Ablation du noyau suprachiasmatique et rythme circadien de la température

Le noyau suprachiasmatique est le générateur du rythme circadien (Hastings et al., 2018). Son ablation provoque une perturbation du rythme veille/sommeil (Drucker-Colin et al., 1984; Lehman et al., 1987; Ralph et al., 1990; Sawaki et al., 1984) et une perturbation du rythme circadien de la température corporelle chez le rat comme chez le tamia (Abe et al., 1979; Ruby et al., 2002). Dans la maladie d’Alzheimer, le rythme circadien est perturbé (réveils nocturnes, siestes diurnes) (Anderson et al., 2013; Weldemichael et al., 2010) et inversement les personnes présentant des perturbations du rythme circadien sont plus à risque pour le déclin cognitif (Foley et al., 2001; Lim et al., 2013a; Tranah et al., 2011). Chez les patients atteints de la MA, la neurodégénérescence est présente dans le noyau suprachiasmatique (Stopa et al., 1999) mais nous ne savons pas si la destruction du rythme circadien est une cause ou la conséquence de la maladie. Pour répondre à cette question, il serait donc intéressant de faire une ablation du noyau suprachiasmatique dans la souris Tg4510, un modèle de pathologie tau pour lequel on sait que le rythme circadien est perturbé

82

(Stevanovic et al., 2017a) et de déterminer si la destruction du noyau suprachiasmatique chez l’animal augmente la pathologie tau.

Nous avons vu en introduction que reproduire les oscillations la température corporelle permet de resynchroniser les oscillateurs périphériques en reprogrammant les gènes de l’horloge dans les tissus en culture comme dans les fibroblastes (Brown et al., 2002; Buhr et al., 2010; Saini et al., 2012). Il est donc probable que le noyau suprachiasmatique dicte le rythme circadien via la régulation de la température corporelle. Également, nous avons démontré dans notre article que la température corporelle joue un rôle important dans le rythme circadien de la phosphorylation de tau. Il serait donc intéressant de déterminer si dans la souris Tg4510, un modèle de pathologie tau pour lequel on sait que le rythme circadien est perturbé (Stevanovic et al., 2017a), après une ablation du noyau suprachiasmatique, le fait de rétablir une oscillation circadienne de la température corporelle de manière artificielle permet de rétablir l’expression des gènes de l’horloge d’une part et de réduire la progression de la pathologie tau d’autre part.

Conclusion générale :

La maladie d’Alzheimer connaît une progression de plus en plus importante. La compréhension des mécanismes menant à la maladie reste superficielle, bien que certains facteurs de risque soient bien identifiés. Les troubles du sommeil font partie des pistes explorées pour comprendre le développement de la pathologie et les mécanismes mettant en lien le sommeil et la MA sont encore peu connus. Nos travaux mettent en évidence que la phosphorylation de tau, un marqueur bien connu de la MA, suit un rythme circadien : tau est hyperphosphorylée pendant le sommeil et déphosphorylée pendant l’activité dans le cortex des souris sauvages. Nous pouvons en déduire que la phosphorylation de tau est dépendante de la température corporelle et du sommeil et que PP2A est probablement impliquée dans la déphosphorylation de tau pendant l’activité. Ce travail pose les bases de la compréhension des mécanismes liant les troubles du sommeil et de la MA.

83

84

Sauna-like conditions induce tau dephosphorylation through hyperthermia

Isabelle Guisle; Séréna Pétry; Françoise Morin; Robert A. Whittington; Frédéric Calon ; Sébastien Hébert; Emmanuel Planel

Abstract:

