Il a été clairement démontré qu’une altération de la neurotransmission sérotonergique n’est pas impliquée dans le risque suicidaire chez les patients schizophrènes, contrairement à ce qui est retrouvé dans la population générale. L’effet bénéfique de la clozapine, un ASG, vis-à-vis du risque suicidaire chez les individus schizophrènes nous amène à nous demander quels mécanismes sont impliqués dans le risque suicidaire chez les patients schizophrènes et dans l’effet protecteur de la clozapine. Ainsi, une meilleure compréhension du phénomène chez les individus schizophrènes permettra d’offrir un meilleur traitement à ces patients. Étant donné que les études cliniques chez les patients schizophrènes comportent plusieurs limites et que le diagnostic du risque suicidaire chez ces gens représente un réel défi pour les psychiatres, il devient nécessaire de développer un modèle in vivo de risque suicidaire dans un contexte de vulnérabilité à la schizophrénie.
2.2.1. Hypothèse de recherche
L’hypothèse de recherche du 2e projet repose encore sur la théorie à double atteinte de la schizophrénie et stipule que des perturbations durant la phase fœtale, ici une
activation immunitaire gestationnelle in vivo (1ère atteinte), rend le SNC plus vulnérable à des perturbations qui ont lieu durant la phase mature et qui sont associés au développement de comportements anxieux et agressifs, plus précisément via l’isolement social (2e atteinte). La combinaison de ces deux atteintes cause donc l’apparition de comportements associés au risque suicidaire dans un contexte de vulnérabilité à la schizophrénie (Figure 11).
Figure 11. Hypothèse à double atteinte pour le développement de comportements associés au risque suicidaire dans un modèle animal de schizophrénie. Des perturbations durant la phase fœtale, comme l’activation immunitaire prénatale (polyIC), rend le système nerveux central (SNC) vulnérable à des perturbations durant la phase mature, comme un isolement social, ce qui amène l’apparition de comportements associés au risque suicidaire dans un contexte de vulnérabilité à la schizophrénie.
2.2.2. Objectifs du projet
Les objectifs du projet sont de :
1- développer un modèle animal permettant l’étude des comportements associés au risque suicidaire dans un contexte préclinique de schizophrénie.
2- étudier de manière distincte les effets antipsychotiques et de « prévention de suicide » du lithium, un régulateur de l’humeur connu pour son efficacité de « prévention de suicide », et de la clozapine, un ASG avec l’effet le plus efficace de « prévention de suicide ».
Les objectifs du 2e projet sont remplis en suivant le protocole détaillé dans l’article 3 de la présente thèse (Deslauriers et al., soumis). Brièvement, chez la souris C57BL/6, une
injection IP de saline ou de polyIC (20 mg/kg) est effectuée au 12e jour de gestation (G12). La naissance des souriceaux survient au 19e ou 20e jour de gestation, ce qui correspond alors au jour postnatal 0 (PN0). Le sevrage est effectué après 3 semaines (PN21) et l’isolement social (ou non) débute pour 4 semaines. Quotidiennement, durant la dernière semaine d’isolement social (PN42-PN49) et 30 minutes avant chaque test comportemental (PN50-PN55) décrit dans l’article 3 (IPP, comportement exploratoire (Open Field), comportement de type anxieux (Elevated Plus Maze), perte d’espoir (Forced Swim Test) ou interactions sociales/agressivité/impulsivité (test du résident-intrus)), une injection IP de véhicule, de chlorure de lithium (200 mg/kg) ou de clozapine (3 mg/kg) est effectuée (Figure 12).
Figure 12. Protocole in vivo du deuxième projet. G, jour de gestation; IPP, inhibition du réflexe de sursaut par stimulus sonore; PN, jour postnatal, polyIC, acide polyinosinique:cytidylique.
ARTICLE 1
Combination of prenatal immune challenge and restraint stress affects prepulse inhibition and dopaminergic/GABAergic markers
Auteurs de l’article: Jessica Deslauriers, Annie Larouche, Philippe Sarret, Sylvain Grignon
Statut de l’article: Publié dans Progress in Neuropsychopharmacology & Biological Psychiatry (2013)
Avant-propos: Cette publication porte sur la caractérisation d’un des deux modèles à double atteinte décrits dans cette thèse. J’ai écrit en entier la première version du manuscrit, en plus de faire entièrement l’analyse et l’interprétation des résultats, sous la supervision de Sylvain Grignon et Philippe Sarret. J’ai aussi fait entièrement les révisions demandées par l’éditeur. En plus, d’avoir mis au point le modèle animal, j’ai effectué 90% des manipulations.
