S100B Inhibitor Pentamidine Attenuates Reactive Gliosis and Reduces Neuronal Loss in a Mouse Model of Alzheimer's Disease

Among the different signaling molecules released during reactive gliosis occurring in Alzheimer's disease (AD), the astrocyte-derived S100B protein plays a key role in neuroinflammation, one of the hallmarks of the disease. The use of pharmacological tools targeting S100B may be crucial to embank its effects and some of the pathological features of AD. The antiprotozoal drug pentamidine is a good candidate since it directly blocks S100B activity by inhibiting its interaction with the tumor suppressor p53. We used a mouse model of amyloid beta- (Aβ-) induced AD, which is characterized by reactive gliosis and neuroinflammation in the brain, and we evaluated the effect of pentamidine on the main S100B-mediated events. Pentamidine caused the reduction of glial fibrillary acidic protein, S100B, and RAGE protein expression, which are signs of reactive gliosis, and induced p53 expression in astrocytes. Pentamidine also reduced the expression of proinflammatory mediators and markers, thus reducing neuroinflammation in AD brain. In parallel, we observed a significant neuroprotection exerted by pentamidine on CA1 pyramidal neurons. We demonstrated that pentamidine inhibits Aβ-induced gliosis and neuroinflammation in an animal model of AD, thus playing a role in slowing down the course of the disease.


Introduction
Alzheimer's disease (AD) is the most common age-related neurodegenerative disorder [1], whose pathologic hallmarks are the deposit of neurofibrillary tangles and senile plaques (beta-amyloid protein deposits) in the brain [2,3]. Increasing evidence demonstrates that inflammation in the brain, specifically neuroinflammation, plays a key role in the development of AD [4,5]. This pathologic event is accompanied by the activation of glial cells in the brain, a phenomenon known as "reactive gliosis" [6]. In fact, it has been shown that amyloidbeta-(A -) induced reactive gliosis and the consequent inflammatory responses with the release of neurotoxic cytokines are present in the AD brain and prominently contribute to the progression of the disease [7]. The two events are thus thoroughly linked and are object of the current research on AD pathophysiology.
The definition of "reactive gliosis" refers to the overexpression of glial-derived factors. Amongst all, one of the most interesting from a pharmacological point of view is the protein S100B [8,9]. This small and soluble protein, belonging to the large family of EF-related Ca ++ -and Zn ++ -binding proteins, plays a dual effect. While at nanomolar concentrations S100B provides to a prosurvival effect on neurons and stimulates neurite outgrowth, at higher (micromolar) concentrations it promotes inflammation and neuronal apoptosis [10]. S100B overexpression has been linked to the typical features of reactive gliosis in AD [11,12]. After release, and only when it reaches micromolar concentrations, the protein accumulates at the RAGE (receptor for advanced glycation end-products) surface [13][14][15]. Such interaction leads to the phosphorylation of mitogen-activated protein kinase (MAPK) and the activation of nuclear factor-kappaB (NF-B). This cascade, in turn, promotes the transcription of proinflammatory cytokines and inducible nitric oxide synthase (iNOS) protein [16]. The possibility of interfering with this harmful cycle, by directly targeting S100B, might therefore represent a novel approach to embank neurotoxicity in AD brain.
Pentamidine isethionate, discovered in 1938 as an antiprotozoal drug and approved in the United States for the treatment of Pneumocystis carinii pneumonia and other protozoal diseases [17], appears to be an intriguing candidate. In fact, in addition to the antiprotozoal activity, pentamidine also inhibits S100B-mediated effects because of its ability to block S100B/p53 interaction [18]. In spite of the several data showing the anti-inflammatory effect exerted by pentamidine due to S100B inhibition [19][20][21][22], no data on the possible effect of pentamidine on gliosis and neuroinflammation in AD models are available so far.
Based on this background, the present study was aimed at evaluating the effect of a daily intrahippocampal administration of pentamidine in a mouse model of AD characterized by A -induced gliosis and neuroinflammation. Because of the capability to inhibit S100B protein, we investigated (1) the effect exerted by pentamidine on reactive gliosis, (2) the molecular mechanism by which pentamidine might interfere with reactive gliosis, and (3) whether pentamidine-mediated inhibition of reactive gliosis may result in the rescue of neuronal loss in AD brain.

