Abnormal cell-intrinsic and network excitability in the neocortex of serotonin-deficient Pet-1 knockout mice

Abstract

Neurons originating from the raphe nuclei of the brain stem are the exclusive source of serotonin (5-HT) to the cortex. Their serotonergic phenotype is specified by the transcriptional regulator Pet-1, which is also necessary for maintaining their neurotransmitter identity across development. Transgenic mice in which Pet-1 is genetically ablated (Pet-1^−/−) show a dramatic reduction (∼80%) in forebrain 5-HT levels, yet no investigations have been carried out to assess the impact of such severe 5-HT depletion on the function of target cortical neurons. Using whole cell patch-clamp methods, two-dimensional (2D) multielectrode arrays (MEAs), 3D morphological neuronal reconstructions, and animal behavior, we investigated the impact of 5-HT depletion on cortical cell-intrinsic and network excitability. We found significant changes in several parameters of cell-intrinsic excitability in cortical pyramidal cells (PCs) as well as an increase in spontaneous synaptic excitation through 5-HT₃ receptors. These changes are associated with increased local network excitability and oscillatory activity in a 5-HT₂ receptor-dependent manner, consistent with previously reported hypersensitivity of cortical 5-HT₂ receptors. PC morphology was also altered, with a significant reduction in dendritic complexity that may possibly act as a compensatory mechanism for increased excitability. Consistent with this interpretation, when we carried out experiments with convulsant-induced seizures to asses cortical excitability in vivo, we observed no significant differences in seizure parameters between wild-type and Pet-1^−/− mice. Moreover, MEA recordings of propagating field potentials showed diminished propagation of activity across the cortical sheath. Together these findings reveal novel functional changes in neuronal and cortical excitability in mice lacking Pet-1.

the monoamine serotonin (5-hydroxytryptamine, 5-HT) is a functionally diverse molecule that plays a critical role within the central nervous system (CNS) as a physiological signal in cell-to-cell communication (Nichols and Nichols 2008) and as a developmental signal guiding the patterning of distinct forebrain structures (Gaspar et al. 2003). A subset of midbrain neurons located within the dorsal and median raphe nuclei distribute 5-HT-containing axonal terminals throughout the entire extent of the forebrain (Tork 1990; Waterhouse et al. 1986). The cortex, in particular, receives dense innervation via the medial forebrain bundle fibers that originate from the raphe nuclei (Blue et al. 1988; Stone and Tork 1990). 5-HT released from their terminals acts on six distinct families of G protein-coupled metabotropic receptors (5-HT₁, 5-HT₂, 5-HT₄, 5-HT₅, 5-HT₆, and 5-HT₇) and one family of ligand-gated ionotropic 5-HT₃ receptors (Dougherty and Aloyo 2011; Nichols and Nichols 2008; Santana et al. 2004). The complexity of serotonergic signaling on cortical neurons is further complicated by heterogeneous expression patterns of specific 5-HT receptors across regions, laminae, and cell types of the cortex (Amargos-Bosch et al. 2004; Blue et al. 1988; Jakab and Goldman-Rakic 2000; Santana et al. 2004; Weber and Andrade 2010) and microcircuit-specific and context-dependent activation of distinct receptors on neurons expressing multiple 5-HT receptor types (Amargos-Bosch et al. 2004). Despite recent advances in understanding the role of 5-HT in cortical function (Andrade 2011; Celada et al. 2013; Nakamura and Wong-Lin 2014), this area of inquiry remains in need of further exploration.

Uncovering the fundamental principles of 5-HT system development and function within the CNS has been aided substantially by the use of transgenic animals in which different components of the 5-HT system are genetically altered to augment or diminish systemic and/or region-specific levels of 5-HT (Deneris and Wyler 2012; Gaspar et al. 2003). The development of a mouse line in which the ETS domain transcription factor Pet-1 is genetically ablated has been particularly useful for understanding the processes underlying specification and maintenance of the serotonergic neuron phenotype (Hendricks et al. 2003; Liu et al. 2010). Constitutive deletion of Pet-1 results in a severe reduction (∼70% loss) of 5-HT-immunoreactive neurons within the brain stem raphe nuclei, which leads to a dramatic decrease (∼80% reduction) in forebrain 5-HT levels (Hendricks et al. 2003; Liu et al. 2010). These deficits are accompanied by a decrease in the expression of genes regulating 5-HT synthesis, reuptake, packaging into vesicles, and storage in the remaining population of 5-HT neurons. The behavioral consequences of such an extreme depletion in systemic 5-HT levels are characterized by increased anxiety and aggression (Hendricks et al. 2003; Liu et al. 2010) as well as altered maternal behavior, i.e., decreased nesting (Lerch-Haner et al. 2008). In addition to the primary changes in expression of genes regulated by Pet-1 and the resulting behavioral phenotypes, mice lacking Pet-1 also exhibit hypersensitive cortical 5-HT₂ receptors and an increased expression and hypersensitivity of cortical 5-HT₁ receptors, perhaps as compensation for decreased 5-HT signaling (Yadav et al. 2011). Whether these perturbations to the 5-HT system correspond to alterations in the function of cortical neurons and networks remains unknown.

In this study, we address the consequences of systemic 5-HT depletion on cortical function and neuronal structure in Pet-1-null (Pet-1^−/−) mice. We employed single-cell electrophysiology in cortical slices from wild-type (WT) and Pet-1^−/− mice to assess potential changes in cell-intrinsic, synaptic, and network excitability. We also performed cytochemistry and three-dimensional (3D) neuronal reconstructions to determine whether deletion of Pet-1 leads to abnormal neuronal morphology in the cortex. Our results show alterations across cell-intrinsic, synaptic, and network levels of cortical excitability in mice lacking Pet-1. We also report changes in dendritic morphology in cortical pyramidal neurons, indicated by a significant decrease in complexity of basal dendritic arbors. In light of the changes in cortical excitability, we also carried out acute behavioral seizure experiments and found no changes in susceptibility to acute seizures in Pet-1^−/− mice, suggesting that morphological changes may compensate for increased excitability in vivo by limiting the propagation of neuronal activity through cortical networks. This hypothesis was corroborated by findings from multielectrode array (MEA) recordings demonstrating reduced propagation of network activity in brain slices from Pet-1^−/− mice compared with their WT littermates. Our study provides a comprehensive analysis of a mouse model of 5-HT depletion, including the molecular, cellular, and network levels.

