Background: Tissue non-specific alkaline phosphatase (TNAP) is an enzyme which functions within the body to catalyze the hydrolysis of pyrophosphate to phosphate, and is a well-known mediator of bone mineralization. It has also been identified as a positive mediator of intracellular lipid accumulation (ICLA) in both murine and human preadipocytes as well as in the hepatocellular cell line HepG2. However, the mechanism through which TNAP functions to control ICLA is not known. Both osteoblasts and adipocytes are both of mesenchymal origin and thus may share conserved mechanisms through which TNAP functions. Within bone, TNAP converts pyrophosphate (which inhibits mineralization) to phosphate. This phosphate is essential to the mineralization process through binding to hydroxyapatite crystals, and it also activates the transcription of genes whose products function in osteoblast differentiation, including NRF2. This thesis therefore aimed to determine the role of both pyrophosphate and TNAP-generated phosphate in ICLA. In addition, it is possible that TNAP may interact with other proteins, as it is known that TNAP is able to dephosphorylate proteins such as tau. This thesis therefore aimed to determine whether TNAP binds to other proteins in the context of ICLA. Lipids are not only stored within hepatocytes and adipocytes, but are also found in cells of the adrenal cortex, and TNAP is known to be expressed within such cells. Therefore, this thesis also aimed to determine whether TNAP is involved in the accumulation of cholesterol esters within lipid droplets in the adrenal cortex. Methods: To determine the effect of high intracellular pyrophosphate levels on ICLA, 3T3-L1 cells (a preadipocyte cell line) were cultured in the presence and absence of probenecid, an inhibitor of the pyrophosphate transporter ANK, and induced to accumulate lipids. Lipid accumulation was monitored through Oil red O staining. The effect of probenecid treatment on TNAP activity and intracellular pyrophosphate levels was also analysed. To determine whether TNAP functions in ICLA by producing phosphate for gene induction, 3T3-L1 cells were stimulated to undergo ICLA in the presence and absence of the TNAP inhibitor levamisole, which in turn blocks ICLA. Levamisole treated cells were also incubated with phosphate to see if this would overcome the inhibitory effect of levamisole on ICLA. The ability of phosphate to induce gene expression of NRF2 was determined through real-time PCR. In addition, an NRF2 expressing plasmid was transfected into cells treated with the TNAP inhibitor levamisole to determine if this would also overcome the block on ICLA caused by TNAP inhibition. In silico analysis identified TRAF2 as a potential binder of TNAP. The expression of TRAF2 during ICLA was determined through real time PCR, and the effect of overexpression of TRAF2 on intracellular lipid accumulation was determined through the transfection of a TRAF2 expressing plasmid in cells induced to undergo ICLA. To determine whether TNAP modulates lipid accumulation in cells of the adrenal cortex, the Y1 murine adrenocortical cell line was cultured in the presence and absence of TNAP inhibitor levamisole, and ICLA measured by Oil Red O staining. The location of TNAP within Y1 cells was identified by histochemical staining. Results: Cells treated with probenecid showed increased pyrophosphate levels (expressed as a % of levels observed at baseline) when compared to untreated controls (155.5 ± 15.1 % vs 51.1 ± 18.9 %; p=0.001) after 24 hours of culture. Increased pyrophosphate levels resulted in ICLA within 3T3-L1 cells surpassing levels seen in untreated controls (507.4 ± 30.4 % vs 337.6 ± 16.17 %; p=0.004). This increase in pyrophosphate was coupled to an increase in TNAP activity within the initial 24 hours (291.5 ± 72.8 % vs baseline of 100%; p=0.038) compared to that seen in control experiments (103.43 ± 24.3 % vs baseline of 100%; p=0.848). Cells treated with levamisole showed minimal ICLA and when exogenous phosphate was added, lipid levels were reconstituted to levels similar to that seen in cells induced to accumulate lipids in the absence of levamisole (284.01 ± 62.52% vs 275.86 ± 35.52%; p= 0.83). In the presence of levamisole plus exogenous phosphate, NRF2 expression was upregulated within 1 hour of treatment