Le lithium est le traitement de référence du trouble bipolaire. Cependant, il peut être responsable d’intoxications dont trois profils différents sont décrits chez l’homme, caractérisés par une expression variable encore inexpliquée de la neurotoxicité. Les objectifs de cette thèse étaient d’étudier la distribution cérébrale du lithium dans trois modèles de rats Sprague-Dawley reproduisant les modes d’intoxication au lithium chez l’homme et de déterminer son implication dans la survenue des complications neurologiques au moyen de tests comportementaux et d’une analyse électroencéphalographique. Dans ce contexte, une étude des relations effet/concentrations du lithium suivant le modèle d’intoxication a été conduite. Enfin, un protocole d’imagerie ex vivo permettant d’explorer la distribution intracérébrale du lithium a été mis en place chez le rat par résonance magnétique nucléaire du lithium-7. Ainsi, nous avons montré que les trois formes d’intoxication au lithium chez le rat diffèrent en termes de pharmacocinétique sanguine et cérébrale, mais aussi en termes d’intensité et de durée des effets neurocomportementaux. Nous avons mis en évidence une accumulation cérébrale de lithium significativement plus élevée suite à un prétraitement répété par lithium et majorée après induction d’une insuffisance rénale. Le surdosage de lithium était constamment à l’origine d’une hypolocomotion chez le rat, dont la profondeur et l’étendue apparaissaient être liées à la durée de l’exposition au lithium, et d’une encéphalopathie dont la sévérité apparaissait plutôt dépendre de la quantité de lithium cérébral accumulée. Ainsi, l’accumulation cérébrale de lithium pourrait engendrer des effets neurotoxiques directs et/ou indirects par modification de l’expression de cibles cérébrales spécifiques du lithium. Enfin, nous avons démontré la faisabilité et la fiabilité de notre technique d’imagerie ex vivo pour explorer la distribution cérébrale du lithium chez le rat, ouvrant dès lors des perspectives à son utilisation future chez l’homme / Lithium is the cornerstone treatment of bipolar disorder. However, lithium may be responsible for poisoning with three various profiles reported in humans and characterized by unexplained variable resulting neurotoxicity. Our objectives were to investigate brain lithium distribution in three Sprague-Dawley rat models mimicking the human intoxication patterns and define its involvement in the occurrence of neurological disorders using behavioral tests and electroencephalographic analysis. The effect/concentration relationships were studied according to the poisoning model. Finally, an ex vivo imaging protocol was established in the rat to investigate brain lithium distribution using the nuclear magnetic resonance of lithium-7. We showed significant differences between the three lithium poisoning patterns in the rat regarding the blood and brain lithium pharmacokinetics as well as the intensity and duration of lithium-induced neurobehavioural effects. We found significantly more marked brain lithium accumulation after an overdose following repeated lithium administration, enhanced after the induction of renal failure. In the rat, lithium overdose consistently induced hypolocomotion whose intensity was related to the duration of lithium exposure and encephalopathy whose severity rather depended on the lithium amount accumulated in the brain. Brain lithium accumulation seems thus able to generate direct and/or indirect neurotoxic effects mediated by the alteration of specific brain lithium target expression. Finally, we demonstrated the feasibility and reliability of our ex vivo imaging technique to investigate brain lithium distribution in the rat, supporting a possible future use in humans

<p>Cetirizine, an H₁-antihistamine, is prescribed for allergic disorders. It exists as a racemic mixture, with levocetirizine being the active enantiomer. The central nervous system side-effects of H₁-antihistamines are caused by their penetration into the brain. In this thesis the plasma pharmacokinetics, transport across the blood-brain barrier (BBB) and H₁ receptor occupancy of cetirizine enantiomers was investigated <i>in vivo</i> in guinea pigs. The transport across the BBB was quantified using the microdialysis technique. Stereoselective brain distribution was investigated by measuring both unbound and total concentrations in plasma and brain. The time aspects of the H₁ receptor occupancy of levocetirizine was studied in the brain and the periphery.</p><p>The plasma pharmacokinetics of cetirizine was stereoselective with clearance and volume of distribution of levocetirizine being approximately half that of dextrocetirizine. This was mainly due to the differences in plasma protein binding of the enantiomers. The stereoselectivity in brain distribution indicated by the partition coefficient K_p (total AUC ratio brain to plasma) was caused by stereoselective plasma protein binding. The transport across the BBB measured in this thesis by the unbound partition coefficient K_p,uu (unbound AUC ratio brain to plasma) was the same for the two enantiomers. Binding within the brain was also not significantly different. The H₁ receptor occupancy of levocetirizine in brain lagged behind the plasma concentrations whereas it was not delayed with respect to the brain concentrations. This indicates that the delayed brain H₁ receptor occupancy of levocetirizine is caused by a slow transport across the BBB.