Reduction of glucose metabolism in brain is one of the main features of Alzheimer’s disease. Thiamine (vitamin B1)-dependent processes are critical in glucose metabolism and have been found to be impaired in brains from patients with Alzheimer’s disease. However, thiamine treatment exerts little beneficial effect in these patients. Here, we tested the effect of benfotiamine, a thiamine derivative with better bioavailability than thiamine, on cognitive impairment and pathology alterations in a mouse model of Alzheimer’s disease, the amyloid precursor protein/presenilin-1 transgenic mouse. We show that after a chronic 8 week treatment, benfotiamine dose-dependently enhanced the spatial memory of amyloid precursor protein/presenilin-1 mice in the Morris water maze test. Furthermore, benfotiamine effectively reduced both amyloid plaque numbers and phosphorylated tau levels in cortical areas of the transgenic mice brains. Unexpectedly, these effects were not mimicked by another lipophilic thiamine derivative, fursultiamine, although both benfotiamine and fursultiamine were effective in increasing the levels of free thiamine in the brain. Most notably, benfotiamine, but not fursultiamine, significantly elevated the phosphorylation level of glycogen synthase kinase-3α and -3β, and reduced their enzymatic activities in the amyloid precursor protein/presenilin-1 transgenic brain. Therefore, in the animal Alzheimer’s disease model, benfotiamine appears to improve the cognitive function and reduce amyloid deposition via thiamine-independent mechanisms, which are likely to include the suppression of glycogen synthase kinase-3 activities. These results suggest that, unlike many other thiamine-related drugs, benfotiamine may be beneficial for clinical Alzheimer’s disease treatment.
glycogen synthase kinase-3
The drastic disturbance of glucose metabolism in the brain is one of the striking features of Alzheimer’s disease. A significant reduction in cerebral glucose consumption can precede overt clinical symptoms or brain atrophy, even for decades (Kennedy et al., 1995; Herholz et al., 2007; Mosconi et al., 2008). Furthermore, type 2 diabetes mellitus, the most common metabolic disorder with a prevalence that increases with ageing just like Alzheimer’s disease, has been identified as a risk factor for Alzheimer’s disease (Arvanitakis et al., 2004). Conversely, patients with Alzheimer’s disease also have a higher risk of developing type 2 diabetes mellitus (Janson et al., 2004). Thus, Alzheimer’s disease has been considered to be an ‘insulin-resistant brain state’ or even a type 3 diabetes mellitus (Steen et al., 2005). These previous studies, therefore, indicate that a disturbance of glucose metabolism in the brain is associated with the pathogenesis of Alzheimer’s disease.
Krebs cycle and the pentose phosphate pathway are two predominant pathways for glucose metabolism in mammalian brain. As a cofactor for several rate-limiting enzymes (alpha-ketoglutarate dehydrogenase and pyruvate-dehydrogenase in the Krebs cycle and transketolase in the non-oxidative branch of the pentose phosphate pathway), thiamine is critical for glucose metabolism of the brain (Gibson et al., 1988; Mastrogiacoma et al., 1996; Zhao et al., 2009). It has been known for a long time that thiamine deficiency and the disturbance of thiamine-dependent processes in glucose metabolism are associated with Alzheimer’s disease (Butterworth and Besnard, 1990; Mastrogiacoma et al., 1996; Gold et al., 1998; Gibson and Blass, 2007). The pathophysiological alterations induced by thiamine deficiency are very similar to that of Alzheimer’s disease, including selective neuron loss, tau protein hyperphosphorylation and neurofibrillary tangle formation, as well as increased β-amyloid secretion with abnormal deposition in hippocampus and its vicinity regions (Calingasan et al., 1995; Cullen and Halliday, 1995). Recent studies have shown that thiamine deficiency increased β-amyloid generation and accumulation in the brain of Alzheimer’s disease mouse models (Karuppagounder et al., 2009; Zhang et al., 2009). Moreover, thiamine-dependent enzyme activities are decreased in Alzheimer’s disease and some other neurodegenerative diseases (Gibson et al., 1988; Mastrogiacoma et al., 1996). Furthermore, a reduction in the serum thiamine level characteristically occurs in Alzheimer’s disease but not in Parkinson’s disease (Gold et al., 1998), supporting the potential use of thiamine in clinical treatment of Alzheimer’s disease.
