INTRODUCTION

Midbrain dopamine neurons have a crucial role in motor functions, motivated behaviors, reward-related learning, and in processing salient signals (Bromberg-Martin et al, 2010; Wise, 2004), as well as in associated pathologies. The GABAergic tail of the ventral tegmental area (tVTA) (Kaufling et al, 2009, 2010a; Perroti et al, 2005), or rostromedial tegmental nucleus (RMTg) (Jhou et al, 2009a, 2009b), is a mesopontine structure, which exerts a major inhibitory control over dopamine cells of the VTA and substantia nigra pars compacta (SNc) (Bourdy and Barrot, 2012; Matsui et al, 2014; Sánchez-Catalán et al, 2014). The tVTA has been implicated in motor functions (Bourdy et al, 2014; Lavezzi et al, 2015), responses to drugs of abuse (Jalabert et al, 2011; Jhou et al, 2012, 2013; Kaufling et al, 2010b; Lavezzi et al, 2012; Lecca et al, 2011, 2012; Matsui and Williams, 2011; Matsui et al, 2014; Melis et al, 2014; Rotllant et al, 2010), and reward prediction error (Hong et al, 2011). In addition, stimulation of the habenulo–tVTA pathway favors avoidance behaviors (Stamatakis and Stuber, 2012).

The tVTA receives inputs from the lateral habenula (LHb) (Balcita-Pedicino et al, 2011; Jhou et al, 2009a; Kaufling et al, 2009), a glutamatergic diencephalic structure involved in processing aversive information and reward prediction error (Matsumoto and Hikosaka, 2009). Thus, the tVTA has been proposed as an intermediate between the LHb and the VTA dopamine neurons, and to be important in promoting behavioral avoidance. Indeed, optogenetic stimulation of LHb terminals in the tVTA induces active, passive, and conditioned behavioral avoidance (Lammel et al, 2012; Stamatakis and Stuber, 2012). Moreover, pharmacological stimulation of the tVTA induces conditioned place aversion (Jhou et al, 2013). Otherwise, ablation of the tVTA inhibits fear-conditioned freezing, passive response to a predator odor, and anxiety-like behavior in an elevated plus maze (Jhou et al, 2009b). It also suppresses cocaine-induced avoidance behavior in a runway operant paradigm (Jhou et al, 2013), and cocaine excitation of LHb-tVTA neurons contributes to cocaine-induced depressive-like behaviors (Meye et al, 2015).

Various psychostimulants recruit the tVTA, as indicated by Fos induction in rats, suggesting that the tVTA is a common target for those drugs (Kaufling et al, 2010b). Similarly, the tVTA displays Fos induction following exposure to series of foot shock (Brown and Shepard, 2013; Jhou et al, 2009b). However, to what extent other aversive stimuli similarly recruit the tVTA remains to be determined. As the tVTA might be part of the brain circuitry implicated in avoidance behaviors, it is important to further explore its response to aversive stimuli of different nature, and more particularly to determine whether the response observed with foot shocks may be generalized and constitutes a hallmark of the structure. Here, we assessed c-Fos expression in the rat tVTA in response to lithium chloride (LiCl), β-carboline, naloxone, lipopolysaccharide (LPS), inflammatory pain, neuropathic pain, foot-shock, loud tone, restraint stress, forced swimming, fox and cat odors, and opiate withdrawal. Moreover, we also assessed the effect of tVTA ablation on physical signs of opiate withdrawal and on LiCl and LPS-induced conditioned taste aversion (CTA).

MATERIALS AND METHODS

Animals

Experiments were performed in adult male Sprague–Dawley rats (Janvier, France) housed under standard conditions (22±1 °C, 12-h light–dark cycle). Rats were habituated to the facilities for at least 1 week and daily handled the week preceding experimentation. They were 7–9 weeks old at testing or surgery time. Experiments were approved by the regional ethical committee (CREMEAS).

