Histaminergic modulation of nonspecific plasticity of the auditory system and differential gating

Abstract

In the auditory system of the big brown bat (Eptesicus fuscus), paired conditioned tonal (CS) and unconditioned leg stimuli (US) for auditory fear conditioning elicit tone-specific plasticity represented by best-frequency (BF) shifts that are augmented by acetylcholine, whereas unpaired CS and US for pseudoconditioning elicit a small BF shift and prominent nonspecific plasticity at the same time. The latter represents the nonspecific augmentations of auditory responses accompanied by the broadening of frequency tuning and decrease in threshold. It is unknown which neuromodulators are important in evoking the nonspecific plasticity. We found that histamine (HA) and an HA3 receptor (HA3R) agonist (α-methyl-HA) decreased, but an HA3R antagonist (thioperamide) increased, cortical auditory responses; that the HA3R agonist applied to the primary auditory cortex before pseudoconditioning abolished the nonspecific augmentation in the cortex without affecting the small cortical BF shift; and that antagonists of acetylcholine, norepinephrine, dopamine, and serotonin receptors did not abolish the nonspecific augmentation elicited by pseudoconditioning. The histaminergic system plays an important role in eliciting the arousal and defensive behavior, possibly through nonspecific augmentation. Thus HA modulates the nonspecific augmentation, whereas acetylcholine amplifies the BF shifts. These two neuromodulators may mediate differential gating of cortical plasticity.

in the central auditory system, conditioned (CS; weak tone bursts) and unconditioned stimuli (US; electric leg shock) elicit either tone-specific plasticity [best-frequency (BF) shifts] or nonspecific plasticity (nonspecific augmentation of auditory responses), depending on whether they are paired or unpaired (Bakin et al. 1992; Ji and Suga 2008, 2009; see Suga et al. 2002 and Weinberger 1998 for reviews). In the big brown bat (Eptesicus fuscus), the CS alone elicits a small short-term BF shift toward the CS frequency when it is ∼5 kHz lower than the BF of a given neuron (Chowdhury and Suga 2000; Gao and Suga 1998; Yan and Suga 1998). Such a small short-term BF shift is termed the “CS-dependent” BF shift (Ji and Suga 2008, 2009). A randomized US activates the somatosensory system and neuromodulatory systems (Boulis and Davis 1989; Magoun 1963) and elicits nonspecific augmentation, i.e., augmentation of auditory responses to tones in a wide frequency range (Ji and Suga 2008). Therefore, Suga (2008) proposed two hypotheses: 1) when the CS-dependent BF shift is increased by pairing the CS and US, the nonspecific augmentation is suppressed or not increased, and 2) when the nonspecific augmentation is increased by pseudoconditioning (unpaired CS and US), the CS-dependent BF shift is suppressed or not increased. Ji and Suga (2008, 2009) have examined the first hypothesis and found that acetylcholine (ACh) does not increase the nonspecific plasticity elicited by the pseudoconditioning, although it plays an important role in eliciting the tone-specific plasticity evoked by the auditory fear conditioning (Bakin et al. 1992; Ji et al. 2001; Weinberger 1998). It has not yet been determined which neuromodulator(s) affect the nonspecific plasticity. Our current study tests the hypothesis that histamine acts to suppress the CS-dependent BF shift when the nonspecific augmentation is increased by pseudoconditioning.

Histamine (HA) is broadly released in the brain and plays an important role in defensive behavior and arousal (see Haas et al. 2008 for reviews). HA3 receptors (HA3R) are abundant in the cerebral cortex and act not only as autoreceptors but also as heteroreceptors for regulating the release of ACh, norepinephrine (NE), dopamine (DA), serotonin (5-HT), glutamate, and GABA (Arrang et al. 1983; Haas and Panula 2003; Takeda et al. 1984; Watanabe et al. 1984).

In our current study, pseudoconditioning delivered to the big brown bat elicited both a prominent nonspecific augmentation and small short-lasting CS-dependent BF shifts of cortical and collicular neurons. The nonspecific augmentation included an overall response increase, a broadening of the frequency tuning, a decrease in threshold, and a shortening of response latency (Ji and Suga 2008, 2009). We report in this article that the antagonists of ACh, NE, DA, and 5-HT receptors applied to the primary auditory cortex (AI) did not block the nonspecific augmentation, whereas HA3R antagonists evoked strong nonspecific augmentation and the HA3R agonist abolished the nonspecific augmentation. The HA3R agonist and antagonist did not affect the CS-dependent BF shifts. We conclude that the histaminergic system plays an important role in eliciting the nonspecific plasticity, but not in the tone-specific plasticity, and that ACh, NE, DA, and 5-HT appear not to be important in eliciting or modulating the nonspecific plasticity.

MATERIALS AND METHODS

Experimental subjects.

Forty-one adult big brown bats (E. fuscus, 18–24 g body weight) were used for the experiments. The Animal Studies Committee of Washington University in St. Louis approved all experimental procedures, which were the same as those previously described (Gao and Suga 2000; Ji and Suga 2008; Ji et al. 2001).

Surgical and recording procedures.

