Skip Navigation


Cerebral Cortex Advance Access originally published online on July 6, 2004
Cerebral Cortex 2005 15(1):1-14; doi:10.1093/cercor/bhh104
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
15/1/1    most recent
bhh104v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (32)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Puig, M. V.
Right arrow Articles by Celada, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Puig, M. V.
Right arrow Articles by Celada, P.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Cerebral Cortex V 15 N 1 © Oxford University Press 2005; all rights reserved

Feature Article

Modulation of the Activity of Pyramidal Neurons in Rat Prefrontal Cortex by Raphe Stimulation In Vivo: Involvement of Serotonin and GABA

M. Victoria Puig, Francesc Artigas and Pau Celada

Department of Neurochemistry, Institut d' Investigacions Biomèdiques de Barcelona (CSIC), IDIBAPS, 08036 Barcelona, Spain

Address correspondence to Francesc Artigas, PhD; Department of Neurochemistry, Institut d' Investigacions Biomèdiques de Barcelona (CSIC), IDIBAPS, Rosselló, 161, 6th floor, 08036 Barcelona, Spain. Phone: +3493–363 8315; Fax: +3493–363 8301; e-mail: fapnqi{at}iibb.csic.es.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 Conclusions
 References
 
Serotonin is involved in psychiatric disorders exhibiting abnormal prefrontal cortex (PFC) function (e.g. major depression, schizophrenia). We examined the effect of the stimulation of the dorsal and median raphe nuclei (DR and MnR, respectively) on the activity of PFC neurons. Electrical stimulation of DR/MnR inhibited 66% (115/173) of pyramidal neurons in the medial PFC (mPFC). The rest of the cases exhibited orthodromic excitations, either pure (13%) or preceded by short-latency inhibitions (20%). Excited neurons had a lower pre-stimulus firing rate than those inhibited. Excitations evoked by MnR stimulation had a shorter latency than those evoked by DR stimulation. WAY-100635 [a 5-hydroxytryptamine1A (5-HT1A) antagonist] and the selective gamma aminobutyric acidA (GABAA) antagonist picrotoxinin partially antagonized DR/MnR-evoked inhibitions, suggesting the involvement of 5-HT1A- and GABAA-mediated components. The presence of a direct DR/MnR-mPFC GABAergic component is suggested by the short latency of evoked inhibitions (9 ± 1 ms), faster than those evoked in the secondary motor area (20 ± 3 ms), and that of antidromic spikes evoked by DR/MnR stimulation in mPFC pyramidal neurons (15 ± 1 ms). Stimulation of the DR/MnR with paired pulses enhanced the duration of inhibitions and turned some excitations into inhibitions. Thus, the DR/MnR control the activity of mPFC pyramidal neurons in vivo in a complex manner, involving 5-HT-mediated excitations and GABA- and 5-HT-mediated inhibitions.

Key Words: 5-HT1A receptors • 5-HT2A receptors • dorsal raphe • medial prefrontal cortex • median raphe


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 Conclusions
 References
 
The ascending monoaminergic systems of the brainstem innervate the prefrontal cortex (PFC) in the mammalian brain and play an important role in the control of higher brain functions (Groenewegen and Uylings, 2000Go; Fuster, 2001Go). In particular, dopamine is critically involved in working memory and cognition through a complex control of the activity of pyramidal neurons (Glowinski et al., 1984Go; Williams and Goldman-Rakic, 1995Go; Goldman-Rakic, 1996Go; Yang and Seamans, 1996Go; Robbins, 2000Go; Tzschentke, 2001Go; O'Donnell, 2003Go; Wang et al., 2003Go). There is also growing evidence that the serotonergic pathways originating in the dorsal and median raphe nuclei (DR and MnR, respectively) may play an important role in prefrontal function. Thus, the PFC of the rodent, primate and human brains contains several 5-hydroxytryptamine (5-HT) receptors, with a particular abundance of the 5-HT1A and 5-HT2A subtypes (Pompeiano et al., 1992Go, 1994Go; Morales and Bloom, 1997Go; Hall et al., 2000Go; Talvik-Lotfi et al., 2000Go; Martinez et al., 2001Go; Arango et al., 2002Go). 5-HT2A receptors in dorsolateral PFC are involved in working memory (Williams et al., 2002Go), and recent work associates allelic variants of this receptor with memory capacity in humans (De Quervain et al., 2003Go). Hallucinogens such as LSD or 1-[2,5-dimethoxy-4-iodophenyl-2-aminopropane] (DOI) are 5-HT2A receptor agonists whereas atypical antipsychotics like clozapine are 5-HT2A receptor antagonists (Kroeze and Roth, 1998Go; Meltzer, 1999Go). On the other hand, 5-HT1A agonists display anxiolytic/antidepressant activity in animal models (De Vry, 1995Go), whereas 5-HT1A receptor antagonists reverse drug-induced cognitive deficits (Harder and Ridley, 2000Go; Mello e Souza et al., 2001Go; Misane and Ogren, 2003Go).

5-HT and selective receptor agonists modulate the excitability of cortical neurons and their discharge rate through the activation of several receptor subtypes: namely 5-HT1A, 5-HT1B, 5-HT2 and 5-HT3 (Ashby et al., 1989Go; Araneda and Andrade, 1991Go; McCormick et al., 1993Go; Tanaka and North, 1993Go; Aghajanian and Marek, 1997Go; Arvanov et al., 1999Go; Zhou and Hablitz, 1999Go; Férézou et al., 2002Go; Puig et al., 2003Go). In vitro and in vivo studies suggest that 5-HT1A and 5-HT2A receptors are key players that exert opposite effects on the excitability and firing activity of pyramidal neurons in the medial PFC (mPFC) (Araneda and Andrade, 1991Go; Ashby et al., 1994Go; Aghajanian and Marek, 1997Go; Puig et al., 2003Go; Amargós-Bosch et al., 2004Go). The activation of 5-HT1A receptors in PFC hyperpolarizes pyramidal neurons, whereas that of 5-HT2A receptors results in neuronal depolarization, reduction of the afterhyperpolarization, and increase of excitatory postsynaptic currents (EPSCs) and of discharge rate (Araneda and Andrade, 1991Go; Tanaka and North, 1993Go; Aghajanian and Marek, 1997Go, 1999Go; Newberry et al., 1999Go; Zhou and Hablitz, 1999Go; Puig et al., 2003Go; Amargós-Bosch et al., 2004Go). 5-HT can also activate excitatory receptors (5-HT2A and 5-HT3) in gamma aminobutyric acid (GABA) interneurons (Morales and Bloom, 1997Go; Jakab and Goldman-Rakic, 2000Go) to increase a synaptic GABA input onto pyramidal neurons (Tanaka and North, 1993Go; Zhou and Hablitz, 1999Go; Férézou et al., 2002Go).