to levels greater than that seen in the absence of phosphate but presence of levamisole (216.64 ± 19.24% vs 98.28 ± 3.79%; p=0.004). Expression of NRF2 (through transfection with an NRF2 expression plasmid) in cells deficient in TNAP activity (via levamisole treatment), and induced to accumulate lipids, was not able to completely reconstitute ICLA when compared to cells not treated with levamisole (193.72 ± 16.51 vs 326.46 ± 47.64; p = 0.019), but ICLA was still greater than that observed at baseline. In silico analysis predicted that TNAP would bind to TRAF2, yet neither band shift assays nor immune co-precipitation showed evidence of this. However, TRAF2 mRNA was down regulated within 3T3-L1 cells during adipogenesis, reaching levels of 15.27 ± 10.27% (p= 0.014) of baseline (levels prior to induction of intracellular lipid accumulation) by day 4 of lipid accumulation. Overexpression of TRAF2 during adipogenesis markedly reduced intracellular lipid accumulation (147.88 ± 11.28% vs 326.46 ± 47.64%; p=0.028 (after 8 days of culture)). In Y1 cells TNAP activity is upregulated during ICLA, reaching 233 ± 37.56% (p=0.019 vs. baseline) of baseline levels within the initial 24 hours. Inhibition of TNAP activity through levamisole treatment resulted in a decrease in ICLA when compared to cells not treated with levamisole. Histochemical analysis showed that TNAP activity was localised to the lipid droplet. Discussion and Conclusions: Within 3T3-L1 cells TNAP mediates intracellular lipid accumulation through the generation of phosphate. The phosphate is able to increase the expression of NRF2, however it is likely that NRF2 is not the only gene whose expression is regulated by TNAP-generated phosphate. It was found that TNAP and TRAF2 do not bind to each other in the context of ICLA; however TRAF2 is a negative mediator of ICLA through a TNAP-independent mechanism. Functional TNAP is necessary for the accumulation of cholesterol esters within the Y1 cell line, suggesting that TNAP is essential for lipid accumulation in cell types that store lipids in intracellular membrane-bound droplets in the form of triglycerides or cholesterol esters. / GR2018

Dans le cerveau, les fonctions de la phosphatase alcaline non spécifique des tissus (TNAP) ne sont pas clairement identifiées. La localisation et l'expression de cette enzyme au niveau neuronal suggère cependant, qu'elle joue un rôle important dans le développement et le fonctionnement du cerveau. Cela est supporté par la présence de graves crises d'épilepsie chez les humains porteurs d'une mutation de la TNAP. Ces crises d'épilepsie sont létales chez les souris KO pour la TNAP. Des études chez la souris montrent que la TNAP pourrait réguler l'inhibition postsynaptique médiée par le GABA et elle pourrait être impliquée dans l'inhibition présynaptique médiée par l'adénosine. L'adénosine est, en partie, synthétisée via la déphosphorylation successive de l'ATP en ADP puis en AMP par des ectonucléotidases. Parmi elles, la TNAP et l'ecto- 5'-nucléotidase (NT5E) catalysent l'hydrolyse de l'AMP en adénosine dans le cortex cérébral. L'adénosine agit principalement au niveau présynaptique par l'intermédiaire des récepteurs A1. Ainsi l'adénosine a une influence sur la transmission synaptique et sur la plasticité synaptique. Ceci pourrait expliquer, en partie, les crises d'épilepsie observées chez les souris KO pour la TNAP. Les deux objectifs principaux de ma thèse ont été : (1) évaluer la contribution de la TNAP dans la production d'adénosine dans le cerveau ; (2) étudier l'influence de l'adénosine sur la plasticité synaptique. Premièrement, l'étude de la contribution de la TNAP dans la production d'adénosine dans le cerveau a été réalisée au moyen de deux approches complémentaires. Une approche métabolomique (spectroscopie RMN du proton) sur des cerveaux entiers de souris KO pour la TNAP a permis de montrer que la TNAP participe, entre autre, à la synthèse d'adénosine dans le cerveau. Une deuxième approche, électrophysiologique sur tranches de cerveaux de souris in vitro, nous permet d'examiner les conséquences de l'inhibition des ectonucléotidases intervenant dans la synthèse de l'adénosine. Elle a révélé que l'inhibition des ectonucléotidases (TNAP et NT5E) ne supprime pas l'effet inhibiteur de l'AMP médiée par les récepteurs A1. Deuxièmement, nous