Under neuroleptanalgesia (Innovar 4.08 mg/kg body weight; Innovar is 0.4 ml of fentanyl citrate in 20 mg/ml droperidol), a 1.5-cm-long metal post was glued on to the dorsal surface of the bat's skull. Two to three days before the experiments, the bat was trained to adapt to staying in a polyethylene foam body mold suspended at the center of a soundproof room maintained at 31°C. Experiments began 3–4 days after the surgery. An awake bat was placed in the body mold. The bat's head was immobilized by fixing the metal post onto a metal rod with set screws to ensure a uniform and stable position of the head, which directly faced a loudspeaker located 80 cm away. A tungsten-wire microelectrode with a tip diameter of ∼7 μm was inserted orthogonally into the AI, and a single-unit recording was made at a depth between 200 and 900 μm. Local anesthetic (5% lidocaine; E. Fougera) and antibiotic (0.2% nitrofurazone; RXV Products) ointments were applied to the surgical wound. The recording session lasted up to 7 h. The data acquisition was disrupted and restarted if the bat in the body mold moved. Only one neuron was studied in a 1-day experiment. Water was provided with a dropper, and lidocaine was reapplied every 2 h. The bat was neither anesthetized nor tranquilized during the experiments. If the bat continued to move, recordings were terminated and the bat was returned to its cage. We gave the animal a 3-day rest between experiments.

Acoustic stimulation.

Acoustic stimuli were 20-ms tone bursts, including a 0.5-ms rise-decay time. They were delivered to the bat at a rate of 4/s from a leaf tweeter with real-time processors (2.1; Tucker-Davis Technologies, Alachua, FL). The frequency of the tone bursts was manually varied or computer controlled to measure the BF and the minimum threshold (MT) of a given neuron. To measure a frequency-response curve and BF, the amplitude of the tone burst was set at 10 dB above the MT of the neuron. The frequency of the tone burst was then randomly varied in 0.5- or 1.0-kHz steps over 10 or 20 kHz. This frequency scan consisted of 21 250-ms-long time blocks, one of which had no sound stimulus. It was repeated 30 times. The frequency scan was delivered by the stimulus control and recording software (Brainware version 8.0). The amplitudes of the tone bursts delivered from the leaf tweeter were calibrated with a microphone (Brüel and Kjær Instruments, Naerum, Denmark) and were expressed in decibels in sound pressure level (dB SPL) referred to 20 μPa (root mean square).

Pseudoconditioning.

The nonassociative auditory learning and nonspecific augmentation are frequency independent. That is, they are equally induced at different frequencies by pseudoconditioning (Bakin et al. 1992; Ji and Suga 2008, 2009). In the big brown bat, pseudoconditioning with the CS at 5.0 kHz lower than the BF of a given neuron elicits CS-dependent BF shifts in addition to the nonspecific augmentation (Ji and Suga 2008, 2009). Therefore, in our current studies, drug effects on the cortical plasticity were studied at a CS frequency 5.0 kHz lower than the BF of a given auditory neuron. We kept all the parameters characterizing the CS for pseudoconditioning identical to those used to elicit auditory fear conditioning (Ji and Suga 2008, 2009). Namely, the CS was a train of tone bursts that were 50 dB SPL, 10 ms long, and 33/s over 1.0 s to mimic the biosonar pulses emitted by a bat at the middle of the approach phase of echolocation, and this CS was delivered every 30 s over 30 min (Gao and Suga 1998, 2000; Ji and Suga 2003; Ji et al. 2001). Pseudoconditioning was delivered to an animal only once a day with at least a 3-day resting period. We did not notice any cumulative or habituation-like effects on the nonspecific augmentation and BF shifts. We had previously shown (Ji and Suga 2008, 2009) that the nonspecific augmentation evoked by pseudoconditioning fully recovers within 4 h.

To explore the neural circuit for the nonspecific augmentation elicited by pseudoconditioning, it is more appropriate to randomize the timing of the US, and not the CS. Accordingly, the unpaired US was abbreviated as the USu. The USu was a 50-ms, 0.1- to 0.4-mA monophasic electric pulse. The USu was randomly delivered to the bat in a time interval between 5 and 25 s after the CS, avoiding the period between 5.0 s before and 5.0 s after the CS. The mean interval of the CS and USu delivery was 14.8 s. There were 60 CS and 60 USu in total per training session. The CS was barely aversive to an animal and did not elicit nonspecific augmentation, whereas the USu was aversive to the animal and elicited it (Ji and Suga 2009).

Drug applications.

The auditory cortex is 4.52 mm² in size. The approximate center of AI was dorsoventrally crossed by a 30-kHz iso-BF line (Dear et al. 1993; Shen et al. 1997). The approximate midpoint of the iso-BF line was located first by recording auditory responses at five to six cortical loci. A hole ∼1.0 mm in diameter was then made there for homolateral drug application to AI. The drug applications were made with or without pseudoconditioning with a 1.0-μl Hamilton syringe. The drug application was made 3–5 min before pseudoconditioning. The dura mater was not removed but was punctured by the penetrations of the recording electrode so that the drug applied diffused into AI through the puncture holes. Pseudoconditioning and/or a drug (i.e., the test session) was delivered to an animal once a day with at least a 3-day resting period.