Drugs

Drugs (Sigma-Aldrich, France; unless otherwise stated) were prepared in 0.9% NaCl (unless otherwise stated) and injected (1 ml/kg, except for λ-carrageenan) at the following doses: D-amphetamine sulfate, 0.25, 0.5, and 1 mg/kg intraperitoneally; LiCl, 60 mg/kg intraperitoneally; β-carboline (FG7142), 15 mg/kg intraperitoneally in 45% cyclodextrin in water; naloxone hydrochloride, 10 mg/kg intraperitoneally or 1 mg/kg subcutaneously; LPS from Escherichia coli 0127-B8, 250 μg/kg intraperitoneally; λ-carrageenan type IV (lot. No. 80K133h), 150 μl of 3% solution, intraplantar; morphine hydrochloride, 10, 15, 20, 30, 40, 60, and 80 mg/kg intraperitoneally (Cooper, France). 2,4,5-trimethylthiazoline (TMT) (PheroTech, Delta, BC, Canada) was dissolved in diethylphtalate.

Aversive Procedures

Drugs producing aversive state were injected in animals: LiCl, β-carboline, naloxone, and LPS. Amphetamine was used as positive control for tVTA Fos induction. Rats were perfused 2 h postinjection, except for LPS (time course: 2, 4, and 8 h).

Inflammatory pain was induced by injection of λ-carrageenan in the right hind paw (Luis-Delgado et al, 2006a), with perfusion being carried out 2, 8, and 24 h later. To induce neuropathic pain, rats were anesthetized (ketamine, 87 mg/kg; xylazine, 13 mg/kg), and a 2 mm cuff of PE-90 polyethylene tubing was placed around the main branch of the right sciatic nerve (Yalcin et al, 2014; Benbouzid et al, 2008). The sham group underwent the same surgery without cuff implantation. The presence of mechanical hypersensitivity was tested using calibrated forceps (Bioseb, France) (Luis-Delgado et al, 2006b). On day 18 after surgery, rats were perfused 2 h after the last measure of mechanical thresholds.

For exposure to predator odors, rats were first habituated to the exposure cages (25 × 16 × 22 cm) for 15 min. On the following day, rats were exposed to the odor for 20 min, starting 3 min after beginning the session. One group was exposed to TMT, a component of red fox feces, by introducing a gauze pad with 20 μl of 10% TMT into the cage (Jhou et al, 2009b). Controls were exposed to the vehicle. Another group was exposed to a 25 mm piece of cat collar previously worn by a domestic cat for 2 weeks (McGregor et al, 2004). Controls were exposed to an unworn collar. Perfusion was carried out 2 h after starting the session.

One hour restraint stress was carried out in transparent cylinders (21 × 6 cm). Forced swimming was carried out by placing rats for 15 min in cylinders (50 × 30 cm diameter) filled to 30 cm with 25 °C water, rats were then removed, dried with towels, and placed in a warmed cage (Pliakas et al, 2001). Perfusion was carried out 2 h after the onset of stress procedure.

As repeated foot shocks have been shown to recruit the tVTA (Brown and Shepard, 2013; Jhou et al, 2009b), we tested them under different procedures. Rats were habituated for 30 min to behavioral chambers (25 × 27 × 18 cm; Med Associates, Saint Albans, Vermont) with a loudspeaker, camera, and grid floor. The following day, rats were exposed to these chambers in a 38 min session. The tone and tone/shock groups received, respectively, five pseudorandom tones (4000 Hz, 15 s) or five pseudorandom tones followed by foot shock (0.5 mA, 0.8 s). The foot-shock group received five pseudorandom shocks with no tone. Freezing behavior was assessed as aversive reaction (Marchand et al, 2003). Rats were perfused 2 h after starting the session. In an independent experiment, we tested exposure to a single foot shock or single tone. Rats were habituated to the chamber (soundproof, 51 × 25 × 24 cm) for 2 min, and tested the following day in a 2-min session. One minute after beginning the session, rats were exposed to a single foot shock (0.5 mA, 0.8 s or 4 s; or 0.8 mA, 0.8 s) or a loud tone (3000 Hz, 1 s), placed in their home cages and perfused 1h30 after starting the test.

Morphine dependence was induced by escalating doses of morphine (2 ml/kg), two times a day for 7 days: 10, 15, 20, 30, 40, 60, and 80 mg/kg. On day 8, 1 h after the morning morphine injection (80 mg/kg), rats received naloxone (1 mg/kg, s.c.) or saline. Opiate withdrawal was assessed in transparent cages (37 × 25 × 30 cm), scoring behavioral and physiological variables over 30 min. Rats were perfused 2 h after naloxone or saline administration.