Disappointingly, thiamine itself has not shown a dramatic benefit in clinical trials of Alzheimer’s disease (Blass et al., 1988; Nolan et al., 1991; Rodriguez-Martin et al., 2000). Nolan et al. (1991) reported that no significant difference was found in the Mini-Mental State Examination, verbal learning and naming scores between the placebo and thiamine groups at any point in a 12-month, double-blind, parallel-group study with 3 g/day thiamine administration. The poor bioavailability of thiamine has been suggested to be a possible explanation for its lack of effect in Alzheimer’s disease patients.
Benfotiamine and fursultiamine are generally considered as lipid-soluble thiamine derivatives with better bioavailability than thiamine (Loew, 1996), although benfotiamine is practically insoluble in oil and several organic solvents (Volvert et al., 2008). Especially, benfotiamine has been shown to be effective on diabetic nerve complications such as diabetic neuropathy and retinopathy (Hammes et al., 2003; Varkonyi and Kempler, 2008). So far, only one clinical study reported that fursultiamine at an oral dose of 100 mg/day had a mild beneficial effect on patients with Alzheimer’s disease in a 12-week open trial (Mimori et al., 1996). To seek experimental evidence for further clinical investigations, here we explored whether benfotiamine and fursultiamine exhibit beneficial effects on cognitive impairment and pathological alterations in a rodent Alzheimer’s disease model: amyloid precursor protein (APP)/presenilin-1 (PS1) transgenic mice (carrying both mutant APP and PS1 transgenes) (Holcomb et al., 1998). Surprisingly, our results suggest that only benfotiamine, but not fursultiamine, is beneficial and the increase in brain thiamine levels does not appear to be sufficient for the beneficial effects.
Materials and methods
APP/PS1 double-transgenic mice used in this study were obtained from the Jackson Laboratory [strain name, B6C3-Tg(APPswe,PSEN1dE9)85Dbo/J; stock number 004462]. These mice express a chimeric mouse/human APP containing the K595N/M596L Swedish mutations and a mutant human PS1 carrying the exon 9-deleted variant under the control of mouse prion promoter elements, directing the transgene expression predominantly to CNS neurons. All male APP/PS1 transgenic mice were produced by the Model Animal Research Centre of Nanjing University. The genotype was confirmed by polymerase chain reaction analysis of tail biopsies. They were individually housed in Plexiglas cages with free access to food and water and maintained on a 12/12 hour light-dark cycle in a temperature-controlled room (22°C). All mouse care and experimental procedures were approved by Medical Experimental Animal Administrative Committee of Fudan University, and by the Institutional Animals Care and Use Committee of the Institute of Neuroscience, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences.
Drugs and treatment
Benfotiamine was purchased from Shanghai Rixin Biomedical Company (Shanghai, China). The batch used in the present study was > 99.0% pure as determined by high performance liquid chromatography (data not shown). Fursultiamine was purchased from Xinyi Jiufu Pharmaceutical Ltd. (Shanghai, China). All other drugs and chemicals in these experiments were purchased from Sigma (St. Louis, MO). Mice were randomly assigned to six groups, with n = 16 per group. Three groups received different doses (50, 100, 200 mg/kg/day) of benfotiamine dissolved in 0.7% carboxymethylcellulose at a volume ratio of 0.2ml/10 g by gastric gavage daily for 8 weeks. One group was administrated with fursultiamine (dissolved in carboxymethylcellulose, 100mg/kg/day). A positive control group was administrated with huperzine A (66.7 µg/kg), a cholinesterase inhibitor (Bai et al., 2000). A solvent control group was treated with equal volume of 0.7% carboxymethylcellulose for 8 weeks.
Morris water maze
The Morris water maze consisted of a circular pool (100 cm diameter, 50 cm deep) filled with water at 24–26°C to a depth of 20 cm. The water surface was covered with floating black resin beads. Yellow curtains were drawn around the pool (50 cm from the pool periphery) and contained distinctive visual marks that served as distal cues. Before training, a 60 s free swim trial was run without the platform. For training, a submerged (1.5 cm below the surface of water, invisible to the animal) platform was fixed in the centre of a quadrant so that the animal had to learn the location of the platform, which was the only getaway from the water. The training session lasted for 4 days (four trials per day). A trial was terminated when the mouse had climbed onto the escape platform or when 60 s had elapsed. Each mouse was allowed to stay on the platform for 30 s. The probe test was performed on the 5th day. The platform was removed and the mouse behaviour was recorded for 60 s. Swimming paths in probe test were monitored using an automatic tracking system. This system was used to record the swimming trace and calculate the latency to the platform and the time spent in each quadrant. Crossing times represent the times the animal crossed the position where the platform was placed during the learning session. The target quadrant occupancy represents the percent time the mouse spent in the quadrant where the platform was placed during the learning session, while the opposite quadrant occupancy means the percent time the mouse spent in the quadrant opposite to the target quadrant.
