Journal of Neurophysiology

Pattern-Specific Synaptic Mechanisms in a Multifunctional Network. II. Intrinsic Modulation by Metabotropic Glutamate Receptors

Steven P. Lieske, Jan-Marino Ramirez


The in vitro respiratory network contained in the transverse brain stem slice of mice simultaneously generates fast (∼15 min-1) and slow (∼0.5 min-1) rhythmic activities corresponding to fictive eupnea (“normal” breathing) and fictive sighs. We show that these two activity patterns are differentially controlled through the modulatory actions of metabotropic glutamate receptors (mGluRs). Sighs were selectively inhibited by agonists of the group III mGluRs according to a pharmacological profile most consistent with activation of mGluR8. Sighs were also blocked by the supposedly inactive L-isomer of the widely used N-methyl-d-aspartate (NMDA) receptor antagonist 2-amino-5-phosphonopentanoic acid (L-AP5, 5 μM), an effect that was abolished in the presence of group III mGluR antagonists. Excitatory postsynaptic potentials (EPSPs) were recorded in pre-Bötzinger Complex neurons after stimulation of the contralateral ventral respiratory group (VRG); evoked EPSP amplitude was variably reduced after bath application of the group III agonist l-serine-O-phospate (L-SOP), with an average reduction of 15%. Therefore although group III mGluRs do play a role in regulating synapse strength, this seems to be only a minor factor in the regulation of synapses made by midline-crossing axons. Intrinsic modulation of the respiratory central pattern generator by mGluRs appears to be an essential component of the multifunctionality that characterizes this network.


A wide variety of behaviors are generated by neuronal networks known as central pattern generators (CPGs), defined by their production of functionally relevant rhythmic activity in the absence of rhythmically varying drive or sensory input. Tonic excitatory or neuromodulatory input is often required to switch the network on (Harris-Warrick and Marder 1991; Marder and Bucher 2001), and in several invertebrate systems, the application of neuromodulators has also been shown to switch the network between multiple functional states (Getting 1989; Harris-Warrick and Marder 1991; Marder and Bucher 2001). These networks are thus capable of producing qualitatively different motor patterns, depending on the neuromodulatory context.

Using the transverse medullary slice preparation of mice, we have shown that a similar reconfiguration plays an important role in the genesis of the respiratory rhythm in mammals (Lieske and Ramirez 2006; Lieske et al. 2000b). In this highly reduced in vitro preparation, three distinct fictive breathing patterns originate from a single network, corresponding to eupnea (“normal” breathing), gasps (in hypoxia), and sighs (“augmented breaths”). Eupnea and sighs are produced simultaneously in control conditions in vitro; fictive eupneic breaths occur at intervals of a few seconds (mean frequency = 0.25 Hz or 15 min-1), whereas sighs interrupt the eupneic rhythm at intervals of one to several minutes (Lieske et al. 2000b).

Although there do not appear to be specialized neurons for eupnea and for sighs, we show in the companion article that distinct glutamatergic synapses may be involved (Lieske and Ramirez 2006). In the present study, we extend this result by examining the effects of metabotropic glutamate receptors (mGluRs) in controlling these two patterns. Eight mGluRs have been identified and placed into three groups on the basis of sequence homology, pharmacological similarity, and associated second-messenger system. The group I receptors (mGluR1 and 5) act via Gq/11 to activate the phospholipase C pathway, whereas group II (mGluR2 and 3) and group III (mGluR4, -6, -7, and -8) receptors act via Gi/o to inhibit adenylate cyclase (Conn and Pin 1997; Nakanishi 1994). All three groups have been implicated in the modulation of synaptic transmission (Cartmell and Schoepp 2000; Schoepp 2001).

Screening both agonists and antagonists of all three groups, we found that activation of the group III receptor mGluR8 abolished sighs for periods of up to several hours. There was also a modest reduction of excitatory postsynaptic potential (EPSP) amplitude, suggesting that activation of this receptor inhibits synaptic glutamate release, as reported for group III receptors in other brain areas (Cartmell and Schoepp 2000; Dube and Marshall 1997; Stefani et al. 1998; Takahashi et al. 1996; Thomas et al. 2001; Wittmann et al. 2001). These results are consistent with those of the companion manuscript, suggesting that—like the involvement of P/Q-type calcium channels—modulation by mGluR8 is very specific to synapses that are essential for sighs but is important at only a minority of the glutamatergic synapses coupling inspiratory neurons on opposite sides of the brain stem.

These results also provide an intriguing parallel to a type of neuromodulation observed in invertebrate CPGs: release of neuropeptides from synaptic terminals that, via colocalized fast neurotransmitters, comprise intrinsic elements of the network itself (Blitz et al. 1999; Wood et al. 2000). It is thought that releasing neuromodulator requires relatively sustained presynaptic activity rather than a single action potential (Nusbaum et al. 2001; Whim and Lloyd 1989). Because metabotropic glutamate receptors in general, and mGluR8 receptors in particular, are often located perisynaptically and activated only with prolonged stimulation (Cartmell and Schoepp 2000; Schoepp 2001), our results suggest that mGluRs might provide an alternative strategy toward producing a similar effect, conferring on these synapses a stimulation dependence that is in turn essential for the production of multiple patterns by a single network (Katz 1998; Katz and Frost 1995, 1996).