Lesion

Rats were anesthetized (sodium pentobarbital, 50 mg/kg, intraperitoneally) for stereotaxic surgery. Ibotenic acid (1% in phosphate-buffered saline (PBS) 0.1 M, 0.2 μl) was injected bilaterally into the tVTA (0.2 μl) (Bourdy et al, 2014). Stereotaxic coordinates relative to bregma were (in mm): anteroposteriority −6.7/6.9, laterality ±1.4/1.5, verticality (from dura) −7.7/7.8, ±6° lateral angle (Paxinos and Watson, 2007). Sham animals underwent the same procedure without lesion. Experiments started 1 to 2 weeks after surgery. After the experiments, rats were anesthetized and perfused to control for tVTA lesion.

Conditioned Taste Aversion

LiCl- or LPS-induced CTA (Slouzkey et al, 2013) was carried out using 30 min daily sessions over 7 days. Following 24 h of water deprivation, rats were trained in their home cage to drink water from two bottles for 30 min during 3 days. On the conditioning day, the two bottles were then filled with 0.1% saccharin solution. Ten minutes following the end of this session, rats received LiCl (Tenk et al, 2006), LPS (Konsman et al, 2008), or saline. The following 2 days, rats were exposed to two bottles of water during the session (baseline consumption). On test day, rats were allowed to drink from one bottle with water and one bottle with 0.1% saccharin, with randomized position. The aversion index was defined as (ml water/(ml water+ml saccharin) × 100) consumed during the test session.

Immunohistochemistry

Rats were anesthetized with pentobarbital overdose and perfused with 100 ml phosphate buffer (0.1 M, pH 7.4), followed by 500 ml of 4% paraformaldehyde in phosphate buffer. Brains were postfixed overnight, cryoprotected in 20% glycerol in PBS, and coronal sections (40 μm) were cut using a cryotome. Immunohistochemistry was performed as described previously (Bourdy et al, 2014; Kaufling et al, 2010b). Primary antibodies targeted: c-Fos (Santa-Cruz; sc-52, lot 1008, 1/10000 for DAB staining, 1/1000 for immunofluorescence), μ-opioid receptor (MOR) (Millipore; AB1724, 1/400 for immunofluorescence), tyrosine hydroxylase (TH) (Millipore-Chemicon; MAB318, 1/2500), or NeuN (Millipore; MAB377, 1/5000). Visualization was performed with biotinylated secondary antibodies and peroxidase/DAB reaction after avidin-biotin-peroxidase amplification (ABC Elite; Vector Laboratories, USA), except for multiple staining performed using Alexa488- and Cy3-conjugated secondary antibodies (1/400; Jackson ImmunoResearch) and DAPI for nuclei detection.

Coronal slices from −5.80 to −7.30 mm from bregma were used to count c-Fos-positive nuclei bilaterally in the tVTA. The tVTA is present between −6 mm and −7 mm from bregma, as identified by amphetamine-induced Fos expression. Rostrally, it starts within the VTA paranigral nucleus, and dorsolaterally to the interpeduncular nucleus. It then extends caudally and dorsally, laterally to the median raphe nuclei, and is partially embedded within the superior cerebellar peduncle fibers (Kaufling et al, 2010a; Sánchez-Catalán et al, 2014). We analyzed a section every 160 μm along the entire tVTA extent; after correction for missing sections, data were expressed per whole unilateral tVTA (Kaufling et al, 2010b). Images were acquired using a microscope (Eclipse E600; Nikon Instruments) with the Neurolucida software.

Analysis of c-Fos/MOR nuclei was performed using the optical fractionator technique with optical dissectors (30 μm height, 5 μm guard zone), under the Stereoinvestigator software (MBF Biosciences). MOR immunostaining allowed delimiting the tVTA (eg, Figure 4e) at × 10 objective, and counting was performed at × 40 on 10 randomly positioned counting frames per section (7–12 sections analyzed per animal). Microphotographs were acquired on a Leica SP5 II confocal microscope.