avons étudié l'influence de l'adénosine sur la plasticité synaptique à courte terme. Nous avons enregistré des potentiels de champs dans la couche Ia du cortex piriforme en réponse à des stimulations électriques (3,125 à 100 Hz) présentée avec des fréquences recouvrant la gamme d'oscillations physiologiques. Nos résultats montrent qu'avec de fortes concentrations d'adénosine, la facilitation est accentuée par rapport à celle observée en situation contrôle. Cet effet est observé pour des fréquences supérieures ou égales à 25 Hz. De plus, cette accentuation est d'autant plus grande que la fréquence est élevée (maximum atteint à 100 Hz pour 100 µM). En bloquant l'action de l'adénosine endogène, l'effet contraire est observé : une facilitation déficitaire par rapport au contrôle et dont le défaut est croissant avec la fréquence de stimulation. Tous ces résultats convergent vers l'hypothèse qu'une déficience en TNAP, traduite par une absence d'adénosine, pourrait contribuer au maintien des processus épileptiques générés par un déséquilibre de l'inhibition et de l'excitation dû à une diminution de GABA. L'effet inhibiteur de l'AMP médié par les récepteurs A1 ne serait pas suffisant pour contrecarrer les crises d'épilepsie observées chez les sujets hypophosphatasiques et les souris KO pour la TNAP. / The functions of Tissue Nonspecific Alkaline Phosphatase (TNAP) in the brain are not clearly identified. The localization and expression of TNAP at the neuronal level, however, suggests that it plays a prominent role in the development and the function in the brain. This is supported by the presence of severe epileptic seizures in humans carrying TNAP mutation. These epileptic seizures are lethal in TNAP KO mice. Studies in mice show that TNAP could regulate GABA-mediated postsynaptic inhibition and may be involved in presynaptic inhibition mediated by adenosine. Adenosine is, partly, synthesized via the successive dephosphorylation of ATP to ADP and then to AMP by ectonucleotidases. Among them TNAP and ecto-5'-nucleotidase (NT5E) are able to hydrolyze AMP into adenosine. Adenosine acts mainly at the presynaptic level via A1 receptors activation. Adenosine has an influence on synaptic transmission and thus on synaptic plasticity. This could partly explain the epileptic seizures observed in TNAP knock-out mice. The two main purposes of my thesis were: (1) to evaluate the contribution of TNAP in adenosine production in the brain; (2) to study the influence of adenosine on synaptic plasticity. Firstly, the study of the contribution of TNAP in adenosine production in the brain was carried out using two complementary approaches. A metabolomic approach (proton NMR spectroscopy) on whole brains of TNAP KO mice showed that TNAP in involved in adenosine synthesis in the brain. In a second approach, in vitro electrophysiological recordings on mouse brain slices allowed us to examine the consequences of the inhibition of the ectonucleotidases involved in adenosine synthesis. This revealed that inhibition of ectonucleotidases (TNAP and NT5E) did not suppress the inhibitory effect of AMP mediated by A1 receptors. Secondly, we studied the influence of adenosine on short-term synaptic plasticity. Field potentials were recorded in response to electrical stimulations (3.125 to 100 Hz) applied with frequencies encompassing the range of physiological oscillation. Our results show that, with high adenosine concentrations, the facilitation is emphasized compared to that observed in the control situation. This effect is observed for frequencies greater than or equal to 25 Hz. In addition, the higher the frequency, the greater the facilitation. Finally, by blocking the action of endogenous adenosine, the opposite effect was observed: a deficient facilitation with respect to the control, whose defect was increasing with stimulation frequency. All these results converge towards the hypothesis that TNAP deficiency, expressed by absence of adenosine, could contribute to the maintenance of the epileptic processes generated by an imbalance of the neuronal inhibition and the excitation due to a decrease of GABA. AMP inhibitory effect mediated by A1 receptors, would not be sufficient to counteract epileptic seizures observed in hypophosphatasic patients and TNAP KO mice.

No description available.