The drugs applied to the surface of AI were the following five neuromodulators and their ligands: 1) 20 mM ACh, 40 mM atropine (muscarinic ACh receptor antagonist), and 10 μM mecamylamine (nicotinic AChR antagonist) (Ji and Suga 2008; Ji et al. 2001); 2) 10 mM DA and 50 mM haloperid (DAR antagonist) (Atzori et al. 2005; Rosenkranz and Grace 1999); 3) 10 mM NE, 3.0 mM phentolamine (α-NER antagonist), and 10 mM propranolol (β-NAR antagonist) (Ferry and McGaugh 2000; Manunta and Edeline 1999); 4) 20 mM 5-HT and 10 mM ritanserin (5-HT antagonist) (Ji and Suga 2007; Stark and Scheich 1997); 5) 0.3 mM HA, 1.0 and 5.0 mM α-methyl-HA (HA3R agonist), or 25 μM thioperamide (HA3R antagonist) (Blandina et al. 1996; Brown and Haas 1999; Di Carlo et al. 2000; Kohler et al. 2011; Stark et al. 2001; Vanni-Mercier et al. 2003). The doses of a neuromodulator and its ligands were based on the articles cited above. All the drugs were purchased from Sigma Chemical (St. Louis, MO). The drugs were dissolved in a 0.9% saline solution. The volume of each drug applied was 0.2 μl. Ritanserin was sonicated for 2 min before use (Ji and Suga 2007).

Data acquisition.

Brainware data acquisition software (Tucker-Davis Technologies) was used for sorting and selecting the action potentials of a single cortical or collicular neuron. At the beginning of the data acquisition, the waveform of an action potential (i.e., template) was stored and then compared with other action potentials obtained during data acquisition. An array of peristimulus time (PST) histograms displaying the responses to the frequency scan repeated 30 times was acquired before and after the onset of pseudoconditioning, as long as action potentials visually matched the template. Frequency-response curves were obtained every 15 min, beginning ∼40 min before a drug or saline application to AI, with or without the pseudoconditioning. Two “control” data were obtained to calculate a percent change in the responses to tone bursts evoked by the drug or saline, with or without the pseudoconditioning. When the pseudoconditioning was delivered to the animal, the artifact of the electric leg stimulus interrupted the recording of action potentials. Therefore, the data acquisition was interrupted during the 30-min-long pseudoconditioning. It was restarted at the end of pseudoconditioning and repeated every 15 min for up to 270 min. The percent change in response was also calculated for a physiological saline solution applied to AI with or without pseudoconditioning. The percent changes in this sham experiment were 0.0 ± 1.4% (n = 10) for the saline solution without pseudoconditioning and 56.7 ± 3.4% (n = 23) for the saline solution with pseudoconditioning and were used as the “baseline” data to compare with the percent changes evoked by the drugs with or without pseudoconditioning.

The BF shift elicited by the CS was detectable at the end of pseudoconditioning, i.e., 35 min after the onset (1st sample), but was not detectable 50 min after the onset (2nd sample). On the other hand, the nonspecific augmentation was largest 65–95 min after the onset of pseudoconditioning (3rd to 5th samples) and then gradually reduced and disappeared ∼240 min after the onset (Ji and Suga 2008). The data for the CS-dependent BF shift in Table 1 and US-dependent nonspecific augmentation shown in Fig. 1 were obtained 35 min after and within 95 min after a drug application with or without pseudoconditioning, respectively. The effect of the drug with or without pseudoconditioning was represented by the largest change in the response to the tone burst at the BF of a given neuron evoked by it.

Table 1.

Numbers of AI neurons showing BF and response changes caused by histaminergic ligands applied to AI with or without pseudoconditioning

				HA			α-Methyl-HA						Thioperamide
	Saline			0.3 mM			1.0 mM			5.0 mM			25 μM
	N	+	−	N	+	−	N	+	−	N	+	−	N	+	−
Without P-cond
No BF shift	10	0	0	0	0	11	8	0	18	0	0	20	0	3	0
BF downward shift	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
BF upward shift	0	0	0	0	0	3	0	0	5	0	0	5	0	18	0
With P-cond
No BF shift	0	0	0	0	6	0	0	0	0	0	0	0	0	0	0
CS-dependent BF shift	0	23	0	0	0	0	10	13	0	0	7	21	0	17	0
BF shift away from CS	0	0	0	0	8	0	0	0	0	0	0	0	0	0	0

Values are numbers of cortical auditory neurons showing a best-frequency (BF) shift and/or a response change elicited by histaminergic ligands applied to the primary auditory cortex (AI) without or with pseudoconditioning (P-cond). CS, conditioned stimulus delivered with the unpaired unconditioned stimulus for P-cond; N, +, and −, no change, augmentation, and suppression in the responses to tone bursts caused by a drug application without or with P-cond, respectively; HA, histamine; α-methyl-HA, an HA type 3 receptor agonist; thioperamide, a HA3R antagonist.

Fig. 1.

Percent changes in the responses to tone bursts evoked by drugs applied to the primary auditory cortex (AI) without (open bars) or with (shaded bars) pseudoconditioning (P-cond). Nicotinic (n) and muscarinic (m) acetylcholine receptor (AChR) antagonists: mecamylamine and atropine, respectively; α- and β-norepinephrine receptor (NER) antagonists: phentolamine and propranolol, respectively; dopamine receptor (DAR) antagonist: haloperidol; serotonin receptor (5-HTR) antagonist: ritanserin; histamine type 3 receptor (HA3R) agonist: α-methyl-HA; and HA3R antagonist: thioperamide. *P < 0.05; **P < 0.01 compared with saline without pseudoconditioning. +P < 0.05; ++P < 0.01 compared with saline with pseudoconditioning. Values are means ± SE. NA (noradrenaline) is the same as NE.