However, despite the wealth of in vitro studies on the actions of 5-HT on cortical neurons, there is little information on the relative balance of inhibitory and excitatory responses elicited by endogenous 5-HT in vivo. Nearly 60% of the neurons in the PFC of the rat and mouse express the mRNAs of 5-HT1A and/or 5-HT2A receptors, with a high degree of co-expression (nearly 80% in most PFC areas; Amargós-Bosch et al., 2004Go). The vast majority of these mRNAs co-localized with vGluT1 mRNA, suggesting a major location in pyramidal neurons (Santana et al., 2004Go). Consistent with these data, the electrical stimulation of the DR can inhibit (via 5-HT1A receptors) or excite (via 5-HT2A receptors) the pyramidal neurons in the mPFC (Puig et al., 2003Go; Amargós-Bosch et al., 2004Go), although the reasons determining the nature of the response (i.e. inhibitory or excitatory) are not fully understood. Here we examined the responses elicited by the physiological stimulation of the DR and MnR in pyramidal neurons of the cingulate and prelimbic areas of the mPFC which, in turn, project to the raphe nuclei, and compared these responses with those elicited in the secondary motor area (MOs) in the vicinity of the mPFC.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 Conclusions
 References
 
Animals

A total of 74 male albino Wistar rats weighing 250–320 g at the time of experiments were used (Iffa Credo, Lyon, France). They were kept in a controlled environment (12 h light:12 h dark cycle and 22 ± 2°C room temperature), with food and water provided ad libitum. Animal care followed the European Union regulations (O.J. of E.C. L358/1 18/12/1986) and experimental procedures were approved by a local Institutional Animal Care and Use Committee. Stereotaxic coordinates were taken from bregma and duramater according to the atlas of Paxinos and Watson (1998)Go. Additionally, we used the brain maps (CD-edition; Swanson, 1998Go) for nomenclature of the cortical areas.

Single Unit Recordings

We examined the responses elicited in pyramidal neurons of the mPFC by the electrical stimulation of the DR and/or MnR in anesthetized rats. Rats were anesthetized (chloral hydrate 400 mg/kg i.p.) and positioned in a David Kopf stereotaxic frame. Additional doses of chloral hydrate (80 mg/kg) were administered i.v. through the femoral vein. Typically, recordings were made between 10 and ~45 min after additional doses of anesthetic to avoid the effects of peak concentrations of chloral hydrate during recordings. Body temperature was maintained at 37°C throughout the experiment with a heating pad. All wound margins and points of contact between the animal and the stereotaxic apparatus were infiltrated with lidocaine solution (5%). In order to minimize pulsation, the atlanto-occipital membrane was punctured to release some CSF.

Bipolar stimulating electrodes consisted of two stainless steel enamel-coated wires (California Fine Wire, Grover Beach, CA) with a diameter of 150 µm and a tip separation of ~100 µm and in vitro impedances of 10–30 K{Omega}. Stimulating electrodes were stereotaxically implanted in either of these coordinates, within the DR (AP –7.8, L 0, DV –6.5; and AP –7.3, L –2.2 with a lateral angle of 20°, DV –6.6 mm) or the MnR (AP –7.8, L 2.0 with a lateral angle of 13°, DV –8.8 mm). These angles resulted in the tip of the electrodes at DV –6.2 and –8.6 mm, respectively in the vicinity of the midline. In most experiments, two electrodes were implanted, one in DR (either location) and another one in MnR. After each implant, the electrodes were secured to the skull with glue and dental cement. Constant current electrical stimuli were generated with a Grass stimulation unit S-48 connected to a Grass SIU 5 stimulus isolation unit. Stimulating current was typically between 0.1 and 2 mA, 0.2 ms square pulses at 0.9 Hz. In some experiments, we recorded the same pyramidal neuron in mPFC after the sequential stimulation of the DR/MnR with single and twin pulses while keeping current intensity (0.5–1.7 mA) and frequency (0.9 Hz). Twin pulses were delivered 7 ms apart. Twin pulse stimulation of the DR has been shown to increase the cortical 5-HT release compared with single pulse stimulation (Gartside et al., 2000Go).

Pyramidal neurons were recorded extracellularly with glass micropipettes pulled from 2.0 mm capillary glass (WPI, Sarasota, FL) on a Narishige PE-2 pipette puller (Narishige Sci. Inst., Tokyo, Japan). Microelectrodes were filled with 2 M NaCl. Typically, in vitro impedance was between 4 and 10 M{Omega}. Single unit extracellular recordings were amplified with a Neurodata IR283 (Cygnus Technology Inc., Delaware Water Gap, PA), postamplified and filtered with a Cibertec amplifier (Madrid, Spain) and computed on-line using a DAT 1401plus interface system Spike2 software (Cambridge Electronic Design, Cambridge, UK). Descents in mPFC were carried out at AP +3.2–3.4, L –0.5 to –1.0, DV –1.0 to –4.0 below the brain surface. We systematically confirmed that only a single pyramidal neuron was recorded by (i) identification by antidromic activation from DR and/or MnR and (ii) collision extinction with spontaneously occurring spikes (Fuller and Schlag, 1976Go). Neurons without antidromic activation or without spontaneous firing activity were not considered. Additionally, recordings were made in neurons of the secondary motor area (MOs; Swanson, 1998Go), at AP +3.2–3.4, L –2.0–2.6, DV between 0.8 and 1.4 mm (see Fig. 1 for localization of the recording areas in PFC). After recording the effects of DR/MnR stimulation on pyramidal activity (see below), we administered the 5-HT1A receptor antagonist WAY-100635 or the GABAA receptor antagonist picrotoxinin (both from Sigma/RBI) and further post-drug recordings were made to evaluate the actions of these drugs on DR/MnR-evoked inhibitory responses. WAY-100635 and picrotoxinin were dissolved in saline at the appropriate concentrations and injected (0.5–1 ml/kg) through the femoral vein. Finally, to determine the latency of antidromic spikes traveling along serotonergic axons projecting to mPFC we recorded serotonergic neurons in the DR during electrical stimulation of the mPFC. The methods are described in full in Celada et al. (2001)Go.



View larger version (31K):
[in this window]
[in a new window]
  		Figure 1. Schematic localization of the recording area in the dorsal anterior cingulate area (ACAd) and prelimbic area (PL) of the medial prefrontal cortex. Shown is also the localization of recordings in the secondary motor area (MOs). In both areas recordings were made mainly in layer V neurons. Section taken from Brain Maps, Swanson (1998)Go.

 
At the end of the experiments, rats were killed by an overdose of anesthetic. The placement of the stimulating electrodes was verified histologically. Rats were transcardially perfused with saline followed by 10% formalin solution (Sigma). Brains were post-fixed, sagitally sectioned (80 µm) and stained with Neutral Red. The data from rats with electrodes implanted outside the DR or MnR were not used.