</p><p>In summary, the results of this thesis emphasize the importance of measuring both the unbound and total concentrations in blood and brain to characterize stereoselective brain distribution. The thesis also emphasize the importance of taking local brain pharmacokinetics into consideration in understanding pharmacokinetic-pharmacodynamic relationships of drugs with central activity.</p>

<p>The pharmacokinetics and pharmacodynamics of oxycodone and morphine was investigated and related to the transport across the blood-brain barrier (BBB) in rats. The influence of a P-glycoprotein (P-gp) inhibitor on the plasma pharmacokinetics and pharmacodynamics of oxycodone was evaluated. Microdialysis experiments were conducted to evaluate the unbound pharmacokinetics, including the rate and extent of transport across the BBB, of oxycodone and morphine. Mathematical models were used to assess the pharmacokinetics and also the relationship between pharmacokinetics and pharmacodynamics of the drugs.</p><p>Oxycodone clearance, volume of distribution at steady-state, half-life, total brain tissue concentrations and tail-flick latency were all unaffected when a P-gp inhibitor was co-administered with oxycodone as compared to a control group. The lack of differences between the groups indicates that oxycodone BBB transport is not affected by P-gp inhibition. Investigating the unbound concentrations of oxycodone in brain and blood using microdialysis revealed an exciting finding. At steady-state, the unbound concentration in brain was 3 times higher than in blood (i.e. a K_p,uu of 3), indicating that active influx is involved in the BBB transport of oxycodone. In contrast, the K_p,uu of morphine was estimated to 0.56, which is an indication that active efflux mechanisms are involved in the BBB transport of morphine. This means that based on the same unbound concentration in blood, an approximately 6-fold higher unbound concentration of oxycodone compared to morphine will be reached in the brain. Using pharmacokinetic-pharmacodynamic modelling, the unbound brain concentrations of oxycodone and morphine were correlated to the tail-flick latency in vivo. The relative potency of the drugs was found to be concentration dependent with an infliction point of 55 nM.</p><p>In summary, this thesis emphasise the importance of taking the local brain pharmacokinetics into consideration when investigating the pharmacokinetics and the pharmacokinetic-pharmacodynamic relationships of centrally acting drugs.</p>

Cetirizine, an H1-antihistamine, is prescribed for allergic disorders. It exists as a racemic mixture, with levocetirizine being the active enantiomer. The central nervous system side-effects of H1-antihistamines are caused by their penetration into the brain. In this thesis the plasma pharmacokinetics, transport across the blood-brain barrier (BBB) and H1 receptor occupancy of cetirizine enantiomers was investigated in vivo in guinea pigs. The transport across the BBB was quantified using the microdialysis technique. Stereoselective brain distribution was investigated by measuring both unbound and total concentrations in plasma and brain. The time aspects of the H1 receptor occupancy of levocetirizine was studied in the brain and the periphery. The plasma pharmacokinetics of cetirizine was stereoselective with clearance and volume of distribution of levocetirizine being approximately half that of dextrocetirizine. This was mainly due to the differences in plasma protein binding of the enantiomers. The stereoselectivity in brain distribution indicated by the partition coefficient Kp (total AUC ratio brain to plasma) was caused by stereoselective plasma protein binding. The transport across the BBB measured in this thesis by the unbound partition coefficient Kp,uu (unbound AUC ratio brain to plasma) was the same for the two enantiomers. Binding within the brain was also not significantly different. The H1 receptor occupancy of levocetirizine in brain lagged behind the plasma concentrations whereas it was not delayed with respect to the brain concentrations. This indicates that the delayed brain H1 receptor occupancy of levocetirizine is caused by a slow transport across the BBB. In summary, the results of this thesis emphasize the importance of measuring both the unbound and total concentrations in blood and brain to characterize stereoselective brain distribution. The thesis also emphasize the importance of taking local brain pharmacokinetics into consideration in understanding pharmacokinetic-pharmacodynamic relationships of drugs with central activity.