Analyses

Data are expressed as mean±SEM. A power analysis was performed by using G*Power3 (version 3.1.9). Other statistical analyses were performed using STATISTICA 7.1 (Statsoft, Tulsa, OK, USA), with Student’s t-test for two-group comparison and ANOVA for multiple comparisons. Post hoc analyses were carried out using either Dunnett or Bonferroni tests. For CTA experiments, a factorial two-way ANOVA was followed by a Duncan test. We performed a test of comparison between the mean and a standard value (50%) to compare aversion indices. Level of significance was p<0.05.

RESULTS

Aversive Drugs

As psychostimulants induce Fos proteins in the rat tVTA (Kaufling et al, 2010b; Perrotti et al, 2005), we used amphetamine as an internal positive control and showed that the amphetamine c-Fos induction is dose-dependent (F3,16=86.7, p<0.001) (Figures 1a and g). Concerning aversive drugs (F3,19=21.9, p<0.001), tVTA Fos induction was observed only with the anxiogenic drug β-carboline (File and Baldwin, 1987) (p<0.001) (Figures 1e and g). This induction was significant, but quantitatively small compared with psychostimulants. No significant induction was present with LiCl (p=0.35), a drug provoking visceral illness (Ossenkopp and Eckel, 1995), with naloxone at high dose (p=1), and with LPS (F3,13=1.3, p=0.31), a bacterial endotoxin inducing sickness behavior (Figure 1) (Castanon et al, 2001; Konsman et al, 2008). Power analysis using the saline and β-carboline groups as a reference to calculate the effect size index (considering parametric t-test for two independent groups with α=0.05 and power=0.80) gave an estimated sample size of three animals per group, indicating that n in our groups was sufficient to detect relevant differences.

Figure 1
figure 1

Aversive drugs barely induce c-Fos expression in the GABAergic tail of the ventral tegmental area (tVTA). (a) Amphetamine dose–response (0.25, 0.5, and 1 mg/kg intraperitoneally, n=3–4 per group) was used as a positive control for c-Fos expression in the tVTA. (b) Acute injection of saline (n=5) has no effect on c-Fos in the tVTA. (c–f) Illustrations for c-Fos in the tVTA after acute injection of naloxone 10 mg/kg (n=4) (c), lithium chloride (LiCl) 60 mg/kg (n=5) (d), β-carboline 15 mg/kg (n=6) (e), and lipopolysaccharide (LPS) 250 μg/kg (f). (g) Quantification of c-Fos-positive nuclei in the tVTA. Rats were perfused 2 h after injection, except for LPS (time course: 2, 4, and 8 h, n=3 per time point). The squares in a1–f1 indicate the regions shown at higher magnification in a2–f2. Scale bars, 500 μm (a1–f1) and 100 μm (a2–f2, a3–a4, and f3–f4). ***p<0.001, **p<0.01, and *p<0.05 (relative to the saline group).

PowerPoint slide

Full size image

Painful Stimuli

Inhibition of the tVTA impacts nocifensive responses (Jhou et al, 2012). However, painful stimuli did not induce c-Fos locally. With λ-carrageenan-induced inflammatory pain (Luis-Delgado et al, 2006a, b), no c-Fos induction was present in the tVTA (Figures 2a–c and m). Similarly, no c-Fos was observed in the tVTA with neuropathic pain (p=0.6) (Figures 2d, e and m), whereas ipsilateral paw withdrawal thresholds were chronically lowered (group, F1,12=219.2, p<0.001; paw, F1,12=297.1, p<0.001; time × group × paw, F6,72=18.5, p<0.001) (Figure 2f).

Figure 2
figure 2

Painful or stressful stimuli do not induce c-Fos expression in the GABAergic tail of the ventral tegmental area (tVTA). (a–c) Intraplantar injection of λ-carrageenan (Carr) does not induce c-Fos expression in the tVTA (n=3 per time point). (d and e) Neuropathic pain (NP) does not induce c-Fos expression in the tVTA. Sham animals underwent the same surgical procedure without cuff implantation (n=4 per group). (f) Cuff implantation results in a chronic ipsilateral allodynia (day 0: surgery, ***p<0.001 relative to the other groups). (g and h) Fox odor exposure (2,4,5-trimethylthiazoline (TMT); control group: TMTc) do not induce c-Fos expression in the tVTA (n=3 per group). (i and j) Cat odor exposure (CAT; control group: CATc) do not induce c-Fos expression in the tVTA (n=3 per group). (k and l) Acute restraint stress (n=5) and forced swimming test (FST, n=5) do not induce c-Fos expression in the tVTA. (m) Quantification of c-Fos-positive nuclei in the tVTA. An amphetamine group (amph, 1 mg/kg) (n=3) was included as a positive control group. Rats were perfused 2 h after starting the behavioral test or after the last measure of mechanical thresholds for neuropathic pain. The squares in (a1–e1 and g1–l1) indicate the regions shown at higher magnification in (a2–l2 and g2–l2). Scale bars, 500 μm (a1–e1 and g1–l1) and 100 μm (a2–l2 and g2–l2).