For immunohistochemistry, mice were sacrificed by transcardial perfusion of 10 ml of saline followed by 50 ml of 4% paraformaldehyde in 0.2 M phosphate buffer (pH 7.2). Brains were removed and post-fixed in 4% paraformaldehyde overnight, followed by immersion first in 75% and then in 85% ethanol before being embedded with paraffin. Coronal paraffin sections (6 μm thick) were deparaffinized in xylene and rehydrated in a series of graded alcohols. After being washed in distilled water followed by phosphate buffered saline at pH 7.4, endogenous peroxidase activity was quenched by immersing the sections in 40% methanol containing 3% H2O2 at room temperature for 30 min. After being washed three times in phosphate buffered saline, antigen retrieval was achieved by boiling sections in 0.01 M citrate buffer (pH 6.0) for 15 min. For examination of amyloid plaque, the sections were incubated in 98% formic acid at room temperature for 2 min. Non-specific binding sites were blocked by treatment with goat serum for 1 h. The tissues were then incubated with primary antibody dilutions (rabbit anti-β-amyloid42, 1:100, Calbiochem, CA, USA; rabbit anti-phosphorylated tau Ser396, 1:100, Santa cruz, CA, USA) at 4°C overnight. The UltraSensitiveTM S-P stain system (Maixin Biotechnology Development Co., Ltd. Fuzhou, China) was applied according to the manufacturer’s instructions. Sections were developed in diaminobenzidine solution and counterstained with haematoxylin. Then, sections were observed by a computerized image system composed of a Leica CCD camera DFC420 connected to a Leica DM IRE2 microscope (Leica Microsystems Ltd, Wetzlar, Germany) and images were captured by the Leica QWin Plus v3 software under the same conditions. The plaque numbers were measured as the total number of six fields (100×) per section per mouse. Phosphorylated tau Ser396 antibody is immunoreactive with intra-neuronal as well as extra-neuronal neurofibrillary tangles. In our study, 6 μm thick coronal paraffin sections were analysed for immunohistochemistry so that typical integral axons were sparsely seen on the thin brain section. Thus, we counted the phosphorylated tau-positive cell numbers in the cortex and measured them as the average of eight cortex fields (200×) per section per mouse.
Detection of levels of brain thiamine and its phosphate esters
The effects of benfotiamine and fursultiamine on the levels of thiamine and its phosphate esters in blood and brain were examined in adult (2-month-old) wild-type mice. Both drugs were administrated at the dose of 100 mg/kg by gastric gavage in two different manners: a single administration and a consecutive 10 day administration. At 1 h after the drug administration, whole blood samples were collected into tubes containing the anticoagulant disodium EDTA (1 mg/ml). The samples were deproteinized with 5% trichloroacetic acid for 20 min on ice, centrifuged at 20 000g at 4°C for 10 min, and then the supernatant was stored at −80°C. The brains were removed after perfusion with ice-cold saline. All tissues were immediately frozen in liquid nitrogen and stored at − 80°C until sample preparation. A high performance liquid chromatography method for determining thiamine and its phosphate esters was used following Lu and Frank (2008) with a minor modification. In brief, tissues were homogenized using a Polytron homogenizer and the proteins were precipitated by trichloroacetic acid for 20 min on ice. After centrifugation at 20 000g at 4°C for 20 min, the samples of supernatant were kept at −80°C until use. Thiamine and its phosphate esters were derivatized using potassium ferricyanide into thiochromes. The derivatives of thiamine and it phosphate esters were separated by gradient elusion with a C18 reversed-phase analytical column (250 × 4.6 mm) and measured by fluoroscope with an excitation wavelength of 375 nm and emission wavelength of 435 nm. For quantification, peak areas were measured.
In the western blotting experiments, cortices were isolated from dissected brains of drug-treated mice, and then homogenized using a Polytron homogenizer. Protein samples were separated by 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes (Amersham Pharmacia Biotech, Buckinghamshire, UK). The membranes were blocked with 3% bovine serum albumin and incubated with specific antibodies for glycogen synthase kinase-3 (GSK-3)-β (1:1000, Cell Signaling, MA, USA), phospho-GSK-3β (Ser9) (1:1000, Cell Signaling), GSK-3α (1:1000, Cell Signaling), phospho-GSK-3α (Ser21) (1:2000, Cell Signaling), Akt (1:1000, Cell Signaling), phospho-Akt (Ser473) (1:1000, Cell Signaling), β-site of APP cleaving enzyme (BACE) (1:500, Millipore) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (1:1000, KangChen, Shanghai, China). The membranes were washed with Tris-buffered saline containing 0.05% Tween-20 and then processed using either a horseradish peroxidase-conjugated anti-rabbit or anti-mouse antibody (Millipore, MA, USA) followed by enhanced chemiluminescence detection (Amersham Pharmacia Biotech). The bands on the radiographic films were scanned and quantified with an image analyser (Imagequant 5.2, GE healthcare, Buckinghamshire, UK).
