International Journal of Molecular Sciences

Article The Dipeptidyl Peptidase-4 Inhibitor Linagliptin Ameliorates High-fat Induced Cognitive Decline in Tauopathy Model Mice

Yuriko Nakaoku 1,2, Satoshi Saito 2,3, Yumi Yamamoto 4, Takakuni Maki 1, Ryosuke Takahashi 1 and Masafumi Ihara 2,*

1 Department of Neurology, Kyoto University Graduate School of Medicine, Kyoto 606-8507, Japan; [email protected] (Y.N.); [email protected] (T.M.); [email protected] (R.T.) 2 Department of Neurology, National Cerebral and Cardiovascular Center, Suita 565-8565, Japan; [email protected] 3 Research Fellow of Japan Society for the Promotion of Science, Tokyo 102-0083, Japan 4 Department of Regenerative Medicine and Tissue Engineering, National Cerebral and Cardiovascular Center, Suita 565-8565, Japan; [email protected] * Correspondence: [email protected]; Tel.: +81-6-6833-5012; Fax: +81-6-6872-7486

 Received: 31 March 2019; Accepted: 21 May 2019; Published: 23 May 2019 

Abstract: Vascular risk factors, such as type 2 diabetes mellitus (T2DM), are associated with the increased risk of Alzheimer’s disease. One of the common T2DM medications, dipeptidyl peptidase (DPP)-4 inhibitors, have a minimum risk for hypoglycemia and have recently been suggested to ameliorate β-amyloid pathology. However, conflicting results have been reported regarding the effects of DPP-4 inhibition on cognitive function and tau pathology. Thus, we investigated whether inhibiting DPP-4 affects tau pathology and cognition in a mouse model of tauopathy with hyperglycemia. Male mice overexpressing the P301S mutant human microtubule-associated tau (PS19) were fed either a low or high-fat diet. PS19 mice were then administered either linagliptin, a DPP-4 inhibitor, or vehicle, from 6 weeks to 8 months of age. Linagliptin-treated mice exhibited higher levels of glucagon-like -1 and decreased fasting blood glucose, compared with the vehicle-treated mice at 8 months. Linagliptin treatment significantly restored spatial reference memory and increased cerebral blood flow without affecting phosphorylation levels of tau or endothelial nitric oxide synthase (eNOS) in the brain. Linagliptin may ameliorate HFD-induced cognitive worsening in tauopathy, at least partially, by increasing cerebral perfusion via the eNOS-independent pathway.

Keywords: dipeptidyl peptidase-4 inhibitors; high-fat diet; spatial reference memory; cerebral blood flow; tau

1. Introduction Alzheimer’s disease (AD) is a common, progressive degenerative brain disease, leading to cognitive deficits, and accounts for 60–80% of global cases [1]. The principle neuropathological hallmarks of AD are amyloid plaques and neurofibrillary tangles (NFTs), which are composed of aggregated β-amyloid (Aβ) peptide and hyperphosphorylated tau, respectively [2]. Many investigators have believed that Aβ is a critical factor for the initiation of the disease (as documented in the amyloid hypothesis) [3], and tau pathology is a secondary effect. Recent studies, however, suggest tau pathology plays a more important role than previously thought as it appears earlier in life, even before Aβ deposition in some individuals [4]. In and parkinsonism linked to 17 (FTDP-17), in microtubule-associated protein tau (MAPT) alone can cause

Int. J. Mol. Sci. 2019, 20, 2539; doi:10.3390/ijms20102539 www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2019, 20, 2539 2 of 11 neurodegenerative disease in the absence of Aβ [5]. Robertson et al. also demonstrated that reducing tau levels can prevent behavioral deficits in an AD mouse model without altering Aβ levels [6]. Indeed, another study found a strong correlation between the amount of neurofibrillary tangles in the neocortex and the degree of cognitive deficits in AD patients [7]. These data suggest tau, not amyloid, is the key player of in AD. Nonetheless, the precise etiology of AD is still unclear and the relationship between Aβ and tau pathology needs to be determined. Type 2 diabetes mellitus (T2DM) is a major risk factor for neurocognitive disorders including AD [8]. In an aging society, an increasing number of people have both AD and T2DM. In the last few decades, the link between diabetes mellitus and AD has been recognized, sharing several common mechanisms, especially defective insulin signaling [9]. Insulin and its receptors are widely expressed in and glial cells throughout the brain, suggesting insulin signaling may play a role in the control of cognition and neuronal function in brains. Several studies have reported significantly decreased neuronal insulin signaling in the cortex and of AD cases [10]. Insulin resistance and deficiency also increase phosphorylation in both cultured neuronal cells [11] and in vivo models [12]. Indeed, a nutritional regimen based on a high-fat diet (HFD) has been shown to render the brain insulin-resistant [13] and exacerbate , tau phosphorylation and behavioral deficits in AD mouse models [14,15]. Some investigators refer to AD as type 3 diabetes or an insulin-resistant brain state [16,17]. Intracerebroventricular administration of the pro-diabetes drug streptozotocin in rats results in cognitive impairment with deficits in spatial learning and memory, brain insulin resistance and deficiency, and AD-type neurodegeneration [16]. Targeted exposure to a pro-diabetes drug can cause neurodegeneration that closely mimics AD pathology. It is therefore not surprising that T2DM treatment has been proposed as a potential therapy for AD [18]. Dipeptidyl peptidase-4 (DPP-4) is an enzyme that rapidly inactivates endogenous glucagon-like peptide-1 (GLP-1), an incretin hormone that induces insulin secretion [19]. Incretin-based therapies (GLP-1 analogs and DPP-4-inhibitors) are thus frequently used for the treatment of T2DM with a minimum risk for hypoglycemia [20]. GLP-1 receptors are abundantly expressed in the vascular endothelium and neurons in the central [21,22]. For this reason, GLP-1 has been suggested to have direct central action in the brain, as well as neurotrophic and neuroprotective effects, and thus, may have beneficial effects for the treatment of dementia [23]. Preclinical studies have pointed to incretin-based therapies as potential novel strategies in the treatment of AD [24]. In a transgenic mouse model of AD, peripherally-administered incretin reduced levels of important markers associated with AD neuropathology, including Aβ plaque load, microglial activation, and central insulin resistance [25,26]. Recently, Kosaraju et al. verified the neuroprotective activity of linagliptin in a 3xTg-AD mouse model (harboring triple PS1, APP and tau mutations) with concurrent β-amyloid and tau pathology [27]. The study demonstrated that linagliptin reduced both amyloid burden and tau phosphorylation. The mechanism by which phosphorylated tau was reduced may be an indirect effect of decreased Aβ. These preclinical studies have mainly focused on the therapeutic potential on amyloidosis-associated ; however, the effect on tau-associated pathology has not been fully elucidated. A tauopathy-specific model by Hansen et al. did demonstrate that a GLP-1 receptor agonist reduced neuronal phospho-tau load in hTauP301L mice but this study did not evaluate cognitive function [28]. In addition, a recent study presented a contradictory result, showing that sitagliptin aggravates tau phosphorylation in a rat model of T2DM and called for caution when administrating DPP-4 inhibitors to AD patients [29]. So far, no consensus has been made regarding the effects of DPP-4 inhibition on tauopathy-specific pathology and related cognitive dysfunction independent of amyloid pathology. In this study, we examined whether the DPP-4 inhibitor linagliptin affects tau pathology and cognition in a mouse model of tauopathy (PS19) challenged with HFD. Int. J. Mol. Sci. 2019, 20, 2539 3 of 11