The abnormal hyper-phosphorylation and aggregation of tau is a hallmark of many neurodegenerative diseases referred to as tauopathies, the most common Alzheimer's disease (AD). Previously we have shown that tau phosphorylation is a dynamic process highly dependent on temperature. Recently, a 20 years prospective study on the effects of sauna bathing revealed an association between moderate to high frequency of sauna bathing and a lower incidence of dementia and Alzheimer's disease however the mechanism underlying this benefit remains uncertain. Effects of sauna include relaxation or improved hemodynamic functions, but also transient hyperthermia. In this study, we tested the hypothesis that sauna-like conditions could lower tau phosphorylation by increasing body temperature. We exposed wild-type and hTau mice (a mouse model of AD-like tau pathology) to mild hyperthermia (39-40°C). We observed a strong and significant de- phosphorylation of tau at multiple epitopes. This effect was correlated with increased phosphatases and lower kinases activities but without obvious change in inflammatory or heat-shock responses. We could confirm the effects on tau phosphorylation in neuronal- like cells in culture. In conclusion, increasing body temperature seems to have beneficial effects on tau pathology and represents a new and promising therapeutic strategy for AD and other tauopathies.

85 Introduction:

Tau protein is a microtubule associated protein that binds to and stabilizes microtubules(Guo et al., 2017). The phosphorylation of tau accounts for its main post translational modification (Martin et al., 2011). In physiological condition, tau phosphorylation regulates several functions including stabilization of microtubules (Weingarten et al., 1975), synaptic plasticity (Kimura et al., 2014),regulation of gene expression etc. (Bukar Maina et al., 2016). The phosphorylation of tau is in turn regulated by many kinases (Martin et al., 2013) and phosphatases (Liu et al., 2005), with GSK3 (Glycogen Synthase Kinase 3 ) (Hanger et al., 2011; Planel, 2002) and PP2A (Protein Phosphatase 2 A) (Liu et al., 2005), being the main tau kinases and phosphatases, respectively.

Aggregation of abnormally hyperphosphorylated tau into neurofibrillary tangles are a feature of several diseases known as tauopathies, the most common being Alzheimer's disease (AD). The hyperphosphorylation of tau has been shown to lead to paired helical filaments and tangle formation (Martin et al., 2011). The neurodegeneration and clinical progression of AD in the brain is correlated with tau pathology (Braak et al., 2016). Thus tau phosphorylation is a valid target for the treatment of AD and other tauopathies. However and despite many efforts, no treatment exists for tauopathies (Graham et al., 2017).

Recently, a 20 years longitudinal study led in a cohort of more than 2000 finish people revealed that people practicing sauna bathing between 4 to 7 times a week are at lower risk to develop AD (Hazard Ratio: 0.35, 95% CI: 0.14-0.90) when compared to people practicing sauna bathing once a week (Laukkanen T.S.K, 2016). The authors suggested that the benefits observed could be due to improved hemodynamic functions conjugated with reduced inflammation. However, the impact of sauna on AD biomarkers was not investigated (Laukkanen T.S.K, 2016). In animal models, hypothermia consistently induces tau hyper- phosphorylation (Bretteville et al., 2012; Planel et al., 2004; Tournissac et al., 2017; Vandal et al., 2016; Whittington et al., 2010). As body temperature rises up during sauna of 0.9ºC in humans (Leppaluoto, 1988), we thus hypothesized that, conversely to what was observed during hypothermia, mild hyperthermia could lead to tau de-phosphorylation. To verify this hypothesis, we explored the immediate effect of moderate acute hyperthermia on phosphorylation of tau and its underlying molecular mechanisms. We report that elevated

86

temperature induces de-phosphorylation of tau in a murine model of tauopathy. To dissect out the effect of temperature from both the effect of hemodynamic functions and inflammation, we exposed neuron like cells expressing human tau to higher temperatures. In agreement with our hypothesis, we observed that exposing cells to hyperthermia led to de- phosphorylation of tau. Dephosphorylation of tau was correlated with inactivation of JNK and activation of PP2A. Altogether, our results suggest that sauna-like conditions can lower tau pathology through hyperthermia and might explain in part the beneficial effects of sauna bathing observed by Laukkanen et al. These finding open new potential therapeutic insights that are discussed here.