Enzyme activity assay
Activity of β-secretase was determined using a commercial kit according to manufacturer’s instructions (Biovision Inc., CA, USA). Briefly, one half of the brain was washed twice in ice-cold phosphate buffered saline and then homogenized with 500 µl extraction buffer. After 10 min incubation on ice, the extract was centrifuged at 10 000g for 5 min. An aliquot (50 µl) of the supernatant was mixed with an equal volume of 2 × reaction buffer and 2 µl substrate in a 96-well microplate. The plate was kept in the dark at 37°C for 1.5 h, and the fluorescence was recorded using FlexStation 384 (Molecular Devices, Sunnyvale, CA, USA). Protein concentrations were determined with the bicinchoninic acid assay (Sigma).
Activities of GSK-3, GSK-3α and GSK-3β were determined by the GENMED kit (Genmed, MA, USA) according to manufacturer’s instructions. Briefly, the whole brains of drug-treated mice were rinsed twice by reagent A and then homogenized with the extraction buffer reagent B. After a 30 min incubation on ice, the homogenized mixture was centrifuged twice at 10 000g for 10 min at 4°C. The supernatant was collected and then assayed for enzyme activity, by mixing 65 µl reagent C, 10 µl reagent D, 10 µl reagent E and 10 µl reagent F and incubation at 30°C for 3 min. Immediately after the addition of 5 µl supernatant to the reagent mixture, the optical density was measured at 340 nm (SpectraMax 190, Molecular Devices, CA, USA) every 30 s for 2 min. The activity was measured as the difference between the absorbance value at 0 and 120 s. The assay was repeated twice for each sample.
All data are shown as the mean ± SEM, with statistical significance assessed by Student’s t-test. All statistical analyses were performed using Origin 7.0 (OriginLab, USA).
Chronic benfotiamine treatment improved cognitive function of APP/PS1 mice
Benfotiamine and fursultiamine were gastrointestinally administered to APP/PS1 mice (20-weeks-old) daily for 8 weeks (Fig. 1A). The use of 20-week-old APP/PS1 mice to receive the drug treatment is based on previous reports demonstrating that these mice begin to have β-amyloid plaques as early as 2.5 months of age and have a high β-amyloid load in hippocampal and cortical sub-areas from 6 months of age (Blanchard et al., 2003; Trinchese et al., 2004). In our study, these mice were examined at 28 weeks of age, when the cognitive impairment and pathological alterations were evident.
Benfotiamine but not fursultiamine restores the impaired spatial memory of APP/PS1 mice in the Morris water maze test. (A) Typical swimming traces during probe test showing the effects of benfotiamine (BT: 50, 100, 200 mg/kg), fursultiamine (FT: 100 mg/kg) and a positive control drug, huperzine A (Hup: 66.7 µg/kg), on the spatial memory of APP/PS1 mice. ‘T’ and ‘O’ indicate the target and opposite quadrant, respectively. Drugs were gastrointestinally administered to APP/PS1 mice (20 weeks old) daily for 8 weeks. (B) Statistical results showing the effects of different drugs on the crossing times of APP/PS1 mice, which represent the times the animals crossed the position where the platform was placed during the learning session. Benfotiamine dose-dependently increased the crossing times, but fursultiamine had no significant effect. n = 10 for each group. *P < 0.05; **P < 0.01; NS = no significant difference; drug-treated groups versus control group (Student’s t-test). (C) Statistical results showing the effects of different drugs on the target or opposite quadrant occupancy, which represent the percentage of time animals spent in target or opposite quadrant, respectively. Benfotiamine dose-dependently increased the preference for target quadrant of APP/PS1 mice, as shown by the increased target quadrant occupancy (**P < 0.01) and reduced opposite quadrant occupancy (##P < 0.01), but fursultiamine had no significant effect. n = 10 for each group; NS = no significant difference; drug-treated groups versus control group (Student’s t-test).
