www.nature.com/scientificreports

OPEN
received: 01 December 2015 accepted: 29 March 2016 Published: 27 April 2016

Fingolimod modulates multiple neuroinflammatory markers in a mouse model of Alzheimer’s disease
Nurgul Aytan1,2,3, Ji-Kyung Choi3, Isabel Carreras1,4, Volker Brinkmann5, Neil W. Kowall1,2, Bruce G. Jenkins3,* & Alpaslan Dedeoglu1,2,3,*
Sphingosine 1-phosphate (SP1) receptors may be attractive targets for modulation of inflammatory processes in neurodegenerative diseases. Recently fingolimod, a functional S1P1 receptor antagonist, was introduced for treatment of multiple sclerosis. We postulated that anti-inflammatory mechanisms of fingolimod might also be protective in Alzheimer’s disease (AD). Therefore, we treated a mouse model of AD, the 5xFAD model, with two doses of fingolimod (1 and 5 mg/kg/day) and measured the response of numerous markers of Aβ pathology as well as inflammatory markers and neurochemistry using biochemical, immunohistochemistry and high resolution magic angle spinning magnetic resonance spectroscopy (MRS). In mice at 3 months of age, we found that fingolimod decreased plaque density as well as soluble plus insoluble Aβ measured by ELISA. Fingolimod also decreased GFAP staining and the number of activated microglia. Taurine has been demonstrated to play a role as an endogenous anti-inflammatory molecule. Taurine levels, measured using MRS, showed a very strong inverse correlation with GFAP levels and ELISA measurements of Aβ, but not with plaque density or activated microglia levels. MRS also showed an effect of fingolimod on glutamate levels. Fingolimod at 1 mg/kg/day provided better neuroprotection than 5 mg/kg/day. Together, these data suggest a potential therapeutic role for fingolimod in AD.

Alzheimer’s disease (AD) is the most common age-related neurodegenerative disease, characterized by progressive memory loss and irreversible cognitive decline.
The extracellular senile plaque deposit of insoluble, aggregated amyloid β  (Aβ ) peptide is the most prominent neuropathological hallmark of AD1. Aβ neurotoxicity is established to be a critical event in AD pathogenesis and correlated with neuronal and synapse loss, which causes synaptic failure resulting in cognitive dysfunction2–4. These events are accompanied by a progressive neuroinflammatory reaction involving the activation of microglia and astrocytes around amyloid plaques in AD pathology5–8.
The ultimate role of Aβ plaque-associated microglial inflammatory response remains controversial. Studies have proposed that in the AD brain, microglia are increased around extracellular Aβ plaques, and activate abnormal production of inflammatory mediators that are neurotoxic, suggesting they promote neuronal degeneration in AD9,10 whereas other studies suggested that microglia play critical roles as mediators of Aβ clearance therefore exerting neuroprotective effects against Aβ toxicity in neurons11,12.
We previously reported the beneficial effects of nonsteroidal anti-inflammatory drug ibuprofen in lowering the Aβ levels in the triple transgenic mouse model of AD (3xTg-AD)13 and in double transgenic Alzheimer mice (APPxPS1)14. Prophylactic treatment of 3xTg-AD mice with ibuprofen at 6 months of age showed a significant decrease in intraneuronal oligomeric Aβ and hyperphosphorylated tau immunoreactivity in the hippocampus,

1Department of Veterans Affairs, VA Boston Healthcare System, Boston, MA 02130, USA. 2Department of Neurology Boston University School of Medicine, Boston, MA 02118, USA. 3A. A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, MA 02114, USA. 4Department of Biochemistry Boston University School of Medicine, Boston, MA 02118, USA. 5Development Franchise Neuroscience, Novartis Pharma AGNovartis Campus Fabrikstrasse 12CH-4056 Basel, Switzerland. *These authors contributed equally to this work. Correspondence and requests for materials should be addressed to A.D. (email: dedeoglu@bu.edu)

Scientific Reports | 6:24939 | DOI: 10.1038/srep24939

1

www.nature.com/scientificreports/

Figure 1. (A) Representative pictures showing Aβ 40 and Aβ 42 immunostained brain sections of 3 months old 5xFAD mice from untreated and 1 mg/kg/day and 5 mg/kg/day of fingolimod treatment groups (Magnification x40). (B) Effects of fingolimod treatments on Aβ plaque burden. Brain sections of representative groups were stained for Aβ plaque using Aβ 40 and Aβ 42 antibodies. 1 and 5 mg/kg/day of fingolimod treatment significantly lowered the Aβ 42 plaque burden after 2 months of treatment in the 3 month-old mice compared to the regular diet as measured by the percent of cortical area. Decreased Aβ 40 plaque burden was detected in the fingolimodtreated groups however the decreases were not significant. *p <  0.05, (n =  8–10 mice/group). The Aβ 40 at 1 mg/ kg/day almost reached significance (omnibus ANOVA F =  3.22, p =  0.058).