Preparation and population recordings

Rhythmic 600–700 μm thick slices containing the PBC were obtained from 6- to 12-day-old CD-1 outbred mice (Charles River Laboratories, Wilmington, MA) as described in detail elsewhere (Ramirez et al. 1997; Telgkamp and Ramirez 1999). These slices were transferred to a 5-ml recording chamber, and superfused (flow rate 10–15 ml/min) with an artificial cerebrospinal fluid (ACSF) containing (in mM) 118 NaCl, 3 KCl, 1.5 CaCl2, 1 MgCl2, 25 NaHCO3, 1 NaH2PO4, and 30 d-glucose, and equilibrated with carbogen (95% O2-5% CO2, pH 7.4). Temperature was maintained at 29–31°C, and KCl was elevated to 8 mM over a span of 30 min before commencing recordings. Except where otherwise noted, the superfused ACSF was continuously recycled (total bath volume: 200 ml). For experiments with CPT-cAMP and (RS)-α-cyclopropyl-4-phosphonophenylglycine (CPPG), the total volume of the recycling bath was reduced to 20 ml, ≥10 min before bath-applying either compound. Some experiments were performed with two slices (from different animals) in a single perfusion chamber and recorded simultaneously. All salts were obtained from Sigma (St. Louis, MO).

For experiments involving population recordings, mass activity was recorded extracellularly from the caudal surface of the slice in a region presumably corresponding to the rostral VRG (in which inspiratory cells predominate) (Merrill 1970; Sun et al. 1998; von Euler 1986). This activity was band-pass filtered between 300 Hz and 3 KHz, rectified, and integrated (tau ∼150 ms). The resulting data were digitized at 500 Hz with a Digidata board (Axon Instruments, Foster City, CA), stored on an IBM compatible PC using Axotape software (Axon), and analyzed off-line using Igor Pro (WaveMetrics, Lake Oswego, OR), Clampfit (Axon), and Prism (GraphPad, San Diego, CA). Only recordings with qualitatively good signal-to-noise ratios (SNR) were used (Lieske and Ramirez 2006) (we estimate <2% of preparations were discarded).

l-serine-O-phospate (L-SOP), L-AP3, and CPT-cAMP were obtained from Sigma. d-AP5, l-AP5, 6-cyano-7-nitroquinoxalene-2,3-dione (CNQX), 6,7-dinitroquinoxaline-2,3-dione (DNQX), and all other metabotropic glutamate receptor agonists and antagonists were obtained from Tocris (Ellisville, MO).


Current-clamp recordings were obtained in the whole cell patch configuration from inspiratory neurons located within 300 μM of the rostral surface of the slice, in a region presumably corresponding to the PBC. Population recordings were obtained from the contralateral VRG for reference, as described in the preceding text except from the rostral surface of the slice. Patch electrodes were manufactured from filamented borosilicate glass (Clark GC150F-10, Warner Instruments, Hamden, CT) and filled with a solution containing (in mM) 140 K-gluconic acid, 1 CaCl2, 10 EGTA, 2 MgCl2, 4 Na2ATP, and 10 HEPES. A liquid junction potential of 12 mV was measured for this solution in ACSF and taken into account. Intracellular recordings were low-pass filtered at 1.8 kHz and sampled at 4.2 kHz.

Once the whole cell patch configuration was obtained, the ACSF was replaced with a fresh solution containing only 3 mM K+ so that the low-amplitude evoked EPSPs could be recorded without a confounding population rhythm. When necessary, constant current was applied to hyperpolarize the recorded cell to a level such that tonic firing activity also ceased (typically <1 nA for an Em of −70 to −80 mV). Stimuli were delivered to the contralateral VRG via the same borosilicate glass electrode from which the population rhythm had been recorded, using an Iso-flex stimulus isolator (AMPI) in constant-voltage mode. Individual stimulation pulses were 0.2–0.3 ms in duration at the lowest amplitude that consistently evoked an EPSP in the recorded cell (typically 200 mV). Stimuli were delivered at 0.5 Hz, and 20 stimulations were averaged for each data point. A complete description of the EPSPs so recorded is given in the companion manuscript (Lieske and Ramirez 2006).