In the Morris water maze test, mice were trained four trials per day for 4 days and then tested on the fifth day. APP/PS1 mice with vehicle administration showed no significant preference for the position of the platform, as reflected by the crossing times and the target quadrant occupancy (Fig. 1). The cholinesterase inhibitor, huperzine A significantly improved spatial memory as shown by the increased crossing times (P < 0.01, compared with the control group, Fig. 1B) and target quadrant occupancy (P < 0.01, Fig. 1C), and the reduced opposite quadrant occupancy (P < 0.01, Fig.1C). Notably, we found that benfotiamine dose-dependently (50, 100, 200 mg/kg) improved the spatial memory in APP/PS1 mice (crossing times, P < 0.05 for 50 mg/kg, P < 0.01 for 100 mg/kg and 200 mg/kg, Fig. 1B; target quadrant occupancy, P < 0.01 for all groups, Fig. 1C; opposite quadrant occupancy, P < 0.01 for all groups, Fig. 1C). Interestingly, fursultiamine (100 mg/kg), another thiamine derivative, had no such effects (P > 0.1 for all groups, Fig. 1). These results indicate that benfotiamine has beneficial effects in the treatment of cognitive impairment in the APP/PS1 mice.
Chronic benfotiamine treatment alleviated the pathological alterations in APP/PS1 mice
To study the effect of benfotiamine on the pathological alterations in Alzheimer’s disease, we first examined the formation of amyloid plaques, which is a classical pathological hallmark of Alzheimer’s disease. By immunochemical staining in cortical slices, we analysed the numbers of amyloid plaques in APP/PS1 mice after drug treatment. We found that chronic treatment of benfotiamine significantly reduced the number of amyloid plaques (P < 0.01 for 50, 100, 200 mg/kg, Fig. 2). By contrast, fursultiamine had no effect on the number of amyloid plaques (P > 0.1 for all groups, Fig. 2). Furthermore, we examined the phosphorylation of tau, and also found that only chronic treatment of benfotiamine (P < 0.01 for 50, 100, 200 mg/kg) but not fursultiamine (P > 0.1 for all groups) significantly reduced the number of the phosphorylated tau-positive cells (Fig. 3). Taken together, these results indicate that benfotiamine, but not fursultiamine, exerts robust neuroprotection against cognitive impairment and pathological alterations in the Alzheimer’s disease mouse model.
Benfotiamine but not fursultiamine reduces the formation of amyloid plaques in APP/PS1 mice. Benfotiamine dose dependently (BT: 50, 100, 200 mg/kg) reduced the number of amyloid plaques in APP/PS1 mice after 8 weeks of treatment. Fursultiamine (FT: 100 mg/kg) had no such effect. The plaque numbers were measured as the total number of six fields (100×) per section per mouse. Arrows indicate typical amyloid plaques (A). Scale bars: 100 µM. (B) Statistical results showing the effects of different drugs on the plaque numbers. n = 6 for each group. *P < 0.05; **P < 0.01; NS = no significant difference; drug-treated groups versus control group (Student’s t-test).
Benfotiamine but not fursultiamine reduces tau hyperphosphorylation in APP/PS1 mice. Benfotiamine dose-dependently (BT: 50, 100, 200 mg/kg) reduced the number of phosphorylated tau-positive cells in APP/PS1 mice after 8-week treatment. Fursultiamine (FT: 100 mg/kg) had no such effect. The phosphorylated tau-positive cell numbers were measured as the average of eight cortex fields (200 × ) per section per mouse. Arrows indicate typical phosphorylated tau-positive cells (A). Scale bars: 10 µM. (B) Statistical results showing the effects of different drugs on the phosphorylated tau-positive cell numbers. n = 6 for each group. **P < 0.01; NS = no significant difference; drug-treated groups versus control group (Student’s t-test).
Both benfotiamine and fursultiamine increased levels of thiamine but not its phosphate esters in the brain
Next, we investigated the effects of benfotiamine and fursultiamine on levels of thiamine and its phosphate esters in blood and brain. Using high performance liquid chromatography examination, we found that in adult wild-type mice, both benfotiamine and fursultiamine significantly increased the blood concentrations of free thiamine, thiamine monophosphate (TMP) and thiamine pyrophosphate (TPP) at 1 h after a single administration or a consecutive 10 day administration, at the dose of 100 mg/kg (Fig. 4A). Benfotiamine showed better bioavailability than fursultiamine, as shown by the significantly higher blood concentrations of thiamine and TPP after benfotiamine treatment than fursultiamine (Fig. 4A). However, in the brain, after a single administration or a consecutive 10-day administration, benfotiamine and fursultiamine only elevated the level of free thiamine, but not that of TMP and TDP (Fig. 4B), consistent with the results shown by previous studies (Volvert et al., 2008; Nozaki et al., 2009). Furthermore, the increased thiamine level in the brain did not differ between benfotiamine and fursultiamine after the chronic administration (Fig. 4B), despite the slightly greater effect of benfotiamine than fursultiamine on the brain thiamine level after the single administration (Fig. 4B). Taken together, these results suggest that the beneficial effect of benfotiamine in APP/PS1 mice cannot be solely attributed to the increase in the brain levels of thiamine and/or its phosphate esters.