2. Results

2.1. Hyperglycemia by HFD was Ameliorated by Linagliptin Treatment Int. J. Mol. Sci. 2019, 20, x 3 of 11 After 5.5 months of HFD treatment, mice exhibited ~200% weight gain (p < 0.001, Figure1A). Glucose andAfter cholesterol 5.5 months levelsof HFD were treatment, elevated mice in exhibited HFD-fed ~200% mice weight (p < 0.001 gain in(p < Figure 0.001, 1FigureB and 1A)p <. 0.01 in FigureGlucose1C). and Linagliptin-treated cholesterol levels were HFD-fed elevated mice in HFD showed-fed mice significantly (p < 0.001 increasedin Figure 1B blood and pGLP-1 < 0.01 in levels comparedFigure to 1C) the. Linagliptin vehicle-treated-treated HFD-fed HFD-fed mice PS19 showed mice (Figure significantly1D), which increased may blood explain GLP- why1 levels fasting bloodcompared glucose levelsto the vehicle in linagliptin-treated-treated HFD-fed HFD-fed PS19 mice mice (Figure was 1D) significantly, which may lower explain than why vehicle-treated fasting blood glucose levels in linagliptin-treated HFD-fed mice was significantly lower than vehicle-treated HFD-fed PS19 mice (p < 0.05, Figure1B). HFD-fed PS19 mice (p < 0.05, Figure 1B).

FigureFigure 1. E ff 1.ects Effects of linagliptin of linagliptin (10 mg(10 / mg/kgkg BW BW/day)/day) on on (A ) (A) body body weight, weight, (B ) (B) fasting fasting blood blood glucose glucose (FBG), (C) total(FBG), cholesterol, (C) total cholesterol, and (D) and glucagon-like (D) glucagon peptide-1-like peptide (GLP-1)-1 (GLP in-1) PS19 in PS19 mice mice being being fed fed aa low-fatlow-fat diet (LFD)diet or high-fat(LFD) or diet high (HFD)-fat diet for (HFD) approximately for approximately 6 months. 6 months. Mean MeanSD. One-way ± SD. One ANOVA,-way ANOVA, followed by ± Bonferronifollowed post by Bonferroni hoc tests waspost used.hoc tests Significant was used. diSignificantfferences differences are indicated are indicated when * pwhen< 0.05, * p < ** 0.05,p < 0.01, or *****p p< < 0.001.0.01, or *** p < 0.001.

2.2. Restoration2.2. Restoration of CBF of CBF in in Linagliptin-Treated Linagliptin-Treated PSPS 1919 Mice ImpairmentImpairment of cerebralof cerebral circulation circulation has has aa significantsignificant role role in in the the onset onset and and progression progression of cognitive of cognitive dysfunctiondysfunction in AD in AD patients patients and and animal animal modelsmodels [30] [30].. To To assess assess the the vascular vascular effect eff ectof linagliptin of linagliptin in PS in PS 19 mice, CBF was measured at 7 months of age using laser speckle flowmetry. Linagliptin-treated 19 mice, CBF was measured at 7 months of age using laser speckle flowmetry. Linagliptin-treated HFD-fed mice exhibited significantly increased CBF compared with vehicle-treated HFD-fed PS19 mice (p < 0.01, Figure 2).

Int. J. Mol. Sci. 2019, 20, 2539 4 of 11

HFD-fed mice exhibited significantly increased CBF compared with vehicle-treated HFD-fed PS19 mice (p < 0.01, Figure2). Int. J. Mol. Sci. 2019, 20, x 4 of 11