and reduced the cognitive decline compared to untreated 3xTg-AD mice13. In a double transgenic mouse model of AD (APPxPS1) treatment with ibuprofen provided significant protection against the neuronal markers N-acetyl aspartate (NAA) and glutamate detected by magnetic resonance spectroscopy (MRS). In that study, we examined the effects of chronic ibuprofen treatment on amyloid plaque deposition, Aβ peptide levels, and the neurochemical profile measured by MRS. Ibuprofen significantly lowered the plaque burden in the 16- to 18-month-old mice as measured by the percent of cortical area. At this age, we found a decrease in the neuronal markers NAA and glutamate and an increase in the astrocytic markers glutamine and myo-inositol in APPxPS1 mice compared to age-matched non-transgenic mice. Ibuprofen provided significant protection against NAA and glutamate loss, however did not significantly affect myo-inositol or glutamine levels14. We similarly found in triple transgenic mice that ibuprofen could protect against NAA and hippocampal loss and that it was protective of myo-inositol increases at early time points in hippocampus, but not at later time points15.
Fingolimod is a functional sphingosine 1-phosphate (S1P) receptor antagonist that has been approved for the treatment for MS (Multiple Sclerosis)16. It has been shown that the anti-inflammatory effect of fingolimod is mediated through S1P1 receptors on lymphocytes, preventing the migration of these cells from peripheral lymphoid organs into the CNS, suggesting it may protect against neuroinflammation16,17. Fingolimod can cross the blood–brain-barrier (BBB) therefore showing a potential effect on CNS cells including neurons, astrocytes, microglia, and oligodendrocytes expressing S1P receptors18. Recently, it has been reported that treatment with fingolimod is associated with microglial neuroprotective effects, including reduction of pro-inflammatory cytokines and increased levels of brain-derived neurotrophic factor (BDNF)19. Doi Y, et al.20 suggested direct effects of the drug on neurons. However, it remains unclear whether modulation of BDNF by the drug in vivo is a direct or indirect effect, and what cell types are involved.
We designed this study to test the effect of fingolimod on amyloid β pathology and neuroinflammation associated with activation of microglia and astrocytes in a mouse model of AD.
Results Effect of fingolimod treatment on Aβ plaque load in 5xFAD mice.  The 5xFAD transgenic mouse
model has been previously reported to exhibit widespread amyloid pathology starting at 2 months old. Here we evaluated the effect of fingolimod treatment at a dose of 1 mg/kg/day and 5 mg/kg/day for 2 months starting at 1 month of age on the deposition of Aβ 40 and Aβ 42 in the frontal cortex (Fig. 1a). Quantitative analyses of the amyloid beta plaques in frontal cortex shows that both dose of fingolimod treatment resulted in significant decreases in Aβ 42 plaque deposition compared with mice given standard drinking water, however the decrease in Aβ 40 plaque load did not reach significance with either dose of fingolimod treatment (Fig. 1b). Effect of fingolimod treatment on the levels of total Aβ in 5xFAD mice.  We further measured the levels of total Aβ 42 and Aβ 40 by ELISA in the frontal cortex of 5xFAD mice after 2 months of fingolimod treatment. We found that the levels of total Aβ 42 and Aβ 40 were statistically lower in 1 mg/kg/day fingolimod-treated group compared to untreated-5xFAD group after 2 months of treatment. However, the levels of total Aβ 42 and Aβ 40 decrease were not statistically significant in the 5 mg/kg/day treatment group (Fig. 2). Effect of fingolimod on intensity of GFAP-positive astrocytes.  Reactive astrocytosis is a well-described pathological process that generally occurs in response to neurodegeneration in AD21,5. To determine the extent of astrocytosis in the 5xFAD mouse brain, we immunostained brain sections from 3 months old 5xFAD mice and fingolimod-treated groups for GFAP. 5xFAD mice exhibit reactive astrocytosis in several regions of the brain, including the cortex, hippocampus, and striatum that correlated with Aβ plaque load (Fig. 3a).
Densitometric analysis of GFAP immunostained brain sections indicates a significant decrease of astrocytosis in the hippocampus of 1 and 5 mg/kg/day fingolimod-treated mice compared to untreated 5xFAD mice (Fig. 3b).

Scientific Reports | 6:24939 | DOI: 10.1038/srep24939

2

www.nature.com/scientificreports/ Figure 2.  Levels of total (soluble and insoluble) Aβ40 and Aβ42 were quantified by ELISA in the frontal cortex. 1 mg/kg/day of fingolimod treatment significantly decreased the levels of Aβ 42 and Aβ 40 however 5 mg/ kg/day of fingolimod treatment did not reach significant decrease. *p <  0.05, (n =  8–10 mice/group).

Figure 3. (A) Immunohistochemical staining for the astrocytic marker glial fibrillary acidic protein (GFAP) in 5xFAD mouse brain sections from the hippocampus of untreated 5xFAD and fingolimod-treated 5xFAD mouse brains. (B) Analysis of GFAP-positive astrocytes in the hippocampus of fingolimod-treated and untreated 5xFAD mice at 3 months of age. 1 and 5 mg/kg/day of fingolimod treatment were significantly decreased the presence of reactive astrocytes in the hippocampus of 5xFAD mice compared to untreated mice. (*p <  0.05), (n =  8–10 mice/group).