Both benfotiamine and fursultiamine increase the level of thiamine but not that of its phosphate esters in the brain. Statistical results showing the effects of benfotiamine (BT) and fursultiamine (FT) on the levels of free thiamine, TMP and TPP in the blood (A) and brain (B). Both drugs were administrated at the dose of 100 mg/kg. The levels of thiamine, TMP and TPP were examined one hour after the last drug administration. Single = single drug administration; multiple = consecutive 10-day drug administration. n = 10 for each group. ***P < 0.001; NS = no significant difference; drug-treated groups versus control group (Student’s t-test).
Chronic benfotiamine treatment had no effect on β-secretase activity
A previous study showed that thiamine deficiency exacerbated the plaque pathology in the mouse model of Alzheimer’s disease through a β-secretase (BACE)-dependent mechanism (Karuppagounder et al., 2009). Thus, we examined further whether BACE is also involved in the beneficial effect of benfotiamine. We found that neither benfotiamine nor fursultiamine had any effect on the expression and activity of BACE (P > 0.05 for all groups, Fig. 5). Together with the finding that benfotiamine and fursultiamine had similar effects on the brain level of thiamine, these results suggest that benfotiamine most likely exerts its beneficial effects through thiamine-independent pathways.
Chronic benfotiamine treatment has no effect on the β-secretase expression and activity. (A) Representative enhanced chemiluminescence radiographs of immunoblots showing the immunoreactivities of BACE and GAPDH in cortical protein samples from APP/PS1 mice with 8-week benfotiamine (BT: 50, 100, 200 mg/kg) or fursultiamine (FT: 100 mg/kg) treatment. (B) Statistical results showing the normalized protein level of BACE. n = 4 for each group. (C) Statistical results showing the normalized enzyme activities of BACE. n = 7 for each group. NS = no significant difference; drug-treated groups versus control group (Student’s t-test).
Chronic benfotiamine treatment inhibited the activities of GSK-3
Recently, GSK-3 has been considered to be a new mechanism in the pathogenesis of Alzheimer’s disease. In particular, inhibition of GSK-3 activity blocked the production of β-amyloid peptides by interfering with APP cleavage at the γ-secretase but not β-secretase step (Phiel et al., 2003). To explore the role of GSK-3 in the beneficial effect of benfotiamine, we examined the phosphorylation levels and enzymatic activities of GSK-3 in APP/PS1 mice with the chronic drug treatment. We found that benfotiamine at 100 and 200 mg/kg significantly increased the ratio of phospho-GSK-3α (Ser21)/GSK-3α (P < 0.05 for 100 mg/kg, P < 0.01 for 200 mg/kg, compared to untreated, Fig. 6A and B) and phospho-GSK-3β (Ser9)/GSK-3β (P < 0.05 for 100 mg/kg, P < 0.01 for 200 mg/kg, Fig. 6C and D). We further examined the phosphorylation level of Akt, an upstream kinase of GSK-3β, and found that benfotiamine at 100 and 200 mg/kg also significantly increased the ratio of phospho-Akt (Ser473)/Akt (P < 0.05 for 100 mg/kg, P < 0.01 for 200 mg/kg, Fig. 6C and E). Increased phosphorylation levels of GSK-3α (Ser21) and GSK-3β (Ser9) indicate decreased activities for both enzymes. Consistently, we found that benfotiamine at 100 and 200 mg/kg significantly decreased the enzyme activity of total GSK-3 (P = 0.06 for 100 mg/kg, P < 0.01 for 200 mg/kg, compared to untreated, Fig. 7A), as well as that of GSK-3α (P < 0.05 for 100 mg/kg, P < 0.01 for 200 mg/kg, Fig. 7B) and GSK-3β (P = 0.08 for 100 mg/kg, P < 0.01 for 200 mg/kg, Fig. 7C). On the contrary, fursultiamine (100 mg/kg) had no effect on the phosphorylation levels of GSK-3α, GSK-3β and Akt (P > 0.1 for all groups, Fig. 6), and did not affect the enzymatic activities of total GSK-3, GSK-3α and GSK-3β (P > 0.1 for all groups, Fig. 7). These results indicate that chronic benfotiamine treatment affects the phosphorylation of both GSK-3α and GSK-3β, and reduces their enzymatic activities, indicative of a possible mechanism underlying the beneficial effect of this drug in APP/PS1 mice.