Int. J. Mol. Sci. 2019, 20, x 4 of 11

FigureFigure 2. Eff 2ects. Effects of linagliptin of linagliptin (10 (10 mg mg/kg/kg BW BW/day)/day) on cerebral blood blood flow flow (CBF). (CBF). Linagliptin Linagliptin restored restored CBF reduction.CBF reduction. (A ) (A) Representative Representative images images showing showing CBF CBF measured measured by by laser laser speckle speckle flowmetry flowmetry in in vehicle-treated mice (left) and linagliptin-treated mice (right) at 7 months of age. ** p < 0.01. vehicle-treated mice (left) and linagliptin-treated mice (right) at 7 months of age. (B) A histogram showingFigure CBF. 2 Error. Effects bars of linagliptin indicate (10SD, mg/kg ** p BW/day)< 0.01. on cerebral blood flow (CBF). Linagliptin restored 2.3. Normalization of Spatial Reference± Memory Impairment in HFD-fed PS19 Mice . CBF reduction. (A) Representative images showing CBF measured by laser speckle flowmetry in 2.3. NormalizationWevehicle evaluated-treated of Spatial whether mice Reference (left) linagliptin and linagliptin Memory (10- treatedmg/kg Impairment mice BW/day) (right) in HFD-fed attreatment 7 months PS19 ofaffected age. Mice ** p spatial< .0.01. learning and reference memory impairment by the Morris water maze test at 8 months of age. Linagliptin-treated We evaluated whether linagliptin (10 mg/kg BW/day) treatment affected spatial learning and HFD2.3-fed. Normalization PS19 mice ofdemonstrated Spatial Reference a gradual Memory improvement Impairment in HFDin learning-fed PS19 during Mice . acquisition trials (days reference memory impairment by the Morris water maze test at 8 months of age. Linagliptin-treated 1–4) andWe exhibited evaluated significantly whether linagliptin shorter escape (10 mg/kg latencies BW/day) than treatment vehicle-treated affected HFD spatial-fed learningPS19 mice and (p < HFD-fed PS19 mice demonstrated a gradual improvement in learning during acquisition trials (days 1–4) 0.05,reference Figure 3B),memory while impairment motor function by the wasMorris not water affected maze (Figure test at 3A).8 months During of age. the Lprobeinagliptin trial- ontreated day 5, and exhibitedtheHFD time- fedspent significantly PS19 in mice the platformdemonstrated shorter quadrant escape a gradual latenciesin the improvement linagliptin than vehicle-treated -intreated learning HFD during-fed HFD-fed groupacquisition was PS19 trialssignificantly mice (days (p < 0.05, Figuregreater3B),1–4) while andthan exhibited those motor in function significantlythe vehicle was-treated shorter not a HFDescapeffected-fed latencies (Figure group than3(pA). < 0.05,vehicle During Figure-treated the 3C). probe HFD These-fed trial results PS19 on daymice indicated 5,(p the< time spentthat in0.05, the linagliptin platformFigure 3B), rest quadrant whileored spatialmotor in function thereference linagliptin-treated was memory not affected impairment (Figure HFD-fed in3A). HFD group During-fed was PS19the significantlyprobe mice. trial on day greater 5, than those inthe the time vehicle-treated spent in the platform HFD-fed quadrant group (inp

Figure 3. Effects of linagliptin on cognitive function as assessed by the Morris water maze test. Swimming speed was unaffected (A), while the time course of escape latency (time to goal) in the Figureacquisition 3.FigureEff ects3 phase. Effects of was linagliptin of reduced linagliptin onmore on cognitive in cognitive the linagliptin function function-treated as as assessed assessed group bythan bythe in the Morris the Morrisvehicle water-treated water maze test.group maze test. (A) Swimming(B).Swimming (C) The speedtime speed spent was was in unaunaffected theff ected,target (A), quadrant (B while) while the in thethetime timeprobe course coursetrial of escapewas of also escapelatency significantly latency(time to goal)higher (time in tointhe goal)the in the acquisitionlinagliptinacquisition-treated phase phase group. was was reduced Error bars more more indicate in the in thelinagliptin ± SD, linagliptin-treated * p- treated< 0.05. group than group in the than vehicle in- thetreated vehicle-treated group group. ((B).C) The(C) The time time spent spent in in the the target target quadrantquadrant in in the the probe probe trial trial was was also alsosignificantly significantly higher higherin the in the linagliptin-treated group. Error bars indicate ± SD, * p < 0.05. linagliptin-treated group. Error bars indicate SD, * p < 0.05. ±

Int. J. Mol. Sci. 2019, 20, x 5 of 11

Int. J. Mol. Sci. 2019, 20, 2539 5 of 11 2.4. ImmunohistochemistryInt. J. Mol. Sci. 2019, 20, x 5 of 11 To evaluate whether improvement in cognitive function resulted from decreased tau deposition 2.4. Immunohistochemistry by linagliptin,2.4. Immunohistochemistry mouse brains were stained with anti-phospho-tau antibody (clone AT8). Figure 4 showsTo evaluaterepresentativeTo evaluate whether whetherimages improvement improve of brain inment sections cognitive in cognitive from function vehicle function resulted and resulted linagliptinfrom decreasedfrom- decreasedtreated tau PS19 deposition tau deposition mice. byPhosphorylated linagliptin,by linagliptin, mouse tau was mousebrains not were brainsreduced stained were in the with stained hippocampus anti-phospho-tau with anti of-phospho linagliptin antibody-tau-treated (clone antibody AT8).PS19 (clone mice. Figure AT8).4 shows Figure 4 representativeshows representative images of brain images sections of from brain vehicle sections and linagliptin-treated from vehicle and PS19 linagliptin mice. Phosphorylated-treated PS19 mice. tau wasPhosphorylated not reduced in tau the was hippocampus not reduced of in linagliptin-treated the hippocampus PS19 of linagliptin mice. -treated PS19 mice.

Figure 4. Immunohistochemical assessment of phosphorylated tau pathology in the hippocampus after the administration of linagliptin. Representative images of phosphorylated tau staining in FigurevehicleFigure 4.-treatedImmunohistochemical 4. (left)Immunohistochemical and linagliptin assessment-treated assessment mice of phosphorylated (right) of phosphorylated at 8 months tau of pathology tauage. pathology Bar in= 100 the µin hippocampusm the hippocampus after theafter administration the administration of linagliptin. of linagliptin. Representative Representative images images of phosphorylated of phosphorylated tau staining tau staining in in µ 2.5. Westernvehicle-treatedvehicle Blotting-treated (left) and (left) linagliptin-treated and linagliptin-treated mice (right) mice (right) at 8 months at 8 months of age. of Bar age.= 100Bar =m 100 µm 2.5. WesternWe also Blotting quantified the amount of phosphorylated tau in vehicle or linagliptin-treated mice by Western2.5. blotting.Western Blotting No significant change in phosphorylated tau was observed between linagliptin- We also quantified the amount of phosphorylated tau in vehicle or linagliptin-treated mice by treated and vehicle-treated HFD-fed PS19 mice (Figure 5). Although an inbred strain was used in this Western blotting.We also No quantified significant the change amount in phosphorylated of phosphorylated tau was tau observed in vehicle between or linagliptin linagliptin-treated-treated mice by study, the amount of phosphorylated tau showed a large variation within the group, which may and vehicle-treatedWestern blotting. HFD-fed No significant PS19 mice change (Figure in5). phosphorylated Although an inbred tau was strain observed was used between in this study,linagliptin- explain why no statistical difference was observed even between LFD- and HFD-fed mice. Linagliptin the amounttreated of and phosphorylated vehicle-treated tau HFD showed-fed PS19 a large mice variation (Figure within 5). Although the group, an whichinbred maystrain explain was used why in this seems to have affected the cognitive function and CBF levels independently of the status of tau no statisticalstudy, the diff amounterence was of phosphorylated observed evenbetween tau showed LFD- a andlarge HFD-fed variation mice. within Linagliptin the group, seems which to may phosphorylation. have aexplainffected thewhy cognitive no statistical function difference and CBF was levels observed independently even between of the LFD status- and of HFD tau phosphorylation.-fed mice. Linagliptin seems to have affected the cognitive function and CBF levels independently of the status of tau phosphorylation.