Atropine applied to AI reduces cortical (AI) auditory responses by ∼10% over 15–30 min after its application, and then its effect disappears in ∼45 (Ji et al. 2001) or ∼90 min after its application (Ji and Suga 2008). A 30-min-long conditioning session elicits a cortical BF shift that gradually increases to a plateau in ∼75 min after the onset of fear conditioning. Such a BF shift is completely abolished by atropine applied to AI 5 min before fear conditioning (Ji et al. 2001). The time course of the development of the nonspecific augmentation elicited by the 30-min-long pseudoconditioning is very similar to that of the BF shift elicited by the 30-min-long conditioning (Ji and Suga 2008). Therefore, in our current experiment, we applied a drug to AI 3–5 min before the pseudoconditioning and examined its effect on the nonspecific augmentation and BF shift elicited by it after the pseudoconditioning.

Off-line data processing.

To obtain the frequency-response curve of a neuron with the frequency scan, the magnitude of the response of the neuron to a tone burst was expressed by the average number of spikes in 30 identical stimuli as a function of the frequency of the tone burst. The BF of the neuron was defined as the frequency at which the frequency-response curve peaked. Because an identical frequency scan was delivered 30 times, there were 30 samples of BFs that could be used to compute the mean ± SE of the BF values and to perform statistical analysis. A P value <0.05 was used to determine whether there were significant differences in the response magnitude between a BF and the adjacent frequencies (ANOVA) or between the BF values obtained before and after pseudoconditioning and/or a drug application (2-tailed unpaired t-test). A percent change in the response at the BF evoked by a drug with or without pseudoconditioning was calculated by referring to the responses in the control conditioning, i.e., the response at the BF obtained before the drug with or without pseudoconditioning. The significance of the percent change was based on the comparison with the baseline data, i.e., the percent change in the response at the BF evoked by a saline application to AI with or without pseudoconditioning.

RESULTS

Neuromodulators potentially block the nonspecific augmentation elicited by pseudoconditioning.

As reported by Ji and Suga (2008), pseudoconditioning elicited two types of changes in the response properties of cortical auditory neurons of the big brown bat: prominent nonspecific augmentation and a small CS-dependent BF shift. These changes were evoked at the same time, but the BF shift disappeared faster than the nonspecific augmentation. The physiological saline solution applied to AI had no effect on the cortical responses to tone bursts. That is, the responses (number of impulses per tone burst) at the BF measured before and after the saline application differed by 0.0 ± 1.4% (n = 10). When its application was accompanied by pseudoconditioning, the cortical response at the BF increased by 56.7 ± 3.4% (n = 23) compared with the control data, i.e., the response obtained before the drug application. That is, nonspecific augmentation was evoked by pseudoconditioning. This cortical nonspecific augmentation gradually developed to a plateau over 65 min after the onset of pseudoconditioning. It then stayed at the plateau for ∼30 min, slowly decreased, and disappeared 240 ± 30 min after the onset of pseudoconditioning. On the other hand, the CS-dependent BF shift was ∼1.0 kHz toward the frequency of the CS and was briefly observed ∼35 min after the onset of pseudoconditioning.

To find modulators that could affect the nonspecific augmentation elicited by pseudoconditioning, we screened cholinergic, noradrenergic, dopaminergic, serotonergic, and histaminergic systems. ACh, NE, and DA applied to AI without pseudoconditioning increased the responses to tone bursts by 60.0 ± 8.1% (n = 20), 66.2 ± 12.1% (n = 20), and 56.9 ± 5.4% (n = 14) compared with the control data, respectively (Fig. 1, open bars). The antagonists of muscarinic (m) and nicotinic (n) AChRs, α- and βNERs, and DAR were applied to AI and reduced the auditory responses by 10.5–35.4% (Fig. 1, open bars). However, their antagonists applied to AI with pseudoconditioning did not reduce the nonspecific augmentation elicited by pseudoconditioning. Namely, the nonspecific augmentation was 50.4 ± 5.2% (n =10) by a mAChR antagonist, 48.6 ± 2.1% (n = 12) by a nAChR antagonist, 46.9 ± 14.2% (n = 5) by a αNER antagonist, 60.1 ± 10.2% (n = 6) by a βNER antagonist, and 60.4 ± 12.4% (n = 5) by a DAR antagonist (Fig. 1, shaded bars). These changes were insignificant compared with the baseline data, i.e., the augmentation (56.7 ± 3.4%, n = 23) evoked by saline with pseudoconditioning (P > 0.05). There was no sign that the cholinergic, noradrenergic, and dopaminergic systems play an important role in eliciting the nonspecific augmentation.

5-HT and its antagonist (ritanserin) reduced the cortical responses to tone bursts by 58.8 ± 2.4% (n = 20) and 20.8 ± 14.3% (n = 16), respectively. However, 5-HT and its antagonist applied to AI with pseudoconditioning evoked nonspecific augmentation of 45.4 ± 9.2% (n = 6) and 50.4 ± 10.2% (n = 6), respectively (Fig. 1). The augmentation did not statistically differ from the baseline data, i.e., 56.7 ± 3.4% (P > 0.05).