Benfotiamine but not fursultiamine increases the phosphorylation of GSK-3 in APP/PS1 mice. (A) Representative enhanced chemiluminescence radiographs of immunoblots showing the immunoreactivities of pGSK-3α (Ser21), GSK-3α and GAPDH in cortical protein samples from APP/PS1 mice with 8-week benfotiamine (BT: 50, 100, 200 mg/kg) or fursultiamine (FT: 100 mg/kg) treatment. (B) Statistical results showing the normalized immunoreactivity ratios of pGSK-3α (Ser21)/GSK-3α. (C) Representative enhanced chemiluminescence radiographs of immunoblots showing the immunoreactivities of pGSK-3β (Ser9), GSK-3β, pAkt (Ser473), Akt and GAPDH. (D, E) Statistical results showing the normalized immunoreactivity ratios of pGSK-3β (Ser9)/GSK-3β and pAkt (Ser473)/Akt. Benfotiamine at 100 and 200 mg/kg significantly increased the ratios of pGSK-3α/GSK-3α (B), pGSK-3β/GSK-3β (D) and pAkt/Akt (E). n = 4 for each group; *P < 0.05; **P < 0.01; NS = no significant difference; drug-treated groups versus control group (Student’s t-test).
Benfotiamine but not fursultiamine reduces GSK-3 enzyme activities in APP/PS1 mice. Statistical results showing the normalized enzyme activities of total GSK-3 (A), GSK-3α (B), GSK-3β (C) in cortical protein samples from APP/PS1 mice with 8-week benfotiamine (BT: 50, 100, 200 mg/kg) or fursultiamine (FT: 100 mg/kg) treatment. Benfotiamine at 100 and 200 mg/kg significantly reduced the enzyme activities of total GSK-3 (A), GSK-3α (B) and GSK-3β (C). n = 7 for each group. *P < 0.05; **P < 0.01; NS = no significant difference; drug-treated groups versus control group (Student’s t-test).
In this study, we examined whether benfotiamine and fursultiamine have beneficial pharmacological effects on the behaviour of APP/PS1 mice. We found that only benfotiamine ameliorated cognitive function of APP/PS1 mice after eight-week administration. Data from histological examinations of amyloid plaques are consistent with the notion that benfotiamine, but not fursultiamine, exerts a unique beneficial effect on APP/PS1 mice. As a thiamine derivative with better bioactivity, benfotiamine has been shown to block three major pathways of hyperglycaemic damage (the hexosamine pathway, the advanced glycation end product formation pathway and the diacylglycerol-protein kinase C pathway), as well as hyperglycaemia-associated nuclear factor-κB activation, by activating transketolase, thus preventing the development and progression of diabetic nerve complications (Hammes et al., 2003). Moreover, all of the mechanisms described for diabetic pathologies are specifically linked to abnormally high concentrations of blood glucose, reflecting a single hyperglycaemia-induced process of superoxide overproduction by the mitochondrial electron transport chain (Nishikawa et al., 2000). Although oxidative stress is also observed in Alzheimer’s disease (Markesbery, 1997), there is no evidence that thiamine and its derivatives could decrease superoxide production in Alzheimer’s disease brain.
A possible mechanism underlying the differential effects of benfotiamine and fursultiamine is that the two drugs may have different pharmacokinetics and different effects on the brain level of thiamine. Thiamine has been shown to be a mild cholinesterase inhibitor (Gjone and Hakon Skramstad, 1955; Alspach and Ingraham, 1977), which is generally known to improve cognitive function in Alzheimer’s disease treatment (Birks, 2006) and might therefore partially underlie the observed effect of benfotiamine. Indeed, we observed a marked beneficial effect of the cholinesterase inhibitor huperzine A on cognitive impairment in the APP/PS1 mice. However, since benfotiamine and fursultiamine showed a similar effect on modulating free thiamine levels in the brain and no effect on TMP and TPP after chronic administration, it is unlikely that thiamine-mediated inhibition of cholinesterases is the reason for the beneficial effects of benfotiamine. Furthermore, inconsistent with recent studies indicating that thiamine deficiency enhanced β-amyloid generation and accumulation in brain by increasing β-secretase activity (Karuppagounder et al., 2009; Zhang et al., 2009), we did not observe any change in the expression and activity of β-secretase after chronic benfotiamine and fursultiamine treatment. Taken together, these results support the idea that the thiamine-dependent pathways are not the main mechanisms underlying the beneficial effect of benfotiamine on cognitive impairment and pathological alterations in the APP/PS1 mice. On the other hand, benfotiamine had better bioavailability than fursultiamine in increasing the blood levels of free thiamine and TPP. Whether this difference contributes to the different effects of benfotiamine and fursultiamine still needs further investigation. Furthermore, in the current study, we have only examined the levels of thiamine, TMP and TPP in wild-type mice, but not in APP/PS1 mice. It is possible that APP/PS1 mice have impaired blood brain barrier permeability, which may alter the accumulation of brain levels of thiamine and its phosphate esters after the drug treatment.