Figure 5. Effects of linagliptin (10 mg/kg BW/day) on tau phosphorylation in PS19 mice at 8 months of age.Figure Representative 5. Effects of linagliptin Western blot (10 for mg/kg ratio ofBW/day) phosphorylated on tau phosphorylation tau (p-tau)/β-actin. in PS19 Values mice are at presented8 months asof the age. mean RepresentativeSD (n = 5 forWestern each group). blot for ratio of phosphorylated tau (p-tau)/β-actin. Values are ± presentedFigure as the5. Effects mean of± SD linagliptin (n = 5 for (10 each mg/kg group). BW/day) on tau phosphorylation in PS19 mice at 8 months To investigateof age. Representative whether linagliptin Western improved blot for CBF ratio via of protection phosphorylated of vascular tau (p endothelial-tau)/β-actin. function, Values we are quantified thepresented amount as ofthe endothelial mean ± SD ( nitricn = 5 for oxide each synthase group). (eNOS) and phosphorylated eNOS in vehicle

Int. J. Mol. Sci. 2019, 20, x 6 of 11

Int. J. Mol.To investigate Sci. 2019, 20, whether 2539 linagliptin improved CBF via protection of vascular endothelial function,6 of 11 we quantified the amount of endothelial nitric oxide synthase (eNOS) and phosphorylated eNOS in vehicle or linagliptin-treated mice by Western blotting. However, no significant change in or linagliptin-treated mice by Western blotting. However, no significant change in phosphorylated phosphorylated eNOS was observed between linagliptin-treated and vehicle-treated HFD-fed PS19 eNOS was observed between linagliptin-treated and vehicle-treated HFD-fed PS19 mice (Figure6). mice (Figure 6).

Figure 6. Effects of linagliptin (10 mg/kg BW/day) on tau phosphorylation in PS19 mice at 8 months of age.Figure Linagliptin 6. Effects didof linagliptin not significantly (10 mg/kg affect BW/day) the ratio on of tau cerebral phosphorylation phosphorylated in PS19 eNOS mice/eNOS. at 8 months Values areof age. presented Linagliptin as mean did notSD ( significantlyn = 5 for each affect group). the ratio of cerebral phosphorylated eNOS/eNOS. Values are presented as± mean ± SD (n = 5 for each group). 3. Discussion 3. Discussion The purpose of the present study was to evaluate whether linagliptin improves or aggravates cognitiveThe purpose impairment of the and present tau pathology study was in to a mouseevaluate model whether of tauopathy linagliptin (PS19). improves PS19 or mice aggravates used in thiscognitive study impairment express mono-transgenic and tau pathology P301S in mutant a mouse human model tau of andtauopathy develop (PS19). filamentous PS19 mice tau used lesions in atthis 6 monthsstudy expr ofess age, mono subsequently-transgenic progressing P301S mutant to tau human accumulation tau and develop in association filamentous with neuronaltau lesions loss at by6 months 9–12 months of age, subsequently of age [31]. As progress this modeling to doestau accumulation not express amyloidin association pathology, with neuronal it is considered loss by ideal9–12 months for tauopathy-specific of age [31]. As this conditions. model does No studiesnot express have amyloid previously pathology, examined it is how considered DPP-4 inhibitors ideal for atauopathyffect tau pathology-specific condition and memorys. No deficitsstudies inhave tauopathy previously models. examined In the how present DPP study,-4 inhibitors we showed affect thattau linagliptinpathology and reversed memory HFD-induced deficits in tauopathy cognitive model declines. andWe previously increased CBF showed in PS19 that mice. linagliptin Unexpectedly, reversed however,HFD-induced phosphorylation cognitive decline of tau an wasd neitherincrease increasedd CBF in nor PS19 decreased, mice. Unexpectedly, as reported in however, previous studiesphosphorylation [27,29]. Such of tau contradictory was neither resultsincreased may nor stem decreased from di, ffaserences reported in animalin previous models, studies medication [27,29]. andSuch experimental contradictory design. result Kosarajus may stem et al. demonstratedfrom differences improved in animal tau pathology models, in medication 3xTg-AD mice and administeredexperimental with design. linagliptin Kosaraju orally et (5, al. 10, demonstrated and 20 mg/kg) improved for 8 weeks tau starting pathology from 9in months 3xTg- ofAD age mice [27]. However,administered the with study linagliptin reported aggravationorally (5, 10, of and tau 20 pathology, mg/kg) for used 8 weeks a 21 week-old starting from T2DM 9 months model rat of andage administered[27]. However sitagliptin, the study orally reported (100 ag mggravation/kg/day) forof tau 12 weekspathology, [29]. Sinceused NFTa 21 week pathology-old T2DM spreads model long beforerat and cognitive administered symptoms sitagliptin appear, orally the oral (100 administration mg/kg/day) for of linagliptin 12 weeks (10[29] mg. Since/kg/day) NFT in pathology this study startedspreads from long abefore much cognitive younger agesymptoms (6 weeks) appear, and for the a oral longer administration period of time of linagliptin (6.5 months). (10 Themg/kg/day) fact that ourin this PS19 study mice started were under from a additional much younger hyperglycemic age (6 weeks) stress and to induce for a longer cognitive period decline of time also (6.5 complicated months). theThe directfact that comparison our PS19 ofmice our were data withunder that additional of previous hyperglycemic studies. Nevertheless, stress to induce our report cognitive may providedecline importantalso complicated insights the for direct the ecomparisonffectiveness of and our long-term data with safety that of of previous DPP-4 inhibitors studies. Nevertheless in dementia, dueour toreport tauopathy. may provide important insights for the effectiveness and long-term safety of DPP-4 inhibitors in dementiaConsidering due to the tauopathy. lack of improvement in tauopathy by linagliptin, we concluded that the restored cognitiveConsidering function the was lack mainly of improvement attributed in to tauopathy increased CBF.by linagliptin, Indeed, previous we concluded studies that have the reportedrestored thatcognitive DPP-4 function inhibitors was improve mainly attributed vascular functionto increased via endothelium-dependentCBF. Indeed, previous studies vasodilatory have reported effects.

