells, the frequency of burst discharges can serve as a selective measure of hippocampal CA3 neural network activity.

All experiments were performed at room temperature (22 ± 1°C). Slices were continually superfused with ACSF containing picrotoxin at 100 μM (to elicit burst discharges) and any applicable receptor antagonists. If no burst discharges were seen after 20 min of perfusion, the slices were determined to be unresponsive and discarded. Once burst discharges were evident, 30 min of baseline data was recorded before any exposure to agonist. Preliminary experiments were conducted to ensure that pharmacological antagonists showed no effect on their own and that each agonist dose produced its maximum effect in the time allotted (data not shown).

Data Analysis. Burst discharge frequency was analyzed using Mini Analysis 6.0 (Synaptosoft, Decatur, GA). Frequency versus agonist concentration data were then entered into GraphPad Prism 4.0 (GraphPad Software Inc., San Diego, CA), and concentration-response curves were constructed using a nonlinear least-squares curve-fitting method. Concentration-response curves for an agonist were plotted as percentage of maximal response. Each curve was fit with a standard (slope = unity) or variable slope, and the best fit was determined using an F-test with a value of P < 0.05. The calculated EC₅₀ value was used as a measurement of agonist potency. Significance between groups was tested using an unpaired two-tailed Student's t test (P < 0.05).

Schild analysis was used to functionally determine apparent equilibrium dissociation constants (pK_b) for atenolol, butoxamine, ICI-118,551, and metroprolol (Arunlakshana and Schild, 1959). For each experiment, cumulative concentration-response curves were performed in adjacent hippocampal slices of the same rat (one concentration-response curve per slice). Dose ratios of EC₅₀ values in the presence and absence of a selective AR antagonist were calculated and Schild plots constructed by graphing the log of the dose ratio – 1 versus the log of the concentration of that AR antagonist (Arunlakshana and Schild, 1959). Linear regression analysis of these plotted points was used to determine the slope and x-intercept of the Schild regressions. Schild regression slopes are expressed as the mean ± 95% confidence interval and were only considered different from unity if the 95% confidence interval did not include the value of 1 (Kenakin, 1997). The pK_b values of subtype-selective βAR antagonists causing competitive inhibition of agonist initiated burst discharge frequencies were calculated from Schild regression x-intercepts. Differences in pK_b values and Schild regression slopes were determined by analysis of covariance with a P < 0.05 level of probability accepted as significant. Calculated values (i.e., EC₅₀ and K_b) are expressed as the mean ± S.E.M. for n experiments.

Results

β1AR Distribution in the Rat Hippocampal Formation. Sagittal sections of p21 rat whole brain displayed abundant and intense β1AR and CaMK-II immunolabeling in the hippocampus. β1AR immunofluorescent puncta were observed along the plasma membrane of both the perikarya and proximal dendrites of putative pyramidal cells in the hippocampal CA3 region (Fig. 1, A and B). Compared with the perikarya, dendrites displayed a lower concentration of β1AR immunofluorescent labeling. As a control, a similar intense β1AR immunofluorescent labeling was observed in fixed sections of the rat heart (Fig. 1A, inset), and alternatively, minimal β1AR labeling was documented in tissue preparations from the rat lung (data not shown). Within the hippocampus, β1AR immunoreactivity was observed on pyramidal cells in all three cytoarchitectural regions, with cells in the CA3 displaying the heaviest labeling (Fig. 1A). CaMK-II immunofluorescence was also observed within pyramidal cells of the stratum pyramidale in the same cytoarchitectural regions of the hippocampus. This labeling was localized within both perikarya and proximal processes (Fig. 1C). These CaMK-II results reflect those reported previously and confirm the utility of CaMK-II as a marker of pyramidal cells in the hippocampus (Goto et al., 1993). In tissue double labeled for CaMK-II and β1AR, CaMK-II-labeled pyramidal cells in the CA3 also displayed intense and abundant somatic and dendritic β1AR immunofluorescent puncta (Fig. 1D). These results suggest that there is a distinct pattern of β1AR expression located predominantly on the hippocampal CA3 pyramidal neurons. These data, however, do not give an indication as to the role this βAR subtype may be performing with respect to hippocampal network activity.

βAR Effects on Hippocampal CA3 Network Activity. To determine possible βAR action on hippocampal activity, we examined the effects of AR agonists on burst discharges recorded extracellularly from the hippocampal CA3 pyramidal cell layer. For these recordings, the electrode was placed precisely in the area of the CA3 region that showed the greatest intensity of β1AR and CaMK-II immunofluorescence (Fig. 2). As illustrated in Fig. 2A, picrotoxin-induced burst discharges seen as sharp biphasic spikes. The depolarizing/hyperpolarizing waveform corresponds to a series of population spikes followed by an afterhyperpolarization in CA3 pyramidal neurons (Traub et al., 1991). Application of the selective βAR agonist ISO caused a concentration-dependent increase in the number of these events (Fig. 2). Using a frequency histogram of the ISO-induced increase in burst discharges (Fig. 2B), a concentration-response curve can be constructed from a plot of maximal burst frequency versus ISO concentration (Fig. 2B, inset). For this experiment, the EC₅₀ value calculated from nonlinear regression analysis was 13.7 nM. Overall, the mean EC₅₀ value for ISO-induced increased burst firing was 13.6 ± 1.3 nM (n = 14) (Fig. 2C), suggesting that ISO initiates an enhancement of hippocampal CA3 network activity most likely through βAR activation.