Résumé:
L’activation immunitaire gestationnelle avec le poly I:C, un antigène de type viral, constitue un modèle in vivo neurodéveloppemental connu de schizophrénie. Cependant, une réponse inflammatoire de la mère durant la grossesse peut augmenter la vulnérabilité du cerveau fœtal en développement à des expositions secondaires. Afin de mieux comprendre la pathophysiologie de la schizophrénie, nous avons développé un modèle animal à double atteinte basé sur l’activation immunitaire prénatal avec le polyIC, suivie d’un stress de contention chez le souriceau à l’âge juvénile. Des souris C57BL/6 ont reçu une injection IP de poly I:C ou de saline au jour gestationnel 12. Après la naissance, les souriceaux subissent, ou non, un stress de contention durant 2 heures par jour, pendant 3 jours consécutifs (jours postnataux 33 à 35). L’inhibition du réflexe de sursaut par stimulus sonore (IPP ou PPI, de l’anglais prepulse inhibition) est couramment mesurée pour évaluer le couplage sensorimoteur, un processus cérébral souvent perturbé chez les patients schizophrènes. Nos résultats ont révélé que, alors que chaque insulte n’a aucun effet, la combinaison de l’activation immunitaire prénatale, avec le polyIC, et du stress de contention à l’âge juvénile induit un déficit de l’IPP chez les souriceaux âgés de 36 jours. En parallèle au déficit de l’IPP, on observe des anomalies dopaminergique et GABAergique dans le cortex préfrontal et le striatum. En effet, la combination de l’exposition gestationnelle au polyIC et du stress de contention postnatal induit une augmentation des niveaux d’ARNm et protéique du récepteur dopaminergique de type 2 (DRD2). De plus, la combination des deux atteintes réduit l’expression de l’ARNm et protéique de la forme 67 kDa de l’acide glutamique décarboxylase (GAD67), dans les mêmes régions cérébrales. À notre connaissance, ce modèle animal à double atteinte est le premier modèle in vivo présentant un déficit d’IPP à l’âge pubertaire. Le modèle à double atteinte présenté dans l’article pourra aider à étudier de nouvelles thérapies innovantes pour améliorer le traitement de la schizophrénie, spécialement durant la phase prodromale.
Abstract
Gestational immune challenge with the viral-like antigen poly I:C is a well- established neurodevelopmental model of schizophrenia. However, exposure to inflammation during early life may sensitize the developing brain to secondary insults and enhance the central nervous system vulnerability. To gain a better understanding of the pathophysiology of schizophrenia, we thus developed a two-hit animal model based on prenatal poly I:C immune challenge followed by restraint stress in juvenile mice. C57BL/6 gestational mice were intraperitoneally injected with poly I:C or saline at gestational day 12. Pups were then submitted or not, to restraint stress for two hours, for three consecutive days, from postnatal days 33 to 35. Prepulse inhibition (PPI) of acoustic startle response is commonly used to assess sensorimotor gating, a neural process severely disrupted in patients with schizophrenia. Our results revealed that the combination of prenatal immune challenge with poly I:C followed by a restraint stress period was able to induce a PPI disruption in 36-day-old pups, as opposed to each insult applied separately. PPI deficits were accompanied by dopaminergic and GABAergic abnormalities in the prefrontal cortex and striatum. Indeed, measurements of cortical and striatal dopamine D2 receptor (D2R) mRNA and protein levels revealed that the combination of gestational exposure to poly I:C
and postnatal restraint stress induced an increase in D2R protein and mRNA levels.
Likewise, the combination of both insults reduced the mRNA and protein expression levels of the 67 kDa form of glutamic acid decarboxylase (GAD67), in those two brain regions. To our knowledge, this two-hit animal model is the first in vivo model reporting PPI deficits at pubertal age. This two-hit animal model may also help in studying innovative therapies dedicated to the treatment of schizophrenia, especially in its early phase.