Effect of fingolimod treatment on Iba1-positive activated microglia.  We performed immunostaining for the Iba1 microglial marker at 3 months of age in 5xFAD mice and after fingolimod treatment (Fig. 4a).
Microglia are found in increased numbers in close proximity to Aβ plaques in AD9. Microglia associated with Aβ plaques display an activated phenotype characterized by enhanced Iba1-immunoreactivity, retracted processes, perikaryal hypertrophy, and amoeboid appearance that contrast with the microglia not associated with Aβ plaques which display a resting morphology, with small compact somata bearing many long thin ramified processes5.
Iba1 immunopositive cell were counted using the optical fractionator method and optical dissector probe. We detected significant differences in the numbers of Iba1-positive activated microglia in the hippocampus of both fingolimod-treated doses in 5xFAD mice group compared to untreated 5xFAD mice.
The number of activated Iba1-positive microglia in the CA1/subiculum region of the hippocampus of fingolimod-treated 5xFAD mice was significantly lower than in the untreated 5xFAD mice group. Moreover, we didn’t observe a significant difference between 1 and 5 mg/kg/day of fingolimod-treated 5xFAD mice group for the number of activated microglia. The number of resting microglia did not show any significant difference between fingolimod-treated and untreated group of 5xFAD mice (Fig. 4b).
We also evaluated the rankings of the various measures we made for classification of the groups. We used a relief-f attribute selector22 to rank the various attributes with regards to classification. The results (Fig. 5a) showed that the GFAP contributed the largest weights, followed by Iba1 (activated/total) or Iba1 (activated/resting), and

Scientific Reports | 6:24939 | DOI: 10.1038/srep24939

3

www.nature.com/scientificreports/

Figure 4. (A) Sections of hippocampus and subiculum immunostained for Iba1 in 5xFAD mice untreated and treated with 1 mg/kg/day, and 5 mg/kg/day of fingolimod. Untreated 3 months old 5xFAD mice showed significant increase in the number of activated microglial cells. (B) Quantitation of total, active and resting variants of microglia in the hippocampus of 5xFAD untreated and fingolimod-treated groups. The number of activated Iba1-positive microglia significantly decreased in the hippocampus of 1 mg/kg/day, and 5 mg/kg/day fingolimod-treated groups compared with untreated group (*p <  0.05), (n =  8–10 mice/group).

Figure 5.  Evaluating the power of the various markers to assess protection with fingolimod. (A) Left) Rankings determined from a relief-f algorithm. GFAP has the highest weighting followed by Iba1 quantified by either activated/total (A/T) or activated/resting (A/R), Aβ 42 plaque density, Iba1 activated microglia, Aβ 42 ELISA, Aβ 40 ELISA, Iba1 (total) or Iba1 (resting). (B) Right) We then used the top four markers (GFAP, Iba1 (A/T), Aβ 42 plaque density and Aβ 42 ELISA to perform a linear discriminant analysis (Wilk’s lambda =  0.122; p <  0.0001 for function 1, not significant for function 2). This analysis shows a larger average distance from regular diet for the 1 mg/kg/day fingolimod treatment (Lo) than for the 5 mg/kg/day dose (Hi).

Aβ 42 plaque density. We then used the top four attributes (GFAP, Iba1 (activated/total), Aβ 42 plaque density, Aβ 42 ELISA) to perform cluster analysis using a simple k-means. This analysis showed that there were only two clusters of which 9/9 regular diet mice were in one cluster with 1 each of the 1 and 5 mg/kg/day and the other 15 fingolimod-treated mice were clustered together. We also performed linear discriminant analysis (LDA), as well as support vector machines (SVM) using a logistic function to demonstrate the separability of the groups. The SVM and the LDA showed similar efficacy with 100% accurate separation of the regular diet from the fingolimod-treated animals and mixing of the two dose groups. However as shown in Fig. 5b, the overall distance of the 5 mg/kg/day group was closer to the regular diet than was the 1mg/kg/day group. The confusion matrix showed mixing between the high and low dose groups. For SVM 9/9 regular diet were classified as regular diet, 5/9 low dose diet were classified as low dose diet and 4/9 were classified as high dose diet, and 1/8 high dose diet was classified as regular dose diet, 5/8 were classified as low dose diet and 2/8 were classified as high dose diet. For LDA 9/9 regular diet were classified as regular diet, 6/9 low dose diet were classified as low dose diet and 3/9 were classified as high dose diet, and 0/8 high dose diet was classified as regular dose diet, 4/8 were classified as low dose diet and 4/8 were classified as high dose diet. Magnetic resonance spectroscopy of changes in brain neurochemicals.  Two MRS studies of 5xFAD mice have appeared, one from our lab23 and one from another lab Mlynarik et al.24. The changes

Scientific Reports | 6:24939 | DOI: 10.1038/srep24939

4

www.nature.com/scientificreports/

Figure 6.  HRMAS of fingolimod treatment in 3 months old mice. (A) Representative HRMAS spectra from hippocampus in 5xFAD normal diet and fingolimod-treated (1 mg/kg/day) a few molecules including N-acetylaspartate (NAA), glutamate, glutamine, creatine, taurine and myo-inositol are labeled. (B) Effect of fingolimod treatment as a function of dose on taurine levels. There is a very significant effect on raising taurine levels in the fingolimod-treated mice. (C) Correlation between taurine levels and GFAP staining in the mice that had both measured. There is a very strong correlation R =  0.83; p <  0.0001.

observed in both studies were very similar (decreased NAA, glutamate and GABA, and increased glutamine

and myo-inositol). Our data showed large numbers of neurochemical changes that could be ameliorated using

treatment with scyllo-inositol and the NSAID flurbiprofen. That study was conducted on mice 8 months of age.