Data analysis

Statistical values are given as means ± SE. Significance was assessed with a paired t-test (normally distributed paired data: burst width, postsigh apnea), the Wilcoxon matched pairs test (paired data which could not be assumed to be normally distributed: irregularity), or a one-sample t-test (ratio data: burst amplitude, and normalized frequencies). Sigh frequencies in (RS)-3,5,-dihydroxyphenylglycine (DHPG) were clearly not distributed normally, so the nonparametric Wilcoxon signed-rank test was used instead of the one-sample t-test. In all cases, significance was defined as P < 0.05. “Burst duration” refers to the full width at half-maximal amplitude. Cycle periods were defined from the beginning of a fictive inspiration to the beginning of the next. The irregularity score was computed for each cycle as Sk = 100% × Abs(Pk –Pk-1)/Pk-1, where Sk is the score of the kth cycle, Pk is its period and Pk-1 is the period of the preceding cycle (Barthe and Clarac 1997)—this represents the breath-by-breath percent change in cycle length. Median irregularity scores were used for comparisons because the distributions were often highly skewed, and a few outliers contributed greatly to the means. Postsigh apneas were excluded from calculations of eupneic frequency and irregularity. Sigh phases were averaged using circular statistics, in which each individual measurement is represented as a unit vector with direction equal to 2× pi × phase. The phase deviation was computed as sqrt[2 × (1 − r)]/(2 × pi), where sqrt is the square root function, and r is the magnitude of the average vector. Phase comparisons were performed by computing pairwise phase differences and applying Watson's test for a specified mean direction of zero, then bootstrapping on the test statistic (10,000 resamples) to improve accuracy given the small sample size (Fisher 1993). For CPPG versus control, the P value obtained from the large sample statistic was P = 0.0302 and that for the bootstrapped test was P = 0.0772. For CPT-cAMP versus control, the large-sample test gave a P value of 0.0061, whereas the bootstrapped test gave a P value of 0.0741. Mean burst frequencies were computed as a time-weighted average, mathematically equivalent to the reciprocal of the mean interburst interval. Neither burst frequency nor cycle length passed normality tests both for eupnea and for sighs, but a log-normal frequency distribution was consistent with both patterns (not shown, see also Lieske et al. 2000b). Statistical comparisons were therefore performed with nonparametric tests or (where necessary) with parametric tests on log-transformed data. For the same reason, the measure of central tendency used in reporting the eupnea:sigh frequency ratio is the geometric mean.

In the companion article, we describe a slow increase in sigh frequency in vitro, independent of any experimental manipulation (Lieske and Ramirez 2006). The rate of increase was dependent on the volume of the recirculating bath. For the 20-ml recirculating volume used for CPPG, sigh frequency increased at an average rate of (0.3343 min-1) per hour (n = 4, not shown). We repeated the control d-AP5 experiments also in the 20-ml bath (n = 6). Sigh frequencies for both were computed as a percent of the predicted frequency given the passage of time, according to the formula %predicted = fEXP/(fCTRL + t × k) where fEXP is the observed sigh frequency for a given dose, fCTRL is the sigh frequency for the same preparation in control conditions (at the start of the experiment), t is the length of time passed since the control sigh frequency was measured, and the constant k is the slope described in the preceding text. There was no significant difference in normalized sigh frequency in d-AP5 alone compared with the results obtained in the large bath (P = 0.8998), so these results were combined to provide the control for the experiments examining the effects of d-AP5 in the presence of CPPG.


Blockade of sighs by L-AP5

We previously described the blockade of sighs by the NMDA receptor antagonist d-AP5 (50 μM) (Lieske et al. 2000a), an effect not observed with the open channel blocker MK-801 (Lieske and Ramirez 2006). We hypothesized that the L-isomer (the less active isomer against the NMDA receptor) might be responsible for this effect. Bath application of the d-isomer of AP5 (25 μM) reproduced the effects of MK-801 on eupnea, decreasing burst duration and frequency, and increasing the irregularity of the rhythm, but inhibited sighs only transiently (Fig. 1A). The L-isomer, by contrast, was without significant effect on eupnea (Fig. 1, B and C), but abolished sighs for periods of up to several hours at concentrations as low as 5 μM (n = 8).

FIG. 1.

Selective blockade of sighs by 2-amino-5-phosphonopentanoic acid (L-AP5) but not d-AP5. A, integrated ventral respiratory group (VRG) population activity shows a transient inhibition of sighs, and a sustained inhibition of eupnea, after bath application of 25 μM d-AP5. Upward excursions represent fictive inspiratory events; smaller excursions corresponding to fictive eupnea and larger excursions corresponding to fictive sighs. B: in another preparation, 5 μM L-AP5 led to a prolonged inhibition of sighs but no apparent effect on eupnea. C: quantification of changes in frequency, irregularity, and burst width of the fictive eupneic rhythm, in d-AP5, L-AP5, and the mixed isomers. Means + SE. Values significantly different from control (*P < 0.05, ***P < 0.001) by 1-sample t-test (frequency), Wilcoxon matched pairs test (irregularity), or paired t-test (burst width). The relatively conservative Wilcoxon test did not quite attain significance for the increase in irregularity in d-AP5, but the trend was quite strong (P = 0.0625).