Data and Statistical Analysis

The responses in prefrontal pyramidal neurons evoked by DR and MnR stimulation were characterized by measuring the magnitude and duration of inhibitory and excitatory responses from peristimulus-time histograms (PSTH) (4 ms width). For a better precision of the onset of inhibitory responses, latencies were calculated with a bin width of 1 ms. Orthodromic excitations elicited spikes with short and variable latencies and a post-stimulus firing rate superior to the mean pre-stimulus firing rate plus two times the standard deviation during at least four bins (Hajós et al., 1998Go). Antidromic spikes had a fixed latency and were produced by the electrical stimulation of axons of mPFC pyramidal neurons projecting to the DR and MnR (Celada et al., 2001Go). Inhibitions were defined by a total cessation of spikes with respect to the pre-stimulus value for at least four successive bins (Hajós et al., 1998Go). The onset of the inhibition was defined as the last bin containing a spike whereas the end of the inhibition was defined as the first of four bins equal to or above the pre-stimulus value. The magnitude of the inhibition was calculated as percentage of firing versus the pre-stimulus (200 ms) firing rate. Drug effects were calculated by comparing 2-min PSTHs at basal and post-drug periods. Data are expressed as the mean ± SEM. Statistical analysis was carried out using independent and paired Student's t-tests. Statistical significance has been set at the 95% confidence level (two tailed).


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 Conclusions
 References
 
Responses in mPFC Pyramidal Neurons Elicited by DR/MnR Stimulation

We performed 173 experiments in which we examined the effect of the stimulation of the DR or MnR on pyramidal neurons of the cingulate and prelimbic areas of the PFC. In 115/173 cases (66%) pure inhibitory responses were recorded, while in 23/173 cases (13%), pure orthodromic excitations were observed. The rest of responses (35/173, 20%) were biphasic, with an orthodromic excitation preceded by an inhibition of short latency and duration. Figure 2 shows representative examples of the three different types of responses obtained after stimulation of the DR and MnR.



View larger version (32K):
[in this window]
[in a new window]
  		Figure 2. Representative examples of the responses evoked in pyramidal neurons of the anterior cingulate and prelimbic areas of the mPFC by the electrical stimulation of the DR and MnR (see Materials and Methods). (A, B) Short latency long duration inhibitory responses. (C, D) Pure excitatory responses, which are often seen in pyramidal neurons with a low firing rate. These orthodromic excitations have a similar duration after DR or MnR stimulation but the latency is significantly lower after MnR stimulation, as in the example seen in the figure (see also Table 1). (E, F) Orthodromic excitations preceded by short latency inhibitions (biphasic responses). Note the presence of antidromic spikes in the units recorded (B, C, E, F). The units in (A) and (D) were antidromically activated at currents higher than those used to evoke a pyramidal response. Each PSTH consists of 110 triggers (2 min). Bin size = 4 ms. The arrow in the abcissa denotes the stimulus artifact (time = 0). The arrowheads indicate antidromic spikes.

 

View this table:
[in this window]
[in a new window]
  		Table 1 Characteristics of the pure inhibitory and excitatory responses evoked in mPFC pyramidal neurons by raphe stimulation

 
Inhibitions had a latency of 9 ± 1 ms and a duration of 161 ± 9 ms (n = 115). Firing rate during the inhibition was reduced on average to 13% of pre-stimulus firing rate. Table 1 shows the characteristics of these inhibitions, classified according to the stimulation site (DR or MnR). There were no significant differences in the characteristics of the inhibitions between those elicited by DR stimulation (n = 70) or MnR stimulation (n = 45). In 26 cases we examined the response of the same pyramidal neuron to the stimulation of the DR and MnR in rats implanted with two stimulating electrodes. When a response was triggered in a pyramidal neuron by DR or MnR stimulation, it was recorded for 2–4 min and then the stimulation current was switched to the stimulation electrode implanted in the other nucleus and the response of the unit was recorded for a similar period of time. The statistical analysis (paired t-test) of this subpopulation of neurons showed no significant differences between the effects of the stimulation of both raphe nuclei.

Some of these neurons (n = 15) were antidromically activated from both the DR and the MnR at the currents used, showing that some pyramidal neurons can simultaneously control the activity of both serotonergic nuclei. In this subgroup, the latency of the antidromic spikes was 16 ± 1 ms and 14 ± 1 ms from the DR and MnR, respectively (non-significant difference).

Pyramidal neurons excited by DR/MnR stimulation (n = 23) were located at the same DV coordinates than those inhibited, i.e. near 2.5 mm below brain surface. However, the units inhibited by the DR/MnR stimulation had a higher pre-stimulus firing rate than those excited (2.3 versus 0.9 spikes/s; n = 115 and 23, respectively; P < 0.0006; Table 1). The duration, pre- and post-stimulus firing rates and success rate of the orthodromic excitations did not differ between stimulation sites (DR versus MnR). However, unlike the inhibitions, the latency of the excitations was significantly lower when stimulating the MnR (45 ± 5 versus 75 ± 11 ms, n = 10 and 13, respectively; P < 0.05, Student's t-test) (Table 1).

A subgroup of pyramidal neurons exhibited biphasic responses to DR/MnR stimulation. In 35 experiments, the orthodromic excitations were preceded by short latency inhibitions. Table 2 shows the characteristics of these responses. There were no significant differences between the responses elicited by DR or MnR stimulation, with the exception of the duration of the excitations, which was slightly but significantly greater when the DR was stimulated (96 ± 6 ms for the DR versus 71 ± 7 ms for the MnR; n = 22 and 13, respectively; P < 0.02; Student's t-test). The success rate was also similar for the DR- or MnR-induced excitations (58 ± 8 versus 55 ± 8%, respectively).


View this table:
[in this window]
[in a new window]
  		Table 2 Characteristics of the biphasic responses evoked in mPFC pyramidal neurons by raphe stimulation

 
A comparison of the characteristics between the pure and biphasic inhibitory responses revealed a similar latency but significantly lower duration of the biphasic responses (79 ± 6 versus 166 ± 10 ms for the DR stimulation; 67 ± 10 versus 153 ± 16 ms for the MnR stimulation; P < 0.01 in both cases). Biphasic excitations had a similar duration but slightly longer latency (90 ± 7 versus 75 ± 11 ms for the DR, n.s.; 88 ± 11 versus 45 ± 5 ms for the MnR, P < 0.004), i.e. as if the preceding inhibition had delayed the excitation in neurons showing a biphasic response (Table 3).


View this table:
[in this window]
[in a new window]
  		Table 3 Comparison of pure and biphasic responses in mPFC neurons

 
As seen in Table 1, the characteristics of the inhibitory and excitatory responses did not differ between stimulation sites, with the exception of the latency of the orthodromic excitations, lower for MnR stimulation. When the type of response elicited (pure excitations versus inhibitions) were compared depending on the stimulation site, a Chi-square analysis revealed no overall differences between the stimulation of the DR at AP –7.8 mm and that at the MnR (also at an AP coordinate of –7.8 mm). The corresponding ratios between inhibitions and pure orthodromic excitations were 70/13 for the DR and 45/10 for the MnR (P = 0.86). However, the stimulations performed in the DR at a more rostral coordinate (AP –7.3 mm) yielded a significantly greater proportion of excitations than those at AP –7.8, both in DR and MnR (4/8 versus 70/13, P < 0.001; 4/8 versus 45/10, P < 0.002).