Benfotiamine is a special S-acyl thiamine derivative, while most other thiamine derivatives including fursultiamine are disulphide derivatives. This special modification may endow benfotiamine with unique pharmacological effects. Recent studies have suggested that benfotiamine increases the phosphorylation of Akt, an upstream kinase of GSK-3β, in counteracting glucose toxicity and ischaemic diabetes (Gadau et al., 2006; Marchetti et al., 2006), which suggests that benfotiamine may also regulate GSK-3 activities. GSK-3 was first identified as a kinase involved in controlling glycogen metabolism. Now, it has been demonstrated as a ubiquitous serine/threonine protein kinase that participates in multiple physiological and pathological processes. Specifically, GSK-3 is involved in insulin signalling cascade and molecular pathogenesis of diabetes. Interestingly, recent evidence shows that GSK-3 also contributes to the pathogenesis of Alzheimer’s disease (Takashima, 2006). Inhibiting GSK-3 activity has been demonstrated to reduce the production and accumulation of β-amyloid in APP overexpressing mice (Rockenstein et al., 2007). Furthermore, GSK-3α regulates the production of β-amyloid by interfering with APP cleavage at the γ-secretase step (Phiel et al., 2003), which represents a β-secretase-independent mechanism. Thus, it is possible that benfotiamine reduces the formation of amyloid plaques through modulating GSK-3 activities. Indeed, our results show that benfotiamine significantly enhanced the phosphorylation levels and reduced enzymatic activities of both GSK-3α and -3β in the APP/PS1 mice, suggesting that a GSK-3-dependent pathway may be involved in the beneficial effects of benfotiamine. Admittedly, our current study only demonstrates a correlation between the beneficial effects and reduced GSK-3 activities in the APP/PS1 mice after the chronic benfotiamine treatment. We cannot rule out other possibilities besides regulating GSK-3 by benfotiamine at the current stage.
The current study has a major clinical implication in preventing and treating Alzheimer’s disease because benfotiamine has been utilized and shown to be without any serious side-effect as a food supplement or a drug for therapies of peripheral nervous disorders in some countries. The purpose of this work is to promote additional studies on benfotiamine for its effects cognitive function. The identification of tentative non-thiamine metabolite(s) of benfotiamine and their mechanisms for modulating Akt/GSK-3 activities and β-amyloid deposits is yet to be done in detail. Also, benfotiamine differs from other water-soluble and lipophilic thiamine derivatives both in its physicochemical properties and its structure, especially in the open thiazole ring and S-acyl. These properties endow benfotiamine with complicated pharmacokinetics. Thus, the possible involvement of this special structure, and the properties of benfotiamine in modulating GSK-3 and Alzheimer’s disease treatment warrants further investigation.
The Science and Technology Commission of Shanghai Municipality (grant nos 07DJ14005, 09DZ1950400 and 09XD1404900); the National Natural Science Foundation of China (grant no 30830035); Key fund for the platform of developing new drugs from State Scientific & Technological Ministry of China (2009ZX09301-011) and opening funds from the Institutes of Brain Science of Fudan University and the State Key Laboratory of Neuroscience.
The authors thank Dr Michael Zhu for critical reading and discussion of the manuscript.
. Neuroprotective effects of regulators of the glycogen synthase kinase-3beta signaling pathway in a transgenic model of Alzheimer's disease are associated with reduced amyloid precursor protein phosphorylation. J Neurosci 2007;27:1981-91.
. Benfotiamine, a synthetic S-acyl thiamine derivative, has different mechanisms of action and a different pharmacological profile than lipid-soluble thiamine disulfide derivatives. BMC Pharmacol 2008;8:10.