For the histaminergic system, either HA, an HA3R agonist, or an HA3R antagonist was applied to AI without or with pseudoconditioning. A total of 173 cortical neurons were studied (Fig. 1, Table 1). Their BF values ranged from 21.0 to 60.0 kHz (38.0 ± 2.5 kHz). The HA3R agonist reduced the responses to tone bursts and nonspecific augmentation elicited by pseudoconditioning, whereas the HA3R antagonist augmented them as described below.

Effects of HA on the cortical nonspecific augmentation and CS-dependent BF shift elicited by pseudoconditioning.

HA (0.3-mM) applied to AI without pseudoconditioning reduced the responses to tone bursts of all 14 cortical neurons studied. The average amount of reduction at the BFs was 42.4 ± 2.6% (n = 14) compared with control data (Fig. 1, open bars). The responses reverted to the control ∼90 min after the application. Eleven of the 14 neurons did not show a BF shift, but the remaining 3 showed an “upward BF shift,” i.e., shift away from the CS frequency. It was opposite to the CS-dependent BF shift in direction (Table 1). Figure 2A shows the PST histograms of the responses to 22.0-kHz tone bursts (a) and the frequency-response curves (b) of a cortical neuron. The responses of this neuron to tone bursts were reduced at all frequencies of the tone bursts delivered within 15 min after a HA application to AI (Fig. 2, Aa2 and Ab2). The responses gradually reverted to the control condition over 90 min after the application (Fig. 2, Aa3, Aa4, Bb3, and Bb4). The reduction of the response at its BF was up to 46% compared with the control shown in Fig. 2, Aa1 and Ab1. The BF of the neuron at 22.0 kHz did not change.

Fig. 2.

Changes in the auditory responses displayed by peristimulus time (PST) histograms and frequency-response curves of 2 cortical neurons evoked by 0.3 mM HA applied to AI without (A) or with (B) pseudoconditioning. The histograms (a) and curves (b) display the responses of the neurons to tone bursts delivered 30 times. The tone bursts were 10 dB above the minimum thresholds of the given neurons. BFc and BFs indicate the BF values of the neurons in control and shifted conditions, respectively. CS and open arrow in Bb indicate the conditioned stimulus (CS) frequency delivered with the unpaired unconditioned stimulus (US) for pseudoconditioning. A1–A4, control and 15, 35, and 55 min after the HA application without pseudoconditioning. B1–B4, control and 35, 60, and 180 min after the HA application with pseudoconditioning. Note the overall reduction of the response in A and augmentation in B.

When HA was applied to AI before pseudoconditioning, all 15 neurons studied showed similar nonspecific augmentation to that elicited by saline with pseudoconditioning [Fig. 1, shaded bars; 53.8 ± 6.8% (n = 15) vs. 56.7 ± 3.4% (n = 23), P > 0.05]. HA did not block the nonspecific augmentation elicited by pseudoconditioning, although HA alone reduced the auditory response by 42.4 ± 2.6%. HA with pseudoconditioning evoked a BF shift away from the CS in 8 of the 15 neurons. These BF shifts were different from the CS-dependent BF shift in direction (Table 1). In Fig. 2B, the effects of HA on the BF shift and nonspecific augmentation were demonstrated with the PST histograms of the responses (a) and frequency-response curves (b) of a cortical neuron tuned to 25.0 kHz. When HA was applied to AI, the auditory response of this neuron was initially reduced by 42.2% at 25.0 kHz but increased by 400% at 28.0 kHz (Fig. 2, Ba2 and Bb2). Accordingly, the BF of the neuron shifted from 25.0 to 28.0 kHz, i.e., away from the 20.0-kHz CS. The nonspecific augmentation developed between 18.0 and 32.0 kHz (Fig. 2, Ba3 and Bb3). The maximum increase in the response was 62.5% at 25 kHz. The neuron's BF reverted to 25.0 kHz ∼65 min after the onset of pseudoconditioning (Fig. 2, Ba4 and Bb4). The nonspecific augmentation disappeared in 180 min. The reduction of the response at 35 min after the drug application was due to the drug's direct effect.

Effects of a HA3R agonist on the cortical nonspecific augmentation and CS-dependent BF shift elicited by pseudoconditioning.

Without pseudoconditioning, the HA3R agonist (1.0 mM α-methyl-HA) applied to AI did not affect the responses to tone bursts in 8 of the 31 neurons studied but reduced the auditory responses in the remaining 23 neurons by 34.0 ± 6.4% at BF compared with the control data, i.e., those obtained before the drug application (Fig. 1, open bar). The responses fully recovered ∼120 min after the application. Of these 23 neurons, 5 showed an upward BF shift that was different from the CS-dependent BF shift. The remaining 18 neurons did not show any BF shift (Table 1).

When 1.0 mM α-methyl-HA was applied to AI with pseudoconditioning, the responses to tone bursts of the 23 neurons studied showed nonspecific augmentation of 23.3 ± 2.3% compared with those obtained before the drug application. This was significantly smaller than that elicited by saline with pseudoconditioning (56.7 ± 3.4%, n = 23). Of the 23 neurons, 13 showed nonspecific augmentation of 41.0 ± 5.3%, whereas the remaining 10 did not (0.0 ± 2.4%). All 23 neurons showed the CS-dependent BF shift (Table 1). Therefore, the dose of α-methyl-HA was increased to 5.0 mM.