MATERIALS AND METHODS

All experimental protocols were approved by the Institutional Animal Care and Use Committee at Case Western Reserve University.

Thalamocortical slice preparation.

Thalamocortical slices (350 μm) of somatosensory cortex were prepared as previously described (Agmon and Connors 1991; Puzerey et al. 2014) from juvenile (P13–P21) C57BL/6 WT and Pet-1^−/− mice. Animals were anesthetized with vapor isoflurane and decapitated with a guillotine. The brain was then submerged in ice-cold artificial cerebrospinal fluid (ACSF) saturated with 95% O₂-5% CO₂ containing the following (in mM): 125 NaCl, 2.5 KCl, 1 MgCl₂, 2 CaCl₂, 25 NaHCO₃, 1.25 NaH₂PO₄, and 25 glucose. Brain slices were cut on a vibratome (Leica VT1200). All chemical salts and reagents were purchased from Fisher Scientific (Pittsburgh, PA) and Sigma-Aldrich (St. Louis, MO). After the vibratome, slices were transferred to a bath containing room temperature ACSF for 20 min to incubate. Subsequently, slices were moved to the recording chamber and perfused with standard ACSF warmed to 31°C with a TC-324B Automatic Temperature Controller (Warner Instrument, Hamden, CT) at a rate of 2 ml/min. Slices were then incubated for 1 h before electrophysiological recordings began.

Single-cell electrophysiology.

Pyramidal cells (PCs) within layer 2/3 were identified by visual inspection at ×63 magnification using Kohler illumination with an upright microscope (Zeiss Axioskop 2 FS+). Patch-clamp recordings under the whole cell current-clamp configuration were collected from single neurons with borosilicate glass electrodes (6–10 MΩ) filled with standard internal solution containing the following (mM): 120 potassium gluconate, 2 KCl, 10 HEPES, 10 sodium phosphocreatine, 4 MgATP, 0.3 Na₃GTP, and 25 QX-314, adjusted to pH 7.4 with KOH. For voltage-clamp recordings, a cesium-based internal solution was used to improve space clamp and contained the following (in mM): 120 cesium gluconate, 2 CsCl, 10 HEPES, 10 sodium phosphocreatine, 4 MgATP, 0.3 Na₃GTP, 20 BAPTA, and 25 QX-314 to block voltage-gated sodium channels and adjusted to pH 7.4 with CsOH. Electrophysiological recordings were amplified with a Multiclamp 700B amplifier (Molecular Devices, Foster City, CA) and digitized at 10 kHz with Digidata 1400 data acquisition interface. Data were low-pass filtered online at 1 kHz.

Spontaneous excitatory postsynaptic currents (sEPSCs) were recorded in the voltage-clamp configuration for 60 s. sEPSCs were recorded by holding the membrane potential at the reversal potential for inhibitory postsynaptic currents (E_IPSC), which was experimentally determined to be −80 mV, consistent with previous reports (Chagnac-Amitai and Connors 1989; Hasenstaub et al. 2005). Neurons whose access resistance exceeded 30 MΩ throughout the duration of the recordings were excluded from analysis. Drift of the resting membrane potential to values more positive than −60 mV was also used as an exclusion criterion. All pharmacological manipulations were carried out via bath application of the drug. Slices were then given 1 h to incubate before recordings began.

Current-clamp recordings were carried out to measure cortical network activity in a disinhibited slice. We added 1–5 μM gabazine (GZN), a selective GABA_A receptor antagonist, to the bath ACSF to partially disinhibit the slice. In WT animals, spontaneous network activity under these conditions manifested as paroxysmal depolarizing shifts (PDSs), which have a consistent voltage profile comprising an initial plateau depolarization (60–80 mV relative to baseline) lasting 400–500 ms succeeded by a decaying tail lasting ∼500 ms. Recordings of cortical network activity lasted 10 min and were obtained without current injection (I_hold = 0 pA). The concentration of GZN was chosen so to elicit on average one or two PDSs per minute, thus making the recordings amenable to statistical analysis. At this concentration, the network was not sufficiently disinhibited to spontaneously exhibit fast run epileptiform oscillations in WT animals but caused fast runs in Pet-1^−/− mice.

One slice was used from a single animal, and five to seven cells were recorded in each slice. Statistical tests were carried out on groups of 30–40 cells (corresponding to 5–6 animals/group).

2D multielectrode array electrophysiology.

High-density perforated MEAs with 120 electrodes (100-μm pitch) and the MEA-2100 System (amplifier and digitizer) from Multichannel Systems (MCS, Reutlingen, Germany) were used to record the emergence and propagation of network activity in disinhibited thalamocortical slices (350 μm thick). The entire vertical width of the somatosensory cortex (∼1 mm) was fixed to the 1.2 × 1.2-mm MEA recording field with 15-mbar suction. The slice was continuously perfused with ACSF at 2 ml/min at 31°C for 1 h before 1–5 μM GZN was washed in to disinhibit the slice. The activity from the disinhibited slice was recorded at a sampling frequency of 50 kHz with a resolution of 16 bits in the range of −2 mV to 2 mV. Data were digitally band-pass filtered from 1 to 40 Hz, and 2D spline interpolation was used to plot activity maps over the MEA field with a spatial resolution of 10 μm. To determine the maximal area of propagation of a PDS and its maximal growth rate, we first set a threshold at 20% of the peak activity (largest potential recorded with the MEA during the PDS). Then we calculated the total area exceeding that threshold as a function of time, so that its time derivative yields the growth rate. The maxima of both the area and its growth rate were then obtained for each propagating PDS.

Seizure induction.

Juvenile C57BL/6 and Pet-1^−/− mice (P18–P24) were injected with the convulsant pentetrazol (PTZ; 80 mg/kg ip) dissolved in 0.1 M phosphate-buffered saline (PBS). The starting point of behavioral seizures was considered as the appearance of generalized clonic convulsions with loss of righting reflex (GCC-LOR) (Loscher et al. 1990). The end point for each GCC-LOR seizure was considered as release from tonus (in animals presenting with tonus) or as cessation of generalized clonic convulsions and recovery of righting. Time elapsed between PTZ injection and the first GCC-LOR seizure was taken as the seizure latency. In most acute seizure experiments the animal died after experiencing several seizures. If the seizures persisted for >1 h after PTZ injection, the animals were euthanized by isoflurane anesthesia and decapitation. Animals that did not experience seizures within an hour after PTZ injection were not included in the data analysis.