The mice studied here were three months of age. At three months of age the neurochemical changes are consid-

erably less profound than at 8 months of age. We compared four groups of mice (WT, 5xFAD regular diet and

1 mg/kg/day and 5 mg/kg/day of fingolimod). We found trends towards decreased NAA, GABA and glutamate

at three months that was barely significant There was a large decrease in taurine that

fworasglhuitgahmlyatseig(npi <fic  0a.n0t5;aFn3d,36w =a s 3p.1r)oatencdtewdabs yprfointegcotleimd boyd

fingolimod. (p <   0.001;

Fst3a,3i6n =in  7g.a3n).dTthaeusreindeatleavaerles

shown in Fig. 6. We in the hippocampus

found of the

an unusually strong inverse mice (R =  0.84; p <  0.0001).

correlation between GFAP We also performed similar

correlations between Iba1 staining and as well as ELISA measures of Aβ 40 and Aβ 42 and Aβ 40/42 plaque den-

sity. The correlation between taurine and Iba1 activated was not very good (R =  0.3), while the correlations with

Aβ 40/42 plaque density were better (R =  0.47; 0.41 respectively) and better still for Aβ 40/42 ELISA (R =  0.65,

0.67 respectively; p <  0.001 for both). These data are of interest since taurine has been shown to act as an endog-

enous anti-inflammatory chemical25.

Discussion
Fingolimod displays potent anti-inflammatory effects in MS patients and animal models of neurodegenerative disease. Therefore we investigated the effect of fingolimod treatment on neuroinflammation, amyloid pathology and neurochemical markers in a transgenic mouse model of AD.
Our results indicate that fingolimod intake reduced the levels of Aβ and plaque burden in 5xFAD mice and correlated with the significant decrease of activated microglia measured using Iba1 and the astrocytic inflammatory marker GFAP.
Fingolimod treatment has been previously reported to improve cognitive function in a rodent model of AD. Chronic administration of fingolimod (1 mg/kg, i.p., 14 days) significantly attenuated the Aβ 42-induced learning and memory impairment and prevented the hippocampus neuronal damage in a rat model of AD26. Consistent with these results, it has been found that treatment with fingolimod protects against Aβ 1-42-induced cognitive impairment in a mouse model of AD19.
Further, neuroprotective effects of fingolimod treatment have also been reported in Aβ 1-42 induced neurotoxicity in vitro in cultured mouse primary cortical neurons20.
On the other hand, 6-days of treatment of APP transgenic mice with fingolimod (0.5 mg/kg/day) has been reported to decrease the levels of soluble Aβ 40 and, interestingly, increase the levels of soluble Aβ 4227. These results were not consistent with observations from their in vivo results, in which fingolimod treatment was shown

Scientific Reports | 6:24939 | DOI: 10.1038/srep24939

5

www.nature.com/scientificreports/

to decrease the levels of Aβ in cultured neuronal cells. They considered that fingolimod might have a distinct

effect on Aβ 42 metabolism through altering the clearance of Aβ 42 by microglial cells in vivo.

Recently, the neuroprotective effect of fingolimod has been shown in cerebral white matter injury fol-

lowing neonatal hyperoxia28. In a neonatal model of hyperoxia, fingolimod (1 mg/kg) treatment reduced

hyperoxia-induced cognitive dysfunctions, microglia activation and associated pro-inflammatory cytokine

expression28.

Although the neuropathological mechanisms between AD and white matter damage remain unclear, white

matter ischemic damage is considered to be an important contributor to the pathology and development of

dementia in AD. Ischemia has also been revealed to increase the production of Aβ and facilitate cognitive impair-

ment in animal models29,30.

The neuroprotective effect of fingolimod on human astrocytes, has been shown through the induction of

neurotrophic factors and inhibition of inflammatory genes31 suggesting, a part of the effects of fingolimod might

be mediated via astrocytes.

Although many experiments strongly suggest a role for microglia in Aβ 42 phagocytosis or clearance, ongoing

controversy exists regarding whether activated microglia contribute to their maintenance or clearance.

Our results demonstrate that administration of 1 mg/kg/day of fingolimod for 2 months results in a significant

decrease in total Aβ 42 plaque burden and the levels of Aβ 40 and 42 in the brains of 3 months old 5xFAD mice

and that lower-dose (1 mg/kg/day) of fingolimod treatment was more effective in decreasing Aβ 40 and 42 levels

compared to treatment with the higher-dose (5 mg/kg/day). However, treatment with either low or the high dose

resulted in significant decrease in Aβ 40 plaques. Effect of the low and the high dose on Aβ 40 and 42 plaques, acti-

vated microglia and astrocytes was comparable. Lower dose was also more effective in increasing taurine levels

than the high dose.

In addition to its effects on lymphocyte egress from lymph nodes, fingolimod has been also shown to mod-

ulate a variety of intracellular signaling pathways in different cell types following the internalization of S1P1

receptors suggesting, this compound has a more complex mechanism32–34.

The fact that lower dose work better than higher dose may indicate that complete removal of S1P1 by fingo-

limod (through internalization and degradation) may not be the one and only beneficial effect. It could be that

some agonistic effects/signaling by drug in one or the other cell type/S1PR could also contribute, and this occurs

more at the lower dose when internalization is incomplete.

The results from fingolimod phosphate (FTY720P)-mediated signal transduction experiment in primary

human umbilical vein endothelial cells shows that despite S1P1 receptors complete internalization induced by a

transient exposure to FTY720P, S1P1 receptors were able to maintain the transmission of the persistent signaling

activity from receptor-proximal G proteins to downstream effector systems34.

These results suggest that FTY720P remains bound to internalized S1P1 receptors and maintains an active

conformation that targets the receptor for atypical signaling, processing and trafficking.

In addition, S1P1 has been shown to interact with the regulator of G protein signaling (RGS)-2 protein,

which is a GTPase-activating protein (GAP), to regulate cell motility in a concentration dependent manner.