Because this effect was also distinct from the effects of non-NMDA glutamatergic blockade (Lieske and Ramirez 2006), we further hypothesized that the receptor associated with the L-AP5 effect might be a metabotropic glutamate receptor. A complete screen of agonists and antagonists for each of the various mGluRs was performed using (RS)-4-carboxyphenylglycine (4-CPG), α-methyl-4-carboxyphenylglycine (MCPG), and l-2-amino-3-phosphonoproprionic acid (L-AP3), as prototypical antagonists for group I and group II mGluRs, CPPG as a prototypical group III antagonist, and DHPG, (2R,4R)-4-aminopyrrolidine-2,4-dicarboxylate (APDC), and l-2-amino-4-phosphonobutyric acid (L-AP4) as the respective group I, II, and III agonists (Cartmell and Schoepp 2000; Schoepp et al. 1999). Although each of these compounds has been reported to have effects on the other groups in addition to the prototypical effect, we wished to examine (Cartmell and Schoepp 2000; Schoepp et al. 1999), these effects require higher concentrations, and screening both agonists and antagonists for all three groups allows for the resolution of any ambiguity.

Effects of mGluR agonists

The group I mGluR agonist DHPG has been reported to increase respiratory frequency in a preparation similar to ours, but that apparently did not exhibit sighs (Mironov and Richter 2000); those authors compared the time course of this effect to the transient frequency increase observed when the preparation is exposed to anoxia (Haddad and Jiang 1993; Lieske et al. 2000b; Mironov and Richter 2000; Neubauer et al. 1990). Our experiments with DHPG (20 μM) reproduced this result (Table 1) and extended it by showing a significant increase in sigh frequency as well (Fig. 2, peak frequency = 2,658 ± 1,665% of control, P = 0.0156; sustained frequency = 499 ± 118% of control, P = 0.0156, n = 6), which is also a feature of the anoxic augmentation (Bartlett 1971; Cherniack et al. 1981; Lieske et al. 2000b). The eupnea:sigh frequency ratio decreased from 31.1 to 4.5 at the peak frequency augmentation with a value of 8.8 after the peak (n = 6, geometric means; P = 0.0004, 1-way ANOVA on log-ratios). There were also significant decreases in the amplitude and duration of the eupneic bursts (Table 1), again consistent with an anoxic-like effect (Lieske et al. 2000b).

FIG. 2.

Effects on eupnea and sighs of the group I mGluR agonist (RS)-3,5,-dihydroxyphenylglycine (DHPG). A: integrated VRG population activity shows a transient increase in the frequency of both eupnea and sighs, and a sustained increase in sigh frequency after bath application of 20 μM DHPG. Tonic activity is also increased, as reflected in the upward shift in the baseline. B: sigh frequency in DHPG, computed as a percentage of that observed in control, quantified over the initial peak, and from 10 to 20 min after application. Note the logarithmic y axis. C: postsigh apnea in control conditions and in DHPG, normalized to eupneic cycle length. Means ± SE. Data from 6 preparations. Values significantly different from control, *P < 0.05.

View this table:

Effects of selected mGluR agonists and antagonists on eupnea

The group II agonist APDC (20 μM) exhibited modest effects on both eupnea and sighs (Fig. 3, A and B) but also failed to mimic the effects of L-AP5. On average, the amplitude of the eupneic rhythm was significantly reduced (Table 1), as was the frequency of sighs (P = 0.0156, n = 6; Fig. 3C). The eupnea:sigh frequency ratio increased from 55.7 to 76.0 (n = 6, geometric means; P = 0.0209).

FIG. 3.

Effects on eupnea and sighs of the group II mGluR agonist (2R,4R)-4-aminopyrrolidine-2,4-dicarboxylate (APDC). A: integrated VRG population activity shows a decrease in eupneic amplitude, coupled with an increase in the amplitude variability following application of 20 μM APDC. B: population activity from a different preparation shows only a modest effect on eupneic amplitude but a clear decrease in sigh frequency and regularity. C: sigh frequency in APDC, computed as a percentage of that observed control. D: postsigh apnea in control conditions and APDC, normalized to eupneic cycle length. Data from 6 preparations. Values significantly different from control, *P < 0.05.

By far the most dramatic effects were observed with the group III agonists L-AP4 and l-serine-O-phosphate (L-SOP), in which sighs were consistently abolished, for periods of up to several hours (Fig. 4A; 1–20 μM; L-AP4, n = 6; L-SOP, n = 21). In one case, the fictive eupneic rhythm was transiently abolished as well, but returned within a few minutes (Fig. 4B). L-SOP (20 μM) led to a statistically significant reduction of eupneic frequency (Table 1); both amplitude and irregularity tended to increase, but neither effect achieved significance (Table 1), and the duration of the eupneic bursts was unaffected.

FIG. 4.

Selective blockade of sighs by group III mGluR agonists l-serine-O-phospate (L-SOP) and L-AP4. A: integrated VRG population activity shows the presence of both fictive eupnea and sighs in control and only eupnea in the presence of 20 μM L-SOP. B: integrated VRG population activity in another preparation shows both a transient suppression of fictive eupnea and a persistent blockade of sighs after bath application of 20 μM L-AP4.