Data analysis and statistics.

sEPSCs were detected with a custom algorithm written in MATLAB (MathWorks). Detection was based on a threshold for the sEPSC derivative with threshold values for events obtained empirically. We provided the algorithm with additional detection criteria based on the event kinetics. Events with rise time exceeding 5 ms and decay time constant longer than 30 ms were excluded; also, the rise time was not allowed to exceed the decay time constant. For verification of successful detection, events were visually inspected after being passed through the exclusion criteria. Network activity in current clamp was detected manually with a custom interactive detection algorithm written in MATLAB. Voltage deflections during PDS network activity typically were between 50 and 80 mV, depending on the resting potential of the neuron (typically −70 mV), and thus could be detected easily by a simple amplitude threshold-based detection. Reverberant afterdischarges during fast runs were detected by setting an absolute amplitude threshold of 30 mV relative to the resting potential and a 10 mV threshold relative to the local minimum (i.e., if the voltage deflection occurred on top of a depolarization plateau). We determined statistically significant differences between distributions with the nonparametric Wilcoxon rank sum test, which tests the null hypothesis that two independent groups of samples come from distributions with equal medians. The power spectrum analysis of PDS is described in detail in our previous publication (Puzerey et al. 2014). In brief, the horizontal baseline of the power spectral density is proportional to the number of PDSs and the peak frequency, when present, corresponds to the frequency of the fast runs.

Biocytin staining, histology, and imaging.

Layer 2/3 PCs were filled with 20 mM biocytin (Sigma-Aldrich) during patch-clamp recordings. At the end of the experiments, slices were preserved in 0.1 M PBS solution containing 4% paraformaldehyde for >24 h. Slices were then subjected to three washes in 0.1 M PBS solution containing 0.5% Triton X-100 for 15 min and left to incubate for 24 h in the same solution containing avidin-biotin complex (Elite ABC Staining Kit, Vector Labs). Slices were then washed three times in the PBS-Triton solution for 15 min and stained with a diaminobenzedine (Elite DAB Peroxidase Substrate Kit, Vector Labs) solution until desired staining intensity was achieved. Slices were rinsed three times for 10 min in 0.1 M PBS solution followed by series of alcohol and xylene washes to dehydrate the tissue. Slices were mounted on glass slides with Permount. Stained neurons were visualized with brightfield illumination through a Zeiss ×63 oil immersion objective and reconstructed in 3D with the Neurolucida software package (MicroBrightField, Colchester, VT). Images were further processed and refined with ImageJ (National Institutes of Health) graphics editing software.

RESULTS

Cell-intrinsic parameters of neuronal excitability are altered in Pet-1 knockout mice.

To determine whether the drastic 5-HT depletion in the forebrain of Pet-1 knockout (KO) mice has any effect on the intrinsic excitability of cortical neurons, we carried out whole cell patch-clamp recordings in layer 2/3 of somatosensory cortex in WT and KO animals from individual PCs. We first tested the passive properties of the neuronal membrane, namely the neuronal input resistance (R_input) and the membrane time constant (τ). Comparison of these parameters between WT and KO mice reveals that both R_input and τ are increased in KO animals (R_input: WT mean ± SD: 126 ± 73 MΩ; KO mean ± SD: 333 ± 134 MΩ; P < 0.01, Wilcoxon rank sum test; τ: WT mean ± SD: 9.5 ± 2.6 ms; KO mean ± SD: 13.2 ± 1.9 ms; P < 0.01, Wilcoxon rank sum test; Fig. 1, A and B). A longer time constant corresponds to slower decay of the membrane potential in response to a transient current and, consequently, longer integration time for synaptic inputs. Larger values of τ confer enhanced summation of excitatory inputs in time, thus leading to increased excitability of the neuronal membrane. R_input dictates the magnitude of the voltage response to an injected current according to Ohm's law. This means that larger values of R_input would on average lead to a larger voltage deflection in response to a transient input current of a fixed amplitude. Therefore, the changes observed in R_input and τ confer enhanced excitability to cortical neurons in KO mice. We also compared the resting membrane potential and report no significant differences between WT (mean ± SD: −67.8 ± 11.3 mV) and KO (mean ± SD: −64.2 ± 9.3 mV; P = 0.28, Wilcoxon rank sum test) mice.

Fig. 1.

Altered cell-intrinsic excitability in mice lacking Pet-1. A: box plots representing input resistance measurements from layer 2/3 cortical pyramidal neurons in wild-type (WT; n = 26 cells) and Pet-1^−/− (n = 30 cells) mice. B: box plots representing the membrane time constants from pyramidal cells (τ) in WT (n = 47 cells) and Pet-1^−/− (n = 24 cells) mice. C: membrane potential in response to a current ramp from −50 to 200 pA over 1 s. D: firing patterns in response to current steps of 1 s. E: box plots of spike latency times in response to ramp current injection from WT (n = 22 cells) and Pet-1^−/− (n = 30 cells) mice. F: box plots of maximal firing frequency in response to step current from WT (n = 22 cells) and Pet-1^−/− (n = 30 cells) mice. G: box plots of input dynamic range, taken as the range of current values that elicit changes in the neuronal firing rate, from cortical pyramidal cells in WT (n = 22 cells) and Pet-1^−/− (n = 30 cells) mice. Single asterisks indicate P values between 0.01 and 0.05, and double asterisks P values <0.01.