Int. J. Mol. Sci. 2019, 20, 2539 7 of 11

Hasegawa et al. showed that linagliptin ameliorated cognitive impairment, which was associated with an increase in CBF, at least in part mediated by an increase in cerebral phospho-eNOS [30]. However, we did not detect any significant changes in phosphorylated eNOS between linagliptin- and vehicle-treated HFD-fed mice. Several recent studies have suggested that the inhibition of DPP4 exerts beneficial pleiotropic effects aside from their anti-hyperglycemic properties. Linagliptin has been shown to improve cerebrovascular dysfunction observed in diabetes through anti-inflammatory and vasodilatory effects in a glucose-independent manner [32]. Alternatively, linagliptin may restore cerebrovascular function via the regulation of hyper-reactivity to endothelin 1 and Toll-like receptors 2 expression [33]. Treatment with sitagliptin was reported to reverse memory impairment in HFD-fed mice through reduced and enhanced neurogenesis [34]. In our study, while treatment with linagliptin in PS19 mice fed with a high-fat diet was able to restore cerebral perfusion, there was no similar improvement in cerebral perfusion in PS19 mice fed with a standard diet. Several limitations in this study should be addressed. Firstly, we conducted behavioral and biochemical analysis at 8 months only, while PS19 mice are reported to develop tau accumulations progressively by 9–12 months of age. Further studies are needed to elucidate whether the amount of phosphorylated tau or phosphorylated eNOS is increased in older age and if linagliptin is more effective at this time point. Secondly, we used a set dose of 10 mg/kg BW/day in this study. Considering that the previous report using 3xTg-AD mice demonstrated dose-dependent effects of linagliptin for cognitive function and tau phosphorylation [27], varying doses of linagliptin should be tested in future studies. Finally, we could not identify the exact mechanism of CBF increase by linagliptin. Further studies are warranted to elucidate the underlying mechanisms of the data presented here.

4. Materials and Methods

4.1. Animals PS19 transgenic mice expressing human (1N4R) with P301S driven by mouse protein promoter (Prnp) [31] were purchased from Jackson Laboratories. PS19 mice display age-related NFTs from 6 months of age, degeneration by 8 months of age, and subsequently behavioral deficits. These mice were maintained on a C57BL/6J background. All animal protocols were approved by the Institutional Animal Care and Use Committee at the National Cerebral and Cardiovascular Center (Permit Number: 15045, 16043, 1 April 2015), and were performed in accordance with the Guidelines for Proper Conduct of Animal Experiments established by Science Council of Japan.

4.2. Study Design Male heterozygous PS19 mice aged 6 weeks were randomly assigned to two groups and received either vehicle or linagliptin (100 mg/L, kindly provided by Boehringer Ingelheim) in their drinking water, at 10 mg/kg BW/day, which is frequently used and reported to be the most effective dose in a previous study [35]. Both groups were fed either with a low-fat diet (LFD; caloric composition, 10% fat, 70% carbohydrate, and 20% protein, Research Diet, Inc, New Brunswick, NJ, USA; vehicle, n = 15; linagliptin, n = 17) or high-fat diet (HFD; caloric composition, 60% fat, 20% carbohydrate, and 20% protein, Research Diet, Inc, New Brunswick, NJ, USA; vehicle, n = 15; linagliptin, n = 17) in order to induce hyperglycemia. At 7 months of age, evaluations of body weight, fasting serum glucose, and cerebral blood flow (CBF) were conducted. At 8 months of age, spatial reference memory was assessed by the Morris water maze test (MWM). After behavioral testing, mice were sacrificed under deep anesthesia to obtain a blood sample (Figure7). PS19 mice were then perfused transcardially with 0.01 mol/L phosphate buttered saline (PBS). Brains (n = 10 per group) were immediately removed and divided along the sagittal line into two hemispheres for Western blotting and immunohistochemistry. Int. J. Mol. Sci. 2019, 20, 2539 8 of 11 Int. J. Mol. Sci. 2019, 20, x 8 of 11