Pharmacological Characterization of DR/MnR-elicited Responses in mPFC

The pyramidal excitations induced by the electrical stimulation of the DR/MnR were reversed by the treatment with the selective 5-HT2A receptor antagonist M100907 (Puig et al., 2003Go; Amargós-Bosch et al., 2004Go). Likewise, inhibitions were partly blocked by WAY-100635 administration (10–60 µg/kg i.v.) (Amargós-Bosch et al., 2004Go) (Fig. 3A). An early component of the inhibitions (up to 69 ± 32 ms; Amargós-Bosch et al., 2004Go) could not be blocked by WAY-100635. The failure to block this earlier component cannot be ascribed to an insufficient dose since higher doses of WAY-100635 (e.g. 100–200 µg/kg i.v.) were even less effective, perhaps as a result of some partial agonist activity of this agent at these doses (Martin et al., 1999Go). We therefore reasoned that the WAY-100635-insensitive, low latency inhibitory responses might be due to an increased GABAergic input onto pyramidal neurons, resulting from various sources, such as the activation of 5-HT receptors in local interneurons or the activation of direct GABAergic inputs from the DR (see Discussion). Likewise, since pyramidal neurons in mPFC project to the raphe nuclei, a GABAergic component might also result from stimulus-evoked antidromic spikes in collaterals of pyramidal axons impinging on local GABAergic interneurons. We preliminarily examined the presence of these possible GABA inputs by various means.



View larger version (31K):
[in this window]
[in a new window]
  		Figure 3. Antagonism of the DR/MnR-evoked inhibitory responses in mPFC by the 5-HT1A receptor antagonist WAY-100635 and the GABAA receptor antagonist picrotoxinin (PIC). Upper panels correspond to PSTHs obtained in basal conditions and lower panels, to those after drug treatment. The response in A was obtained by DR stimulation at 1 mA and shows an antidromic spike at 10 ms. This inhibitory response had a latency of 9 ms. The firing rate, expressed as percentage of pre-stimulus value, was 19% in basal conditions and 215% after WAY-100635 administration. The inhibitory response of this unit was particularly sensitive to WAY-100635 administration. On average, WAY-100635 administration did not reverse the inhibitory response between 0 and 69 ms (n = 8; Amargós-Bosch et al., 2004Go). The response in B was obtained by DR stimulation at 2 mA and shows an antidromic spike at 14 ms. This inhibitory response had a latency of 13 ms. The firing rates, expressed as percentage of pre-stimulus value were 27 and 153% in basal conditions and after picrotoxinin, respectively. Note the overall increase in firing rate after picrotoxinin administration. Each PSTH consists of 110 triggers. Bin size = 4 ms.

 
The presence of a GABAergic component is supported by the antagonism of the DR/MnR-induced inhibitions by the GABAA receptor antagonist picrotoxinin (1–2.5 mg/kg i.v., mean = 2.0 mg/kg, n = 4; Fig. 3B). Picrotoxinin significantly reduced the duration of the inhibition elicited by DR/MnR stimulation from 114 ± 10 to 32 ± 4 ms (P < 0.004). Higher picrotoxinin doses could not be used due to its reversal of the anesthesia, which prevented to examine the effects of a full GABAA receptor blockade on these responses.

Responses Evoked by DR/MnR Stimulation in the Secondary Motor Area (MOs)

We examined the effects of the stimulation of the DR and MnR on neurons of MOs, which, in common with the cingulate and prelimbic areas of the PFC, contains a large abundance of 5-HT1A and 5-HT2A receptors in pyramidal neurons (Santana et al., 2004Go). Recordings were made at DV –0.8 to –1.4 mm, corresponding mainly to layer V. A total of 65 recordings were made, of which only one was a pure orthodromic excitation and five were biphasic responses. The excitations in these biphasic responses had a greater latency than the 5-HT2A-mediated excitations observed in mPFC [306 ± 33 ms, range = 232–424 (n = 5) versus 75 ± 11 ms for DR- and 45 ± 5 ms for MnR-evoked excitations]. The rest were inhibitions, whose characteristics are shown in Table 4. Figure 4 shows representative PSTHs of neurons in mPFC and MOs inhibited by the stimulation of DR and MnR. Only two antidromic responses were observed.


View this table:
[in this window]
[in a new window]
  		Table 4 Comparison of the characteristics of inhibitory responses in mPFC and secondary motor area (MOs)

 


View larger version (23K):
[in this window]
[in a new window]
  		Figure 4. Comparison of the inhibitory responses evoked in mPFC (A, B) and MOs (C, D) by the electrical stimulation of the DR and MnR. Characteristics are as follows: (A) stimulation at 2 mA; latency = 4 ms; duration = 140 ms; 5% of pre-stimulus firing; (B) stimulation at 1 mA; latency = 11 ms; duration = 148 ms; 0% of pre-stimulus firing; (C) stimulation at 1 mA; latency = 31 ms; duration = 232 ms; 19% of pre-stimulus firing; (D) stimulation at 1 mA; latency = 20 ms; duration = 156 ms; 3% of pre-stimulus firing. Note the antidromic spikes (arrowheads; latency = 13 and 18 ms in A and B, respectively) in the units recorded in mPFC (but not MOs), a consequence of the reciprocal connectivity of the mPFC and the raphe nuclei (only two cells out of a total of 65 in MOs were found to be antidromically activated from the raphe nuclei; see Fig. 5C). For a better visualization of the inhibitions, PSTHs have been drawn between –100 and +400 ms. Each PSTH consists of 110 triggers. Bin size is 4 ms.

 


View larger version (24K):
[in this window]
[in a new window]
  		Figure 5. PSTHs of a midbrain raphe-evoked inhibitions in mPFC (A, B) and MOs (C) (lower panels show an enlargement of the time period around the stimulus). (A) The stimulation of the DR at 1 mA evoked a short latency, short duration inhibition in mPFC (32 ms; 8% of pre-stimulus firing rate). The latency of the inhibition was 11 ms, whereas the antidromic spike (arrowhead) had a latency of 15 ms. PSTH made of 718 triggers (~13 min recording). (B) Short latency inhibition evoked in mPFC by DR stimulation at 1 mA (duration 78 ms, 3% of pre-stimulus firing rate). The latency of the inhibition was 12 ms and that of the antidromic spike was 15 ms. PSTH made of 617 triggers (~11 min). (C) Inhibition evoked in an identified projection neuron in MOs by MnR stimulation at 1 mA (duration 257 ms, 38% of pre-stimulus firing rate). Despite this neuron projected to midbrain, as denoted by the antidromic spike (latency 9 ms), the latency of the inhibition was 19 ms. PSTH made of 102 triggers (~2 min). Note the absence of spikes before the antidromic spike in the inhibitions recorded in pyramidal neurons of the mPFC, which precludes that the onset of these inhibitory responses is triggered by antidromic invasion of pyramidal collaterals and further activation of local GABA interneurons. In contrast, the two projection neurons found in MOs (out of 65) had a latency greater than the antidromic spike, as in the example shown in (C). Bin size = 1 ms.