We then investigated the active intrinsic membrane properties underlying the generation of action potentials in cortical PCs of both WT and KO mice. We first compared differences in spike threshold by injecting ramp currents (from −50 to 200 pA over 1 s) in the current-clamp configuration (Fig. 1C) and observed no difference between the two groups (WT mean ± SD: −31.8 ± 8.4 mV, KO mean ± SD: −32.1 ± 6.7 mV; P = 0.33, Wilcoxon rank sum test). Ramp current injection also allows for measurement of the latency to the onset of the first spike, which serves as a proxy for neuronal excitability. Figure 1, C and E, shows that KO mice have substantially delayed spike responses to injection of ramp current compared with their WT littermates (WT mean ± SD: 447 ± 80 ms, KO mean ± SD: 636 ± 233 ms; P < 0.01, Wilcoxon rank sum test). We also compared frequency-input (F-I) curves, which determine neuronal gain, and observed no significant differences between the two groups (data not shown). Despite seeing no differences in neuronal gain, we observed that cortical PCs from KO animals tend to reach a lower maximal firing frequency (MFF) in response to step current injection (Fig. 1, D and F). Quantifying the MFF for WT and KO groups showed that, indeed, PCs in KO mice reach a lower MFF on average than those from WT mice (WT mean ± SD: 47.3 ± 9.3 Hz, KO mean ± SD = 32.8 ± 6.9 Hz; P < 0.05, Wilcoxon rank sum test; Fig. 1F). To understand why PCs from KO mice have lower MFF, we compared the range of values of the input current that elicit changes in the firing rate (i.e., input dynamic range) and report that this range is on average lower in KO animals (WT mean ± SD: 439 ± 119 pA, KO mean ± SD: 273 ± 113 pA; P < 0.01, Wilcoxon rank sum test; Fig. 1G). Combined together these results demonstrate an overall decrease in the active properties of cell-intrinsic excitability, perhaps as compensation for changes in passive membrane properties in the opposite direction.

Motivated by 5-HT's established role in guiding neuritic development (Bou-Flores et al. 2000; Rebsam et al. 2002; Smit-Rigter et al. 2011), we analyzed the detailed neuronal morphology of DAB-labeled cortical PCs by generating their 3D morphological reconstructions. Gross neuronal morphology appeared largely preserved in Pet-1 KO animals (Fig. 2A, left and center); however, Scholl analysis of dendritic morphology revealed a significant reduction in dendritic complexity at the basal arbors of PCs (Fig. 2A, right). These differences present at distances 15–55 μm away from the soma. Scholl analysis of apical dendrites showed no significant differences in dendritic complexity between WT and KO mice (data not shown). Additionally, total length for both apical (mean ± SD: WT: 1,066 ± 424 μm, KO: 1,044 ± 704 μm; P = 0.59, Wilcoxon rank sum test) and basal (mean ± SD: WT: 860 ± 282 μm, KO: 658 ± 398 μm; P = 0.0762) dendrites was comparable between the two groups, although basal dendrites of PCs from KO mice trended toward shorter lengths. Furthermore, the mean surface area of PC somas in Pet-1 KO mice was statistically indistinguishable from that in WT mice (mean ± SD: WT: 174 ± 33 μm², KO: 140 ± 54 μm²; P = 0.18, Wilcoxon rank sum test). These findings show that the aforementioned changes in cell-intrinsic excitability in Pet-1 KO mice are associated with significantly altered dendritic morphology.

Fig. 2.

Dendritic morphology and spontaneous excitatory postsynaptic currents (sEPSCs) in cortical pyramidal cells from Pet-1 knockout mice. A, left and center: images of reconstructed layer 2/3 pyramidal cells from WT (left) and Pet-1^−/− (center) mice. Right: Scholl analysis of basal dendrites in cortical pyramidal cells from WT (n = 6) and Pet-1^−/− mice (n = 14). Asterisks correspond to significant differences in mean number of intersections at a 5% significance level. Error bars represent SE. B: raw traces of sEPSCs in layer 2/3 pyramidal neurons from WT mice, Pet-1^−/− mice, and Pet-1^−/− mouse brain slices treated with the 5-HT₃ receptor antagonist granisetron (GSN, 2 μM) and the 5-HT₂ receptor antagonist ketanserin (KSN, 10 μM). C: distribution of sEPSC amplitudes recorded in cortical slices from WT control slices (n = 48 cells), Pet-1^−/− slices (n = 20), and Pet-1^−/− slices treated with GSN (n = 42) and with KSN (n = 42). D: distribution of sEPSC interevent intervals (IEIs) recorded in cortical slices from the cells in C. All groups in C and D have significantly different medians, except for those pairs denoted by n.s.

Cortical pyramidal cells exhibit increased spontaneous synaptic activity in Pet-1 knockout mice.

One possible outcome of 5-HT depletion in the cortex is a change in synaptic transmission between PCs, which might arise owing to 5-HT's capacity to modulate or directly mediate synaptic transmission (Aghajanian and Marek 1999; Beique et al. 2007; Foehring et al. 2002; Lambe et al. 2000; Roerig et al. 1997; Zhou and Hablitz 1999). To test this possibility, we recorded sEPSCs from layer 2/3 cortical PCs in voltage-clamp configuration from WT and KO mice. Our results show a dramatic and significant increase in the amplitude of sEPSCs in KO animals compared with WTs (P < 0.01, Wilcoxon rank sum test, Fig. 2, B and C). Surprisingly, this increase was normalized to near control levels with an antagonist of 5-HT₃ receptors, granisetron (GSN, 2 μM). This unexpected finding suggests that despite a substantial (∼80%) depletion of forebrain 5-HT levels (Hendricks et al. 2003), there may be a compensatory upregulation of synaptic serotonergic signaling, a finding consistent with other models of serotonin depletion in which signaling through 5-HT receptors was altered (Liu et al. 2010; Moya et al. 2011; Yadav et al. 2011). The effect on the frequency of sEPSCs is less obvious. As shown in Fig. 2D, shorter (<100 ms) interevent intervals (IEIs) are decreased while longer (>100 ms) IEIs are increased in Pet-1 KO animals, a finding that is difficult to interpret. Since the IEI is a proxy for presynaptic vesicle release probability, this finding suggests a complex effect of 5-HT on regulation of spontaneous presynaptic release that is perhaps regulated by more than one receptor type (Gothert 1990). GSN treatment in KOs decreased the IEI relative to both WT and KO groups without pharmacological treatment, but only the former was statistically significant (P < 0.01, Wilcoxon rank sum test; Fig. 2D). Furthermore, we tested the contribution of 5-HT₂ receptors to spontaneous synaptic activity in KO mice, since these receptors are known to exhibit hypersensitive responses to 5-HT (Yadav et al. 2011) and regulate synaptic transmission within the cortex (Aghajanian and Marek 1999; Beique et al. 2007; Foehring et al. 2002; Lambe et al. 2000; Roerig et al. 1997; Zhou and Hablitz 1999). We find that blockade of 5-HT₂ receptors with a selective antagonist, ketanserin (KSN; 10 μM), dramatically reduced the amplitude (P < 0.05, Wilcoxon rank sum test; Fig. 2C) and increased the IEI (P < 0.05, Wilcoxon rank sum test; Fig. 2D) of sEPSCs. These results suggest that aberrant 5-HT signaling through 5-HT₂ receptors results in abnormal spontaneous synaptic transmission in cortical neurons.