1. Introduction
Schizophrenia is a neuropsychiatric disorder affecting 1% of the general population. Manifestations of illness are characterized by three core symptoms: positive (delusion, hallucinations, disturbances of thoughts and paranoia), negative (lack of motivation, alogia and social withdrawal) and cognitive (attention and memory deficits) symptoms (Flaum and Andreasen, 1991). The neurodevelopmental hypothesis of schizophrenia posits that a cerebral insult during early (prenatal or perinatal) brain development increases the vulnerability for the subsequent emergence of clinical symptoms, manifesting itself between adolescence and early adulthood (Weinberger, 1987; Rapoport et al., 2005).
Schizophrenia is a significantly heritable disease and likely implicates multiple susceptibility genes, which are involved, among others, in neuronal differentiation, survival, apoptosis, as well as neurotransmission (Harrison and Weinberger, 2005). Beside genetic factors, the risk of schizophrenia is increased by environmental factors, which occurs at prenatal or early postnatal periods, two critical neurodevelopmental stages (McDonald and Murray, 2000; Opler and Susser, 2005). Among environmental factors, prenatal and perinatal environmental events, like maternal stress, maternal infection or obstetric complications, have been linked to an increased prevalence of schizophrenia (van Os et al., 2010).
Maternal viral infection during the second trimester of pregnancy has been associated with increased risk of schizophrenia (Torrey et al., 1988; O'Callaghan et al., 1994; Suvisaari et al., 1999; Brown et al., 2000; 2004). This phenomenon has been replicated in animal models and many studies have used prenatal influenza in mice as an experimental animal model of schizophrenia (Fatemi et al., 1999; 2000; 2002; Shi et al., 2003; Fatemi et al., 2004; 2008; 2009). Subsequently, the use of the synthetic double strand RNA, viral-like antigen, poly-inosinic/cytidylic acid (poly I:C) during the gestational period has also been established as a neurodevelopmental model of schizophrenia (Shi et al., 2003). Poly I:C is a toll-like receptor 3 (TLR3) agonist, mimicking the acute phase of viral infection by inducing inflammatory responses accompanied by the presence of pro- inflammatory cytokines, such as interleukin-1β (IL-1β), interleukin-6 (IL-6) and tumor necrosis factor-α (TNF-α) (Meyer et al., 2006b). Prenatal immune challenge with poly I:C
also induced several behavioral changes and neurochemical abnormalities in adult offspring (Zuckerman et al., 2003; Meyer et al., 2005; Zuckerman and Weiner, 2005; Ozawa et al., 2006; Smith et al., 2007; Meyer et al., 2008b). However, these behavioral alterations have been mostly reported in adult rodents. Conversely, schizophrenia typically begins during adolescence or early adulthood, especially in males (Picchioni and Murray, 2007; Eranti et al., 2013) while cognitive and functional decline, as well as prodromal symptoms, occur even earlier (Murray and Lewis, 1987; Weinberger, 1987; 1995; Gogtay et al., 2011).
Maynard and colleagues (Maynard et al., 2001) have suggested the “two-hits” hypothesis of schizophrenia, because of the late adolescent/early adulthood onset of the disease. Accordingly, genetic or environmental factors disrupt early central nervous system (CNS) development and these early insults produce long-term vulnerability to a “second hit”, that occurs later in life, therefore leading to the onset of overt schizophrenia symptoms. Accordingly, it has been suggested that maternal infection interacts with other etiologic factors to trigger or aggravate the development of schizophrenic symptoms (Boksa, 2008). Notably, stressors, such as life adversity, urban upbringing and psychostimulant or cannabis use, can influence neuronal responsiveness and result in prefrontal cortex (PFC) dysfunction, thus contributing to the development of cognitive deficits observed in schizophrenia (Arnsten, 2009).