S1P1 responds to S1P ligand stimulation with a typical bell-shaped dose curve, and the RGS2–S1P1 complex

dissociated at higher S1P concentrations in S1P1-expressing CHO cells. Similar to these results, they observed

that FTYP-induced cell motility followed a similar bell-shaped dose–response curve and cell mobility was also

decreased at higher concentrations of FTYP. The S1P1–RGS2 complex was stable upon stimulation with lower

FTYP concentrations, whereas the complex dissociated at higher concentrations. These results suggest that the

decreased S1P1-mediated cell motility at high doses of FTYP might be through a mechanism acts independently

of S1P1 internalization32.

At low dose (10 nM), chronic fingolimod treatment (72 h) has been shown to stimulates the differentiation of

oligodendrocyte precursor cells (OPCs) and expression of myelin basic protein (MBP)-positive cells compare to

higher dose (1μ M) of fingolimod treatment in cultured rat OPCs, indicating that the drug may have a dose and

time dependent effect on developing oligodendrocytes33.

Based on the literature, our results support the hypothesis that the effect of fingolimod on both Aβ 42 and

inflammation with different doses could be mediated via different signaling mechanisms through S1P receptors

in neuron and glial cells.

In conclusion, these studies showing the beneficial effect of fingolimod treatment in different cell types rep-

resent that a dose and time dependent treatment of fingolimod could reflect a distinct property of this agonist/

receptor pair resulting in a sustained activation of receptor internalization and active signaling following expo-

sure to fingolimod.

In contrast to our previous study of 5xFAD mice at 8 months of age23 the mice in this study were only three

months of age. Although there is already some early pathology in these mice at this age (for instance increase

Aβ 42 plaque deposition)35 there is a dramatic increase in many pathological markers from 3 months to

8 months including an increase in Iba1 staining (ratio of activated/total microglia) from 0.58 ±  0.07 at 3 months

to 0.78 ±  .043 at 8 months (unpaired t-test, t value − 6.19, p <  0.0001) and in GFAP staining from 0.106 ±  0.005 in

WT tests

to 0.245 ±  0.007 at 3 months to 0.375 ±  0.018 at show all three groups are significantly different

8 months of age from each other

(aFt2p,2 0< =  0   8.06070.313),(pse  <e   F0.i0g0. 071).;

Tukey In line

post-hoc with this

age dependence, we found a total decrease in the number of large neurochemical changes in the hippocampus at

3 months of age compare to 8 months of age that we reported earlier23. At three months of age the largest change

was in taurine levels (see Fig. 6). Even more striking, there was a very strong correlation between taurine levels

and GFAP staining (Fig. 6). These data suggest that fingolimod increases taurine levels and that this may be one

mechanism behind the neuroprotection noted. Taurine is the amino acid with the highest concentration in the

body and has numerous different roles including osmoregulation as well as acting as an indirect anti-oxidant

and endogenous anti-inflammatory agent36,37,25. Given that cells exposed to high oxidative stress also show high

Scientific Reports | 6:24939 | DOI: 10.1038/srep24939

6

www.nature.com/scientificreports/

Figure 7.  Effects of age on inflammatory markers in hippocampus in 5xFAD mice. Left) GFAP staining shows a large increase from WT to 3 months and then 8 months old mice (P <  0.0001 for differences between all groups using ANOVA with a Tukey-HSD post-hoc test). Right) Increased activated/total microglia ratio for 3 and 8 months old animals (p <  0.0001 for difference between groups. There were no activated microglia noted in the WT mice.

taurine levels, it is logical to assume that the inverse correlation between GFAP staining and taurine levels may reflect protective effects of taurine. Oral supplementation with taurine has shown some success for neuroprotection in models of traumatic brain injury38. Whether fingolimod increases taurine levels, or alters the cellular distribution of taurine containing cells is a topic for further investigation. Since rodents have much higher concentrations of taurine than do humans39, it would be of interest to see if such relations also hold for human tissue between brain taurine levels and neuroinflammatory markers. The fact that we observed strong correlations with taurine for GFAP as well as ELISA measurements of Aβ 40/42, but not with activated microglia or with plaque density measures suggests that taurine is more closely associated with astrocytic activity rather than activated microglia. Since taurine is readily measureable using MRS in both humans and mice, it may provide a non-invasive marker for modulation of neuroinflammation in the brain.
Material and Methods Mice.  Female 5xFAD transgenic mice35 and non-transgenic female littermates as controls were used in this
study. 5xFAD mice coexpress the human APP and PS1 genes harboring a total of 5 FAD mutations [APP K670N/ M671L (Swedish) +   I716V (Florida) +  V717I (London) and PS1 M146L +  L286V] under the control of the murine Thy-1 promoter. The study comprised 4 groups of mice: transgenic mice treated at 2 different doses of fingolimod, untreated 5xFAD mice, and untreated non-transgenic (WT) mice (n =  10). Mice were housed on a 12 h light:12 h dark schedule. All mice were given access to food and water ad libitum. All animal experiments were carried in accordance with the NIH Guide for the Care and Use of Laboratory Animals and were approved by research animal care committee at VA Boston Healthcare System. The Medical Center is fully AAALAC accredited as an approved research facility and the hospital has on file an Assurance of Compliance with Public Health Service regulations and requirements and provisions of the Animal Welfare Act. Diet and drug protocol.  Fingolimod was kindly provided by Volker Brinkmann, Novartis Pharma AG Basel, Switzerland. Fingolimod, added into the drinking tap water, was tested at two different doses, 1 mg/kg/ day and 5 mg/kg/day. Water was changed twice weekly. Food and water consumption were monitored weekly. We monitored mice for general well being, and no side effects related to either treatment were observed. Tissue collection.  At 3 months of age, mice were euthanized and brains removed for analysis. The left hemisphere was post-fixed with the 4% paraformaldehyde solution for 24 h and cryoprotected in a graded series of 10% and 20% glycerol/2% DMSO solution for histological analysis. The right hemisphere was dissected at the frontal level and saved for the Enzyme-linked immunosorbent assay (ELISA) assay. Tissue punches of freshly frozen hippocampus were collected for MRS studies. Histology/Immunohistochemistry.  Brains were serially cut at 50 μm on a freezing microtome. Immunohistochemical procedures were performed as previously described40. In brief, free-floating sections were incubated overnight in primary antibody followed by PBS (Phosphate buffered saline) washes and incubation in peroxidase-conjugated secondary antibody followed by development using 3,3′ -diaminobenzidine tetrahydrochloride (DAB) as a chromagen. The antibodies used were Aβ 1–40 and Aβ 1–42 (Invitrogen, Grand Island, NY) to define Aβ deposits, ionized calcium binding adaptor molecule 1(Iba1) (Wako Chemicals, USA) to stain microglia, and glial fibrillary acidic protein (GFAP) (Chemicon, USA) to stain astrocytes. Quantitative analysis of Aβ deposits and intensity of GFAP-positive astrocytes.  For the estimation of cortical Aβ plaque burden, three serial sections per mouse brain were analyzed blindly using a custom