Activation of local cortical networks is enhanced in Pet-1 knockout mice.

As a next step to understanding effects of central 5-HT depletion on cortical neurophysiology, we investigated cortical network activity in a partially disinhibited cortical slice (Castro-Alamancos and Rigas 2002; Hablitz 1987; Puzerey et al. 2014). The slice was disinhibited with the GABA_A receptor antagonist GZN (1–5 μM). Under these conditions, cortical networks in slices from WT animals undergo massive spontaneous depolarization plateaus known as paroxysmal depolarizing shifts (PDSs; Fig. 3A) (Hablitz 1987). The voltage profile of a PDS consists of an early plateau phase lasting ∼500 ms followed by a decay phase that lasts up to 2 s. The PDS events correspond to synchronized activity in local cortical networks (Johnston and Brown 1984; McCormick and Contreras 2001). Therefore, one can use single-cell recordings from cortical PCs as a readout of network activity. We thus compared the voltage profile and statistics of cortical network activity in slices from WT and KO mice. We found that unlike the slices from WT animals, which manifest network activity as sparse (∼1 PDS/min; Fig. 3A) and temporally random PDS, the network activity in KO mice had a voltage profile that was typically characterized by a massive PDS succeeded by a series of afterdischarges with a preferred frequency of 10–15 Hz and a burst duration of ∼40 ms (Fig. 3B). The afterdischarges were typically superposed on a long depolarizing plateau. The emergence of these afterdischarges resulted in a substantially larger number of network events recorded over a 10-min period (P < 0.01, Wilcoxon rank sum test; Fig. 3E). It is important to note that the afterdischarges were not a result of intrinsic spiking in recorded neurons since they persisted in the presence of a voltage-gated sodium channel blocker, QX-314, in the intracellular solution. Thus the afterdischarges were of synaptic origin, a finding that is verified by previous studies of such patterns of activity in both cortex and hippocampus (Lee and Hablitz 1991; Schmitz et al. 1997). The pattern of network activity in disinhibited cortical slices from Pet-1 KO mice has been previously reported in disinhibited rat cortex in vivo (Castro-Alamancos 2000) and is an electrographic signature of epileptic seizures in certain childhood-onset epilepsies (Camfield 2011). We will henceforth refer to it as “fast runs,” a term designating this pattern of activity in previous studies (Steriade et al. 1998; Timofeev et al. 1998).

Fig. 3.

Enhanced cortical network excitability in mice lacking Pet-1. Raw traces of disinhibition-induced cortical network activity recorded from layer 2/3 pyramidal cells under control conditions (1 μM gabazine) in slices from WT mice (A; n = 30 cells), in control conditions in Pet-1^−/− mice (B; n = 32), in the presence of the 5-HT₂ receptor antagonist KSN in slices from Pet-1^−/− mice (C; n = 19 cells), and in the presence of the 5-HT₃ receptor antagonist GSN in slices from Pet-1^−/− mice (D; n = 11 cells). Top trace in A–D corresponds to the entire 10-min recording period, and bottom trace inside the rectangle shows a zoomed trace of a single network event. E: box plots showing quantification of network activity as the number of network bursts recorded over 10 min. All groups have significantly different medians, except the WT and Pet-1^−/− with KSN. F: power spectral density plot as measured from the 3 groups mentioned above. Note that the presence of the oscillatory fast runs corresponds to a peak in the 10–15 Hz band, while the baseline of the function denotes baseline activity levels in each group.

In a previous paper, we showed that elevated endogenous 5-HT signaling in WT animals transforms PDS into fast runs via augmented signaling through 5-HT₂ receptors (Puzerey et al. 2014). Given that Pet-1 KO mice express hypersensitive 5-HT₂ receptors (Yadav et al. 2011), we hypothesized that fast runs observed in Pet-1 mice could result from excessive 5-HT₂ receptor signaling and could, therefore, be abolished by blockade of this receptor. Consistent with our hypothesis, bath application of the selective 5-HT₂ receptor antagonist KSN (10 μM) reduced the number of network events (P < 0.01, Wilcoxon rank sum test; Fig. 3, C and E) and abolished fast runs in slices from KO mice. The voltage profile of network activity in KOs treated with KSN transformed to single PDS events similar to those seen in control WT slices (Fig. 3C). The absence of fast runs after KSN treatment was also apparent in the lack of the 10–15 Hz peak in the power density plot (Fig. 3F). Therefore, our results support the idea that hypersensitive 5-HT₂ receptors in Pet-1 mice increase the excitability of cortical networks. We also tested the contribution of 5-HT₃ receptor signaling in the modulation of cortical network activity. We bath-applied GSN (2 μM) to disinhibited cortical slices and observed an increase in the number of network events, also in the form of fast runs (Fig. 3, D and E) with a slightly lower frequency peak in the power spectral density (Fig. 3F). Such an increase could potentially be accounted for by reduced activation of 5-HT₃ receptor-expressing inhibitory interneurons within the cortex, leading to reduced inhibition and higher likelihood of entering into an active state (Jakab and Goldman-Rakic 2000; Lee et al. 2010; Morales and Wang 2002).

Susceptibility to convulsant-induced seizures is unchanged in Pet-1 knockout mice.