Synaptic neurotransmission (i.e. dopaminergic, GABAergic and glutamatergic) in some brain circuits may be disturbed in schizophrenia pathophysiology (Maynard et al., 2001). Notably, it has been reported that early exposure to environmental factors increases mesolimbic dopaminergic transmission (Lieberman et al., 1997) and tends to elicit schizophrenia-like phenotypes in vivo (van Os et al., 2010). Additionally, there is now clear evidence, from neuroimaging studies, of increased presynaptic dopamine function in schizophrenia. The results regarding the levels of dopamine D2 receptor (D2R) are, however, more equivocal and likely influenced by previous antipsychotic treatments (Goldsmith et al., 1997; Laruelle, 1998; Howes et al., 2012; 2013). Reduced GABA synthesis has also been observed in schizophrenia (Akbarian et al., 1995). Specifically, expression of the 67 kDa form of glutamic acid decarboxylase (GAD67), the limiting enzyme in GABA synthesis has consistently been shown to be decreased in the
hippocampus (Benes et al., 2007) and the PFC of schizophrenic patients (Mirnics et al., 2000; Reynolds et al., 2001; Hashimoto et al., 2003; Torrey et al., 2005).
These considerations led us to hypothesize that prenatal and postnatal factors interact to play a role in the onset of schizophrenia symptoms in juvenile animals. Thus, we developed and characterized a two-hit model in mice, based upon the well established prenatal immune challenge with poly I:C (PIC) followed by juvenile restraint stress (RS). We report here that the combination of both insults induces an impairment in the prepulse inhibition (PPI) of acoustic startle, a measure of the sensorimotor gating function, which is altered in schizophrenia (Braff et al., 1999; 2001). We also found that these sensorimotor gating deficits are accompanied by neurochemical dopaminergic (dopamine D2 receptor up-regulation) and GABAergic abnormalities (decrease of GAD67 expression) in the PFC and striatum, reminiscent of the pathophysiology of schizophrenia.
2. Material and methods 2.1. Animals
Male and female C57BL/6 mice were obtained from Charles River laboratories (QC, Canada) at age of eight to ten weeks for mating. Animals were kept at 20oC environmental temperature and on a 10-h light/14-h dark cycle. They had ad libitum access to food and water. Experimental protocol was approved by the institutional Animal Research Ethics Review Board at the Université de Sherbrooke, in compliance with the policies of the Canadian Council on Animal Care. Mating was done on site. Timed pregnant mice were injected intraperitoneally (IP) with saline or 20 mg/kg poly I:C (Sigma- Aldrich, ON, Canada), as previously described at gestational day 12 (G12) (Shi et al., 2003), a dose associated with significant brain and behavioral impairments in adulthood but not in juvenile animals. On the postnatal day 22 (PN22), pups were weaned and, from postnatal days 33 to 35 (PN33-35), were submitted, or not, to restraint stress (RS) for 2 h per day, for 3 consecutive days, as previously described (Kim et al., 2005). Restraint stress was performed in a 50 mL conical tube. From 14 saline-treated dams (mean: 8.6 per litter; n = 90) and 17 poly I:C-treated dams (mean: 4.5 per litter; n = 100), the pups were randomized into four groups: (1) saline (n = 43), (2) saline with restraint stress (saline+RS) (n = 47), (3) poly I:C (n = 47) (PIC) and (4) poly I:C with restraint stress (PIC+RS) (n = 54). Also, pups from each litter were divided equally in each group and for behavioral assessment (n = 48) and for biochemical analysis (n = 71 and n = 72 for 3 h and 24 h after the last restraint stress period, respectively). There was no significant difference in male/female ratios between saline-treated and poly I:C-treated dams’ litters and all male and female pups were divided equally for behavioral observations and biochemical analysis.
2.2. Prepulse inhibition of acoustic startle reflex
Behavioral observations were done 24 h after the last restraint stress period and testing took place between 10:00 and 17:00. Prepulse inhibition (PPI) of acoustic startle was evaluated using the SR-LAB apparatus (San Diego Instruments, CA, USA). The animal enclosure (Plexiglas cylinder; 5" length x 1 ½ " height) was resting on a
piezoelectric transducer that detected the vibrations caused by the movements of the animals. As previously described (Swerdlow et al., 2008), testing began with a 5 min acclimatization period with 71 dB background noise, followed by 4 sessions. During the first session, 6 pulse alone trials (120 dB for 40 msec) were presented. For the second session, five pulse alone trials (120 dB for 40 msec), five null trials (no stimulation) and five prepulse+pulse trials at four different intensities were presented in a randomized order. The prepulse+pulse trials was a 20 msec prepulse of 75, 79, 83 or 87 dB (4, 8, 12 or 16 dB over background noise, respectively), followed by a 80 msec delay and a startle pulse (120 dB for 40 msec). A 60 sec break preceded the third session, which consisted again of five pulse alone trials (120 dB for 40 msec), five null trials (no stimulation) and five prepulse+pulse trials at the four different intensities presented in a randomized order. Finally, the last session was presented with 6 pulse alone trials (120 dB for 40 msec). PPI is evaluated as PPI %, calculated as (1-[startle amplitude on prepulse+pulse trial/mean startle amplitude on pulse alone trials])×100.