Scientific Reports | 6:24939 | DOI: 10.1038/srep24939

7

www.nature.com/scientificreports/

software written in Matlab. The most rostral section analyzed was at the anterior commissure level (~0.1 mm anterior to bregma), and each successive section was at 0.34 mm increments. Images of the Aβ stained sections were acquired using a Nikon Optiphot microscope at 4x magnification. The cortical region of interest was processed in Photoshop together with an area from the same image that did not contain plaques in order to control for background stain. Using both images in the Matlab program we determined the percent plaque area.
Similarly, the percentage of GFAP-positive astrocytes was analyzed using the Mat Lab program in the subiculum/CA1 region of the hippocampus in 3 sections per mice.

Quantitative analysis of Iba1-positive microglia cell number.  To quantify microglia, we applied the Optical Fractionator probe in the StereoInvestigator software (MBF Bioscience, Williston, VT) using two different markers to count resting microglia and activated microglia based on their morphological differences. Tree sections from the hippocampus between bregma levels − 2.5–3.0 mm were selected for each mouse. The region of interest, the subiculum/CA1 area of the hippocampus was delineated at 4x magnification. Cell counting was performed at 40x magnification in a counting frame of 350 ×  350 μ m.

ELISA assay.  Dissected frontal brain tissue was homogenized with 8 ×   5 M guanidine HCl buffer to analyze the level of total (soluble +  insoluble) Aβ . To determine Aβ levels, human Aβ 40 and Aβ 42 ELISA kits (Invitrogen, Grand Island, NY) were used according to manufacturer’s specifications. Briefly, homogenized samples were added into the wells of a 96-well plate, samples were then mixed with a cleavage-specific antibody to either Aβ 40 or Aβ 42. After overnight incubation at 4 °C, plates were washed and incubated with the secondary antibody for 30 minutes at 25 °C. Washed wells were developed by the addition of a substrate. The substrate reaction was then stopped and color intensity was measured at 450 nm.

High resolution magic angle spinning spectroscopy (HRMAS).  In vitro MRS was collected as pre-

viously published41,14. We collected high resolution magic angle spinning (HRMAS) spectra on Bruker 14T

(Billerica, MA). We obtained tissue punches of freshly frozen hippocampus from mice. The punches were 1 mm

in diameter and were taken from hippocampus. The dissected tissue sample was placed into a glass cylinder

positioned in a 3 mm zirconium oxide MAS rotor (volume 50 μ L). HRMAS measurements were performed using

a sample spinning rate of 3.6 kHz selected to push the spinning side bands outside the frequency region of the

metabolites. The experiments were performed at 4 °C to minimize tissue degradation.

rseemquDoeavntecatehw[ee9lr0iep −iad  c(/τqm u −airc  1er8od0mu −osil  nτec g −ual  Aarrcorqteo)snro]nswyainnthcchetsrwoaonndidzietfhfdee,rseThn2o-tfreitlftfTeeErcetirdveetCaTianErsrs–t(hP1eu0mr0c .meTlslh–/e1M0in emtiebsro)p.ouTmlhse–eGdloeinllalgy(e,CrτPT , MiEs sGsye)nrvcpehusrlotsoe-

nmwiaazsteed3ly6to09.5tTh%heoerfostthhooerrfttorτetqa vluaselpnuececytr,reaamnl idonvitseens2s7ail2tlyμt osh.feTaThll2em*n-evltiaakbleuoeeliftffeoesrctotshf oeinnrteetlrhaetesivtliecnloyemsshhpaoaprreetsdT. t2TofhitlehteelrosnwhgoasrTt32T6fEialt.neOdrtyfhoieerrltdahsceqalupoipnsrigtoiTxoiEn-

parameters were a 90° pulse of 5–10 μ s, a spectral width of 8 kHz, 16 K complex points, 256 averages and a TR of