PowerPoint slide

Full size image

Stressful Stimuli

Predator odors induce unconditioned fear responses in rats, and tVTA lesion changes these responses (Jhou et al, 2009b). However, exposure to fox or cat odors did not induce c-Fos in the tVTA (TMT, p=0.69; cat odor, p=0.79) (Figures 2g–j and m). Similarly, restraint stress, or forced swimming as classically used in depression research, also failed to recruit the tVTA (Figures 2k–m).

The only stress procedure that led to significant c-Fos induction in the tVTA was the repeated exposure to mild foot shocks (F3,22=23.4, p<0.001) delivered in a protocol usually used to elicit fear conditioning. Freezing behavior during the shock session was observed in foot-shock-exposed animals (F3,21=28.7, p<0.001) (Figure 3e). This c-Fos induction was observed whether a predictive tone was present or not before each shock exposure (Figures 3a–d and f). However, a single foot shock did not evoke c-Fos in the tVTA. Indeed, no induction was observed with a single shock of intensity as used in fear conditioning (0.5 mA/0.8 s), with a single shock of intensity as used in fear conditioning but of a duration equivalent to the sum of five shocks (0.5 mA/4 s), or with a single shock of higher intensity (0.8 mA/0.8 s) (Figures 3g–l). Moreover, single salient information (arousing but non-nociceptive), such as a loud auditory stimulus, also failed to recruit the tVTA (Figure 3h). These findings support the idea that repetition of the stimulus was critical for inducing c-Fos in the tVTA as a consequence of foot shocks.

Figure 3
figure 3

Repeated foot-shock exposure induces c-Fos expression in the GABAergic tail of the ventral tegmental area (tVTA). The context (a) and the tone (b) exposure hardly induces c-Fos expression in the tVTA, whereas (c and d) repeated foot shocks (0.5 mA, 5 × 0.8 s) induce c-Fos expression in the tVTA, under the presence of the tone or not (n=5–6 per group). (e) Freezing behavior in the test session following repeated foot-shock exposures. (f) Quantification of c-Fos-positive nuclei in the tVTA following repeated foot-shock exposures. ***p<0.001 (relative to the control group). (g–k) Single foot-shock exposure fails to induce c-Fos in the tVTA. Compared with controls (context, (g)), no c-Fos expression was present with 0.5 mA/0.8 s (i), 0.8 mA/0.8 s (j), or 0.5 mA/4 s (k) (n=4 per group). The exposure to a salient loud auditory stimulus also failed to recruit the tVTA (h) (n=4). (l) Quantification of c-Fos-positive nuclei in the tVTA following single foot-shock exposure or a loud auditory stimulus. Rats were perfused 2 h after the beginning of the session for repeated foot-shock experiments, or after 1h30 for single foot-shock experiments. The squares in (a1–d1 and g1–k1) indicate the regions shown at higher magnification in (a2–d2 and g2–k2). Scale bars, 500 μm (a1–d1, g1–k1) and 100 μm (a2–d2, g2–k2).

PowerPoint slide

Full size image

Opioid Withdrawal

The tVTA is a direct target of opiates (Jalabert et al, 2011), even though morphine itself does not induce Fos locally (Perrotti et al, 2005; Kaufling et al, 2010b). Here we show that precipitated opiate withdrawal (Supplementary Table S1) induces c-Fos in the tVTA (F3,16=406.9, p<0.001) (Figures 4a and b). Morphine treatment per se did not affect the total number of MOR-positive neurons in the tVTA (data not shown). Withdrawal-induced c-Fos was present (15%) in MOR-expressing neurons, but mostly (85%) in neurons not expressing this receptor (Figures 4c–e). In a separate experiment, we also show that a bilateral lesion of the tVTA did not significantly change the individual physical signs of withdrawal or the global withdrawal score (Figure 4f and Supplementary Table S2, p=0.91).