2.3. Biochemical analysis
For biochemical analysis, mice were sacrified by decapitation 3 h or 24 h after the last restraint stress period and the whole brain was kept and frozen at -80oC until use. To extract prefrontal cortex (PFC) and striatal tissues, 3 mm slices from bregma +4 mm to +1 mm and from bregma +1 mm to -2 mm (Franklin and Paxinos, 2008), respectively, were cut out in a brain matrix.
2.3.1. Real Time-Polymerase Chain-Reaction (RT-PCR)
As described previously (Deslauriers et al., 2011), RNA isolation from brain tissue was performed using TRIzol reagent (Invitrogen, ON, Canada), homogenized by mechanical disruption and quantified with Nanodrop spectrophotometer (Nanodrop Technologies, DE, USA). Reverse transcription was done as follows: RNA (1 µg) was incubated with 0.02% (w/v) random primer (Roche Applied Science, QC, Canada) for 5 min at 70oC followed by an incubation at 42oC for 60 min and at 94oC for 5 min with 1 mM dNTPs (Roche Applied Science), 20 U RNAse inhibitor (Invitrogen) and 30 U AMV reverse transcriptase (Roche Applied Science). Then, cDNA (1 µL) was amplified with
12.5 µL SYBRGreen (Qiagen, ON, Canada), 10.5 µL water and 0.5 µL of each primer. For D2R amplification, the primers were 5’-TATGCCCTGGGTCGTCTATC-3’ (forward) and 5’-AGGACAGGACCCAGACAATG-3’ (reverse) and, for GAD67 amplification, the primers were 5’-TCCTCCAAGAACCTGCTTTCCTGT-3’ (forward) and 5’- CCATGAGAACAAACACGGGTGCAA-3’ (reverse) (Integrated DNA Technologies, IA, USA). For the tyrosine hydroxylase (TH) amplification, the primers were 5’- TCTCCCAGGACATTGGACTT-3’ (forward) and 5’-CGGCTGGTAGGTTTGATCTT-3’ (reverse) (Integrated DNA Technologies, IA, USA). β-actin was used as the reference gene: 5’-TGGAATCCTGTGGCATCCATGAAAC-3’ (forward) and 5’- TAAAACGCAGCTCAGTAACAGTCCG-3’ (reverse) (Integrated DNA Technologies). The cycling parameters were 95oC 10 min followed by 40 cycles (95oC 15 sec, 60oC 30 sec and 72oC 45 sec).
2.3.2. Western blotting
Total protein were prepared from brain tissue extracts, as previously described (Deslauriers et al., 2011). RIPA buffer (20 mM Tris pH 7.5, 150 mM NaCl, 1 mM EDTA, 1.0% (v/v) Nonidet-P40, 0.5% (w/v) Na-deoxycholate and 0.1% (w/v) SDSm) supplemented with SIGMAFASTTM protease inhibitor cocktail 1X (containing 4-(2- aminoethyl)benzenesulfonyl fluoride hydrochloride (AEBSF), bestatin hydrochloride, leupeptin, E-64, aprotinin, pepstatin A, and phosphoramidon disodium salt) (Sigma- Aldrich) was added to brain tissue. Brain tissue was then homogenized by mechanical disruption, lysed on ice for 15 min and centrifugated at 13,000g for 15 min. The supernatant was kept and stored at -20oC until use. Protein concentrations were measured with Coomassie Plus (Bradford) Protein Assay Reagent (Thermo Fisher Scientific, IL, USA), using bovine serum albumin (BSA) as a standard. For western blotting, soluble proteins (10-30 µg) were migrated by SDS-PAGE electrophoresis (10% sodium dodecyl sulfate-polyacrylamide gel) for 90 min at 120 V. Proteins were then transferred onto immobilon-P polyvnylidene fluoride (PVDF) transfer membrane (Millipore, MA, USA) for 55 min at 120 mA. Membranes were blocked in 8% (w/v) milk powder in Tris-buffered saline (TBS) with 0.2% (v) Tween-20 (TBST) for 90 min at room temperature. Membranes were incubated with primary antibodies diluted in blocking solution at 4oC overnight.