京都大学 / 0048 / 新制・課程博士 / 博士(生命科学) / 甲第20069号 / 生博第358号 / 新制||生||47(附属図書館) / 京都大学大学院生命科学研究科統合生命科学専攻 / (主査)教授 永尾 雅哉, 教授 佐藤 文彦, 教授 河内 孝之 / 学位規則第4条第1項該当 / Doctor of Philosophy in Life Sciences / Kyoto University / DFAM

L’arthrose est la pathologie articulaire la plus répandue et, avec l’allongement de l’espérance de vie, sa prévalence ne cesse d’augmenter. Elle se caractérise par une dégénérescence du cartilage articulaire associée à une inflammation synoviale et un remodelage anormal de l’os sous-chondral, qui résultent en une perte progressive de mobilité et des douleurs très handicapantes. Dans le cartilage, le chondrocyte est le seul type cellulaire et il est responsable de la synthèse des composants de la matrice extracellulaire (collagènes, protéoglycanes). Au cours de l’arthrose, le phénotype du chondrocyte est altéré et la balance synthèse/dégradation des composants matriciels est déséquilibrée en faveur de la dégradation du cartilage. Il n’existe actuellement aucun traitement permettant de ralentir efficacement l’évolution du processus arthrosique, de sorte que la recherche de biomarqueurs pertinents et de cibles thérapeutiques potentielles est en pleine effervescence depuis l’explosion de l’étude des microARNs. Les microARNs sont des petits ARNs non codants régulant négativement l’expression des gènes. On estime que 50% des gènes sont potentiellement régulés par les miARNs. Les miARNs semblent impliqués dans tous les processus biologiques majeurs tels que la différenciation cellulaire, l’apoptose ou encore la cancérisation. Ces petits ARN non codants sont donc des biomarqueurs potentiels très intéressants. Au cours de ces travaux de thèse l’implication des miARN dans la régulation du phénotype chondrocytaire a été étudiée. A partir d’un modèle de perte du phénotype chondrocytaire différencié, provoquée par des repiquages successifs ou une stimulation par l’IL-1β les variations du profil d’expression des miARNs ont été analysées par l’utilisation de puces dédiées. Ces données ont permis de mettre en évidence 43 miARNs candidats dont le cluster miR-23~27b~24-1 et miR-29b. L’étude de la régulation de la production différentielle des miARNs de ce cluster a été entreprise, sans que nous parvenions toutefois à apporter une réponse formelle sur les mécanismes impliqués. Néanmoins, nous avons identifié miR-29b comme un régulateur négatif de l’expression du gène codant Col-IIa1 au cours de la perte du phénotype différencié, ainsi que chez les chondrocytes « arthrosiques ». Enfin, comme il a été montré au laboratoire que l’équilibre entre les concentrations extracellulaires de pyrophosphate/phosphate inorganique (ePi/ePPi) était essentiel au maintien du phénotype chondrocytaire différencié, nous nous sommes intéressés à la régulation des gènes codant les acteurs protéiques impliqués dans cette balance (ANK, PC1, Pit-1 et TNAP). A partir de prédictions de cibles par analyse in silico, un panel de 4 miARNs candidats a été établi : let7e, miR-9, miR-188 et miR-219. Nos travaux avec des systèmes rapporteurs ont démontré l’implication de miR-9 en tant que régulateur négatif de l’expression des gènes PC-1, Pit-1 et TNAP, de façon cohérente ou non avec les prédictions bio-informatiques. / Osteoarthritis (OA) is the most frequent joint disease and its prevalence still grows with the increase in lifespan. OA is characterized by articular cartilage degeneration, together with synovitis and abnormal subchondral bone remodeling, leading to progressive loss of mobility and pain. Chondrocyte is the unique cell type in cartilage which accounts for the synthesis of extracellular matrix (ECM) components (collagens, proteoglycans). During OA, chondrocyte phenotype is altered and the balance between ECM synthesis and degradation is impaired towards cartilage degradation. To date no treatment can efficiently reduce OA progression so that the search for reliable biomarkers and potential therapeutic targets is very active, particularly since the discovery of microRNAs. miRNAs are estimated to regulate 50% of cellular genes. They contribute to major cellular processes such as cell differentiation, apoptosis or tumorigenesis. Therefore, miRNAs are interesting putative biomarkers. During this PhD thesis, we studied the contribution of miARNs to the control of chondrocyte phenotype. Using a model of chondrocyte differentiated phenotype loss induced by extensive subculturing or IL-1β challenge we studied changes in miRNAs profile with microarrays. We determined a panel of 43 varying miRNA including the miR-23~27b~24-1 cluster and miR-29b. The differential production of miRNAs from this cluster has been investigated, but we didn’t succeed in identifying the underlying mechanisms. However, we identified miR-29b as a negative post-transcriptional regulator of Col-IIa1 during differentiated phenotype loss and OA. Finally, as equilibrium between extracellular levels of inorganic phosphate and pyrophosphate (ePi/ePPi) was previously shown in the laboratory to be crucial for the maintenance of a differentiated chondrocyte phenotype, we studied the regulation of the genes encoding the 4 proteins regulating this balance (ANK, PC1, Pit-1 and TNAP). From in silico analysis, we selected a panel of 4 miRNAs: let7e, miR-9, miR-188 and miR-219. Using reporter assays, we showed that miR-9 was a negative regulator of PC-1, Pit-1 and TNAP, according or not to bioinformatics prediction