Figure 4
figure 4

Opiate withdrawal induces c-Fos expression in the GABAergic tail of the ventral tegmental area (tVTA). (a) Animals receiving repeated saline injections and a saline (S/S) or naloxone (1 mg/kg) (S/N) challenge, or repeated morphine injections and a saline challenge (M/S), do not display c-Fos induction in the tVTA, whereas animals receiving repeated morphine injections and a naloxone challenge (M/N) exhibited opioid withdrawal symptoms (Supplementary Table S1) and high levels of c-Fos-positive nuclei in the tVTA. (b) Quantification of c-Fos-positive nuclei in the tVTA (n=4–6 per group). (c) Quantification of c-Fos-positive nuclei in MOR-positive and -negative cells of the tVTA (n=5 per group). ***p<0.001 (relative to the control group). (d) After opiate withdrawal, c-Fos induction (in green) in the rat tVTA was present in MOR-expressing neurons (in red) (as indicated by the arrows), and in neurons not expressing them (as indicated by the arrow heads). Cell nuclei are labeled with DAPI (4',6-diamidino-2-phenylindole) (in blue). Rats were perfused 2 h after the naloxone or saline administration. Scale bars, 100 μm (a) and 10 μm (d). (e) Illustration of one of the sections used for stereological analysis (data in c). The square in left picture indicates the region shown at higher magnification in right pictures (top: MOR in orange; bottom: c-Fos in green). The dotted area in right pictures indicates the tVTA limits used for stereological analysis; at this anteroposterior level, the tVTA is embedded within fibers of the superior cerebellar peduncle. Scale bars, 500 μm (left picture) and 100 μm (right pictures). (f) The lesion of the tVTA did not significantly alter the physical signs or the global score (Gellert and Holtzman, 1978) of naloxone-precipitated opiate withdrawal (sham n=7; lesion n=6). A full color version of this figure is available at the Neuropsychopharmacology journal online.

PowerPoint slide

Full size image

Conditioned Taste Aversion

The lack of tVTA c-Fos induction in response to most of the tested aversive stimuli raised the question of tVTA requirement for avoidance behaviors. We assessed the effect of tVTA ablation on the LiCl- and LPS-induced CTA, aversive-like behaviors not yet tested in tVTA studies. CTA being a robust conditioning, single conditioning procedures were used (Slouzkey et al, 2013). Only rats with lesions encompassing the tVTA were included in the results, TH immunostaining was used to control for the absence of VTA dopaminergic lesion (Figures 5a and b). Sham control rats (Sal-sham) displayed a preference for saccharin in the two-bottle choice test (Figure 5c). tVTA ablation did not abolish saccharin preference or affect the LiCl- or LPS-induced CTA (treatment, F2,32=21.3, p<0.001; surgery, F1,32=2.5, p=0.12; treatment × surgery, F2,32=0.9, p=0.41). Indeed, rats showed decreased preference for the saccharin solution previously paired to LiCl or LPS injection in both sham (p<0.001 and p<0.01, respectively) and lesion groups (p<0.001). Moreover, the aversive index differed from chance level (50%) in LPS-lesion, LiCl-lesion and LiCl-sham groups, even though it failed to differ from chance level in the LPS-sham group, because of high variability (Figure 5c).

Figure 5
figure 5

GABAergic tail of the ventral tegmental area (tVTA) ablation does not affect the lithium chloride (LiCl)- and lipopolysaccharide (LPS)-induced conditioned taste aversion (CTA). (a) Histological evidence for the tVTA sham group with NeuN (a1) and tyrosine hydroxylase (TH) (a2) immunostaining. (b) Histological evidence for the tVTA lesion group with NeuN (b1) and TH (b2) immunostaining. (c) Aversion index for the LiCl and LPS-induced CTA in sham controls and lesion animals (n=5–8 per group). Scale bars, 200 μm (a1 and b1) and 500 μm (a2 and b2). **p<0.01, ***p<0.001 (relative to their respective saline group). #p<0.05 and ###p<0.001 (referred to 50%).