Effects of mGluR antagonists

Several antagonists of the group III receptors were tested, including (S)-2-amino-2-methyl-4-phosphonobutanoic acid (MAP4), (RS)-α-methylserine-O-phosphate (MSOP), CPPG, and (RS)-α-methyl-4-phosphonophenylglycine (MPPG). Both MAP4 (100–200 μM, n = 3) and MSOP (100–200 μM, n = 3) led to a decrease in sigh frequency (Fig. 5, A and C). This was not observed with CPPG (100 μM, n = 16, Fig. 5, B and C) or MPPG (200–300 μM, n = 3, not shown). The variability in the phase-locking of the sigh to the eupneic rhythm was increased in CPPG (phase deviation = 0.0623 ± 0.0139 vs. 0.0319 ± 0.0085, P = 0.0119, n = 16) but not in MSOP/MAP4 (P = 0.6875, n = 6), while the duration of the postsigh apnea was not affected by either (Fig. 5D). The phase value itself tended to increase in CPPG, from 0.079 ± 0.053 to 0.097 ± 0.064 (P = 0.0772). Neither the amplitude nor the duration of the sigh burst was affected by CPPG (n = 16; amplitude, P = 0.2718; duration, P = 0.1456). MSOP and MAP4 were without significant effect on the eupneic rhythm (Table 1), while CPPG led to a significant increase in eupneic frequency and significant decreases in amplitude and burst duration (Table 1). The eupnea:sigh frequency ratio increased in MSOP/MAP4 (40.9 vs. 29.0, P = 0.0048, paired t-test on log-ratios, n = 6) but did not change in CPPG (22.2 vs. 19.9, P = 0.2208, n = 16). For completeness, antagonists of the group I and group II mGluRs including 4-CPG, L-AP3, and MCPG were also screened; blockade of these receptors did not lead to any obvious effect on sighs (n = 4, not shown).

FIG. 5.

Effects on eupnea and sighs of the group III mGluR antagonists (S)-2-amino-2-methyl-4-phosphonobutanoic acid (MAP4) and (RS)-α-cyclopropyl-4-phosphonophenylglycine (CPPG). A: integrated VRG population activity shows no obvious effect on either eupnea or sighs after bath application of 200 μM MAP4. B: in another preparation, 100 μM CPPG was also without obvious effect. C: sigh frequency in (RS)-α-methylserine-O-phosphate (MSOP)/MAP4 (pooled data, n = 3 MSOP, 3 MAP4) computed as a percentage of that observed in control, and in CPPG (n = 16), computed as a percentage of that predicted after the passage of time in the 20-ml bath (see methods). D: postsigh apnea in control conditions and in MSOP/MAP4 or CPPG normalized to eupneic cycle length. Means ± SE, Values significantly different from control (statistical tests as in Fig. 2), ***P < 0.001.

L-AP5 acts as mGluR agonist

To test whether L-AP5 might also be acting as an agonist at a group III mGluR, we examined whether its effects persisted in the presence of various group III antagonists. Both MSOP and MAP4 (100–200 μM) dramatically reduced the duration of sigh blockade evoked by L-AP5 (Fig. 6, A and C; P = 0.0007, n = 6). In the presence of higher concentrations of antagonist, the effect of L-AP5 appeared to be completely blocked (Fig. 6B, 600 μM MPPG). Similar results were obtained using CPPG (100 μM) to antagonize the effect of L-SOP (20 μM; not shown, n = 5).

FIG. 6.

Blockade of sighs by L-AP5 due to agonistic action at a group III mGluR. A: integrated VRG population activity shows the preservation of sighs in the presence of both L-AP5 and the group III mGluR antagonist MAP4. Note that sighs, although not blocked, nonetheless decrease in frequency in MAP4 + L-AP5 compared with MAP4 alone. B: in a separate preparation, a very high concentration of the group III mGluR antagonist (RS)-α-methyl-4-phosphonophenylglycine (MPPG) prevents L-AP5 from inhibiting sighs frequency, at concentrations of ≤40 μM. C: sigh frequency in the presence of L-AP5 + MSOP or MAP4 as compared with L-AP5 alone. □, “instantaneous” frequency corresponding to the initial pause in sighs, obtained by taking the reciprocal of the 1st inter-sigh interval after application of L-AP5. ▪, mean sigh frequency obtained over the interval from 10 to 30 min after application of drug. In L-AP5 alone, sighs were abolished altogether; the bar shown estimates the maximum sigh frequency possible, given by the reciprocal of the recording duration. D: sigh frequency in the presence of d-AP5 + CPPG, as compared with d-AP5 alone. CPPG experiments were performed in the 20-ml bath, and thus all frequencies are normalized for the passage of time and presented as percentage predicted (see methods). □ and ▪ as in C. Significant differences between d-or L-AP5 in the presence of antagonist and AP5 alone: *P < 0.05, **P < 0.01, ***P < 0.001) by unpaired t-test and significant differences from 100% (i.e., control) by 1-sample t-test: #P < 0.05, ##P < 0.01, ###P < 0.001). Note that for both L-AP5 and d-AP5, the inhibition of sighs was partly prevented by the presence of group III mGluR antagonists (*), but, at the concentration of antagonist used, some inhibition of sigh frequency persisted (#). Blockade of sighs by L-AP5 due to agonistic action at a group III mGluR.