Material and methods: Animals

8 week-old C57BL6 and 12 month-old hTau female mice (Andorfer et al., 2003) were used for heat exposure (n=11-12). The htau mice (Andorfer et al., 2003) were obtained by crossing mice that express the six isoforms of human non mutated tau (8c mice, (Duff et al., 2000)) with mice knock out for murine tau (Tucker et al., 2001). Tau knock out mice were generated by insertion of Enhanced Green Fluorescent Protein cDNA into exon 1 of murine tau (Tucker et al., 2001). The founders of our htau and tau knock out colonies were derived on a C57BL6/J background (B6.Cg-Mapttm1(EGFP)Klt-Tg(MAPT)8cPdav/J, Jackson

Laboratories, Bar Harbor, ME, USA). htau mice have hyperphosphorylated and aggregated

tau at 2month of age (Andorfer et al., 2003). The Animals were handled according to procedures approved by the “Comité de Protection des Animaux du CHU” under the guidelines of the Canadian Council on Animal Care. Mice were killed by decapitation without anesthesia, as anesthesia can lead to hypothermia-induced tau phosphorylation (Planel et al., 2007).

Heat treatment

To induce mild hyperthermia (39-40˚C rectal temperature), the mice were kept in a ventilated incubator at 42°C for one hour. Control mice where kept at ambient temperature (23°C). All mice had access to water and food ad libitum. Core body temperature was assessed at the end

87

of the incubation with a rectal probe (RET-3, Brain Tree Scientific Inc, Braintree, MA, USA) just prior to sacrifice.

Cell culture

Human neuroblastoma cells stably expressing human tau 3 repeat isoform 2+3-10-(generous gift from Luc Buée) was used for this experiment (Delobel et al., 2003). Cells were handled as previously described (Delobel et al., 2003). Briefly, cells were cultured in Dubleco’s modified eagle medium high glucose without pyruvate (11965-092, Life Technologies, Carlsbad, CA, USA) supplemented with 10% of bovine growth serum (SH30541.3, GE Health Care life Science, Marlborough, MA) (inactivated by heat exposure at 56°C for 30min) and a combination of penicillin 100u/mL-streptomicin 100µg/mL (15140-122, Life Technologies, Carlsbad, CA, USA) in a 5% CO2 incubator for 24h either at 34°C, at 37°C or at 40°C. At the end of the incubation, cells were rapidly put on ice, the culture medium was removed and the cells were washed once with phospho-buffered saline 1x. They were collected in 100µL of Radioimmunoprecipitation Assay (RIPA) (50 mM Tris-HCl, pH 7.4, 1% NP-40, 150 mM NaCl, 0.25% Na-deoxycholate, 1 mM EDTA, 1 mM Na3VO4, 1 mM

NaF, 1 mM PMSF, 10 μl/ml of Proteases Inhibitors Cocktail (P8340, Sigma-Aldrich, St. Louis, MO, USA), and sonicated. Samples were then centrifuged for 20 min at 20,000g at 4 °C. The supernatant was recovered, diluted in sample buffer (NuPAGE LDS; Invitrogen, Carlsbad, CA) containing 5% of 2-β-mercapto-ethanol, 1 mM Na3VO4, 1 mM NaF, 1 mM

PMSF, 10 μl/ml of Proteases Inhibitors Cocktail (P8340; Sigma-Aldrich, St. Louis, MO, USA), and heated for 10 min at 95 °C. 10 μg of protein were analyzed as described previously (Petry et al., 2017; Petry et al., 2014).