PowerPoint slide

Full size image

DISCUSSION

By exploring the impact of aversive stimuli, we found that most of them failed to significantly induce c-Fos in the tVTA. Induction was only observed with repeated foot shocks and precipitated morphine withdrawal, and mildly with β-carboline at a high dose. Opiate withdrawal Fos response in the tVTA concerned MOR-expressing and -non-expressing cells. Last, we show that bilateral ablation of the tVTA did not affect physical signs of opiate withdrawal, or LPS- and LiCl-induced CTA.

Implication of the tVTA, which receives afferents from the LHb (Balcita-Pedicino et al, 2011; Jhou et al, 2009a; Kaufling et al, 2009) and controls dopamine cell activity (Bourdy et al, 2014; Jalabert et al, 2011; Jhou et al, 2009b; Matsui and Williams, 2011), has been described in active, passive, and conditioned avoidance behaviors and in the processing of aversive information (Jhou et al, 2009b, 2013; Lammel et al, 2012; Meye et al, 2015; Stamatakis and Stuber, 2012). To date, however, the most striking Fos responses in this structure were observed with psychostimulants (Cornish et al, 2012; Jhou et al, 2009a, 2013; Geisler et al, 2008; Kaufling et al, 2010b; Lecca et al, 2011; Perrotti et al, 2005; Zahm et al, 2010) (see Supplementary Table S3 and Supplementary Data). Our first aim was thus to test whether aversive stimuli would display a common molecular impact on the tVTA, as revealed by induction of c-Fos proteins. Surprisingly, the exposure to aversive drugs such as LiCl, naloxone at high dose, or LPS did not induce c-Fos expression in the tVTA. The anxiogenic agent β-carboline, used at a high dose to model panic attacks (Thiébot et al, 1988), resulted in a significant induction, although quantitatively mild compared with the response observed with amphetamine. More sustained aversive experiences, such as inflammatory pain and neuropathic pain, also failed to change c-Fos expression locally, even though inhibition of the tVTA by local microinjection of endomorphin or muscimol has an analgesic action in the formalin test (Jhou et al, 2012). Lesion of the tVTA prevents the freezing response to TMT odor, while increasing other defensive behaviors (Jhou et al, 2009b), which suggests that the tVTA is important to passive defensive behavior. However, we did not evidence c-Fos induction when testing unconditioned fear through predator odor exposure, or after stress elicited by immobilization or forced swimming. Importantly, a lack of c-Fos expression does not imply a lack of functional recruitment of tVTA cells. For example, there is evidence for an electrophysiological impact on tVTA activity in the absence of Fos induction with drugs that inhibit the tVTA, such as opiates and cannabinoids (Kaufling et al, 2010b; Lecca et al, 2011). However, avoidance behaviors are supposedly associated with activation, and not inhibition, of the tVTA.

Contrary to most stressful stimuli described above, c-Fos expression in the tVTA was present following exposure to repeated mild foot shocks, similarly to previous reports (Brown and Shepard, 2013; Jhou et al, 2009b). The mechanism underlying this expression depends on LHb inputs. Indeed, a lesion of the fasciculus retroflexus impairs this response to mild foot shocks (Brown and Shepard, 2013), and acute unpredictable foot shocks potentiate glutamate release from the LHb to the tVTA (Stamatakis and Stuber, 2012). However, these procedures relied on a series of foot shocks, that is, the repetition of identical discrete aversive events. On the contrary, exposure to a single foot shock does not induce c-Fos in the tVTA, even though the duration of the single shock is extended to 4 s (mimicking the cumulated duration of the shock series), or if the intensity is increased to reach levels usually used for learned helplessness procedures (0.8 mA). These findings suggest that the effect of a series of foot shocks on the tVTA is unlikely related to their aversive impact per se, nor to the salient, arousing, or attentional properties of the shocks. These conclusions are also supported by the lack of c-Fos response following exposure to a single loud auditory tone. The present data more likely support the participation of the tVTA in learning and prediction processing associated with the repetition of discrete events. In this context, the tVTA has been shown, in monkeys, to process reward prediction error, which is mirroring the implication of the LHb in the same processing (Hong et al, 2011). It might be hypothesized that the response observed with a series of foot shocks may reflect attempts to predict, and/or failures in predicting, repeated occurrences of the shocks.