Effects of Endogenous Catecholamines on Hippocampal CA3 Network Activity. Since NE and EPI are the endogenous agonists for βAR-mediated responses in the rat hippocampus, we compared the effects of these nonselective AR agonists to ISO on hippocampal CA3 burst activity frequency in the presence of αAR blockade. After pretreatment of slices with 10 μM phentolamine, application of ISO, NE, or EPI again caused a concentration-dependent increase in the frequency of hippocampal burst discharges (Fig. 3). The potency of ISO in the presence of phentolamine (14.4 ± 1.8 nM; n = 5) was not significantly different from ISO EC₅₀ values calculated in the absence of αAR blockage (13.6 ± 1.3 nM; n = 14). This illustrates that the ISO-mediated effects on the hippocampus are caused by selective βAR activation and not due to a nonspecific AR response. Conversely, the EC₅₀ of NE (307 ± 64 nM; n = 5) and EPI (231 ± 21 nM; n = 5) for increasing CA3 network activity were similar compared with each other but were significantly less potent when evaluated against ISO. This rank order of potency (ISO > NE = EPI) is consistent with results by others who historically characterized catecholamine-initiated responses caused by β1AR activation (Lands et al., 1967). As expected for full AR agonists, there were no significant differences between the maximal effects produced by NE, EPI, or ISO (data not shown).

Effect of Selective β1AR Competitive Antagonists on the ISO-Mediated Increase in Discharge Frequencies. Functional determination of subtype-selective βAR antagonist affinity values were used to characterize the βAR subtype mediating increased burst frequency in the hippocampal CA3 region. Initially, CA3 field potential recordings generated by increasing amounts of ISO in the absence and presence of fixed selective β1AR antagonist concentrations were used for Schild regression analysis (Fig. 4). Hippocampal slices that had been pretreated with 200, 500, 1000, and 2000 nM atenolol produced 4-, 8-, 16-, and 30-fold parallel rightward shifts of the fitted ISO concentration-response curve (Fig. 4A). Dose ratios calculated for individual runs were plotted against each atenolol concentration to generate a straight line using linear regression analysis (Fig. 4B). The Schild regression slope included the value of unity (1.1 ± 0.5), and the x-intercept of the regression line represents the atenolol equilibrium dissociation constant (pK_b) for the βAR subtype mediating the increased burst frequency. The apparent K_b of 85 ± 36 nM correlates to previously published value where atenolol was used to identify the β1AR subtype (Table 1).

Similar experiments using 10, 30, 100, and 300 nM of metoprolol again produced 3-, 6-, 26-, and 56-fold parallel rightward shifts of the fitted ISO concentration-response curve (Fig. 4C). The Schild regression slope for metoprolol (Fig. 4D) included the value of unity (1.0 ± 0.2) and the calculated apparent K_b (3.9 ± 1.7 nM) correlates to published values of metoprolol for the β1AR subtype (Table 1). The high-affinity values calculated for these subtype-selective βAR antagonists to inhibit ISO initiated burst activity suggests that β1AR activation is the specific AR mediating this response in the hippocampal CA3 region.

Effect of Selective β2AR Competitive Antagonists on the ISO-Mediated Increase in Discharge Frequencies. Apparent affinity values of selective β2AR competitive antagonists ICI-118,551 and butoxamine were also calculated to corroborate results obtained using selective β1AR competitive antagonists. In these experiments, hippocampal slices pretreated with 500, 1000, 3000, and 10,000 nM ICI-118,551 produced 3-, 5-, 9-, and 27-fold parallel rightward shifts of the fitted ISO concentration-response curve (Fig. 5A). Dose ratios calculated for each ICI-118,551 concentration were used for Schild regression analysis (Fig. 5B). The slope of the regression line included the value of unity (0.9 ± 0.2), and the apparent K_b of 222 ± 61 nM correlates to previously published values where ICI-118,551 was used to identify the β1AR subtype (Table 1).

Comparable experiments using 10,000, 20,000, and 30,000 nM butoxamine once more produced 2-, 4-, and 5-fold parallel rightward shifts of the fitted ISO concentration-response curve (Fig. 5C). The Schild regression slope for butoxamine (Fig. 5D) included the value of unity (1.1 ± 0.3), and the calculated apparent K_b (9268 ± 512 nM) correlates to published values of butoxamine for the β1AR subtype (Table 1). The low-affinity values calculated for these selective β2AR antagonists to inhibit ISO initiated burst activity indicates that β1AR activation is the specific AR mediating this CA3 response.