Ethics Statement.
All the experiments were performed in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, 1996) and those of the Italian Ministry of Health (D.L. 116/92). The study was approved by the Institutional local Animal Care and Use Committees.

Animals.
Experiments were performed in C57BL/6J mice (3-5 months old, weight range: 35-40 g; Harlan, Udine, Italy). Animals were housed under controlled illumination (12 h light/12 h dark cycle) and standard environmental conditions (room temperature 20-22 ∘ C, humidity 55-60%). Food and water were available ad libitum. All surgery and experimental procedures were performed during the light cycle. All efforts were made to reduce the number of animals used and the suffering during surgical experiments.

Surgical Preparation and Intrahippocampal Injection.
Mice (total = 40) were anesthetized i.p. with pentobarbital (40 mg/kg). They were then placed in a stereotaxic frame and injected in the hippocampi (CA1 area) with human A (1-42) peptide (Tocris Cookson, UK). The coordinates for the injection were −1.58 mm posterior from bregma, ±1.2 mm lateral and 1.60 mm ventral to the skull surface. A peptide was dissolved in ALZET artificial cerebrospinal fluid according to the manufacturer's instructions (ALZET-company, Cupertino, CA, USA). The final concentration was 10 g/mL and a volume of 3 L was injected using an ALZET microdialysis pump by keeping the flow at the constant speed of 0.5 mL/min. Control mice (vehicle-treated group, = 8) were injected with an equivalent volume of artificial cerebrospinal fluid. Starting at the third day after surgery and using the previously implanted cannula, three groups of mice ( = 8 per group) received intrahippocampal infusion of pentamidine (0.05-5 g/mL/day) for consecutive 7 days. At the end of treatments, the cannula was removed and the animals, to prevent damage to the scalp sutures, were kept in individual cages until they were killed for tissue processing.

Immunohistochemistry and Immunofluorescence
Analyses. Immunohistochemistry analysis was performed on hippocampal coronal sections (adjacent to the site of the injection) obtained from the brains of vehicle-, A -, and pentamidine-treated mice. Sections were incubated for 2 hours with blocking buffer (PBS containing 15 mM NaN 3 , 10% albumin, and 0.25% Triton X-100) and then with mouse anti-GFAP antibody (1 : 400, Sigma-Aldrich, Milan, Italy) overnight at 4 ∘ C. Biotinylated secondary antibody (1 : 200; Vector Laboratories, Peterborough, UK) and the preformed avidin biotinylated peroxidase complex (VECTASTAIN ABC kit; Vector Laboratories) were then added and the reaction was revealed by 3,3-diaminobenzidine tetrahydrochloride (Sigma-Aldrich). Representative pictures were captured using a high-resolution digital camera (Nikon Digital Sight DS-U1).
For the immunofluorescence, hippocampal coronal sections obtained from the brains of vehicle-, A -, and pentamidine-treated mice were blocked in 10% albumin bovine serum 0.1% Triton-PBS solution for 90 min and subsequently exposed for 1 h to rabbit anti-GFAP antibody (1 : 1000, Abcam, Cambridge, UK) and mouse anti-p53 antibody (1 : 500, Abcam). Sections were then incubated in the dark with the proper secondary antibody: fluorescein isothiocyanate-conjugated anti-rabbit (1 : 1000, Abcam) or Texas Red-conjugated anti-mouse (1 : 1000, Abcam), respectively. Immunofluorescence was analyzed with a Nikon Eclipse 80i microscope (Nikon Instruments Europe, Kingston upon Thames, UK) and images were captured by a high-resolution digital camera (Nikon Digital Sight DS-U1). The number of GFAP+ or p53+ cells was then calculated in every tenth coronal section spanning the ipsilateral hippocampus at the injection site using unbiased stereology (Stereo Investigator, MBF, Williston, VT, USA). According to the manufacturer's protocol, a counting frame (15 × 15 × 20 m) was placed at the intersection of a matrix (200 × 200 m) randomly superimposed by the software onto the region of interest.