Tissue extraction

Brains were immediately removed and tissues were dissected on ice. Tissues were frozen on dry ice and kept at − 80 °C until they were homogenized. Cortices samples were homogenized without thawing in 5 times volume-weight of RIPA with a mechanical homogenizer (14-261-01, Fisher Scientific, Waltham, MA, USA). Hippocampi samples were homogenized without thawing by sonication in 150µL of RIPA. Samples were then centrifuged for 20 min at 20,000g at 4 °C. The supernatant was recovered, diluted in sample

88

buffer (NuPAGE LDS; Invitrogen, Carlsbad, CA) containing 5% of 2-β-mercapto-ethanol, 1 mM Na3VO4, 1 mM NaF, 1 mM PMSF, 10 μl/ml of Proteases Inhibitors Cocktail (P8340;

Sigma-Aldrich, St. Louis, MO, USA), and heated for 10 min at 95 °C. 10 μg-20µg of protein were analyzed as described previously (Petry et al., 2017; Petry et al., 2014).

Analysis of aggregated tau

Tau aggregates were extracted following previously published protocols derived from Greenberg and Davis (Greenberg et al., 1990; Julien et al., 2012). Briefly, the RIPA supernatant was adjusted to 1% sarkosyl (N-lauroylsarcosine), incubated for 60 min at 37°C with constant shaking, and centrifuged at 100,000×g for 1 h at 20 °C. The pellet containing sarkosyl-insoluble aggregated was resuspended and diluted in Sample buffer (NuPAGE

LDS) containing 5% of 2- β -mercapto-ethanol, 1 mM Na3VO4, 1 mM NaF, 1 mM PMSF,

10 µl/ml of proteases inhibitors cocktail (P8340, Sigma-Aldrich), sonicated and boiled for 5 min, and stored at −20 °C.

SDS-PAGE, Western blot analysis and antibodies

SDS-PAGE and Western blot analysis was done as previously described (Gratuze et al., 2017a; Planel et al., 2001). Briefly, 10µg of brain homogenates were separated on a SDS- 10% polyacrylamide gel, transferred onto nitrocellulose membranes (Amersham Biosciences, Pittsburgh, PA, USA). Membranes where then saturated, hybridized with antibodies and revealed as described in (Gratuze et al., 2017a). All antibodies used in this study, in addition to their dilution, are listed in Table 1. For Western blot of tau protein, the signal was normalized to total tau for all epitopes. For Western blot regarding kinases, signal was normalized to the total corresponding protein. Two lanes from representative immunoblots are displayed for each condition. Dividing lines represent areas where lanes from the same blot were removed and the remaining lanes were spliced together. Brightness levels were adjusted as needed.

89 Brain Protein Phosphatase 2 A (PP2A) activity

A 2 months old C57BL6 mice was killed by decapitation without anesthesia. Brain was removed, rapidly dissected on ice and frozen at -80°C until they were used. The frozen cortex was homogenized following the recommendations of Duo set IC active PP2A kit (reference DYC3309, R&D systems, Minneapolis, USA) has previously described (Gratuze et al., 2017b) and the lysate was processed at 2µg/µL (n=3) with the same kit at 37°C, 40°C and 42°C.

Brain tau kinase activity

Brain tau kinase assay was done according to our previous protocol (Planel et al., 2004) with some modifications. 1µg of full length recombinant tau (reference T1001-2, R Peptide, Watkinsville, GA, USA) was prepared in sample buffer containing 5µg of RIPA cortex lysate (from a 2 month-old C57BL6 male mouse, see tissue extraction part), 14 mM HEPES (pH7.7), 10 mM NaCl, 2 mM MgCl2, 0.4 mM EDTA, 0.004% Triton-X 100, 1 mM PMSF, 10 μl/ml of Proteases Inhibitors Cocktail (P8340, Sigma-Aldrich, St. Louis, MO, USA), 1 mM Na3VO4, 1 mM NaF,ATP 10µM (Planel et al., 2004). This reaction was incubated for

60 minutes in triplicates for each temperature (37°C, 40°C and 42°C) and terminated by adding one volume of sample buffer (NuPAGE LDS; Invitrogen, Carlsbad, CA) containing 5% of 2-β-mercapto-ethanol, 1 mM Na3VO4, 1 mM NaF, 1 mM PMSF, 10 μl/ml of Proteases