5 s. Samples were and shimming.

placed

in

the

rotor

with

a

small

amount

of

D2O

(Sigma-Aldrich,

Milwaukee,

WI)

for

locking

Data were analyzed using the Chenomx (Edmonton, Alberta, Canada) package fitting the entire metabolite

spectrum for each neurochemical. HRMAS data were reported as molar ratios to creatine since our prior studies

of the absolute concentrations in multiple different AD transgenic mouse models showed no change in total cre-

atine between WT and any of the AD models41,14,42. In a paper recently published by Mlynarik et al.24 examining

the 5xFAD mice using a water normalization method, no significant change in the creatine concentrations was

noted in the 5xFAD mice compared to WT. Statistical analyses of the MRS data were performed using a one-way

ANOVA with Tukey HSD post-hoc tests for between group comparisons for each metabolite. Classification of the

data was performed using Weka43.

Statistical analyses.  Statistical analyses of the data were performed using a one-way ANOVA with Tukey HSD post-hoc tests for between fingolimod-treated and untreated groups.

References
1.	 Mott, R. T. & Hulette, C. M. Neuropathology of Alzheimer’s disease. Neuroimaging clinics of North America 15, 755–765 (2005). 2.	 Selkoe, D. J. Soluble oligomers of the amyloid beta-protein impair synaptic plasticity and behavior. Behavioural brain research 192,
106–113 (2008). 3.	 Shankar, G. M. et al. Amyloid-beta protein dimers isolated directly from Alzheimer’s brains impair synaptic plasticity and memory.
Nat Med 14, 837–842 (2008). 4.	 Koffie, R. M. et al. Oligomeric amyloid beta associates with postsynaptic densities and correlates with excitatory synapse loss near
senile plaques. Proc Natl Acad Sci USA 106, 4012–4017 (2009). 5.	 Akiyama, H. et al. Inflammation and Alzheimer’s disease. Neurobiology of aging 21, 383–421 (2000). 6.	 El Khoury, J. & Luster, A. D. Mechanisms of microglia accumulation in Alzheimer’s disease: therapeutic implications. Trends in
pharmacological sciences 29, 626–632 (2008). 7.	 Heneka, M. T. & O’Banion, M. K. Inflammatory processes in Alzheimer’s disease. Journal of neuroimmunology 184, 69–91 (2007). 8.	 Aguzzi, A., Barres, B. A. & Bennett, M. L. Microglia: scapegoat, saboteur, or something else? Science 339, 156–161 (2013). 9.	 Meyer-Luehmann, M. et al. Rapid appearance and local toxicity of amyloid-beta plaques in a mouse model of Alzheimer’s disease.
Nature 451, 720–724 (2008). 10.	 Shah, P. et al. Neuronal and axonal loss are selectively linked to fibrillar amyloid-{beta} within plaques of the aged primate cerebral
cortex. The American journal of pathology 177, 325–333 (2010). 11.	 El Khoury, J. et al. Ccr2 deficiency impairs microglial accumulation and accelerates progression of Alzheimer-like disease. Nat Med
13, 432–438 (2007). 12.	 Hickman, S. E., Allison, E. K. & El Khoury, J. Microglial dysfunction and defective beta-amyloid clearance pathways in aging
Alzheimer’s disease mice. J Neurosci 28, 8354–8360 (2008).