Nissl
Staining. Hippocampal coronal sections ( = 5) obtained from the brains of vehicle-, A -, and pentamidinetreated mice were sequentially dipped in different alcohol solutions of decreasing percentage to remove lipids from the tissue, stained with 2% cresyl violet solution for 5 minutes, and dehydrated with a series of baths with increasing alcohol percentage solutions. Sections were analyzed by a blind observer through a Nikon Eclipse 80i microscope. Representative pictures were captured using a high-resolution digital camera (Nikon Digital Sight DS-U1) and analyzed using NIS-Elements software (Nikon Instruments Europe). The extent of neuronal damage was expressed as the ratio between the number of nonstained (death) neurons and the total number of neurons per mm of length of CA1 area in injected ipsilateral hippocampi, according with the following formula: death neurons per CA1 mm 2 area total neurons per CA1 mm 2 area = extent of CA1 damage (%) .
(1) 2.6. Fluoro-Jade B Staining. To further evaluate neuronal loss/rescue in the hippocampus, Fluoro-Jade B (FJB) staining was performed on hippocampal coronal sections obtained from the brains of vehicle-, A -, and pentamidine-treated mice. Sections were immersed in a basic alcoholic solution for 6 minutes and 0.06% KMnO 4 for 15 minutes. Sections were then incubated in 0.0004% FJB (Histo-Chem, Jefferson, AR, USA) for 20 minutes, washed in distilled water, and then dried. To quantify neuronal death, every tenth coronal section spanning hippocampus was analyzed from each animal ( = 5). A blinded observer counted the number of FJBpositive neurons in the hippocampal CA1 from ipsilateral hemispheres to the injection site. Mean counts of FJBpositive neurons from each region were used for the statistical analysis.

Nitrite
Assay. NO was measured as nitrite (NaNO 2 ) accumulation in mice hippocampal homogenates, obtained from the brains of vehicle-, A -, and pentamidine-treated mice, by using the Griess method [23]. Briefly, Griess reagent (1% sulphanilamide, 0.1% naphthylethylenediamine in H 3 PO 4 ) was added to an equal volume of tissue homogenate and the absorbance of the reaction product was measured at 550 nm. Nitrite concentration (nM) was thus determined using a standard curve of NaNO 2 .

Lipid Peroxidation Assay.
Malonyl dialdehyde (MDA) was measured by the thiobarbituric acid colorimetric assay in mice hippocampal homogenates obtained from the brains of vehicle-, A -, and pentamidine-treated mice. Briefly, 1 mL trichloroacetic acid 10% was added to 450 L of tissue lysate. After centrifugation, 1.3 mL thiobarbituric acid 0.5% was added and the mixture was heated at 80 ∘ C for 20 min. After cooling, MDA formation was recorded (absorbance 530 nm and absorbance 550 nm) in a PerkinElmer (Waltham, MA, USA) spectrofluorometer and the results were presented as ng MDA/mL.

Statistical Analysis.
Results were expressed as mean ± SEM of experiments. Statistical analysis was performed using analysis of variance (ANOVA) and multiple comparisons were performed by Bonferroni's test, with < 0.05 considered as significant.