 
Onset of Inhibitory Responses in mPFC and MOs

When comparing the latencies of inhibitory responses we observed a marked difference between both areas: 21 ± 3 in MOs versus 9 ± 1 ms in mPFC, n = 59 and 115, respectively; P < 0.0001) (Fig. 4). This difference was also statistically significant when considering the inhibitions evoked by the DR or MnR independently (Table 4). The rest of characteristics (duration, percent of basal firing, etc.) were comparable in both recording areas.

The latency of inhibitions in mPFC was significantly lower than that of (i) antidromic spikes evoked in DR 5-HT neurons by mPFC stimulation (24 ± 1 ms); (ii) the antidromic spikes evoked in mPFC pyramidal neurons by DR/MnR stimulation (15 ± 1 ms); and (iii) the inhibitions evoked by DR/MnR stimulation in the MOs (21 ± 3 ms) (Table 5). Moreover, the latencies of inhibitions in the two identified projection neurons found in MOs were 48 and 20 ms, above the latencies of the respective antidromic spikes, 7 and 9 ms, respectively (Fig. 5).


View this table:
[in this window]
[in a new window]
  		Table 5 Comparison of the latencies of antidromic and orthodromic responses in the mPFC-DR/MnR circuit

 
We conducted 9 additional experiments in which a higher number of triggers was used per each PSTH (342 ± 66 on average, corresponding to 6.3 ± 1.2 min per each PSTH). This procedure would drastically reduce the possibility that the lower latency in mPFC were due to a random lack of spikes around the stimulus artifact and antidromic spikes. The latency of antidromic spikes was 15 ± 1 ms (n = 9) and that of the inhibition was 13 ± 1 ms (n = 9) (P < 0.0006, paired Student's t-test). Figure 5 shows two representative examples.

Effects of WAY-100635 on Inhibitory Responses in mPFC and MOs

A second difference between the inhibitory responses in mPFC and MOs was the sensitivity to 5-HT1A receptor blockade with WAY-100635. In the mPFC, an earlier component (up to ~70 ms) remained insensitive to WAY-100635 (Amargós-Bosch et al., 2004Go). However, in MOs, inhibitions were more sensitive to WAY-100635. Of the six units examined, three inhibitions were partially reversed with 20–30 µg/kg WAY-100635 (from 163 ± 17 to 111 ± 23 ms) whereas the other three were fully blocked with 40–80 µg/kg WAY-100635 (from 205 ± 54 ms to 0 ms) (Fig. 6).



View larger version (22K):
[in this window]
[in a new window]
  		Figure 6. Effect of 5-HT1A receptor blockade on a MnR-evoked inhibition in MOs (1 mA). (A) PSTHs in basal conditions (upper panel) and after WAY-100635 administration (40 µg/kg i.v.) (lower panel). Inhibition latency: 46 ms, duration: 160 ms, 13% of pre-stimulus firing during inhibition. (B) A raster display shows the inhibition and the rapid effect of WAY-100635 administration. The horizontal line in (B) corresponds to the vertical line in (A) and denotes the stimulus artifact. Note the full blockade of the MnR-evoked inhibition by WAY-100635. Each PSTH corresponds to 110 triggers. Bin size = 4 ms. The time of administration of WAY-100635 (WAY) is shown by a vertical arrow.

 
Single versus Twin Pulse Stimulation

Previous data (Amargós-Bosch et al., 2004Go) indicate that pyramidal neurons in the mPFC can respond with excitations or inhibitions depending on the stimulation site in the raphe nuclei. This suggests that, despite 5-HT1A and 5-HT2A receptors are largely co-expressed in pyramidal neurons, certain neuronal subgroups within the raphe complex may project to 5-HT1A or 5-HT2A receptor-rich areas in pyramidal neurons. In support of this view, here we observed that a more rostral location of the stimulating electrode within the DR resulted in a higher proportion of excitations in mPFC. However, since the affinity of 5-HT for 5-HT2A receptors is lower than for 5-HT1A receptors (Peroutka and Snyder, 1979Go; Hoyer et al., 1985Go) the type of response could also be determined by the concentration of 5-HT reached in mPFC after raphe stimulation. It should be noted that the mean currents of inhibitory and excitatory responses were 1.20 ± 0.05 mA (n = 115) and 1.20 ± 0.13 mA (n = 23), respectively. Therefore, we performed a series of experiments in which the same pyramidal neurons in mPFC were recorded after stimulation of the DR/MnR with single and twin pulses while keeping stimulation frequency (0.9 Hz) and intensity. A total of 32 experiments were performed. Sixteen stimulations were performed in the DR (–7.8 mm), yielding seven inhibitions, three pure excitations and six biphasic responses after single pulse stimulation. Twin pulse stimulation markedly enhanced the duration of the inhibitions (Table 6) and converted two pure excitations and two biphasic responses into inhibitions (Fig. 7). The rest of responses were unaltered. Of the sixteen single pulse stimulations performed in the MnR, 12 resulted in inhibitions and four in biphasic responses. When the same neurons were recorded after twin pulse stimulation, the inhibitory responses were enhanced, and two of the biphasic responses were converted into inhibitions (Table 6, Fig. 7).


View this table:
[in this window]
[in a new window]
  		Table 6 Effect of single and twin pulse stimulation on DR/MnR-evoked inhibitions in mPFC pyramidal neurons

 


View larger version (44K):
[in this window]
[in a new window]
  		Figure 7. Enhancement of the DR- and MnR-evoked inhibitions in mPFC by twin pulse stimulation. Twin pulses were given 7 ms apart. The pyramidal neuron in (A) was inhibited by stimulation of the DR at 1.7 mA (latency = 19 ms, duration = 70 ms, 13% of pre-stimulus firing during inhibition). Twin pulse stimulation at the same intensity evoked a longer duration response (181 ms, 0% pre-stimulus firing; see lower panel in A). Similarly, the unit in (B) was inhibited by MnR stimulation at 1 mA (latency = 14 ms, duration = 170 ms, 8% of pre-stimulus firing during inhibition). This unit was antidromically activated at a higher current. Twin pulse stimulation evoked a longer duration response (318 ms, 11% pre-stimulus firing; see lower panel in B). Each PSTH consists of 120 triggers. Bin size = 4 ms. Arrows denote the stimulus artifact and arrowheads show the presence of antidromic spikes. (C, D) Raster displays showing the reversibility of the responses after single (S) and twin (T) pulse stimulation. (C) The enhancement of an inhibitory response after DR stimulation (1 mA) by twin pulse stimulation. This unit was antidromically activated from MnR at 1 mA. (D) The abolishment of an excitation in a biphasic response by twin pulse stimulation in the MnR (1 mA, latency of antidromic spike = 9 ms).

 
We examined the effect of WAY-100635 administration on the inhibitions evoked by twin pulse stimulation (n = 5). The pre-stimulus firing was 1.6 ± 0.4 spikes/s, the latency of the inhibitory response was 11 ± 1 ms and the duration was 180 ± 76 ms. Upon twin pulse stimulation, the duration of the inhibition increased up to 407 ± 82 ms (P < 0.01 versus single pulse). The administration of WAY-100635 (40–80 µg/kg i.v.) significantly reduced the duration of the inhibition evoked by twin pulse stimulation to 237 ± 67 ms (P < 0.003). Likewise, the magnitude of the inhibition was reduced from 93 ± 2% to 19 ± 2% (P < 0.02). When considering the earlier part of the inhibition (not blocked by WAY-100635; up to 237 ± 67 ms), the change was non-significant (from 93 ± 2% to 88 ± 7%). As observed with single pulse stimulations, WAY-100635 could not fully abolish the inhibitory response in mPFC evoked by twin pulse stimulation of the DR (Fig. 8).