Inhibitors Cocktail (P8340; Sigma-Aldrich, St. Louis, MO, USA), and heated for 10 min at 95 °C. A negative control without ATP was processed in parallel following the same procedure in duplicate for each temperature. Samples were analyzed by SDS-PAGE and western blotting as described above. Immunoblot was reveled using anti tau PS404 antibody (see section antibodies) since this epitope is phosphorylated by most of the major tau of kinases(Emmanuel Planel, 2002). The signal was normalized to total tau. Immunoreactive bands were visualized using ImageQuant LAS 4000 imaging system (Fujifilm USA, Valhalla, NY, USA), and densitometric analyses were performed with Image Gauge analysis software (Fujifilm, Valhalla, NY, USA). Images levels were adjusted.

90 Statistical analysis

Most of the statistical analysis was done by Man Whitney testing using GraphPad Prism software 5.0 as data were not normally distributed (Graphpad Software Inc., La Jolla, CA, USA). Significant results are given for indicated p-values and level of significance  (level of significance, refers to the probability to be right testing the null hypothesis that the 2 groups tested are equal). Figure 5A was analyzed by Kruskall Wallis test (p<0.05) followed by Dunn’s post hoc test for pair-wise sample comparison. Significant results were given for p<0.05 (**). For all figures, data are given as mean ± Standard Deviation.

91 Results

Mild hyperthermia induces de-phosphorylation of tau at multiple epitopes in wild type mice and hTau mice

We first wanted to assess whether mild and transient hyperthermia could affect tau phosphorylation in vivo in mice. We thus explored the effect of one hour of hyperthermia on tau phosphorylation in C57B6 mice and in a mouse model of human tauopathy, htau mouse. Both B6 (Ctrl: 38.22 ± 0.13°C; Hyper: 40.08 ± 0.29°C), and htau (Ctrl: 37.45 ± 0.49°C; Hyper: 39.12 ± 0.36°C) mice exposed to heat were hyperthermic at sacrifice compared to the animals housed at room temperature (fig.1). Interestingly, we observed that tau was dephosphorylated at the AT8, CP13, and Tau1 epitoes in the cortex of C57B6 and htau mice (fig.2). We next wanted to investigate whether the de-phosphorylation of Tau was the same in the hippocampus, one of the first regions to display tau pathology in AD. Again we observed a significant de-phosphorylation of tau protein at epitopes AT8, CP13, PHF1 and Tau1 in hippocampi of B6 mice (fig3.A) and at epitopes AT8, CP13 and Tau1 in hippocampi of hTau mice (fig3.B). However, the treatment didn’t impact tau aggregation as evaluated by sarkosyl extraction (data not shown). Overall our data demonstrate that mild hyperthermia can lead to extensive tau dephosphorylation.

Mild hyperthermia does not trigger heat-shock response or subsequent inflammatory response

We next wanted to assess whether mild hyperthermia had the potency to trigger heat-shock response and subsequent inflammation. Indeed, Heat shock can induce markers of heat- shock and inflammation such as HSP70 (Heat Shock Protein 70), TLR2 (Toll Like Receptor 2) and GFAP (Glial Fibrillary Acidic Protein) (Belay et al., 2003; Brown, 1983; Hayward et al., 2014; Lee et al., 2015; Miller et al., 1987; Schiaffonati et al., 2001), that can modulate tau pathology (Bolos et al., 2017). Our results show that none of the three markers were increased, in B6 mouse or in htau mouse (Fig.4), suggesting that at 39-40°C hyperthermia was not enough to induce heat shock or inflammatory response in our experimental conditions. In turn, these result suggest that neither heat shock response nor subsequent inflammatory response are responsible for tau de-phosphorylation in our mice.