Our findings in the partially disinhibited slice show a hyperexcitable cortical phenotype and spontaneous emergence of epileptiform oscillations in Pet-1 KO mice that are absent in WT mice, suggesting that the cortex of the mutant mice is more susceptible to generating epileptiform activity. The fast run oscillations observed in the slice have been previously reported during chemically induced seizures in rats in vivo (Castro-Alamancos 2000) and in spontaneous and electrically induced seizures in cats (Steriade et al. 1998) and are a hallmark of epileptic seizures in certain forms of childhood-onset epilepsy (Camfield 2011). Findings from our previous publication (Puzerey et al. 2014) suggest that 5-HT₂ receptor activity modulates the latency to the onset of acute epileptic seizures induced by the convulsant PTZ, leading us to hypothesize that Pet-1 KO mice may be more susceptible to convulsant-induced epileptic seizures (i.e., may have a lower seizure threshold, hence shorter seizure latencies). To test this hypothesis, we induced epileptic seizures in WT and KO mice with PTZ (80 mg/kg ip) and measured the time to the onset of the first epileptic seizure characterized by GCC-LOR as well as the mean duration of seizures in each animal (Fig. 4A). Surprisingly, no changes were observed in either the seizure latency or seizure duration latency (latency: WT mean ± SD: 313 ± 139 s, KO mean ± SD: 273 ± 186, P = 0.31, Wilcoxon rank sum test; duration: WT mean ± SD: 78 ± 36 s, KO mean ± SD: 129 ± 103, P = 0.15, Wilcoxon rank-sum test; Fig. 4, B and C). These results led us to conclude that despite the sufficiency of 5-HT₂ receptor hypersensitivity in generating epileptiform oscillations in KO mouse slices in vitro, it is not sufficient to alter the susceptibility to epileptic seizures in vivo, and perhaps compensatory molecular mechanisms involving upregulation of other 5-HT receptors (e.g., 5-HT₁ receptors) or structural changes in neuronal morphology, cortical lamination, or bundling neurites (Smit-Rigter et al. 2011) are sufficient to prevent generalization of runaway excitation in local cortical circuits.

Fig. 4.

Seizure susceptibility and propagation of network activity in mice lacking Pet-1. A: experimental paradigm for inducing acute epileptic seizures with the convulsant pentetrazol (PTZ). Mice were administered a vehicle control solution 1 h before PTZ injection. Electric bolts correspond to onset of spontaneous epileptic seizures. B: box plots of latency to the first generalized clonic convulsions with loss of righting reflex (GCC-LOR) seizure in WT (n = 17 mice) and Pet-1^−/− (n = 13 mice) mice. C: box plots of mean seizure duration from the mice in B. D: multielectrode array (MEA) on top of thalamocortical slice. E: local field potential recorded with the MEA in cortical slices from WT (left) and Pet-1^−/− (right) mice. Field potentials in WT and larger and with more pronounced gradients. F: maximal area covered by propagating paroxysmal depolarizing shifts (PDSs) in slices from WT and Pet-1^−/− mice. Each dot corresponds to a single PDS. G: maximal propagation rate for those events. Double asterisks indicate P values <0.01.

Diminished propagation of cortical activity across distant cortical fields in mice lacking Pet-1.

In the previous sections we showed an increase in local cortical excitability in mice lacking Pet-1; however, the finding that these mice do not exhibit an increased propensity to epileptic seizures suggests potential compensatory mechanisms that would prevent local increases in excitability from becoming generalized across the cortex. The reduced dendritic complexity could potentially limit the spread of cortical excitation. We directly tested this hypothesis using high-density MEAs (Fig. 4D). Again, we used partially disinhibited cortical slices to record network activity from WT and Pet-1 KO mice. MEAs record voltage changes occurring near each electrode and enable the quantification of the spatiotemporal patterns of activity (Fig. 4, E–G). We quantified this spread of activity and found that the maximal area of activation (mean ± SD: WT: 0.43 ± 0.06 mm², KO: 0.32 ± 0.11, P ≪ 0.01; Fig. 4F) and its maximal growth rate (mean ± SD: WT: 33 ± 9 mm²/s, KO: 22 ± 12 mm²/s, P ≪ 0.01; Fig. 4G) were dramatically reduced in Pet-1 KO mice. These results lend support to the idea that despite local hyperexcitability, the generalization of paroxysmal activity is limited through reduced spread of excitation.

DISCUSSION

The use of transgenic animals has been an invaluable tool for understanding the function of specific genes; however, the lack of a complete phenotypic profile of a transgenic animal may conceal physiologically relevant biological compensations. In this study, we demonstrate that mice with a constitutive deletion of the ETS domain transcription factor Pet-1, which normally confers the serotonergic phenotype to midbrain raphe neurons, exhibit significant changes in the cell-intrinsic and network excitability of the cortex. First, our results point to altered intrinsic neuronal excitability in pyramidal neurons of layer 2/3 in the somatosensory cortex. We observed that some parameters of intrinsic excitability were biased toward a hyperexcitable phenotype while others showed a change toward decreased excitability. The direction of these changes could be predicted based on whether they related to passive or active membrane properties. More specifically, measures of passive membrane properties, the membrane τ and the R_input, in Pet-1^−/− mice are increased, thus leading to enhanced responsiveness to transient synaptic inputs. On the other hand, measures of active membrane excitability like MFF, input dynamic range, and spike latency to ramp current injection all exhibited a significant tendency toward decreased excitability. These findings show that systemic depletion of forebrain 5-HT levels produces a complex phenotype in cell-intrinsic excitability of cortical PCs. We also demonstrate that these changes are accompanied by an alteration in synaptic excitability characterized by an increased rate and amplitude of sEPSCs. The normalization of the sEPSC amplitudes to control levels with the 5-HT₃ receptor antagonist GSN suggests that Pet-1^−/− mice exhibit increased synaptic signaling through these receptors on cortical pyramidal neurons. Furthermore, we find that 5-HT₂ receptors provide a substantial contribution to spontaneous synaptic activity in mutant mice, consistent with their hypersensitive phenotype in these mice (Yadav et al. 2011). We also report dramatic differences in cortical network excitability characterized by a 5-HT₂ receptor-dependent increase in local network activity. This result is in agreement with our previous study on the effects of enhanced serotonergic signaling on cortical network activity of WT mice (Puzerey et al. 2014). Surprisingly, the increased network excitability recorded from single cells in cortical slices did not correlate with changes in susceptibility to convulsant-induced seizures in vivo. A possible explanation for this discrepancy is offered by the changes in cell morphology. Indeed, the significant reduction in complexity of basal dendrite arborizations of the KO animals suggests a potential compensatory mechanism through which the spread of excitation becomes limited because of decreased connectivity among neighboring neurons. Consistent with this idea, we find that the propagation of cortical network activity across distant cortical fields is diminished in mice lacking Pet-1, providing a potential mechanism for the lack of generalization of runaway excitation in cortical circuits. Combined together, our results show significant changes in cell-intrinsic, synaptic, and network excitability in the cortex of mice lacking the transcription factor Pet-1 and reveal a complex set of biological adaptations in response to systemic 5-HT depletion.