View larger version (19K):
[in this window]
[in a new window]
  		Figure 8. Enhancement of an inhibitory response in a pyramidal neuron of the mPFC by twin pulse stimulation of the DR (1 mA; the neuron was antidromically activated from DR at 2 mA with a latency of 17 ms). Basal inhibition is shown in (A) (latency = 12 ms, duration = 282 ms, 18% of pre-stimulus firing during inhibition). Twin pulse stimulation enhanced the duration of the inhibition up to 412 ms (3% of pre-stimulus firing). Note the two stimulus artifacts (7 ms apart; indicated by arrows in the abcissa). The panel in (C) shows the partial reversal of the effect of twin pulse stimulation by the 5-HT1A receptor antagonist WAY-100635 (80 µg/kg i.v.), which reduced the duration of the inhibition to 268 ms. Each PSTH is composed of 110 triggers. Bin size = 4 ms. (D) Raster display of the PSTHs shown in (A–C). The time of administration of WAY-100635 (WAY) is shown by vertical arrows (40 and 80 µg/kg i.v. cumulative doses).

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
Conclusions
 References
 
The present study confirms and extends previous in vivo observations on the reciprocal control of the PFC and the midbrain raphe nuclei, which are involved in psychiatric diseases like major depression, schizophrenia or some anxiety disorders. Indeed, the DR/MnR project to several rat cortical areas, with an enrichment in the frontal pole (Azmitia and Segal, 1978Go; O'Hearn and Molliver, 1984Go; Blue et al., 1988Go). On the other hand several groups have reported that the mPFC projects to the raphe nuclei (Aghajanian and Wang, 1977Go; Sesack et al., 1989Go; Takagishi and Chiba, 1991Go; Hajós et al., 1998Go; Peyron et al., 1998Go) and the electrical stimulation of the mPFC exerts a profound influence on DR 5-HT neurons (Hajós et al., 1998Go; Celada et al., 2001Go).

The mean firing rate of the pyramidal neurons found in this study was similar to that found in other studies recording PFC cells extracellularly (Ceci et al., 1993Go; Hajós et al., 2001Go) but lower than in some studies using intracellular recordings. Thus, Lewis and O'Donnell (2000)Go and Trantham et al. (2002)Go reported mean values of ~4 spikes/s for pyramidal neurons in the same area. The range of values in these studies (0–28 spikes/s) suggests that different pyramidal types were recorded. Indeed, cells with a regular discharge pattern are the majority (~70%) and have a firing rate of <1 spike/s, lower than that of burst-firing neurons when recorded in vitro (Dégenètais et al., 2002Go). Moreover, there seems to be a relationship between firing pattern and area of projection (e.g. PFC cells projecting to the nucleus accumbens exhibit a burst-firing mode; Yang et al., 1996Go). Hence, it cannot be excluded that the antidromic identification from midbrain in the present study may have resulted in a selection of slowly, regular firing pyramidal cells. Additionally, methodological differences between extra- and intracellular recordings may also contribute to this difference.

The present observations agree with previous in vitro observations on the control of the activity of PFC neurons (see Introduction). Earlier in vivo observations indicated an inhibitory effect of DR and MnR stimulation on rat prefrontal neurons (Mantz et al., 1990Go). More recent studies show that the electrical stimulation of the DR inhibits -via 5-HT1A receptors- and excites -via 5-HT2A receptors- pyramidal neurons in the rat mPFC (Puig et al., 2003Go; Amargós-Bosch et al., 2004Go). Also, a recent study reported 5-HT1A receptor-mediated inhibitory responses in putative pyramidal neurons of the infralimbic area after DR/MnR stimulation (Hajós et al., 2003Go). These in vivo observations are consistent with the high density of these serotonergic receptors in rat mPFC (Pompeiano et al., 1992Go, 1994Go; López-Giménez et al., 1997Go). Nearly half of the neurons in PFC co-express 5-HT1A and 5-HT2A receptor mRNAs (Amargós-Bosch et al., 2004Go). To a large extent, these receptor mRNAs are present in cells also expressing vGLUT1 mRNA, which suggests a predominant pyramidal localization (Santana et al., 2004Go). Light and electronic microscope studies have shown a preferential localization of 5-HT2A receptors in apical dendrites and cell bodies of cortical pyramidal neurons (Jakab and Goldman-Rakic, 1998Go, 2000Go; Jansson et al., 2001Go; Martín-Ruiz et al., 2001Go; Miner et al., 2003Go; but see Cornéa-Hébert et al., 1999Go). However, conflicting results have been reported for 5-HT1A receptors. Using different antibodies, some groups have reported a homogenous distribution in pyramidal neurons (Kia et al., 1996Go; Riad et al., 2000Go) while others have shown a preferential localization in the axon hillock of cortical and hippocampal pyramidal neurons (Azmitia et al., 1996Go; De Felipe et al., 2001Go; Czyrack et al., 2003Go; Cruz et al., 2004Go).

Responses Evoked by DR/MnR Stimulation in PFC Neurons

Three types of responses were observed in mPFC: pure inhibitory, excitatory and biphasic, with a predominance of the former responses. Inhibitions appear to be mediated by two main components: (i) a 5-HT1A receptor-mediated, WAY-100635-sensitive inhibition; and (ii) a GABAA receptor-mediated, picrotoxinin-sensitive inhibition. Additionally, part of these inhibitions may be mediated by the after-hyperpolarization period (Yang et al., 1996Go) evoked by the antidromic spike in mPFC pyramidal neurons, which may perhaps explain the inability of WAY-100635 and picrotoxinin to fully suppress the DR/MnR-evoked inhibitions.

On the other hand, pure excitations have been shown to be mediated by the activation of 5-HT2A receptors (Puig et al., 2003Go; Amargós-Bosch et al., 2004Go). Serotonergic neurons contain vGluT3 (Gras et al., 2002Go; Herzog et al., 2004Go) and can make glutamatergic synapses in vitro (Johnson, 1994Go). Therefore, it is possible that DR/MnR stimulation evoked glutamate-mediated excitations. However, this seems unlikely in our experimental conditions since these excitations were blocked by the selective 5-HT2A antagonist M100907. Also, their latency and duration was greater than that expected from a glutamatergic input, even taking into account the slower conduction velocity of serotonergic neurons (Maurice et al., 1998Go; Celada et al., 2001Go).