Figure 7. Scheme of experimental procedure. Figure 7. Scheme of experimental procedure. 4.3. Blood Testing 4.3. Blood Testing Blood samples for blood glucose level were collected from the tail vein after 6-hour fasting. BloodBlood glucose samples level for was blood measured glucose by level the were glucose collected dehydrogenase from the tail method vein after using 6- Glutesthour fasting. Neoα Blood(Sanwa glucose Kagaku level Kenkyusho, was measured Japan). by the For glucose the measurement dehydrogenase of GLP-1 method and using total cholesterol, Glutest Neoα serum (Sanwa was Kagakuprepared Kenkyusho, from blood samples Japan). collected For the from measurement the inferior vena of GLP cava-1 during and totalanimal cholesterol, sacrificing. Active serum GLP-1 was preparedconcentrations from blood in plasma samples were collected measured from bythe theinferior Active vena form cava Assay during -IBL animal (Immuno-Biological sacrificing. Active GLPLaboratories,-1 concentrations Gumma, in Japan). plasma Plasma were total measured cholesterol by the level Act wasive measuredform Assay by kit automated-IBL (Immuno clinical- Biologicalchemistry Laboratories, analyzer Fuji dri-chemGumma, 7000Japan). with Plasma Fuji dri-chem total cholesterol slide TCHO-PIII level was (Fujifilm, measured Tokyo, by automated Japan). clinical chemistry analyzer Fuji dri-chem 7000 with Fuji dri-chem slide TCHO-PIII (Fujifilm, Tokyo, Japan).4.4. Measurement of Cerebral Blood Flow Relative CBF of Tg-PS19 mice was measured by laser speckle flowmetry (Omegazone-2, 4Omegawave,.4. Measurement Fuchu, of Cerebral Japan), Blood as Flow previously reported [36]. Anesthesia was induced with 2%, andRelative maintained CBF with of 1.5% Tg-PS19 isoflurane. mice was The scalp measured was removed by laser by speckle midline flowmetry incision to expose(Omegazone the skull-2, Omegawave,throughout CBF Fuchu, evaluation. Japan), CBF as previously was measured reported in identically-sized, [36]. Anesthesia round was regions induced of interest with 2%, (1 mm and in maintaineddiameter), locatedwith 1.5% bilaterally isoflurane. at 1 mmThe posteriorscalp was and removed 2 mm lateral by mid fromline theincision bregma, to expose corresponding the skull to throughoutregions around CBF Heubner’s evaluation. anastomoses. CBF was measured Average in CBF identically values in-sized, the bilateral round hemispheres regions of interest were recorded. (1 mm in diameter), located bilaterally at 1 mm posterior and 2 mm lateral from the bregma, corresponding 4.5. Behavioral Analysis: Morris Water Maze Test to regions around Heubner’s anastomoses. Average CBF values in the bilateral hemispheres were recorded.After 6.5 months of linagliptin treatment, spatial reference learning and memory were evaluated by the Morris water maze test, consisting of a circular pool (diameter, 120 cm; depth, 40 cm) and 4a.5 set. Behavioral of video Analysis: analysis Morris systems water (EthoVision Maze Test XT5; Noldus, Wageningen, Netherlands), as previously describedAfter 6.5 [36 months]. During of training,linagliptin a platform treatment, (10 spatial cm in reference diameter) learning was submerged and memory 1 cm were below evaluated the water bysurface the Morris in the water center maze of one test, quadrant consist ofing the of pool a circular (target pool quadrant). (diameter, During 120 thecm; first depth, 4 days, 40 cm) all miceand a were set ofgiven video four analysis trials per systems day with (EthoVision a 30-minute XT5; interval Noldus, between Wageningen, attempts (acquisition Netherlands) phase)., as previously Mice were describedplaced at the[36] starting. During position training, (the a platform quadrant (10 adjacent cm in todiameter) the target) was and submerged released into 1 cm the bel water.ow the For water each surfacetrial, the in mousethe center was of given one 60quadrant seconds of to the reach pool a hidden(target quadrant). platform. TheDuring latency the tofirst reach 4 days, the platformall mice were(escape given latency), four trials swim per distance, day with and a 30 mean-minute swimming interval speedbetween were attempts recorded. (acquisition During the phase). probe Mice trial wereon the placed fifth at day, the mice starting were position allowed (th toe quadrant swim without adjacent the to platform the target) for and 60 secondsreleased tointo search the water. for it. ForThe each duration trial, ofthe swimming mouse was in given each quadrant60 seconds was to recorded.reach a hidden platform. The latency to reach the platform4.6. Western (escape Blotting latency), swim distance, and mean swimming speed were recorded. During the probe trial on the fifth day, mice were allowed to swim without the platform for 60 seconds to search for it. TheThe leftduration hemisphere of swimming was cut in coronally each quadrant into 2-mm-thick was recorded. slices (bregma +2 mm to +4 mm) and homogenized in RIPA buffer containing inhibitor cocktail and phosphatase inhibitor cocktail 4(all.6. Western from Nacalai Blotting Tesque, Kyoto, Japan). Ten micrograms of total protein was resolved by 10% SDS-PAGE, transferred onto Immobilon-P PVDF membranes (Millipore), and analyzed by Western The left hemisphere was cut coronally into 2-mm-thick slices (bregma +2 mm to +4 mm) and homogenized in RIPA buffer containing protease inhibitor cocktail and phosphatase inhibitor cocktail (all from Nacalai Tesque, Kyoto, Japan). Ten micrograms of total protein was resolved by

Int. J. Mol. Sci. 2019, 20, 2539 9 of 11 blotting. Primary antibodies used in this study included AT8 mouse monoclonal anti-human tau pSer202/Thr205 (Thermo Fisher Scientific, Rockford, IL, USA), rabbit polyclonal eNOS (Cell Signaling), rabbit monoclonal phospho-eNOS (Ser1177) (Cell Signaling), and rabbit polyclonal β-actin (Abcam). The densitometric value of each band was quantified using NIH-based Image-J software.

4.7. Immunohistochemistry Dissected right hemisphere samples (n = 10 per group) were post-fixed in 4% paraformaldehyde in 0.1M phosphate buffer overnight, embedded in paraffin, and sliced into 6-µm-thick sagittal sections starting from the position 1 mm lateral from the midline using a microtome (Leica, Nussloch, Germany). Sections were treated for heat-mediated antigen retrieval in 0.01M citrate buffer (pH 6.0) for 10 min. Endogenous peroxidase activity was inactivated by 3% hydrogen peroxide in methyl alcohol for 10 minutes. Immunostaining with mouse monoclonal anti-human tau pSer202/Thr205 antibody (Thermo Fisher Scientific) was performed with Vector M.O.M. Immunodetection Kit (Vector Laboratories). Visualization of bound peroxidase was achieved by incubation in TBS containing 0.05% diaminobenzidine (DAB) and 0.045% hydrogen peroxide.

4.8. Statistical Analysis All values are expressed as means SD unless stated otherwise in the figures. Individual ± comparisons were analyzed by a Student’s t test or ANOVA followed by post hoc Turkey tests. Differences with a probability value of p < 0.05 were considered statistically significant in all analyses. Statistical analysis was conducted using SPSS Statistics version 23.0 (IBM, Armonk, NY, USA).