When 5.0 mM α-methyl-HA was applied to AI without pseudoconditioning, the responses to tone bursts of all 25 neurons studied reduced. The mean reduction was 46.0 ± 5.4% (n = 25) compared with those obtained before the drug application. Five of the 25 neurons showed a BF shift that was opposite in direction to the CS-dependent BF shift. The remaining 20 neurons did not show any BF shift (Table 1). Figure 3A shows the PST histograms of the responses at 47.0 kHz (a) and the frequency-response curves (open circles in b) of a cortical neuron tuned to 47.0 kHz. The auditory responses of the neuron were first reduced by 5.0 mM α-methyl-HA applied to AI (Fig. 3, Aa2 and filled cycles in Ab) and then recovered ∼120 min after its application (Fig. 3, Aa4 and dashed line in Ab). The reduction at the BF was 58% ∼35 min after the application. The BF did not change (arrow in Fig. 3Ab).

Fig. 3.

Changes in the auditory responses displayed by the PST histograms (a) and frequency-response curves (b) of 3 cortical neurons evoked by an HA3R agonist applied to AI without (A) or with (B and C) pseudoconditioning. Note the augmentation of the responses in B but not in C. C also shows a CS-dependent BF shift. Same conventions as Fig. 2.

When 5.0 mM α-methyl-HA was applied to AI with pseudoconditioning, 21 of the 28 neurons studied reduced their responses to tone bursts by 42.6 ± 3.5% (Fig. 3C) and 7 neurons increased their responses by 37.5 ± 6.7% compared with the control data (Fig. 3B). The average reduction of the 28 neurons was 21.0 ± 4.0% (Fig. 1). All of the 28 neurons showed a CS-dependent BF shift for pseudoconditioning with 5.0 mM α-methyl-HA. Figure 3B shows the responses to tone bursts of a neuron tuned to 55.0 kHz. When 5.0 mM α-methyl-HA was applied to AI during pseudoconditioning, its responses to tone bursts were augmented between 51.0 and 62.0 kHz (filled circles in Fig. 3Bb). The maximum increase in response was 34.7% at 55.0 kHz and 108% at 54.0 kHz. The CS-dependent BF shift was 0.5 kHz. The nonspecific augmentation disappeared ∼210 min after the onset of pseudoconditioning (Fig. 3, Ba4 and dashed line in Bb). In Fig. 3C, a neuron tuned to 38.0 kHz (BFc in Fig. 3Ca1 and open circles in Cb) reduced its response at 38.0 kHz by 80% but temporarily increased it at 37.0 kHz by 18% (Fig. 3, Ca2 and filled circles in Cb) 35 min after a 5.0 mM α-methyl-HA application to AI with pseudoconditioning. Its BF shifted from 38.0 to 37.0 kHz, toward the 33.0-kHz CS. The maximal reduction at BFc was 66.7%. All the changes in the responses and frequency-tuning curve reverted to the control ∼270 min after the onset of pseudoconditioning (Fig. 3, Ca4 and dashed curve in Cb). Thus 5.0 mM α-methyl-HA blocked the development of nonspecific augmentation, whereas the CS-dependent BF shift was not affected.

Effects of a HA3R antagonist on the cortical nonspecific augmentation and CS-dependent BF shift elicited by pseudoconditioning.

When the HA3R antagonist, 25 μM thioperamide, was applied to AI without pseudoconditioning, responses to tone bursts increased in all 27 neurons studied. The amount of the increase at BF was 174.5 ± 11.6% compared with the control data (n = 21, Fig. 1, open bar). The responses reverted to the control ∼120 min after its application. Of the 21 neurons, 18 showed an upward BF shift and the remaining 3 did not show any BF shift (Table 1). Figure 4A shows the PST histograms of the responses to 30.0- and 31.0-kHz tone bursts (a) and the frequency-response curves (b) of a cortical neuron tuned to 30.0 kHz. The responses to the tone bursts at all frequencies as well as the spontaneous discharges increased after the HA3R antagonist application (Fig. 4, Aa2 and Ab2). The amount of the increase in the response was 72% at 30.0 kHz and 264% at 31.0 kHz (Fig. 4Aa2). As a result, the BF of the neuron became 1.0 kHz higher (arrow in Fig. 4Ab2). The BF returned to 30.0 kHz in 65 min. The response increase evoked by the HA3R antagonist disappeared ∼120 min after the application (Fig. 4, Aa4 and Ab4).

Fig. 4.

Changes in the auditory responses displayed by PST histograms (a) and frequency-response curves (b) of 2 cortical neurons evoked by an HA3R antagonist applied to AI without (A) or with (B) pseudoconditioning. Note the overall increase in the responses and an upward BF shift in A and a CS-dependent BF shift in B. Same conventions as Fig. 2.

When the HA3R antagonist was applied to AI before pseudoconditioning, the cortical nonspecific augmentation was much larger than that evoked by saline with pseudoconditioning [Fig. 1, shaded bars; 186.0 ± 10.3% (n = 17) vs. 56.7 ± 3.4% (n = 23); P < 0.01]. However, it was only 11.5% larger than 174.5% evoked by the drug alone. The difference was statistically insignificant. The CS-dependent BF shifts were elicited by pseudoconditioning with the drug in all 17 neurons studied. In Fig. 4B, a cortical neuron was tuned to 59.0 kHz (b1). The HA3R antagonist application to AI with pseudoconditioning increased its responses to tone bursts between 48.0 and 66.0 kHz. The increase in response was largest at 57.0–59.0 kHz (Fig. 4, Ba3 and Bb3). The BF of the neuron shifted from 59.0 to 57.0 kHz, i.e., toward the 54.0-kHz CSu (arrow in Fig. 4Bb2). This CS-dependent BF shift reverted to 59.0 kHz in 60 min, whereas the nonspecific augmentation reverted to the control in 210 min.