Intriguingly, the observed proportion of inhibitory and excitatory responses is discordant with the very large (~80%) co-localization of 5-HT1A and 5-HT2A receptor mRNAs in pyramidal neurons of several PFC areas, such as the dorsal anterior cingulate and prelimbic areas, where recordings have been made (Amargós-Bosch et al., 2004Go). This inconsistency cannot be explained by an incomplete translation of the 5-HT2A receptor mRNA into the corresponding protein, since the rat mPFC contains a very high receptor density, as labeled with the selective antagonist [3H]MDL 100907 (López-Giménez et al., 1997Go). The preferential inhibitory action of 5-HT was also observed in early microiontophoretic and stimulation studies (Mantz et al., 1990Go; Ashby et al., 1994Go; for a review, see Jacobs and Azmitia, 1992Go) and may perhaps reflect the localization of 5-HT1A receptors in the pyramidal axon hillock (see above). Such a localization, coincident with the cortical GABAergic axo-axonic synapses between chandelier cells on the pyramidal axon hillock (Somogyi et al., 1998Go; De Felipe et al., 2001Go), would assign a prominent inhibitory role to 5-HT1A receptors in the control of pyramidal activity, as observed in the present study. Yet this interpretation must await the settling of the existing controversy on the cellular localization of the 5-HT1A receptors.

A second, possibly GABAA receptor-mediated, inhibitory component was involved in the DR/MnR-evoked inhibitions of mPFC-pyramidal cells. Three different sources of GABA might account for these results. First, 5-HT has been shown to activate excitatory 5-HT2A and 5-HT3 receptors in mPFC GABAergic interneurons, thus increasing a GABAergic input onto pyramidal neurons (Ashby et al., 1989Go, 1990Go; Tanaka and North, 1993Go; Zhou and Hablitz, 1999Go; Férézou et al., 2002Go; Puig et al., 2004Go). Second, due to the reciprocal anatomical connectivity of the mPFC and the raphe nuclei, the stimulation of descending pyramidal fibers projecting to the DR/MnR may result in antidromic invasion of pyramidal collaterals in mPFC and the subsequent activation of local GABA inputs onto the recorded pyramidal neurons. Third, the stimulation of the DR/MnR may stimulate non-serotonergic inhibitory afferents to the mPFC. While not discarding the first two possibilities, the present data support the latter possibility. Hence, 20% in the DR and 40% in the MnR of cortically projecting cells and one-third of raphe-cortical axons are non-serotonergic (O'Hearn and Molliver, 1984Go; Kosofsky and Molliver, 1987Go). More recent studies have also identified these DR-containing projection cells (Li et al., 2001Go) and the presence of a GABAergic projection from the DR to the mPFC has been described (Jankowski and Sesack, 2002Go). This GABAergic projection would be analogous to that existing from the ventral tegmental area to the mPFC, as evidenced by electrophysiological and anatomical studies (Pirot et al., 1994Go; Carr and Sesack, 2000Go). Interestingly, in the former study, the stimulation of the ventral tegmental area evoked a subgroup of inhibitory responses in pyramidal neurons of the cingulate and prelimbic areas with latencies ≤8 ms, in analogy with those found herein.

We show here that the electrical stimulation of the DR/MnR evokes a short latency inhibitory response in pyramidal neurons that cannot be accounted for by the latency of serotonergic axons. On the other hand, this early inhibitory component seems unlikely to be due to antidromic invasion of the mPFC and further activation of local GABA neurons because the latency of antidromic spikes evoked in pyramidal cells was significantly greater than the latency of DR/MnR-evoked inhibitions (see Table 5). Interestingly, the inhibitory responses evoked in the MOs, which contains a density of cells expressing 5-HT1A receptors comparable to that in mPFC (Amargós-Bosch et al., 2004Go) did not show this early component (latency of 21 ± 3 ms in MOs versus 9 ± 1 ms in mPFC), a difference that cannot be explained by the short distance between both recording areas. On the other hand, the inhibitory responses in MOs were more sensitive to WAY-100635 administration. Actually, three of these inhibitions were completely reversed by 5-HT1A receptor blockade, a fact never observed in mPFC. Interestingly, a recent study (Hajós et al., 2003Go) reported that DR/MnR- evoked inhibitions in putative pyramidal neurons of the rat infralimbic cortex had a latency of >25 ms and were fully blocked by WAY-100635, as observed here in MOs, but not in the prelimbic or cingulate areas. These functional differences possibly reflect the distinct afferent and efferent projections of the prelimbic and infralimbic areas of the PFC (Groenewegen and Uylings, 2000Go; Vertes, 2004Go), and suggest that the DR/MnR-mPFC GABAergic input (Jankowski and Sesack, 2002Go) is restricted to the cingulate and prelimbic areas.

Overall, these observations suggest the presence of a non-serotonergic, possibly GABAergic, component evoked by the DR/MnR stimulation in the mPFC. The observed inhibitory latency is consistent with that of projection GABAergic neurons in other brain areas (Paladini et al., 1999Go). However, due to the complex nature of these in vivo experiments, we could not increase the picrotoxinin dose to >2.5 mg/kg i.v. since it reversed the effects of anaesthesia so that only a partial blockade of the inhibitory response was achieved. Further experiments are required to assess the presence of this putative GABAergic control of mPFC neurons by the raphe nuclei.

The characteristics of the DR/MnR-evoked excitations in mPFC are totally similar to those previously shown to be reversed by the 5-HT2A receptor antagonist M100907. The mechanisms involved have been discussed elsewhere (Puig et al., 2003Go; Amargós-Bosch et al., 2004Go; see above). Interestingly, the latencies of the MnR-evoked excitations were lower than those observed after stimulation of the DR whereas the duration was similar. The exact reasons for this difference are unknown but may lie in anatomical differences between DR and MnR serotonergic neurons, which also exhibit a different sensitivity to neurotoxins (Kosofsky and Molliver, 1987Go; Mamounas and Molliver, 1988Go; O'Hearn et al., 1988Go). In the latter study, the selective lesion of fine axons by 3,4-methylenedioxyamphetamine unveiled a marked overlapping with beaded axons in the same territories of frontal cortex, in agreement with the present observation that PFC neurons are under control of DR and MnR neurons.

Biphasic responses possibly reflect the coexistence of 5-HT1A and 5-HT2A receptors in the same pyramidal neurons (Amargós-Bosch et al., 2004Go) and the temporal summation of DR/MnR-evoked inhibitory (very short latency, long duration) and excitatory responses (longer latency, shorter duration). However, as also observed for the pure orthodromic excitations, the proportion of biphasic responses is much lower than the observed 80% co-expression of the corresponding mRNAs, which again supports the predominance of inhibitory responses. Additionally, the presence of a GABAergic component cannot be ruled out.

What Determines the Type of Response?

Despite 5-HT being able to excite or inhibit pyramidal neurons, the reasons determining the emergence of one or other type of response are unclear. One of the limitations of the present study lies in the reciprocal connectivity of the mPFC and raphe nuclei (see above). One might argue that a GABAergic component of the inhibitions, resulting from the antidromic invasion of pyramidal axon collaterals, may artifactually increase the proportion of inhibitory responses. While the presence of such component cannot be excluded, the inhibitions/excitations ratio in MOs (almost devoid of projections to DR; see Peyron et al., 1998Go; this study) was even greater, which suggests that this is not a major determinant of the higher proportion of inhibitions in mPFC in our experimental conditions. A second limitation in the study is the use of anesthesia, which may alter cortical activity and hence the excitability of pyramidal neurons to incoming inputs, thereby altering the proportion of inhibitory versus excitatory responses. We tried to minimize fluctuations in the level of anesthetic by performing the recordings within a time span after administration of additional doses of chloral hydrate.

