Les variables dépendantes de toutes les expériences étaient quantitatives. Toutes les analyses statistiques ont été réalisées avec des tests paramétriques. L’homoscédasticité et la normalité des données étaient vérifiées avant toute analyse. En cas de violation des conditions d’application, une transformation non linéaire, logarithmique ou racine carrée, était réalisée afin d’atteindre au minimum l’équivalence de variances. La normalité des données n’était pas la principale préoccupation de la violation des conditions d’application. En effet, en absence d’une importante asymétrie, le nombre suffisamment grand de souris utilisées permettaient de ne pas se préoccuper de la normalité de distribution de nos données. En effet, le théorème central limite stipule que lorsque l’échantillon est
122
de grande taille, la distribution d’échantillonnage tend vers une distribution normale. Les tests paramétriques sont robustes et résistent aux écarts minimes de normalité. De plus, la grande majorité de nos tests statistiques ont été réalisés sur des plans équilibrés. Selon Howel, dans les conditions des plans équilibrés et d’égalité des variances entre les groupes, la normalité est une préoccupation secondaire (Howell, 2008).
Globalement, toutes les expériences consistaient en une succession de la procédure de sensibilisation locomotrice suivie de la procédure du TSPT ou l’inverse. Le test t de Student ou l’ANOVA simple étaient systématiquement utilisés pour la première procédure. Lors de la seconde procédure, les données étaient analysées avec une ANOVA à deux facteurs. L’utilisation d’une ANOVA à trois facteurs n’a été utilisée que pour une seule expérience. Le test post-hoc de Student-Newman-Kheuls était utilisé à chaque fois que le test omnibus décelait une différence entre au moins deux moyennes.
Les graphiques ont été réalisées à l’aide du logiciel R (R Core Team, 2020). Les données sont représentées sous forme de moyennes plus ou moins l’erreur type de la moyenne.
123
124
125
Chapitre 7
Investigating the reciprocal relationships between locomotor sensitization to ethanol and PTSD-like clusters in DBA/2J mice
Thierry Matonda-ma-Nzuzi1,2,3*, Vincent Didone1,*, Vincent Seutin2, Ezio Tirelli1, Étienne Quertemont1
1Psychology & Neuroscience of Cognition – PsyNCogn, ULiège, Belgium
2Laboratory of Neurophysiology, GIGA Neurosciences, all at Liège University, B-4000 Sart Tilman, Liège
3Département de psychiatrie, Faculté de médecine, Université de Kinshasa, Democratic Republic of the Congo
*These authors contributed equally to the work.
Published in Behav Brain Res (2019). DOI: 10.1016/j.bbr.2019.111909.
126
Abstract
Background: Post-traumatic stress disorder (PTSD) and alcohol use disorder (AUD) are two conditions that co-occur frequently. The mechanistic explanations of this co-morbidity are still unclear. The goal of this study was twofold. First to investigate whether PTSD reduces the threshold for the acquisition of ethanol sensitization in an animal model of PTSD. Then to investigate whether ethanol sensitization modulates the expression of PTSD.
Methods: 152 female inbred DBA/2J mice were submitted to an inescapable footshock paradigm to induce a PTSD-like condition (PTSDLC) and to a paradigm of locomotor sensitization to ethanol. In a first experiment, mice were submitted to the PTSDLC and then repeatedly injected with either saline, 1 g/kg ethanol or 2 g/kg ethanol. Their sensitization to the locomotor stimulant effects of ethanol was then tested in an open field. In a second experiment, mice were first sensitized to the locomotor stimulant effects of ethanol and then tested for their behavioral response to PTSDLC.
Results: In the first experiment, PTSDLC failed to induce a significant locomotor sensitization at the subthreshold dose of 1g/kg ethanol. However, with 2 g/kg ethanol, a stronger ethanol sensitization was observed in mice submitted to the footshock relative to the control group. In the second experiment, ethanol sensitization increased only some of the behavioral clusters of PTSDLC, namely the fear generalization in a new context.
Conclusion: PTSDLC did not reduce the dose threshold for the acquisition of ethanol sensitization but strengthened the development of ethanol sensitization with effective doses. This suggests that PTSD might interact with one of the mechanisms underlying the development of alcohol sensitization.
When the relationship between ethanol sensitization and PTSDLC is tested in the reverse direction, the present study only shows a significant effect of ethanol administration on the “sensitized fear”
PTSD cluster.
Key words: post-traumatic stress disorder; sensitization to ethanol; mice
127
7.1. Introduction
Post-traumatic stress disorder (PTSD) and substance use disorder (SUD), particularly alcohol use disorder (AUD), are two psychiatric conditions that frequently co-exist in patients. Nearly half of the persons who suffer from PTSD have SUD [1] and conversely PTSD occurs in up to 42.5% persons who suffer from SUD [2]. The relationships between PTSD and AUD still remain unclear, although it is known that each condition worsens the other [3,4]. Human studies on reciprocal relationships between PTSD and SUD are usually based on self-reported histories [3] and are therefore limited by memory biases and possibly by impaired cognitive functions [5]. Also, most human studies include subjects with two or more SUD [3,6- 8] which can sensitize each other (cross-sensitization) and make the interpretation of results difficult. This is why animal studies can help to better understand this comorbidity.