Although the blockade of sighs by d-AP5 is thus attributable primarily to the effects of the L-isomer, the d-isomer did result in a weak inhibition of sighs (Figs. 1A and 6D). The inhibition of sighs by d-AP5 was still significant in the presence of 100 μM CPPG, but the extent of this inhibition was significantly reduced relative to d-AP5 alone (Fig. 6D).

Selective group III mGluR agonists implicate mGluR8

To see if a specific group III receptor could be implicated, we went on to assay a variety of additional group III mGluR agonists with varying selectivities for mGluR4, -6, -7, and -8, including L-AP4, L-SOP, HomoAMPA, l-2-aminoadipic acid (L-2-AA), (RS)-phosphonophenylglycine (PPG), and (S)-3,4-dicarboxyphenylglycine (DCPG). The reported potencies for each of these drugs at each of the group III receptors are given in Table 2. L-SOP and L-AP4 were tested at concentrations as low as 1 μM; both inhibited sighs at least somewhat at this lowest concentration tested (1st interval ranged from 1.6- to 7.4-fold longer than control, n = 3). HomoAMPA was tested at concentrations ranging from 200 to 600 μM; sighs were inhibited at 400 μM and abolished at 600 μM (n = 2). l-2-aminoadipic acid (L-2-AA) did not block sighs at any concentration ≤3 mM (n = 2). PPG inhibited sighs in every preparation at the lowest concentration tested, as low as 1 μM (n = 5). The pharmacological profile for the blockade of sighs established by these experiments strongly suggests that it is mGluR8 that is responsible for this effect (see discussion). This hypothesis was tested with the highly selective mGluR8 agonist (S)-3,4-DCPG (EC50 = 31 nM for mGluR8, >3 μM for mGluR4, -6, -7) (Thomas et al. 2001). Sighs were consistently abolished for ≥1 h after bath application of 100 nM DCPG (n = 6, Fig. 7), indicating that it is indeed activation of mGluR8 that leads to suppression of sighs.

FIG. 7.

The specific group III receptor responsible for the blockade of sighs is mGluR8. Integrated VRG population activity shows both fictive eupnea and sighs in control conditions and only eupnea after bath application of the highly selective mGluR8 agonist DCPG at a concentration of 0.1 μM. Blockade of sighs persisted for ≥1 h in all examined preparations (n = 6).

View this table:

Reported potencies of group III mGluR agonists and antagonists, and their effects on sighs

Dose dependence and time course

The blockade of sighs induced by group III mGluR agonists was not absolute: lower agonist concentrations led to a reduction in sigh frequency rather than a complete abolition (Fig. 8A), and sighs did eventually return even in higher concentrations when recordings were maintained for several hours (Fig. 8B). This recovery was not affected if the ACSF was replaced (keeping the drug concentration constant) rather than recycled thus was not due to a breakdown or sequestering of drug (Fig. 8B). The recovery of sighs was also dose dependent: sighs were again blocked if the concentration of drug was increased (Fig. 8A). These results are consistent with a desensitization of the receptor but could also represent an independent effect: the frequency of in vitro fictive sighs slowly increases in the absence of any experimental manipulation (Lieske and Ramirez 2006), thus it is possible that a greater activation of the mGluR would be necessary to block sighs at later time points.

FIG. 8.

Recovery and dose dependence of sigh blockade by group III mGluR agonists. A: integrated VRG population activity shows (top) a reduction in sigh frequency after application of 1 μM (RS)-phosphonophenylglycine (PPG), blockade of sighs after an increase in PPG concentration to 2 μM. Sighs recover after ∼45 min in the continued presence of 2 μM PPG (bottom) but are again blocked following an increase in PPG concentration to 5 μM. B: in a different preparation, bath application of 10 μM L-SOP leads to a blockade of sighs lasting ∼3 h. After sighs recover (in the continuous presence of 10 μM L-SOP), the recycled artificial cerebrospinal fluid (ASCF) is replaced with fresh ACSF also containing 10 μM L-SOP and sighs persist.

Cellular effects

At the cellular level, there were no obvious effects of L-SOP application, aside from the blockade of sighs. In other systems, group III mGluRs have been shown to inhibit presynaptic voltage-activated calcium channels, reducing synapse strength (Cartmell and Schoepp 2000; Dong and Feldman 1999; Dube and Marshall 1997; Takahashi et al. 1996; Thomas et al. 2001; Trombley and Westbrook 1992; Wittmann et al. 2001). To test the hypothesis that mGluR8 acts by a similar mechanism here, we examined the effects of L-SOP on excitatory postsynaptic potentials (EPSPs) recorded intracellularly from inspiratory neurons within the pre-Bötzinger Complex (PBC). EPSPs were evoked by stimulation of the contralateral nucleus as described in the companion paper (Lieske and Ramirez 2006). On average, L-SOP (20 μM) led to a reduction in EPSP amplitude of only ∼15% (Fig. 9A; n = 7), with a great deal of variability between individual preparations (Fig. 9, B and D). No change was observed in membrane potential (−69.7 ± 4.2 vs. −69.8 ± 4.4 mV, P = 0.7430, n = 7). In one experiment in which the reduction in EPSP amplitude was relatively pronounced, the magnitude of the reduction could be seen to diminish with time, decaying with a time constant of ∼20 min (Fig. 9C).