Interestingly, the units excited had a significantly lower pre-stimulus firing rate than those inhibited, which suggests that 5-HT may physiologically increase the firing of pyramidal neurons with a low activity and depress the activity of neurons with a higher activity. Since 5-HT2A and 5-HT1A receptors may be present in different cellular compartments (see above), we reasoned that certain serotonergic axons, passing near 5-HT2A receptor-rich apical dendrites (Jansson et al., 2001Go), may increase the excitability of pyramidal neurons, as observed in vitro (Aghajanian and Marek, 1997Go), rendering them more sensitive to incoming glutamatergic inputs. On the contrary, 5-HT released by axons passing near the axon hillock of pyramidal neurons would exert a dramatic inhibitory effect, switching off the propagation of action impulses (Amargós-Bosch et al., 2004Go). In that study, we observed that some pyramidal neurons in mPFC exhibited different responses to the stimulation of the DR or MnR at different coordinates. Here we observed that there was no difference between the mean effects (inhibition/excitation ratio) of the stimulation of DR and MnR (both at AP –7.8 mm) whereas the stimulation of the DR at a more rostral coordinate (AP –7.3 mm) yielded a significantly increased number of cells exhibiting excitations. Both observations tend to support the view that certain 5-HT neurons or neuronal subgroups may have a more excitatory effect than others on mPFC pyramidal neurons. However, the validity of this conclusion seems limited by the intrinsic complexity of the in vivo approach, since some DR fibers pass through the MnR and the stimulation of either nuclei may have also activated fibers en passage nearby. Despite this experimental limitation, we observed clear differences (e.g. latency of excitations in the present study or different responses in the same cell in Amargós-Bosch et al., 2004Go) which suggest the presence of a topologically defined connectivity between DR/MnR neurons and compartments of mPFC pyramidal neurons enriched in inhibitory or excitatory receptors.

The possibility that a higher release of 5-HT in mPFC would result in an increased number of cells exhibiting excitations (as suggested by the lower affinity of 5-HT for 5-HT2 versus 5-HT1A receptors; Peroutka and Snyder, 1979Go; Hoyer et al., 1985Go) is not supported by the present study. On the contrary, twin pulse stimulation, which enhances 5-HT release, increased the magnitude of the inhibitions, as previously observed (Gartside et al., 2000Go), and turned some excitations into inhibitions. This was a fully reversible effect, which indicates that it is determined by the actual extracellular concentration of 5-HT released by nerve impulses. As previously observed with single pulses, twin pulse-evoked inhibitions had a very short latency and were only partly blocked by WAY-100635 administration, suggesting the presence of a non-serotonergic component of inhibitions that was also enhanced by twin-pulse stimulation.

Finally, since the recorded pyramidal neurons were antidromically identified from midbrain, we cannot discard the possibility that cells projecting to other brain areas (e.g. nucleus accumbens, thalamus, amygdala) may respond to DR/MnR stimulation in a different manner. Hence, despite the large co-expression of 5-HT1A and 5-HT2A receptor mRNAs, it could be that the final response depends also on the firing pattern of the cells recorded (e.g. slow, regular spiking cells might be more easily inhibited than those firing in bursts).


   Conclusions
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusions
References
 
Despite the presence of excitatory 5-HT2A receptors in a large percentage of PFC neurons and their co-expression with 5-HT1A receptors, the present study shows that the main effect of physiologically released 5-HT in PFC is inhibitory, perhaps due to a suggested localization of 5-HT1A receptors in the axon hillock of pyramidal neurons. A greater increase of 5-HT release in mPFC, as elicited by twin pulse stimulation of the DR and MnR does not increase the proportion of excitatory responses but, on the contrary, transforms them into inhibitions. Lastly, inhibitory responses evoked in mPFC involve 5-HT1A receptor- and GABAA receptor-mediated components, whereas in MOs only the serotonergic component appears to be present. Various sources of GABA may account for the inhibitions in mPFC, yet the very short latency of the DR/MnR-evoked inhibitions suggests the presence of a GABAergic projection from the DR/MnR to the cingulate and prelimbic areas of mPFC.


   Acknowledgments
 
This work was supported by grants SAF 2001-2133 and Fundació La Marató TV3. P.C. is recipient of a Ramón y Cajal contract from the Ministry of Science and Technology. M.V.P. is recipient of a predoctoral fellowship from IDIBAPS. Support from the CIEN network (IDIBAPS-ISCIII RTIC C03/06) and Generalitat de Catlunya (2001-SGR00355) is also acknowledged. We thank Judith Gallart for skillful technical assistance.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusions
 References
 
Aghajanian GK, Marek GJ (1997) Serotonin induces excitatory postsynaptic potentials in apical dendrites of neocortical pyramidal cells. Neuropharmacology 36:589–599.[CrossRef][ISI][Medline]

Aghajanian GK, Marek GJ (1999) Serotonin–glutamate interactions: a new target for antipsychotic drugs. Neuropsychopharmacology 21:S122–S133.[CrossRef]

Aghajanian GK, Wang RY (1977) Habenular and other midbrain raphe afferents demonstrated by a modified retrograde tracing technique. Brain Res 122:229–242.[CrossRef][ISI][Medline]

Amargós-Bosch M, Bortolozzi A, Puig MV, Serrats, Adell A, Celada P, Toth M, Mengod G, Artigas F (2004) Co-expression and in vivo interaction of serotonin1a and serotonin2a receptors in pyramidal neurons of prefrontal cortex. Cereb Cortex 14:281–299.[Abstract/Free Full Text]

Araneda R, Andrade R (1991) 5-Hydroxytryptamine-2 and 5-hydroxytryptamine-1A receptors mediate opposing responses on membrane excitability in rat association cortex. Neuroscience 40:399–412.[CrossRef][ISI][Medline]

Arango V, Underwood MD, Mann JJ (2002) Serotonin brain circuits involved in major depression and suicide. Prog Brain Res 136:443–453.[Medline]

Arvanov VL, Liang X, Magro P, Roberts R, Wang RY (1999) A pre- and postsynaptic modulatory action of 5-HT and the 5-HT2A/ 2C receptor agonist DOB on NMDA-evoked responses in the rat medial prefrontal cortex. Eur J Neurosci 11:2917–2934.[CrossRef][ISI][Medline]

Ashby CR, Edwards E, Harkins K, Wang RY (1989) Characterization of 5-hydroxytryptamine3 receptors in the medial prefrontal cortex: a microiontophoretic study. Eur J Pharmacol 173:193–196.[CrossRef][ISI][Medline]

Ashby CR, Jiang LH, Kasser RJ, Wang RY (1990) Electrophysiological characterization of 5-hydroxytryptamine-2 receptors in the rat medial prefrontal cortex. J Pharmacol Exp Ther 252:171–178.[