5. Conclusions Our investigation indicates that the administration of linagliptin improves HFD-induced cognitive dysfunction in PS19 mice, rather than aggravating tau pathology. The present work thus suggests DPP-4 inhibitors may be suitable in the management of T2DM patients at risk of cognitive impairment, as a result of its beneficial effects on cerebral perfusion and cognitive function.

Author Contributions: Conceptualization, M.I.; methodology,Y.N., S.S. and T.M.; investigation and writing—original draft preparation, Y.N.; writing—review and editing, Y.Y. and S.S.; funding acquisition, and writing—review and editing M.I.; supervision, R.T. Funding: This research was partially supported by a grant from Boehringer Ingelheim. Acknowledgments: We thank Ahmad Khundakar for scientific input and editorial assistance. Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Abbreviations

T2DM Type 2 diabetes mellitus DPP-4 Dipeptidyl peptidase-4 HFD High-fat diet LFD Low-fat diet AD Alzheimer’s disease Aβ β-amyloid NFT neurofibrillary tangle GLP-1 glucagon-like peptide-1 FTDP-17 frontotemporal dementia and parkinsonism linked to chromosome 17 CBF cerebral blood flow MAPT microtubule-associated protein tau Int. J. Mol. Sci. 2019, 20, 2539 10 of 11

Prnp prion protein promoter PCR polymerase chain reaction MWM Morris water maze test NO nitric oxide EDHF endothelium-derived hyperpolarizing factor eNOS endothelial nitric oxide synthase PBS phosphate buttered saline RIPA radio-immunoprecipitation assay PVDF polyvinylidene difluoride DAB diaminobenzidine

References

1. Kumar, A.; Singh, A.; Ekavali. A review on Alzheimer’s disease pathophysiology and its management: An update. Pharmacol. Rep. 2015, 67, 195–203. [CrossRef] 2. Braak, H.; Braak, E. Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol. 1991, 82, 239–259. [CrossRef] 3. Hardy, J.; Selkoe, D.J. The amyloid hypothesis of Alzheimer’s disease: progress and problems on the road to therapeutics. Science 2002, 297, 353–356. [CrossRef] 4. Braak, H.; Del Tredici, K. The pathological process underlying Alzheimer’s disease in individuals under thirty. Acta Neuropathol. 2011, 121, 171–181. [CrossRef] 5. Hutton, M.; Lendon, C.L.; Rizzu, P.; Baker, M.; Froelich, S.; Houlden, H.; Pickering-Brown, S.; Chakraverty, S.; Isaacs, A.; Grover, A.; et al. Association of missense and 5’-splice-site mutations in tau with the inherited dementia FTDP-17. Nature 1998, 393, 702–705. [CrossRef] 6. Roberson, E.D.; Scearce-Levie, K.; Palop, J.J.; Yan, F.; Cheng, I.H.; Wu, T.; Gerstein, H.; Yu, G.Q.; Mucke, L. Reducing endogenous tau ameliorates -induced deficits in an Alzheimer’s disease mouse model. Science 2007, 316, 750–754. [CrossRef] 7. Arriagada, P.V.; Growdon, J.H.; Hedley-Whyte, E.T.; Hyman, B.T. Neurofibrillary tangles but not senile plaques parallel duration and severity of Alzheimer’s disease. Neurology. 1992, 42, 631–639. [CrossRef] 8. Koekkoek, P.S.; Kappelle, L.J.; van den Berg, E.; Rutten, G.E.; Biessels, G.J. Cognitive function in patients with diabetes mellitus: guidance for daily care. Lancet Neurol. 2015, 14, 329–340. [CrossRef] 9. Alam, F.; Islam, M.A.; Sasongko, T.H.; Gan, S.H. Type 2 Diabetes Mellitus and Alzheimer’s Disease: Bridging the Pathophysiology and Management. Curr. Pharm. Des. 2016, 22, 4430–4442. [CrossRef] 10. Bedse, G.; Di Domenico, F.; Serviddio, G.; Cassano, T. Aberrant insulin signaling in Alzheimer’s disease: current knowledge. Front Neurosci. 2015, 9, 204. [CrossRef] 11. Gupta, A.; Bisht, B.; Dey, C.S. Peripheral insulin-sensitizer drug metformin ameliorates neuronal insulin resistance and Alzheimer’s-like changes. Neuropharmacology 2011, 60, 910–920. [CrossRef] 12. El Khoury, N.B.; Gratuze, M.; Papon, M.A.; Bretteville, A.; Planel, E. Insulin dysfunction and Tau pathology. Front. Cell. Neurosci. 2014, 8, 22. [CrossRef] 13. Yue, J.T.; Lam, T.K. Lipid sensing and insulin resistance in the brain. Cell Metab. 2012, 15, 646–655. [CrossRef] 14. Levin-Allerhand, J.A.; Lominska, C.E.; Smith, J.D. Increased amyloid- levels in APPSWE transgenic mice treated chronically with a physiological high-fat high-cholesterol diet. J. Nutr. Health Aging 2002, 6, 315–319. 15. Leboucher, A.; Laurent, C.; Fernandez-Gomez, F.J.; Burnouf, S.; Troquier, L.; Eddarkaoui, S.; Demeyer, D.; Caillierez, R.; Zommer, N.; Vallez, E.; et al. Detrimental effects of diet-induced obesity on τ pathology are independent of insulin resistance in τ transgenic mice. Diabetes 2013, 62, 1681–1688. [CrossRef] 16. Lester-Coll, N.; Rivera, E.J.; Soscia, S.J.; Doiron, K.; Wands, J.R.; de la Monte, S.M. Intracerebral streptozotocin model of type 3 diabetes: relevance to sporadic Alzheimer’s disease. J. Alzheimers Dis. 2006, 9, 13–33. [CrossRef] 17. de la Monte, S.M. Insulin resistance and Alzheimer’s disease. BMB Rep. 2009, 42, 475–481. [CrossRef] 18. Yarchoan, M.; Arnold, S.E. Repurposing diabetes drugs for brain insulin resistance in Alzheimer disease. Diabetes 2014, 63, 2253–2261. [CrossRef] 19. Ahrén, B.; Schmitz, O. GLP-1 receptor agonists and DPP-4 inhibitors in the treatment of type 2 diabetes. Horm. Metab. Res. 2004, 36, 867–876. [CrossRef] Int. J. Mol. Sci. 2019, 20, 2539 11 of 11