Mouse monoclonal DRD2 (B-10) antibody (1:500) (Santa Cruz Biotechnology, CA, USA), mouse monoclonal GAD67 (K-87) antibody (1:5000) (Abcam Inc., MA, USA), mouse monoclonal TH (F-11) antibody (1:500) (Santa Cruz Biotechnology, CA, USA) and mouse monoclonal β-actin antibody (1:10 000) (Sigma-Aldrich) were used as primary antibodies. The last antibody was used as normalization control. Three washes of 10 min each with TBST were done and membranes were incubated with secondary antibody goat anti-mouse IgG-HRP (1:10 000) (Santa Cruz Biotechnology) for 90 min at room temperature. After three additional washes, membranes were developed using Western Lightning Chemiluminescence Reagent Plus (PerkinElmer, MA, USA) and Hyperfilm ECL (Amersham Biosciences, QC, Canada). Densitometric analysis was done using NIH ImageJ software.
2.4. Data analysis
For measured PPI, data were presented as mean ± SD and analyzed using three-way analysis of variance (ANOVA), evaluating the interaction of prepulse intensity, prenatal treatment and juvenile restraint stress, followed by Bonferroni multiple comparisons. For measured neurochemical changes, data were presented as mean ± SD and were analyzed using two-way ANOVA, evaluating the interaction of prenatal treatment and juvenile restraint stress, followed by Bonferroni multiple comparisons.
3. Results
3.1. Effects of prenatal immune challenge and restraint stress on prepulse inhibition of acoustic startle
In the present study, we examined whether the combination of prenatal immune challenge with poly I:C (PIC) followed by restraint stress (RS) period was able to induce a PPI disruption in juvenile mice. Four-way ANOVA (Sex × Prepulse × Prenatal treatment × Restraint stress) revealed that there was no significant effect of sex on prepulse inhibition; therefore the results of male and female offspring were collapsed. There was no significant effect of either gestational immune challenge or juvenile restraint stress or the combination of both conditions on startle amplitude (Fig. 1A). As expected, PPI increased as the prepulse intensity increased from 75 dB to 87 dB, regardless of the experimental condition. Indeed, three-way ANOVA (Prepulse × Prenatal treatment × Restraint stress) revealed a significant main effect of prepulse intensity on PPI in mice (F3,232 = 11.393; P < 0.0001) (Fig. 1B). For example, the startle response to a 120 dB acoustic stimulation in normal saline treated mice was inhibited by a prepulse acoustic stimulation at 75 dB by 31.68 ± 14.59 %. The magnitude of PPI was increased to 54.04 ± 13.00 %, when the prepulse acoustic stimulation was adjusted to 87 dB. Three-way ANOVA revealed that, as compared with prenatal injection of saline, juvenile restraint stress (RS) or prenatal immune challenge with PIC applied separately had a significant effect on PPI magnitude in 36-day-old pups (ANOVA: F1,140 = 4.848; P < 0.05 and F1,104 = 6.730; P < 0.01, respectively) (Fig. 1B). There was no interaction between prenatal immune challenge with polyI:C and juvenile restraint stress. However, Bonferroni comparisons revealed, in mice submitted to both insults (PIC+RS), a mean PPI impairment of 23%, as compared to prenatal saline alone (ANOVA: F1,116 = 14.08; P < 0.001), and 18%, as compared to juvenile restraint stress condition applied alone (ANOVA: F1,128 = 5.543; P < 0.05) (Fig. 1C). Specifically, in mice submitted to both conditions, PPI was significantly reduced to 27.19 ± 16.99 % (P < 0.05) and 40.96 ± 9.10 % (P < 0.01), respectively, in response to 79-120 dB and 87-120 dB