Conclusions
Novel therapeutic approaches for the treatment of AD progression should direct towards the (re)discovery of new molecules able to have an impact on several pathological (1) Vehicle (2) A 10 g/mL (3) A 10 g/mL + pentamidine 0.05 g/mL/day (4) A 10 g/mL + pentamidine 0.5 g/mL/day (5) A 10 g/mL + pentamidine 5 g/mL/day (1) Vehicle (2) A 10 g/mL (3) A 10 g/mL + pentamidine 0.05 g/mL/day (4) A 10 g/mL + pentamidine 0.5 g/mL/day (5) A 10 g/mL + pentamidine 5 g/mL/day  pathways that together converge to the progressive neurological decline characteristic of the disease. Inflammation, more specifically neuroinflammation, has been widely known as an accompanying and key feature in AD [24][25][26]. In fact, both in vitro and in vivo studies have shown that A , the major constituent of the senile plaques in the AD brain, can directly or indirectly activate the secretion of proinflammatory cytokines [27]. Therefore, the search for new drugs should be based on diverse targets in the attempt to blunt the inflammatory scenario in the AD brain and not only to replace the neurotransmission failure.
Here we show that pentamidine, an ancient antiprotozoal drug that inhibits S100B protein, ameliorates gliosis and neuroinflammation in a mouse model of A -induced AD. Many studies have been addressed in the attempt to enlarge the pharmacological knowledge on pentamidine and its novel therapeutic effects in disorders characterized by S100B upregulation, such as melanoma [28], glioblastoma [29], and colitis [19]. This has led to the discovery that, besides being an antiprotozoal drug, pentamidine also inhibits S100B activity by blocking the interaction at the Ca ++ /p53 site of the protein. S100B is a unique glial-derived factor in the sense that it is responsible for the establishment of neuroinflammation and neurodegeneration [30]. In fact, in AD brains, S100B is released by reactive astrocytes, a phenomenon known as "reactive gliosis, " and promotes the formation of neurofibrillary tangles in a RAGE-dependent manner [31]. Once released, S100B accumulates at the RAGE [15,32] and this interaction leads to the induction of lipid peroxidation and to MAPK phosphorylation that in turn converge to NF-B activation. By triggering this pathway, S100B induces the transcription of proinflammatory proteins and cytokines, such as iNOS protein, IL-1 , and TNF [33,34]. It is thus conceivable that, by specifically targeting the RAGE/S100B interaction in the brain, it would be possible to inhibit S100Bdependent neuroinflammation in AD. Different studies have suggested that a possible therapeutic approach might be the inhibition of the binding of S100B to the V domain of RAGE by using specific antibodies or small molecules [35]. However, since RAGE is not the sole receptor mediating S100B effects, it seems more logic to inhibit the protein itself before it binds to any target. The results of this study demonstrate that pentamidine, via direct inhibition of S100B protein, attenuates 1/reactive gliosis and neuroinflammation induced by A in mouse hippocampi and 2/neuronal loss in the CA1 area of the brain. Specifically, pentamidine caused a dose-dependent decrease of GFAP protein expression, a sign of gliosis, in mice hippocampal homogenates. This was accompanied by the dose-dependent inhibition of iNOS, COX-2, and p-p38 MAPK protein expression. Consequently to S100B inhibition, pentamidine indirectly interferes with S100B-RAGE interaction, leading to a marked inhibition of RAGE protein expression, which was upregulated after A injection. This result caused the interruption of the downstream RAGEdependent effects such as NF-B mobilization in the cytosol and the consequent induction of transcription of proinflammatory signaling molecules/cytokines. At confirmation of the amelioration of the inflammatory scenario, pentamidine was also able to reduce the release of proinflammatory cytokines, namely, PGE2 and IL-1 . Moreover, we demonstrated that pentamidine inhibited other proinflammatory events like lipid peroxidation and nitric oxide release. According to our (1) Vehicle (2) A 10 g/mL (3) A 10 g/mL + pentamidine 0.05 g/mL/day (4) A 10 g/mL + pentamidine 0.5 g/mL/day (5) A 10 g/mL + pentamidine 5 g/mL/day (2) A 10 g/mL (3) A 10 g/mL + pentamidine 0.05 g/mL/day (4) A 10 g/mL + pentamidine 0.5 g/mL/day (5) A 10 g/mL + pentamidine 5 g/mL/day ∘∘∘ ∘∘∘  * * * ∘ (1) Vehicle (2) A 10 g/mL (3) A 10 g/mL + pentamidine 0.05 g/mL/day (4) A 10 g/mL + pentamidine 0.5 g/mL/day (5) A 10 g/mL + pentamidine 5 g/mL/day previous observations [29], S100B protein release was upregulated by A injection but its level was not affected by pentamidine treatment. To prove that all the above discussed antiinflammatory effects, together with the reduction of reactive gliosis, were due to the inhibition of S100B/p53 binding, we evaluated p53 expression in the different experimental conditions. We found that the treatment with pentamidine induced p53 expression on infiltrating astrocytes in mouse hippocampi, as sign of enhanced apoptosis. Together with reduced gliosis, we also observed the rescue of neuronal loss in the damaged area of the brain. Though preliminary, our data identifies in pentamidine a novel potential drug for the treatment of AD features. However, future studies are needed to investigate whether, together with its anti-inflammatory and neuroprotective activity, pentamidine may also improve mnemonic and cognitive performances in experimental models of AD. However, one of the limiting factors of pentamidine resides in its pharmacokinetic profile, which is characterized by low blood brain barrier crossing. Thus, new pharmacokinetic approaches aimed at increasing the delivery of pentamidine into the brain, in combination with a suitable compliance in terms of way of administration, look very intriguing.