The pharmacokinetics and pharmacodynamics of oxycodone and morphine was investigated and related to the transport across the blood-brain barrier (BBB) in rats. The influence of a P-glycoprotein (P-gp) inhibitor on the plasma pharmacokinetics and pharmacodynamics of oxycodone was evaluated. Microdialysis experiments were conducted to evaluate the unbound pharmacokinetics, including the rate and extent of transport across the BBB, of oxycodone and morphine. Mathematical models were used to assess the pharmacokinetics and also the relationship between pharmacokinetics and pharmacodynamics of the drugs. Oxycodone clearance, volume of distribution at steady-state, half-life, total brain tissue concentrations and tail-flick latency were all unaffected when a P-gp inhibitor was co-administered with oxycodone as compared to a control group. The lack of differences between the groups indicates that oxycodone BBB transport is not affected by P-gp inhibition. Investigating the unbound concentrations of oxycodone in brain and blood using microdialysis revealed an exciting finding. At steady-state, the unbound concentration in brain was 3 times higher than in blood (i.e. a Kp,uu of 3), indicating that active influx is involved in the BBB transport of oxycodone. In contrast, the Kp,uu of morphine was estimated to 0.56, which is an indication that active efflux mechanisms are involved in the BBB transport of morphine. This means that based on the same unbound concentration in blood, an approximately 6-fold higher unbound concentration of oxycodone compared to morphine will be reached in the brain. Using pharmacokinetic-pharmacodynamic modelling, the unbound brain concentrations of oxycodone and morphine were correlated to the tail-flick latency in vivo. The relative potency of the drugs was found to be concentration dependent with an infliction point of 55 nM. In summary, this thesis emphasise the importance of taking the local brain pharmacokinetics into consideration when investigating the pharmacokinetics and the pharmacokinetic-pharmacodynamic relationships of centrally acting drugs.

La barrière hémato-encéphalique (BHE), localisée au niveau des capillaires cérébraux, contrôle les échanges entre le sang et le compartiment cérébral et assure ainsi le maintien de l'homéostasie du système nerveux central (SNC). La présence de la BHE est un atout lors du développement de médicament à visée périphérique. En effet, en limitant le passage de nombreuses molécules, la BHE protège le SNC des effets potentiellement neurotoxiques de ces molécules. Toutefois, l‟exposition des cellules endothéliales des capillaires cérébraux à des agents chimiques est susceptible d‟engendrer une augmentation transitoire de la perméabilité de la BHE. Cette augmentation peut perturber l‟homéostasie cérébrale et permettre l‟entrée massive de molécules potentiellement neurotoxiques dans le SNC. La prise en compte de la BHE en amont de l‟étude de la neurotoxicité d‟un médicament est donc un élément important. De plus, la majorité des médicaments sont utilisés de façon chronique et les effets secondaires indésirables résultant d‟une administration chronique sont fréquemment liés à une atteinte cérébrale. Afin de répondre à cette problématique, notre modèle in vitro de BHE, qui consiste en une co-culture de cellules endothéliales de capillaires cérébraux et de cellules gliales, a été adapté à l‟étude de la toxicité de molécules lors d‟un traitement prolongé. Les propriétés protectrices de la BHE deviennent une contrainte importante lors du développement de médicament à visée cérébrale. En effet, la présence de la BHE explique en partie les taux de succès très faibles des molécules lors du développement de médicaments à visée cérébrale. Afin de limiter les taux d‟échec, il est nécessaire de prédire efficacement