Present and previous data show that opiates do not induce Fos in the tVTA (Kaufling et al, 2010b; Perrotti et al, 2005). However, acute morphine withdrawal strongly induces c-Fos in this brain region. Behavioral studies support a role for the tVTA in opioid responses, as rats self-administer and develop conditioned place preference when endomorphin-1 is delivered into the tVTA (Jhou et al, 2012). In fact, part of the GABAergic tVTA neurons in rats express MOR (Jalabert et al, 2011; Jhou et al, 2009a, 2012; Kaufling and Aston-Jones, 2015). As a consequence, opioid administration decreases the electrophysiological activity of tVTA neurons (Jalabert et al, 2011; Kaufling and Aston-Jones, 2015; Lecca et al, 2011; Matsui and Williams, 2011), which enhances the activity of VTA dopaminergic neurons (Jalabert et al, 2011; Matsui and Williams, 2011). Under chronic morphine, partial tolerance develops (Matsui et al, 2014), and animals display altered activity of tVTA and VTA neurons, which is reversed by acute precipitated withdrawal (Kaufling and Aston-Jones, 2015). Following chronic morphine exposure, adaptive molecular changes that participate to opponent processes and were described in other brain regions (Nestler, 2015) might also be present in the tVTA and explain the c-Fos expression observed in MOR-expressing cells. This induction could thus be linked to a rebound induced by naloxone in the activity of MOR-expressing cells. On the contrary, the expression observed in MOR-non-expressing cells is more likely to be indirect, through tVTA inputs and even polysynaptic circuits. In this regard, the LHb has, for example, been shown to participate to tVTA Fos induction in response to other stimuli (Brown and Shepard, 2013; Cui et al, 2014). Notably, tVTA lesion had no impact on the physical signs of opiate withdrawal. However, it does not preclude its involvement in the aversiveness of opiate withdrawal.

Despite observations of c-Fos induction with a series of foot shocks or with precipitated opiate withdrawal, the lack of response to most tested conditions raises the question of the exact role of the tVTA in avoidance responses (see Supplementary Table S3 in Supplementary Data). The optogenetic stimulation of the mouse LHb recruits tVTA GABAergic neurons and induces conditioned place avoidance (Lammel et al, 2012). Moreover, the stimulation of terminals from the LHb in the tVTA induces passive and conditioned place avoidance, and can also decrease the reinforcing property of sucrose (Stamatakis and Stuber, 2012). Similarly, a direct S-AMPA stimulation of the tVTA induces conditioned place avoidance (Jhou et al, 2013). These findings from various groups converge to evidence that stimulation of the tVTA promotes avoidance behavior. However, this does not imply that the tVTA is necessary for the physiological processing of all information leading to avoidance behavior. Indeed, we observed that the absence of the tVTA does not prevent LiCl- and LPS-induced CTAs. It shows that some forms of conditioned aversion can develop and be expressed in the absence of information processing through the tVTA.

The tVTA discovery was seminal, fostering research from various research groups. With evidence pointing to the tVTA as a major inhibitory control of dopamine systems, these studies converge to highlight the influence of the tVTA on dopamine-related functions, from responses to drug of abuse, to reward-related behaviors, reward prediction error, and control of motor responses. Generalization in trying to attribute simple functions to this brain region, while attractive for the sake of clarity, may however become a pitfall and mislead future research. The complexity of tVTA roles probably mirrors the complexity of dopamine systems themselves, which is not yet fully elucidated. This work participates to present efforts in understanding those functions of the tVTA. Foot-shock data suggest that the tVTA may be sensitive to the repetition of discrete emotional events, which could be related to previous evidence of reward prediction error processing by this brain region. The avoidance-inducing consequence of stimulating the tVTA reflects one of the functions of this brain region. Present data, however, suggest that all aversive situations may not be necessarily processed through the tVTA, and that some avoidance behavior can develop in the absence of the tVTA.

FUNDING AND DISCLOSURE

This research was supported by the Centre National de la Recherche Scientifique, Université de Strasbourg and its Neuropôle, and Agence Nationale de la Recherche (ANR-11-sv4-002). The authors declare no conflict of interest.