FIG. 9.

Amplitude of intracellularly recorded evoked excitatory postsynaptic potentials (EPSPs) reduced by the group III mGluR agonist L-SOP. A: 20 μM L-SOP, a concentration sufficient to abolish sighs for ≥1 h in every preparation examined, reduced EPSP amplitude by ∼15% on average. Points represent mean (±SE) over all examined preparations (n = 7) including both apparent “responders” and “nonresponders” as shown in B. Values for individual preparations represent the amplitude of a mean EPSP obtained by averaging the response to 20 stimulations at each time point. Stimulations were 0.2-ms duration in duration and delivered to the contralateral VRG at a rate of 0.5 Hz. B: data for individual experiments contributing to averages in A. Note the variation observed between preparations. C: in a preparation in which EPSP amplitude was markedly reduced by 20 μM L-SOP, this effect rapidly diminished over the subsequent hour. Fit line is an exponential decay, τ = 20.1 min. D: in another preparation, EPSP amplitude was decreased minimally or not at all in 20 μM L-SOP but eliminated by application of 200 μM cadmium to block voltage-activated calcium channels.

The adenylate cyclase activator forskolin has been described to stimulate the respiratory network in vitro (Mironov et al. 1999; Shao et al. 2003), an effect that can be used to distinguish between the blockade of sighs (unrecovered in forskolin) and the inhibition of eupnea (recovered in forskolin) caused by certain calcium channel toxins (Lieske and Ramirez 2006). Because the group III metabotropic glutamate receptors are commonly found to act via inhibition of adenylate cyclase (Conn and Pin 1997; Nakanishi 1994), we investigated whether the effects of mGluR agonists could be mimicked, or prevented, by interfering with the cAMP pathway directly.

Bath application of the membrane-permeable cAMP analogue CPT-cAMP (1 mM) led to a significant increase in the frequency of both eupnea and sighs (341 ± 37% of control, P = 0.0003; and 1078 ± 93% of control, P = 0.0001, respectively). Sigh frequency was increased more than eupneic frequency: the eupnea:sigh ratio decreased from 16.6 to 5.2 (n = 6, geometric means; P < 0.0001). In a seventh preparation, sigh frequency increased so dramatically that the ratio fell <2.0 and could no longer be computed. Nonetheless, the blockade of sighs by L-SOP was not prevented in the presence of CPT-cAMP (Fig. 10, n = 6) nor did there appear to be a primary effect on the triggering of sighs as there was no significant change in either phase (P = 0.0741, n = 6) or phase deviation (P = 0.0764). Similar results were also obtained with the adenylate cyclase activator forskolin (10 μM, n = 6, not shown). Conversely, neither of the cAMP-dependent protein kinase inhibitors Rp-cAMPS (100–200 μM, n = 3) or H-89 (30–100 μM, n = 3) abolished sighs (not shown).

FIG. 10.

The membrane permeable cAMP analogue CPT-cAMP stimulates the frequency of both eupnea and sighs but does not prevent the blockade of sighs by L-SOP. Top: integrated VRG population activity corresponding to designated portions of bottom. Bottom: amplitude of integrated activity during sighs (○) and eupnea (•) plotted against time of occurrence. Units are arbitrary.


Potential role of mGluR8 receptors in the generation of sighs

Here we report that sighs were selectively abolished by L-AP5. L-AP5 is a low-affinity NMDA receptor antagonist (Ki = 40 μM) (Chen et al. 2002; Davies and Watkins 1982; Olverman et al. 1988) and has also been described to act as an antagonist at the group II metabotropic glutamate receptor mGluR2 (Kb vs. glutamate = 205 μM) and as a weak agonist at the group III receptor mGluR4 (EC50 ∼1 mM) (Brauner-Osborne and Krogsgaard-Larsen 1998). It is inactive at the group I receptors mGluR1 and -5 (Brauner-Osborne and Krogsgaard-Larsen 1998), and to our knowledge, the presence or absence of activity at the other group III receptors (mGluR6, -7, and -8) or at the group II receptor, mGluR3 has not been reported. Because both the identity of the receptor involved and the extent of activation necessary to prevent sighs were unknown, it was necessary to compare the effective concentrations of a variety of group III agonists with differing pharmacological profiles. Comparing effective concentrations with EC50 values reported in the literature (Table 2), the L-AP4 and L-SOP results are consistent with activation of mGluR4, -6, or -8. Of these, mGluR4 is argued against by its high EC50 for L-AP5, and mGluR6 is argued against by the homoAMPA and L-2-AA data. Elimination thus suggests mGluR8 as the receptor involved, a hypothesis confirmed by the blockade of sighs by both PPG and DCPG at concentrations above their respective EC50's for mGluR8, and below their EC50's for each of the other receptors. While the activity of homoAMPA at mGluR8 has not been described, the blockade of sighs at concentrations several-fold higher than its EC50 for mGluR6 suggests that this compound may have a weak agonistic effect on mGluR8 as well.