The incentive sensitization theory [9] is one of the leading paradigms used to explain findings in animal models of SUD. It states that drugs of abuse have the capacity to induce brain changes. These cerebral alterations disrupt the brain motivational processes that in turn give salience to these substances. This phenomenon is called sensitization. In practice, sensitization to a substance of abuse can be measured, inter alia, by the enhancement of the locomotor activity induced by repeated administration of the drug at the same dose. However, acting concomitantly with others factors (e.g.
administration pattern, rodent strain, age, context, etc.), a drug’s ability to induce locomotor sensitization is dose-dependent [10-12]. In mice, the sensitizing ethanol dose is usually equal or above 1.5 g/kg [13,14]. In rodent studies, drug-induced locomotor sensitization is considered as homologous to the intensification of drug craving in addicts and the paradigm is therefore used to model various features of drug addiction, including its brain bases. For example, several studies reported correlations between ethanol locomotor sensitization and activation of the reward system, among others through dopaminergic and glutamatergic neurons, with an increased release of dopamine [11,15,16].
In humans, PTSD is defined by the occurrence of four symptom clusters that occur in the aftermath of an extreme stress: (1) intrusion defined as re-experiencing of the traumatic event in the form of spontaneous flashback memories or recurrent vivid dreams, (2) hyperousal characterized by difficulties sleeping and concentrating and hypervigilance, (3) avoidance of trauma reminders and (4) numbing of affects [17]. In animal studies, fear conditioning, a pavlovian conditioning, is the main
128
paradigm used to model human PTSD especially in rodents. One of the most used stressors is a brief inescapable and unpredictable electric footshock. Although clearly far from perfect, this model of PTSD has shown face and predictive validity [18]. In previous studies, the fear conditioning procedure was also used to mimic the symptom clusters of PTSD in rodents [18]. Rodents are tested for conditioned fear long after the electrical footshock in the same environment. The persistence of fear reactions (freezing and increased startle) is used to mimic the intrusion cluster of PTSD.
Generalization of fear reactions in a new context, i.e. sensitized fear, is used to mimic the hyperarousal cluster of PTSD. A neutral tone is used in the new context in order to test for the generalization of sensitized fear to a new stimulus. However in the present study, freezing is recorded both before and during the neutral tone because previous studies reported that some strains of mice are more reactive to the stimulus, whereas other strains show higher levels of freezing in the new context before the tone [19]. Finally, conditioned fear is often paired with procedures used to measure anxiety-like behaviors, such as the elevated plus-maze or the black and white box. They are used to mimic the avoidance cluster of the PTSD [18,20]. From here on, we will use the term “PTSD-like condition” (PTSDLC) to describe findings in rodents.
Several animal studies have shown that stress and/or PTSDLC increase both locomotor sensitization to ethanol and alcohol consumption in rodents [21-23]. However, the mechanisms of this relationship remain poorly understood. One hypothesis to explain this modulation by PTSDLC could be by the reduction of the dose necessary to induce a transition to AUD. The first aim of the present study is to test this hypothesis using the “locomotor sensitization” paradigm. We will evaluate whether PTSDLC can elicit ethanol-induced locomotor sensitization with a dose that is unable to induce locomotor sensitization to ethanol in control conditions.
Most of the studies on the PTSD-AUD comorbidity are based on the self-medication theory [3,4].
Therefore, they have focused on the influence of PTSD on the subsequent development of AUD. Few studies have focused on the inverse relationship and even fewer have focused on how SUD affects the PTSD clusters [3,6-8]. In general it is observed that SUD worsens the PTSD clusters [3,4,6-8]. But results often are controversial. For example, hyperarousal and avoidance are overexpressed in patients with AUD according to some studies [6,8] but not to others [7]. One of the few animal studies assessing the effects of prior drug administration on the later development of PTSD symptoms, shows that chronic administration of cocaine increases the hyperarousal induced by single prolonged stress
129
in rats [24]. The reason for these discrepancies is unclear, but may involve a variable interference of the drug withdrawal syndrome. To our knowledge no animal studies have focused on the influence of locomotor sensitization to ethanol on the subsequent development of PTSDLC.
The goals of the present study were twofold: (i) to assess whether PTSDLC modulates the later development of ethanol-induced locomotor sensitization; (ii) to assess whether ethanol-induced locomotor sensitization modulates the subsequent development of PTSDLC. We hypothesized that PTSDLC can reduce the threshold of acquisition and expression of ethanol-induced locomotor sensitization. Secondly, we postulated that ethanol-induced locomotor sensitization would modulate the subsequent development of a PTSDLC.