The specification, maturation, and maintenance of the central 5-HT neuronal phenotype is a complex process characterized by distinct developmental phases that proceed under guidance of specific transcriptional programs. We refer the reader to a comprehensive review on the development of the 5-HT system and its relevance to neuropsychiatric disorders (Deneris and Wyler 2012) and only focus here on the known functions of the Pet-1 gene. The importance of this gene for various physiological functions is underscored by several studies pointing to dysregulation of circadian rhythms (Ciarleglio et al. 2014; Paulus and Mintz 2012), respiration and thermoregulation (Cummings et al. 2010, 2011; Erickson et al. 2007; Hodges et al. 2011), maternal behavior (Lerch-Haner et al. 2008), adult neurogenesis (Diaz et al. 2013), and neonatal growth (Narboux-Neme et al. 2013) following its constitutive genetic deletion. Such a broad range of physiological defects resulting from systemic 5-HT depletion highlights the importance of the serotonergic system in controlling developmental and autonomic processes. Despite the breadth of systems affected by 5-HT depletion in the studies above, no investigation has yet aimed to understand the consequences of Pet-1 deletion on the function of the cortex, a forebrain structure densely innervated by serotonergic afferents. Our study provides a compelling piece of evidence for the importance of the 5-HT system, specified and maintained by the Pet-1 gene, for the normal physiology of the cortical neurons, synapses, and networks as well as proper development of dendritic morphology.

To understand how global alterations in cortical 5-HT levels may affect cell-intrinsic, synaptic, and network function in Pet-1^−/− mice, we speculate on some potential pathways by which these changes may be actualized through the known functions of 5-HT receptors. We address first the changes in cell-intrinsic parameters of neuronal excitability. Our findings show that membrane τ and R_input are increased in Pet-1^−/− mice. These two parameters of passive membrane excitability are directly related to the membrane capacitance (via τ = R_inputC_m), which is typically determined by the size of the neuronal membrane. Given that R_input is significantly higher in Pet-1^−/− mice, the increase in τ is likely to be accounted for by changes in this parameter. How this change comes about in PCs from Pet-1^−/− mice remains to be elucidated. A likely explanation is that loss of some 5-HT receptor-modulated conductance would effectively increase R_input. Although 5-HT₁ and 5-HT₂ receptors are the predominant 5-HT receptors expressed on cortical neurons, their contribution to this effect is unlikely considering previous reports of increased signaling through these receptors in Pet-1^−/− animals (Yadav et al. 2011).

The finding that Pet-1^−/− mice exhibit an increase in the amplitude of 5-HT₃ receptor-mediated sEPSCs was unexpected given that these animals have a severe reduction (∼80%) in forebrain 5-HT levels. However, Pet-1^−/− mice also show decreased levels in the expression of the 5-HT transporter SERT, which mediates reuptake of 5-HT into presynaptic axon terminals, and a parallel decrease in the expression of 5-HT₁ receptors on raphe neurons, which provide an autoinhibitory signal to regulate 5-HT release (Hendricks et al. 2003; Liu et al. 2010). The overall effect of diminished expression of these proteins would result in increased synaptic concentrations of the remaining 5-HT that persists despite Pet-1^−/− deletion, which could account for the increased amplitude of 5-HT₃ receptor-dependent sEPSCs onto cortical neurons. An alternative explanation for the increased sEPSC amplitude would be elevated expression of postsynaptic 5-HT₃ receptors on cortical neurons. Support for this idea remains to be found, although transgenic mice lacking the serotonin transporter SERT have been shown to exhibit adaptive changes in 5-HT₃ receptor expression, suggesting that similar adaptations could be possible in Pet-1^−/− mice (Mossner et al. 2004). We note also that 5-HT₃ receptor expression on cortical pyramidal neurons has been previously contended given the prevailing notion that 5-HT₃ receptors are expressed exclusively on cortical inhibitory interneurons. We refer the reader to our previous publication in which we address the controversy regarding 5-HT₃ receptor expression on cortical PCs (Puzerey et al. 2014).

Another surprising finding of this study was the difference in cortical network dynamics in the disinhibited slice between WT and Pet-1^−/− mice. While the disinhibition-induced network activity in WTs manifested as individual PDS, the 5-HT-deficient mice exhibited fast run epileptiform oscillations that could be abolished via blockade of 5-HT₂ receptors. This is unexpected since our intuition would lead us to think that an 80% drop in forebrain 5-HT levels would result in decreased signaling through 5-HT₂ receptors. However, a possible explanation for this finding comes from a previous study that utilized Pet-1^−/− transgenic mice to study mechanisms of clozapine action. Yadav and colleagues revealed a significant increase in intrinsic signaling of 5-HT₂ receptors in mice lacking Pet-1 (Yadav et al. 2011). Such changes in sensitivity of 5-HT₂ receptors have been reported in other models of 5-HT depletion (Narboux-Neme et al. 2013). Combining this hypersensitivity of 5-HT₂ receptors with the aforementioned decreases in the expression of SERT and 5-HT₁ receptors in raphe neurons, both of which would act to increase synaptic concentrations of the remaining 5-HT (Hendricks et al. 2003; Liu et al. 2010), we hypothesize that the emergence of fast runs in local cortical circuits of Pet-1^−/− mice results from these compensatory increases in postsynaptic 5-HT signaling. This hypothesis is consistent with our previous study in which elevation of endogenous cortical 5-HT levels with the serotonin reuptake inhibitor fluoxetine led to the emergence of fast runs in a 5-HT₂ receptor-dependent manner in cortical slices from WT mice (Puzerey et al. 2014).

Our study employed thalamocortical slices for the sake of preserving corticofugal feedback connections as well as bottom-up thalamocortical inputs. This requires consideration of the potential effects that serotonergic signaling has on thalamic neurons. For instance, 5-HT is known to suppress tonic firing in thalamic neurons through 5-HT₁ receptor-dependent inhibition of primary relay neurons and direct excitation of local inhibitory interneurons (Monckton and McCormick 2002). A separate study found that the hyperpolarizing responses were seen in a minority of higher-order thalamic nuclei (cortico-recipient) and that a large majority of primary and secondary thalamic nuclei were depolarized by 5-HT (Varela and Sherman 2009). These findings suggest that serotonergic modulation of thalamic nuclei could in principle contribute to cortical network activity. However, both patterns of paroxysmal synaptic activity seen in our slice experiments have been shown to be generated and maintained in the absence of thalamic input (Castro-Alamancos and Rigas 2002; McCormick and Contreras 2001; Steriade and Contreras 1998), suggesting that thalamic mechanisms are unlikely to contribute to the generation of these patterns of activity. This line of thought is consistent with the existing hypothesis that paroxysmal activity in cortical networks emerges within cortical layer 5 (McCormick and Contreras 2001).