Discussion

There is currently little information regarding the morphological or functional localization of ARs in the hippocampus. In this study, we aim to offer evidence for a specific βAR subtype-mediated enhancement of hippocampal CA3 network activity. Previous studies have suggested β1AR activation may mediate powerful excitatory actions on pyramidal cells in the hippocampus. Therefore, we first investigated an anatomical basis for this excitatory action by examining the abundance and distribution of immunohistochemically localized β1AR on the somatic and dendritic plasma membrane of CaMK-II-identified pyramidal cells (Goto et al., 1993). β1AR labeling was present on the perikarya and proximal dendrites of pyramidal cells in all three cytoarchitectural regions in the rat hippocampus. The heaviest labeling was observed along the plasma membrane and within the cytoplasm of cells within the CA3 region. Less abundant β1AR labeling was observed in the CA1 pyramidal cell layer compared with CA3, which is indicative of a distinct pattern for βAR subtype expression. This observation is supported by molecular and functional studies that show a lack of β1AR expression and functionality from CA1 pyramidal cells (Hillman et al., 2005a,b). This distinct and diverse βAR subtype expression pattern would provide a means for AR signaling specificity in the hippocampus.

Before introduction of subtype-specific AR antagonists, βAR subtypes were traditionally identified using a profile of AR agonists (Lands et al., 1967). In the present study, application of the selective βAR agonist ISO or the nonselective AR agonists, NE or EPI, in the presence of αAR blockade caused a concentration-dependent increase in the frequency of hippocampal CA3 burst discharges. For this response, ISO was found to have the highest potency of these tested AR agonists, whereas NE was equipotent to EPI. The high potency of ISO to initiate an increased CA3 burst discharge implies that β3AR activation is not mediating this function (Hoffmann et al., 2004). Furthermore, a rank order of potency demonstrating ISO > NE = EPI implies that the β1AR population identified from immunohistochemical analysis may be responsible for this hippocampal CA3 hyperexcitability (Lands et al., 1967; Bylund et al., 1994).

Additional support of our hypothesis for a β1AR mediation of hippocampal CA3 hyperexcitability comes from functional determination of subtype-selective βAR antagonist affinity values using the method of Arunlakshana and Schild (1959). Increasing concentrations of all βAR antagonists used in this study caused parallel rightward shifts in the ISO concentration-response curves without significantly reducing maximal effects, demonstrating the competitive property of these receptor antagonists. Low apparent equilibrium dissociation constants calculated for the selective β1AR antagonists atenolol and metoprolol support our belief that β1AR activation enhances CA3 network activity. Moreover, high apparent K_b values determined for the selective β2AR antagonists ICI-118,551 and butoxamine maintain our assertion that activated β1AR subtypes are mediating this hippocampal CA3 response. The pharmacological profile in our studies is comparable with the K_i values of these subtype-selective βAR antagonists for β1ARs as determined by others using radio-ligand binding studies (Quast and Vollmer, 1984; Juberg et al., 1985). In addition, the rank order of affinity values for these subtype-selective βAR antagonists (metroprolol > atenolol » ICI-118,551 > butoxamine) is consistent with the characteristics for a β1AR subtype described in other investigations (Table 1). Together, these results indicate that ISO activation of β1AR subtypes is mediating the increased network activity in the CA3 hippocampus.

Until recently, little was known about the role of hippocampal CA3 activity in the cognitive functions of attention and memory, except that LTP had been demonstrated at the recurrent collateral-CA3 synapses. However, recent evidence indicates that the CA3 pyramidal neurons, which mediate network activity, play a crucial role in associative memory recall (Nakazawa et al., 2002) and rapid memory acquisition (Nakazawa et al., 2003). Furthermore, new in vivo studies with mutant mice lacking NE has confirmed a role for NE in memory retrieval (Murchison et al., 2004). Moreover, selective use of AR agents also suggested that this NE function required signaling through a hippocampal β1AR subtype. Since βAR activation has been previously shown to decrease calcium-activated potassium conductance by a cAMP-dependent mechanism (Madison and Nicoll, 1986a,b), we speculate that β1AR activation enhances hippocampal CA3 network activity by increasing glutamatergic neurotransmission between recurrent collaterals of pyramidal cells, most likely by decreasing calcium-activated potassium conductance in a cAMP-dependent manner. Additional experiments are currently being performed to confirm this hypothesis.

In conclusion, we provide functional evidence that β1AR activation increases hippocampal CA3 burst activity. We hypothesize that this β1AR-mediated enhancement of CA3 network activity could be the underlying mechanism through which NE reinforces certain cognitive processes such as memory retrieval in the hippocampus.