92

Mild hyperthermia induces de-phosphorylation of tau at multiple epitopes in neuroblastoma cells

Sauna bathing can improve hemodynamic functions (Hannuksela et al., 2001; Keast et al., 2000; Laukkanen et al., 2015; Luurila, 1992), and decrease blood pressure of people with hypertension (Hannuksela et al., 2001; Winterfeld et al., 1993). Variations in blood pressure can modulate tau phosphorylation (Liu et al., 2014; Marfany et al., 2018). To isolate temperature from confounding factors such as blood pressure, we explored the immediate effect of mild hyperthermia (37˚C vs 40°C) on the phosphorylation of tau protein in neuron like cells expressing tau protein (SY5Y cells stably expressing 3 repeat human tau). We observed that tau protein was significantly dephosphorylated at the same epitopes as in the mice (AT8, CP13, Tau1, and PHF1) (fig.5). We also used a positive control for tau hyperphosphorylation by exposing cells to hypothermia. As previously described(Bretteville et al., 2012), at 34°C, tau protein was hyper phosphorylated at all epitopes tested in comparison to 37°C. These results suggest that higher temperature alone is able to induce tau de-phosphorylation.

Mild hyperthermia induces the inhibition of c Jun N-terminal kinase (JNK)

We next wanted to understand the underlying molecular mechanisms of tau dephosphorylation. Since tau phosphorylation is a balance between the activation state of tau kinases and phosphatases (Iqbal et al., 2016; Martin et al., 2013; Wang et al., 2007), we speculated that hyperthermia could induce a misbalance of the system towards kinases inhibition and/or phosphatases activation. We thus analyzed the activation state (phosphorylation) of some of the major kinases involved in the regulation of tau phosphorylation: JNK (cJun N-terminal Kinase), GSK-3Glycogen Synthase Kinase 3), AMPK (Adenosine Mono Phosphate activated protein Kinase), PKA (Protein Kinase A), ERK (Extra-cellular Responsive Kinase), CAMKII (Calcium Calmodulin dependent protein Kinase II), and AKT (Protein Kinase B) (Cai et al., 2012; Iqbal et al., 2016; Martin et al., 2013; Wang et al., 2007). We also analyzed the activation state of PP2A (protein phosphatase 2 A by methylation), a phosphatase involved in AD (Iqbal et al., 2016; Martin et al., 2013; Wang et al., 2007) and responsible of 70% of tau phosphatase activity in human brain (Liu et al., 2005). We observed that JNK was inactivated in cortex of B6 and htau mice by heat

93

exposure (Fig.6), while there was no change in the activation state of all the other kinases or PP2A. Importantly, the antibodies directed against the total forms of these enzymes did not show any change (data not shown). Our results show that JNK gets inactivated during hyperthermia which might explain in part tau dephosphorylation.

De-phosphorylation of tau induced by hyperthermia can be explained by an imbalance between tau phosphatases and kinases activity

From our previous work, we know that lower temperatures can directly and differentially affect the activities of phosphatases and kinases, resulting in tau hyper-phosphorylation (Planel et al., 2004), we thus explored whether higher temperature could also modulate kinases and phosphatases. We assessed PP2A activity (a major phosphatase of tau representing 70% of tau phosphatases activity (Liu et al., 2005)) and total tau kinases activity in brain lysates, when exposed to high temperatures (40°C and 42°C) in comparison to controls at 37°C. Interestingly, we observed an increase in PP2A activity at 40°C and 42°C (fig.7.A) while kinases activity at the same temperatures was decreased (fig.7.B). Thus increasing body temperature can promote an imbalance between kinases and phosphatases favoring tau de-phosphorylation.

94 Discussion

In this article, we investigated the effects of mild hyperthermia on tau phosphorylation in vitro and in vivo. We report that hyperthermia induces tau de-phosphorylation in vitro (SH cells) and in vivo, in cortices and hippocampi of both wild type B6 mouse and htau mouse. We also found that it could be explained by an imbalance between phosphatases and kinases activities favoring tau de-phosphorylation.

Tau hyperphosphorylation and aggregation is associated with the progression of AD (Simic et al., 2016). Here, we showed that mild hyperthermia led to de-phosphorylation of tau at