Scientific Reports | 6:24939 | DOI: 10.1038/srep24939

8

www.nature.com/scientificreports/

13.	 McKee, A. C. et al. Ibuprofen reduces Abeta, hyperphosphorylated tau and memory deficits in Alzheimer mice. Brain research 1207, 225–236 (2008).
14.	 Choi, J. K. et al. Anti-inflammatory treatment in AD mice protects against neuronal pathology. Experimental neurology 223, 377–384 (2010).
15.	 Choi, J. K. et al. The effects of aging, housing and ibuprofen treatment on brain neurochemistry in a triple transgene Alzheimer’s disease mouse model using magnetic resonance spectroscopy and imaging. Brain research 1590, 85–96 (2014).
16.	 Brinkmann, V. et al. Fingolimod (FTY720): discovery and development of an oral drug to treat multiple sclerosis. Nature reviews. Drug discovery 9, 883–897 (2010).
17.	 Mehling, M., Johnson, T. A., Antel, J., Kappos, L. & Bar-Or, A. Clinical immunology of the sphingosine 1-phosphate receptor modulator fingolimod (FTY720) in multiple sclerosis. Neurology 76, S20–27 (2011).
18.	 Blaho, V. A. & Hla, T. An update on the biology of sphingosine 1-phosphate receptors. Journal of lipid research 55, 1596–1608 (2014). 19.	 Fukumoto, K. et al. Fingolimod increases brain-derived neurotrophic factor levels and ameliorates amyloid beta-induced memory
impairment. Behavioural brain research 268, 88–93 (2014). 20.	 Doi, Y. et al. Fingolimod phosphate attenuates oligomeric amyloid beta-induced neurotoxicity via increased brain-derived
neurotrophic factor expression in neurons. PloS one 8, e61988 (2013). 21.	 Olabarria, M., Noristani, H. N., Verkhratsky, A. & Rodriguez, J. J. Concomitant astroglial atrophy and astrogliosis in a triple
transgenic animal model of Alzheimer’s disease. Glia 58, 831–838 (2010). 22.	 Robnik-Sikonja, M. & Kononenko, R. Theoretical and Empirical Analysis of ReliefF and RReliefF. Mechine Learning 53, 23–69
(2003). 23.	 Aytan, N. et al. Combination therapy in a transgenic model of Alzheimer’s disease. Experimental neurology 250, 228–238 (2013). 24.	 Mlynarik, V. et al. Proton and phosphorus magnetic resonance spectroscopy of a mouse model of Alzheimer’s disease. Journal of
Alzheimer’s disease: JAD 31 Suppl 3, S87–99 (2012). 25.	 Marcinkiewicz, J. & Kontny, E. Taurine and inflammatory diseases. Amino acids 46, 7–20 (2014). 26.	 Asle-Rousta, M., Kolahdooz, Z., Oryan, S., Ahmadiani, A. & Dargahi, L. FTY720 (fingolimod) attenuates beta-amyloid peptide
(Abeta42)-induced impairment of spatial learning and memory in rats. Journal of molecular neuroscience: MN 50, 524–532 (2013). 27.	 Takasugi, N. et al. FTY720/fingolimod, a sphingosine analogue, reduces amyloid-beta production in neurons. PloS one 8, e64050
(2013). 28.	 Serdar, M. et al. Fingolimod protects against neonatal white matter damage and long-term cognitive deficits caused by hyperoxia.
Brain. Behav. Immun. 52, 106–119 (2016). 29.	 Sun, X. et al. Hypoxia facilitates Alzheimer’s disease pathogenesis by up-regulating BACE1 gene expression. Proc Natl Acad Sci USA
103, 18727–18732 (2006). 30.	 Guglielmotto, M. et al. The up-regulation of BACE1 mediated by hypoxia and ischemic injury: role of oxidative stress and
HIF1alpha. Journal of neurochemistry 108, 1045–1056 (2009). 31.	 Hoffmann, F. S. et al. Fingolimod induces neuroprotective factors in human astrocytes. Journal of neuroinflammation 12, 184 (2015). 32.	 Kohno, T. & Igarashi, Y. Attenuation of cell motility observed with high doses of sphingosine 1-phosphate or phosphorylated
FTY720 involves RGS2 through its interactions with the receptor S1P. Genes Cells 13, 747–757 (2008). 33.	 Jung, C. G. et al. Functional consequences of S1P receptor modulation in rat oligodendroglial lineage cells. Glia 55, 1656–1667 (2007). 34.	 Mullershausen, F. et al. Persistent signaling induced by FTY720-phosphate is mediated by internalized S1P1 receptors. Nature
chemical biology 5, 428–434 (2009). 35.	 Oakley, H. et al. Intraneuronal beta-amyloid aggregates, neurodegeneration, and neuron loss in transgenic mice with five familial
Alzheimer’s disease mutations: potential factors in amyloid plaque formation. J Neurosci 26, 10129–10140 (2006). 36.	 Green, T. R., Fellman, J. H., Eicher, A. L. & Pratt, K. L. Antioxidant role and subcellular location of hypotaurine and taurine in
human neutrophils. Biochimica et biophysica acta 1073, 91–97 (1991). 37.	 Marcinkiewicz, J., Grabowska, A., Bereta, J., Bryniarski, K. & Nowak, B. Taurine chloramine down-regulates the generation of
murine neutrophil inflammatory mediators. Immunopharmacology 40, 27–38 (1998). 38.	 Su, Y. et al. Taurine improves functional and histological outcomes and reduces inflammation in traumatic brain injury. Neuroscience
266, 56–65 (2014). 39.	 Choi, J. K., Dedeoglu, A. & Jenkins, B. G. Application of MRS to mouse models of neurodegenerative illness. NMR in biomedicine
20, 216–237 (2007). 40.	 Kowall, N. W. et al. MPTP induces alpha-synuclein aggregation in the substantia nigra of baboons. Neuroreport 11, 211–213 (2000). 41.	 Choi, J. K., Carreras, I., Dedeoglu, A. & Jenkins, B. G. Detection of increased scyllo-inositol in brain with magnetic resonance
spectroscopy after dietary supplementation in Alzheimer’s disease mouse models. Neuropharmacology 59, 353–357 (2010). 42.	 Dedeoglu, A., Choi, J. K., Cormier, K., Kowall, N. W. & Jenkins, B. G. MRS analysis of AD mouse brain that express mutant human
APP shows altered neurochemical profile. Brain research 1012, 60–65 (2004). 43.	 Mark Hall, E. F. Geoffrey Holmes, Bernhard Pfahringer, Peter Reutemann, Ian H. Witten. The WEKA Data Mining Software: An
Update; SIGKDD Explorations. volume 11 (2009).
Acknowledgements
This research is supported by grants from the Department of Veteran Affairs (Merit Award) and Alzheimer Drug Discovery Foundation (ADDF) to A. Dedeoglu. Fingolimod was provided by Novartis.
Author Contributions
A.D., N.W.K., V.B. and N.A. designed research; N.A., I.C. and J.-K.C. performed research; V.B. contributed reagents/materials; A.D., N.A., J.-K.C. and B.G.J. analyzed data; N.A., B.G.J. and A.D. wrote the paper. All authors reviewed the manuscript.
Additional Information
Competing financial interests: The authors declare no competing financial interests. How to cite this article: Aytan, N. et al. Fingolimod modulates multiple neuroinflammatory markers in a mouse model of Alzheimer’s disease. Sci. Rep. 6, 24939; doi: 10.1038/srep24939 (2016).
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Scientific Reports | 6:24939 | DOI: 10.1038/srep24939

9