The collicular nonspecific augmentation and CS-dependent BF shift affected by HA applied to AI.

Pseudoconditioning evoked similar plastic changes in the central nucleus of the inferior colliculus (ICc) as observed in AI. Inactivation of AI by muscimol abolished the collicular CS-dependent BF shift elicited by pseudoconditioning without affecting the collicular nonspecific augmentation (Ji and Suga 2009). We studied the effect of HA (0.3 mM) applied to AI on the collicular nonspecific augmentation with or without pseudoconditioning. HA applied to AI reduced the responses to tone bursts of all 11 collicular neurons studied by 40.9 ± 6.2% compared with those obtained before the drug application. When HA was applied to AI with pseudoconditioning, all the collicular neurons showed augmentation. The amount of the nonspecific augmentation was 50.3 ± 11.8% (n = 11), which was not different from the augmentation (45.8 ± 6.8%, n = 34) evoked by pseudoconditioning without the HA application (Ji and Suga 2009) (P > 0.05). The collicular BF shift was not evoked under the cortical HA application. In Fig. 5, a collicular neuron was tuned to 22.0 kHz. When HA (0.3 mM) was applied to AI, its response to tone bursts was reduced to 56% at its BF at 5 min after the application (Fig. 5, A2 and B2). The responses recovered about 40 min later (Fig. 5, A3 and B3). After recovery, HA was applied to AI with pseudoconditioning. Nonspecific augmentation appeared between 18.0 and 25.0 kHz (Fig. 5, A4 and B4) and then gradually disappeared in 210 min (Fig. 5, A6 and B6). No BF shift was evoked.

Fig. 5.

Changes in the auditory responses displayed by PST histograms (A) and frequency-response curves (B) of a neuron in the central nucleus of the inferior colliculus (ICc) evoked by 0.3 mM HA applied to AI without pseudoconditioning and then with pseudoconditioning. Note a decrease in the response by HA and an increase in the responses by HA with pseudoconditioning.

DISCUSSION

Upward BF shifts evoked by HA and its ligands.

As shown in Table 1, HA and an HA3R agonist both evoked upward BF shifts and reduced responses to tone bursts of AI neurons. On the other hand, an HA3R antagonist evoked upward BF shifts, but increased responses to tone bursts. HA plus pseudoconditioning evoked nonspecific augmentation and upward BF shifts. These upward BF shifts were unexpected, since the HA3R agonist plus pseudoconditioning did not evoke any upward BF shift, but the CS-dependent BF shifts of all 51 neurons studied. The HA3R antagonist plus pseudoconditioning also evoked the CS-dependent BF shift and nonspecific augmentation of all 17 neurons studied. These intriguing phenomena remain to be further studied.

Proposed role of the histaminergic system in modulating cortical nonspecific plasticity.

AChR antagonists block BF shifts but do not affect nonspecific augmentation (Ji and Suga 2008, 2009). An ACh or DA application to AI with pseudoconditioning may evoke the large auditory responses due to both ACh or DA and pseudoconditioning. However, such large auditory responses do not necessarily mean that the nonspecific augmentation is facilitated by ACh or DA, because AChR or DAR antagonists applied to AI reduced the auditory responses but not the nonspecific augmentation (Fig. 1). Therefore, both ACh and DA may not be involved in evoking the nonspecific augmentation, although they both play an important role in evoking tone-specific plasticity, BF shifts, in the auditory system (Bao et al. 2001; see Suga and Ma 2003 and Weinberger 1998 for reviews). NE and a NER antagonist showed effects on the auditory response and nonspecific augmentation similar to those of ACh, DA, and their antagonists and may also not be involved in evoking the nonspecific augmentation (Fig. 1).

One might interpret that 5-HT facilitated the nonspecific augmentation by 104% (the difference between the 59% decrease in the response without pseudoconditioning and the 45% increase in the response with pseudoconditioning). One should then expect that a 5-HTR antagonist would abolish this facilitation. On the contrary, the 5-HTR antagonist with pseudoconditioning evoked a 50% increase in response (Fig. 1). Therefore, one may conclude that 5-HT did not play a role in evoking the nonspecific augmentation. 5-HT modulates the cortical tone-specific plasticity (Ji and Suga 2007).

The response change evoked by the HA3R (5 mM) agonist was −46% without pseudoconditioning and −21% with it. Since the response change evoked by pseudoconditioning alone was 57%, the 78% decrease (the difference between 57% and −21%) in the augmentation by pseudoconditioning may mostly be due to the direct effect of the drug.

HA without pseudoconditioning reduced the auditory response by 42%, but HA with pseudoconditioning increased the response by 54%. Therefore, one may interpret that HA evoked a 96% increase in response. One may then expect that a HA3R antagonist would abolish this increase. On the other hand, the HA3R antagonist applied with or without pseudoconditioning greatly increased the auditory response. The conceivable function of HA3R is discussed below.