Also consistent with our pharmacological data are the recent demonstration of mGluR8-positive neurons within the medulla of rats (Pamidimukkala et al. 2002). Although these authors did not specifically identify the VRG, mGluR8 immunoreactivity was found in the ventrolateral medulla in and adjacent to the lateral reticulate nucleus and nucleus ambiguus in the area which the VRG occupies.

Cellular level effects

In expression systems, the group III mGluRs exert their downstream effects via the cAMP second-messenger system, specifically via the inhibition of adenylate cyclase by Gi/o (Conn and Pin 1997; Nakanishi 1994). We found, however, that both the adenylate cyclase activator forskolin and the membrane permeable cAMP analogue CPT-cAMP, while dramatically increasing the frequency of both eupnea and sighs, failed to prevent the blockade of sighs by L-SOP. The presence of a significant pause in fictive sigh activity despite the relatively intense stimulation of the cAMP pathway suggests that this inhibition of sighs is not dependent on a reduction in intracellular cAMP.

One possible alternative mechanism by which this receptor might be acting is via the blockade of α1A-containing (P/Q-type) voltage-activated calcium channels. In the companion article, we describe a selective elimination of sighs after the blockade of α1A calcium channels with ω-agatoxin TK or ω-conotoxin MVIIC (Lieske and Ramirez 2006), and these channels have been shown to be directly inhibited by G-protein βγ subunits (Herlitze et al. 1996; Ikeda 1996; Miller 1998). Inhibition of non-NMDA glutamate receptors also abolishes sighs, suggesting that it is in a synaptic role that α1A channels are important (Lieske and Ramirez 2006). The present results are also consistent with this hypothesis, as mGluR8 is well known as a presynaptic modulator of synaptic strength (Cartmell and Schoepp 2000; Wittmann et al. 2001), although we cannot preclude the possibility that mGluR8 activation inhibits sighs through an additional (postsynaptic) mechanism. The relatively weak reduction of EPSP amplitude by L-SOP (∼15%) suggests that either a different stimulation protocol is necessary to elicit EPSPs that are highly sensitive to mGluR8 activation or there is an anatomical distinction such that the particular synapses at which mGluR8 activation abolishes sighs is distinct from those on the crossing connections studied herein (or both). The hypothesis of distinct subpopulations of glutamatergic synapses is consistent with the differential roles observed for different subtypes of calcium channels (Lieske and Ramirez 2006) and might also explain the disparity between the time constant of the desensitization of EPSP inhibition (20 min) and the duration of sigh blockade (typically greater than an hour for 20 μM L-SOP).

Network level effects

Bath application of exogenous agonists, in addition to abolishing sighs, also led to a variable but on average significant decrease in eupneic frequency, similar to what has been described in the lamprey locomotor system (El Manira et al. 2002). This indicates that mGluR8 receptors are present either postsynaptically, on eupnea-generating cells, or presynaptically, on cells providing either tonic or phasic (inspiratory) input to those cells. If these receptors are routinely activated endogenously—either tonically or phasically—bath application of antagonists would be expected to also have significant effects, in the opposite direction of the effects of the agonists. Consistent with this, a significant increase in eupneic frequency was observed in the presence of 100–200 μM CPPG. Surprisingly, CPPG did not lead to an increase in sigh frequency. Thus although exogenous activation of mGluR8 suffices to abolish sighs, endogenous activation may not normally contribute to determining sigh frequency. MSOP and MAP4 had slightly different effects than CPPG: there was a significant, although modest, reduction in sigh frequency; and although eupneic frequency tended to increase, this effect was less dramatic than in CPPG and did not achieve significance. In addition, CPPG had effects on eupneic amplitude and burst duration that were not observed with MSOP/MAP4 (Table 1). These differences could be consistent with subtle differences in the effects of these molecules, including partial agonist activity (Knopfel et al. 1995; Laurie et al. 1997; Sekiyama et al. 1996).

The signaling pathway through which activation of mGluR8 abolishes sighs appears to be independent of the cAMP pathway, and could represent a direct interaction between the Gβγ subunits of the associated G protein and the P/Q-type calcium channel identified in the companion article as essential for sighs (Lieske and Ramirez 2006). Further experiments will be necessary to identify those neurons that possess the synaptic properties as suggested by our results.


This work was supported by National Institutes of Health Grant 60120 and 5 T32 GM-07281.


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


View Abstract