Our study focused particularly on the activity of the principal cortical neuron—the PC. While PC excitability is subject to neuromodulatory regulation by 5-HT, inhibitory interneurons within the cortex also express 5-HT receptors and receive serotonergic innervation. Thus we briefly speculate on the potential effects of 5-HT depletion on the activity of cortical interneurons in Pet-1^−/− mice. Under normal conditions, 5-HT can modulate GABAergic synaptic transmission. For instance, activation of 5-HT₂ and 5-HT₃ receptors leads to a sustained and a transient increase in spontaneous inhibitory postsynaptic currents (sIPSCs) onto PCs within the cortex, respectively (Zhou and Hablitz 1999). The transient increase in sIPSCs is likely to be accounted for by a massive class of cortical inhibitory interneurons expressing 5-HT₃ receptors (Jakab and Goldman-Rakic 2000; Lee et al. 2010; Morales and Wang 2002). Diminished serotonergic signaling through this class of neurons could lead to dampened cortical inhibition, potentially resulting in the increased local network excitability we see in our experiments.

Despite the increased cortical network excitability in Pet-1^−/− mice, we observed no significant differences in seizure susceptibility compared with their WT counterparts. This is surprising considering our previous findings, which showed that elevated 5-HT₂ receptor signaling has a proepileptic role in cortical networks (Puzerey et al. 2014). One possible explanation may come from our findings on reduced dendritic complexity in cortical neurons. Local network excitability is increased in Pet-1^−/− mice as shown in Fig. 3 but may be unable to propagate to distant cortical sites, presumably from decreased synaptic connectivity secondary to altered dendritic structure. This would allow epileptiform activity to be expressed in local cortical networks but limit its spread across larger areas of the cortex. To test this hypothesis, we simultaneously measured disinhibition-induced network activity across distant cortical fields in a Pet-1^−/− brain slice to determine whether long-distance propagation of neural activity is different in these mice. Using 2D MEAs that can sample various cortical columns across distant sites, we found that propagation of cortical activity was substantially attenuated in mice lacking Pet-1. To put these findings in context, we must integrate these findings with the previous literature investigating the role of 5-HT in the generation of epileptiform activity cortical activity and seizures. To this day, there is no solid consensus on the relationship between 5-HT and paroxysmal activity. Several lines of evidence point to a proconvulsant role of 5-HT within the CNS (Bercovici et al. 2006; Freitas et al. 2006; O'Dell et al. 2000; Wada et al. 1992), while others suggest that 5-HT may act to suppress the spread of paroxysmal excitation (Bagdy et al. 2007; Buchanan et al. 2014; Supornsilpchai et al. 2006; Trindade-Filho et al. 2008; Yan et al. 1995). Such conflicting reports are likely to arise out of differences in experimental seizure paradigms, animal models utilized in the study, as well as pharmacological agents used to target the 5-HT system (Loscher et al. 1990). For instance, Buchanan et al. showed that mice lacking 99% of their 5-HT neurons through the deletion of the transcription factor Lmx1b (Buchanan et al. 2014; Zhao et al. 2006) actually exhibit higher seizure thresholds and reduced seizure-related mortality, a finding supported by a similar effect with acute depletion of 5-HT in WT mice. This study lends support for an anticonvulsive role of 5-HT; however, a direct comparison between these findings and our findings in 5-HT-deficient Pet-1 KO mice is complicated because of differences in seizure paradigms, measurements and categorization of seizure onset, and the degree of 5-HT depletion in each model. Differences in 5-HT depletion between the two models are likely to have a significant effect on cortical excitability, since the higher residual levels of 5-HT in the Pet-1 KO mice could act on highly sensitive 5-HT receptors (Yadav et al. 2011) and thereby affect cortical excitability. Although both animal models offer interesting insight into the effects of central 5-HT depletion, one must be careful in comparing them side by side since the full ranges of biological compensations in both mouse models are not known and may have an influence over factors that determine seizure susceptibility. For instance, the Tph2 KO model of 5-HT depletion spares normal 5-HT₁ autoreceptor function (Araragi et al. 2013), while signaling through the same receptors in mice lacking Pet-1 is severely disrupted (Liu et al. 2010), suggesting that 5-HT depletion alone is not predictive of the direction of biological compensations in mice with altered 5-HT systems.

The study of transgenic animal models for the purpose of understanding the function of a biological system is undoubtedly useful but may be confounded by secondary biological compensations that are often unexpected. These compensations may arise as a result of inherent correlations in the expression of certain genes that result from their embedding into highly interconnected transcriptional or protein networks (Marder 2011) or from intrinsic protein properties. Such compensations are certainly familiar in the study of the 5-HT system since several transgenic mouse lines in which distinct components of this system are genetically altered exhibit correlated compensations in expression or sensitivity of related proteins (Araragi et al. 2013; Fabre et al. 2000; Goodfellow et al. 2012; Hendricks et al. 2003; Liu et al. 2010; Mossner et al. 2004; Moya et al. 2011; Narboux-Neme et al. 2011; Veenstra-VanderWeele et al. 2012; Yadav et al. 2011). Our study builds upon this idea and shows new evidence for a functional increase in signaling through 5-HT₂ and 5-HT₃ receptors in Pet-1^−/− mice that corresponds to altered cortical network activity and intrinsic cell excitability.

GRANTS

This work has been supported by The Ruth L. Kirschstein National Research Award NS067431 (P. A. Puzerey), The Mt. Sinai Health Care Foundation, The Alfred P. Sloan Foundation, and The Hartwell Foundation (R. F. Galán).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: P.A.P., N.X.K., and R.F.G. conception and design of research; P.A.P., N.X.K., and R.F.G. performed experiments; P.A.P., N.X.K., and R.F.G. analyzed data; P.A.P., N.X.K., and R.F.G. interpreted results of experiments; P.A.P., N.X.K., and R.F.G. prepared figures; P.A.P., N.X.K., and R.F.G. drafted manuscript; P.A.P. and R.F.G. edited and revised manuscript; P.A.P., N.X.K., and R.F.G. approved final version of manuscript.