la distribution cérébrale des composés en prenant en compte la BHE. Or, il est admis que l‟effet pharmacologique est lié à la concentration libre du médicament au niveau de sa cible. Ainsi, les nouvelles approches visent à prédire la concentration libre que la molécule atteindra dans le cerveau. Toutefois, les méthodes existantes pour prédire ce paramètre reposent sur une méthodologie in vivo et ne présentent pas un débit suffisant pour être utilisées lors des phases précoces du développement de médicaments. Une méthodologie in vitro pour obtenir le ratio de concentrations libres d‟une molécule entre le cerveau et le sang a été développée pour répondre à ce besoin. Le travail réalisé a permis de développer deux méthodologies in vitro. La première permet de prédire la toxicité chronique des molécules. En prédisant le ratio des concentrations libres entre le compartiment cérébral et sanguin des composés, la seconde facilite la sélection des médicaments candidats lors du développement de médicaments à visée cérébrale. Ces méthodologies pourront donc contribuer à diminuer les taux d‟échecs lors des phases précliniques et cliniques du développement de médicaments. / The blood-brain barrier (BBB), located at the level of brain capillaries, is responsible for brain homeostasis maintenance by tightly controlling blood-borne substances access to the brain. The presence of the BBB is an asset during peripheral drug development. Indeed, the BBB protects the central nervous system (CNS) against potential neurotoxic effects of compounds by strongly limiting their passage. However, exposure of brain capillaries endothelial cells to chemical agents is likely to cause a transient increase in BBB permeability. This increase can disrupt brain homeostasis and allow the massive entry of potentially neurotoxic molecules in the CNS. Hence, taking into account BBB toxicity in alternative neurotoxicity studies is important. In addition, the CNS side effects of several drugs used chronically could be at least partly attributed to their toxicity at the level of the BBB causing unwanted, indirect effect on brain cells. To address this issue, our in vitro BBB model, which consist of a co-culture of brain capillary endothelial cells and glial cells, has been adapted to the evaluation of repeated-dose toxicity at the BBB. The protective properties of the BBB become a major hurdle during CNS drug development. One way to reduce theimportant attrition rate, consists in predicting the CNS distribution of drug candidates early in CNS drug discovery programs. The use of unbound brain concentrations has been shown to provide the best correlations with pharmacological data. Hence, new approaches aim to predict the free brain concentration of compounds. However, the determination of free brain / free plasma ratios requires both in vitro and in vivo experiments that are both animal and time consuming. Consequently, we have explored the possibility to directly generate free brain / free plasma ratios under steady-state and non-steady state conditions in our in vitro BBB model, thereby greatly simplifying existing experimental procedures.. The work presented herein aimed to develop two in vitro methodologies. The first one allows the study of repeated-dose BBB toxicity. The second one allows free brain / free plasma ratios assessment using an in vitro model of the blood brain barrier, which can drive the selection of CNS drug candidates with the most favourable target engagement. The use of these two methodologies may help to reduce attrition rates in drug discovery and development by appreciating the eventual central toxicity of systemic drug associated with BBB dysfunction and by identifying centrally acting-compounds with a desirable in vivo response in the CNS early on in the drug discovery process.