The antagonists of AChR, NER, DAR, and 5-HTR all reduced the auditory response, but none of them reduced the nonspecific augmentation. Therefore, the action of HA3R was different from the other four receptors. It appeared to be specifically involved in nonspecific augmentation.

Histaminergic neurons originating from the hypothalamic tuberomammillary nuclei widely distribute in the brain for modulation of awake/sleep and defensive behavior (see Brown et al. 2001 and Schwartz et al. 1991 for reviews). The inhibition of histaminergic neurotransmission decreases cortical activation and arousal, and has been hypothesized to suppress fear and anxiety (Dere et al. 2010). The hypothalamus projects to AI through the ascending reticular activating system (Rouller et al. 1989a, 1989b). The ascending reticular activating system and brain aversion system both are activated in stressful situations, such as electric foot shock (Schwartz et al. 1991).

Interpretation of the effects of HA3R agonist and antagonist on cortical nonspecific plasticity.

HA neurons are linked to maintaining wakefulness (Passani et al. 2007; Siegel 2002). HA excites the cortex mainly due to the activation of HA1R and HA2R (Haas and Panula 2003). HA3R is located on the histaminergic and other neuronal soma, dendrites, and axonal varicosities for regulating signal transduction pathways (Haas and Panula 2003; Lin et al. 2011; Passani and Blandina 2011; Passani et al. 2000, 2004, 2007). HA3R acts as an autoreceptor and also as a heteroreceptor, providing feedback inhibition of HA synthesis and inhibiting the release of ACh, DA, GABA, glutamate, NE, and 5-HT (Haas et al. 2008). HA3R modulates a variety of neurotransmitter systems (Brown et al. 2001; Cenni et al. 2004; Passani et al. 2000).

Our data show that the HA3R agonist strongly suppresses the responses to tone bursts and the nonspecific augmentation elicited by pseudoconditioning. This suppression may be due to the reduction of HA synthesis and release caused by the activation of HA3 autoreceptors. On the other hand, the HA3R antagonist evokes the nonspecific augmentation of auditory responses. It may block the action of the HA3 autoreceptor and enhance HA release, leading to the activation of postsynaptic HA1R and HA2R which cause the augmentation of the neural activity for alertness (Haas et al. 2008).

Cortical vs. collicular nonspecific augmentation elicited by pseudoconditioning.

The hypothalamus, amygdala, dorsal periaqueductal gray, and IC comprise the brain aversion system (Brandão et al. 2003). The IC has reciprocal connections with the hypothalamus associated with defensive behavior, such as startle responses, freezing, and vocalization (see Brandão 1993 for review). The link between the IC and fear or anxiety has been demonstrated by behavioral, electrophysiological, and immunohistochemical studies (LeDoux 2003). Exposing animals to an aversive environmental stimulation causes histochemical changes in the IC (Silveira et al. 1993, 1995, 2001; see Brandão et al. 2003 for review). In our current study, the cortical application of HA suppresses the collicular auditory responses and the CS-dependent collicular BF shift through corticofugal feedback. However, it does not suppress the collicular nonspecific augmentation elicited by pseudoconditioning. Unlike the BF shift, the collicular nonspecific augmentation appears not to be mediated by the corticofugal projection (Ji and Suga, 2009).

Differential gating by the neuromodulators for tone-specific and nonspecific plasticity.

Electric stimulation of AI generally excites the neurons in the ventral medial geniculate body (MGBv) but inhibits the neurons in the medial MGB (MGBm) (Yu et al. 2004; Tang et al. 2012). ACh depolarizes MGBv neurons but hyperpolarizes MGBm neurons (Mooney et al. 2004). ACh plays a major role in the development of the tone-specific plasticity elicited by auditory fear conditioning (Suga and Ma 2003, 2005; Weinberger 1998) but not in the development of the nonspecific plasticity elicited by pseudoconditioning, because the muscarinic AChR antagonist applied to AI blocks the CS-dependent BF shifts but not the nonspecific augmentation (Ji and Suga 2008, 2009). HA plays a role in evoking nonspecific plasticity but not in evoking tone-specific plasticity (current study). These findings favor the presence of a differential gating mechanism by which the nonspecific plasticity (nonspecific augmentation) is suppressed or not enhanced when the tone-specific plasticity (large, long-term BF shift) is elicited, whereas the tone-specific plasticity is suppressed or not enhanced when the nonspecific augmentation is elicited (Suga 2008).

HA release in the cortex may depend on the level of stress and arousal evoked by the randomized electric leg stimulation. We hypothesize that activation of postsynaptic HA3R induces the release of GABA, which inhibits the depolarization-induced ACh release. Similarly, the HA3R agonist may act to reduce ACh release, leading to specific object-recognition impairment (Blandina and Passani 2006; Blandina et al. 2004). Thus the histaminergic system may suppress ACh release during pseudoconditioning, and HA release may be regulated by the mAChR (Gulat-Marnay et al. 1989; Kohler et al. 2011). Therefore, the cholinergic and histaminergic systems may differentially gate specific and nonspecific plasticity.

GRANTS

This work was supported by National Institute on Deafness and Other Communication Disorders Grant DC-000175.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

W.J. and N.S. conception and design of research; W.J. performed experiments; W.J. analyzed data; W.J. and N.S. interpreted results of experiments; W.J. prepared figures; W.J. drafted manuscript; W.J. and N.S. edited and revised manuscript; W.J. and N.S. approved final version of manuscript.