20. Sebastião, I.; Candeias, E.; Santos, M.S.; de Oliveira, C.R.; Moreira, P.I.; Duarte, A.I. Insulin as a Bridge between Type 2 Diabetes and Alzheimer Disease - How Anti-Diabetics Could be a Solution for Dementia. Front. Endocrinol. (Lausanne). 2014, 5, 110. [CrossRef] 21. Hamilton, A.; Patterson, S.; Porter, D.; Gault, V.A.; Holscher, C. Novel GLP-1 mimetics developed to treat type 2 diabetes promote progenitor cell proliferation in the brain. J. Neurosci. Res. 2011, 89, 481–489. [CrossRef] [PubMed] 22. Harkavyi, A.; Whitton, P.S. Glucagon-like peptide 1 receptor stimulation as a means of neuroprotection. Br. J. Pharmacol. 2010, 159, 495–501. [CrossRef][PubMed] 23. Holscher, C. Incretin analogues that have been developed to treat type 2 diabetes hold promise as a novel treatment strategy for Alzheimer’s disease. Recent Pat. CNS Drug Discov. 2010, 5, 109–117. [CrossRef] [PubMed] 24. Talbot, K.; Wang, H.Y. The nature, significance, and glucagon-like peptide-1 analog treatment of brain insulin resistance in Alzheimer’s disease. Alzheimers Dement. 2014, 10, S12–S25. [CrossRef][PubMed] 25. Long-Smith, C.M.; Manning, S.; McClean, P.L.; Coakley, M.F.; O’Halloran, D.J.; Holscher, C.; O’Neill, C. The diabetes drug liraglutide ameliorates aberrant localisation and signalling in parallel with decreasing both amyloid-β plaque and glial pathology in a mouse model of Alzheimer’s disease. Neuromolecular Med. 2013, 15, 102–114. [CrossRef][PubMed] 26. McClean, P.L.; Hölscher, C. Liraglutide can reverse memory impairment, synaptic loss and reduce plaque load in aged APP/PS1 mice, a model of Alzheimer’s disease. Neuropharmacology. 2014, 76 Pt A, 57–67. [CrossRef] 27. Kosaraju, J.; Holsinger, R.M.D.; Guo, L.; Tam, K.Y. Linagliptin, a Dipeptidyl Peptidase-4 Inhibitor, Mitigates Cognitive Deficits and Pathology in the 3xTg-AD Mouse Model of Alzheime’s Disease. Mol. Neurobiol. 2017, 54, 6074–6084. [CrossRef][PubMed] 28. Hansen, H.H.; Barkholt, P.; Fabricius, K.; Jelsing, J.; Terwel, D.; Pyke, C.; Knudsen, L.B.; Vrang, N. The GLP-1 receptor agonist liraglutide reduces pathology-specific tau phosphorylation and improves motor function in a transgenic hTauP301L mouse model of tauopathy. Brain Res. 2016, 1634, 158–170. [CrossRef] 29. Kim, D.H.; Huh, J.W.; Jang, M.; Suh, J.H.; Kim, T.W.; Park, J.S.; Yoon, S.Y. Sitagliptin increases tau phosphorylation in the hippocampus of rats with type 2 diabetes and in primary neuron cultures. Neurobiol. Dis. 2012, 46, 52–58. [CrossRef] 30. Hasegawa, Y.; Hayashi, K.; Takemoto, Y.; Cheng, C.; Takane, K.; Lin, B.; Komohara, Y.; Kim-Mitsuyama, S. DPP-4 inhibition with linagliptin ameliorates the progression of premature aging in klotho-/- mice. Cardiovasc. Diabetol. 2017, 16, 154. [CrossRef] 31. Yoshiyama, Y.; Higuchi, M.; Zhang, B.; Huang, S.M.; Iwata, N.; Saido, T.C.; Maeda, J.; Suhara, T.; Trojanowski, J.Q.; Lee, V.M. loss and microglial activation precede tangles in a P301S tauopathy mouse model. Neuron 2007, 53, 337–351. [CrossRef][PubMed] 32. Kröller-Schön, S.; Knorr, M.; Hausding, M.; Oelze, M.; Schuff, A.; Schell, R.; Sudowe, S.; Scholz, A.; Daub, S.; Karbach, S.; et al. Glucose-independent improvement of vascular dysfunction in experimental sepsis by dipeptidyl-peptidase 4 inhibition. Cardiovasc. Res. 2012, 96, 140–149. [CrossRef][PubMed] 33. Hardigan, T.; Abdul, Y.; Ergul, A. Linagliptin reduces effects of ET-1 and TLR2-mediated cerebrovascular hyperreactivity in diabetes. Life Sci. 2016, 159, 90–96. [CrossRef][PubMed] 34. Gault, V.A.; Lennox, R.; Flatt, P.R. Sitagliptin, a dipeptidyl peptidase-4 inhibitor, improves recognition memory, oxidative stress and hippocampal neurogenesis and upregulates key involved in cognitive decline. Diabetes Obes. Metab. 2015, 17, 403–413. [CrossRef] 35. Ma, M.; Hasegawa, Y.; Koibuchi, N.; Toyama, K.; Uekawa, K.; Nakagawa, T.; Lin, B.; Kim-Mitsuyama, S. DPP-4 inhibition with linagliptin ameliorates cognitive impairment and brain atrophy induced by transient cerebral ischemia in type 2 diabetic mice. Cardiovasc. Diabetol. 2015, 14, 54. [CrossRef][PubMed] 36. Hattori, Y.; Enmi, J.; Kitamura, A.; Yamamoto, Y.; Saito, S.; Takahashi, Y.; Iguchi, S.; Tsuji, M.; Yamahara, K.; Nagatsuka, K.; et al. A novel mouse model of subcortical infarcts with dementia. J. Neurosci. 2015, 35, 3915–3928. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).