Les glioblastomes sont les tumeurs cérébrales les plus fréquentes avec une incidence en France de l'ordre de 4 nouveaux cas par an et pour 100 000 habitants (2400/ an). Le traitement standard pharmacologique des glioblastomes nouvellement diagnostiqués consiste en première ligne en une administration de témozolomide (75 mg/m2/j) pendant la radiothérapie, suivie d’une consolidation de 6 cycles. Cependant, malgré ce traitement, la médiane de survie n’est que de 15 mois et de 3 à 9 mois pour les rechutes. De nouvelles approches thérapeutiques sont donc indispensables. Parmi elles, ont été évalués le recours à d’autres chimiothérapies (irinotecan) et à l’inhibition de l’angiogénèse. L'angiogenèse est en effet un processus critique dans la progression GBM. L'inhibition de l'angiogenèse, induisant une réduction des vaisseaux sanguins, permet une diminution de l’apport des nutriments et d'oxygène à la tumeur. De manière générale, l’efficacité des traitements du glioblastome est soumise, dans un premier temps, à leur passage intra-cérébral au travers de la barrière hémato-encéphalique (BHE). L’objectif de notre travail était d’une part d’étudier l’impact de l’expression du transporteur d’efflux ABCB1 sur la distribution cérébrale du témozolomide (TMZ) et de l’irinotecan (CPT-11), et d’autre part, d’évaluer le rôle du bevacizumab (BVZ)(inhibiteur de l’angiogénèse) dans la modulation du passage intra-cérébral et intra-tumoral du TMZ et du CPT-11. A l'aide d'une étude pharmacocinétique comparative chez les souris CF1 mdr1a (+/+) et les souris CF1 mdr1a (-/-), nous avons mis en évidence un efflux actif du TMZ, du CPT-11 et de son métabolite actif le SN-38 du cerveau vers le plasma, impliquant ABCB1 au niveau de la BHE. Nous avons également démontré in vivo que le TMZ s'accumule dans la tumeur cérébrale et que le prétraitement par BVZ augmente la distribution tumorale de TMZ. En revanche, le BVZ n’a montré aucun effet sur la distribution cérébrale ou tumorale du CPT-11. Le BVZ apparaitrait donc comme un moyen intéressant d’augmenter la distribution intratumorale du TMZ. Dans ce même objectif, une collaboration initiée dans le cadre de cette thèse a permis de mettre en évidence l’intérêt de l’utilisation d’ultrasons dans l’ouverture de la BHE et l’amélioration de la distribution cérébrale des médicaments. / Glioblastomas are the most common brain tumors occurring in France with an incidence of 4 new cases per year per 100 000 population (2400/year). The gold standard pharmacological treatment of newly diagnosed glioblastoma relies on temozolomide administration (75 mg/m2/d) concomitant to radiotherapy, followed by six cycles consolidation. However, despite this treatment, the median survival is only 15 months and relapse occurs within 3 to 9 months. New therapeutic approaches are needed. Among them, other chemotherapies (irinotecan) and inhibition of angiogenesis were explored. Angiogenesis is a critical process in GBM progression. Inhibition of angiogenesis, inducing a reduction of the blood vessels, reduces supply of nutrients and oxygen to the tumor. The effectiveness of GBM treatment is subjected to intra-brain diffusion through the blood-brain barrier. The objective of this study was firstly to study the impact of efflux transporter ABCB1 brain expression on temozolomide (TMZ) and irinotecan (CPT-11) brain distribution, and secondly, to assess the role of bevacizumab (BVZ)(angiogenesis inhibitor) in the modulation of TMZ and CPT-11 brain and tumor distribution. Using a comparative pharmacokinetic study in CF1 mdr1a (+/+) and CF1 mdr1a (-/-) mice, we demonstrated an active efflux of TMZ, CPT-11 and its active metabolite SN-38 from the brain to the plasma involving ABCB1. We also demonstrated in vivo that TMZ accumulates in brain tumor and BVZ pretreatment increased TMZ tumor distribution. However no effect of BVZ on CPT-11 brain or tumor distribution was evidenced. Therefore BVZ would appear to be an interesting way to increase TMZ tumor distribution. The same objective was pursued through a different approach using ultrasound unfocused to open the BBB (Carthera collaboration).