3,6′-Dithiothalidomide, a new TNF-α synthesis inhibitor, attenuates the effect of Aβ_1-42 intracerebroventricular injection on hippocampal neurogenesis and memory deficit

Abstract

Evidence indicates altered neurogenesis in neurodegenerative diseases associated with inflammation, including Alzheimer’s disease (AD). Neuroinflammation and its propagation have a critical role in the degeneration of hippocampal neurons, cognitive impairment and altered neurogenesis. Particularly, tumor necrosis factor (TNF)-α plays a central role in initiating and regulating the cytokine cascade during an inflammatory response and is up-regulated in brain of AD patients. In this study, we investigated the effects of a novel thalidomide-based TNF-α lowering drug, 3,6′-dithiothalidomide, on hippocampal progenitor cell proliferation, neurogenesis and memory tasks after intracerebroventricular (i.c.v.) injection of β-amyloid (Aß)_1-42 peptide. Seven days after Aβ_1-42 injection, a significant proliferation of hippocampal progenitor cells and memory impairment were evident. Four weeks after Aβ_1–42 peptide injection, elevated numbers of surviving BrdU cells and newly formed neurons were detected. Treatment with 3,6′-dithiothalidomide attenuated these Aβ_1-42 provoked effects. Our data indicate that although treatment with 3,6′-dithiothalidomide in part attenuated the increase in hippocampal neurogenesis caused by Aβ_1-42-induced neuroinflammation, the drug prevented memory deficits associated with increased numbers of activated microglial cells and inflammatory response. Therefore, 3,6′-dithiothalidomide treatment likely reduced neuronal tissue damage induced by neuroinflammation following Aβ_1-42 injection. Understanding the modulation of neurogenesis, and its relationship with memory function could open new therapeutic interventions for AD and other neurodegenerative disorders with an inflammatory component.

Introduction

Within the adult brain, neurogenesis constitutively occurs in the subventricular zone (SVZ) of the lateral ventricles and the subgranular zone (SGZ) of the hippocampus (Doetsch 2003; Russo et al. 2011a). After proliferation, hippocampal progenitor cells migrate from the SGZ into the granule cell layer (GCL) of the dentate gyrus (DG) where they differentiate into granule cells (Kempermann et al. 2003; Russo et al. 2011b; Russo et al. 2011a) and, thereafter, morphologically and functionally integrate into the existing hippocampal circuitry (van Praag et al. 2002). Hippocampal neurogenesis thereby plays an important role in learning, memory and repair processes, whereas abnormalities in neural progenitor cell capacity may well contribute to the pathogenesis of diseases such as Alzheimer’s disease (AD) (Haughey et al. 2002).

A primary neuropathological hallmark of AD is the accumulation of Aß peptide in brain (Glenner and Wong 1984), which is derived from the proteolytic processing of β-amyloid precursor protein. Soluble oligomeric assemblies of Aß (Shankar et al. 2008; Ashe and Zahs 2010) and Aß–derived diffusible ligands (ADDLs) (De Felice et al. 2007) have been reported to target synapses, induce neuronal dysfunction and impair memory. Furthermore, the administration of Aß_1–42, the most toxic and pro-aggregation form of Aß (Li et al. 1999; Takeda et al. 2009) into the lateral ventricle of mice has been used to model neuroinflammation, synaptic dysfunction, neuronal death, and memory impairment (Yankner 2000; Hardy and Selkoe 2002; Yamada et al. 2005; Choi and Bosetti 2009), all of which are observed during the progression of AD (Yamada and Nabeshima 2000; Van Dam and De Deyn 2006). It has been suggested that amyloid deposition may also impact processes that regulate neurogenesis (Hardy and Selkoe 2002). In this regard, numerous growth factors that are potent modulators of neural stem cell activity (Neeper et al. 1996; Gomez-Pinilla et al. 1997; Fabel et al. 2003) have been found up-regulated in proximity of amyloid plaques (Tarkowski et al. 2002; Burbach et al. 2004). Furthermore, inflammation accompanying amyloid abnormal production and deposition (Akiyama et al. 2000) may negatively affect neurogenesis (Monje et al. 2003; Russo et al. 2011b; Russo et al. 2011a). The inflammatory response and its propagation have a critical role in the degeneration of hippocampal neurons and the progression of AD (Hoozemans et al. 2008; Choi and Bosetti 2009), and ultimately lead to massive degeneration of pyramidal neurons and cognitive impairment (Haughey et al. 2002).

The pro-inflammatory cytokine TNF-α, which plays a central role in initiating and regulating the cytokine cascade during an inflammatory response (Makhatadze 1998; Sutton et al. 1999), has been found up-regulated in post-mortem brain from AD patients (Perry et al. 2001). However, its role in the modulation of the neurogenic niche remains unclear (McAlpine et al. 2009). In this study, we investigated the activity of a novel thalidomide based TNF-α lowering agent, 3,6′-dithiothalidomide (Scheme 1A). This compound, similar to but more potent than thalidomide (Baratz et al. 2011), readily enters the brain and lowers the rate of synthesis of TNF-α post-transcriptionally via the 3′-untranslated region of TNF-α mRNA (Zhu et al. 2003; Greig et al. 2004; Tweedie et al. 2009). Recently, 3,6′-dithiothalidomide was shown to reverse hippocampus-dependent cognitive deficits in a model of neuroinflammation induced by chronic LPS infusion (Belarbi et al. 2012), and attenuated neuroinflammation, AD pathology and behavioral deficits in animal models of neuroinflammation and AD (Tweedie et al. 2012, in press). The use of 3,6′-dithiothalidomide allowed us to evaluate the role of TNF-α on hippocampal neurogenesis during Aβ_1-42-induced neuroinflammation, in relation to hippocampal progenitor cell proliferation, survival and phenotypic differentiation, as well as actions on memory, and additionally to assess the agent as a potential therapeutic for specific aspects of AD neuropathology.

Scheme 1.

A: Chemical structure of thalidomide and 3,6-dithiothalidomide.

B & C: Animal experimental protocol indicating the time-course for various interventions utilized during the experiment. Drug (56mg/kg)/vehicle was administered daily, for the indicated time (from 14 up to 35 days). Aβ_42-1 or Aβ_1-42 (400 pmol) was administered once at the time point indicated. BrdU (90 mg/kg) was administered daily for 7 consecutive days from day 7 onward of drug administration.

Materials and Methods

Animals: stereotaxic intracerebroventricular Aβ_1-42 and peripheral 3,6′-dithiothalidomide administration

All animal procedures were approved by the National Institutes of Health (NIH) Animal Care and Use Committee in accordance with NIH guidelines on the care and use of laboratory animals. Three-month-old male C57BL/6 mice (purchased from Taconic), were housed at 25 °C in our facility with a 12 h light/dark cycle with free access to food and water. Studies were aimed at examining the effects of Aβ peptide administration on (i) hippocampal progenitor cell proliferation and (ii) progenitor cell survival and generation of newly derived neurons. To assess these features two different experimental time courses were followed.

For the former (i), the experimental duration was for 14 days to allow quantification of Aβ_1-42 -induced hippocampal progenitor cell proliferation and accompanying memory deficits (Scheme 1, B). Mice received drug or vehicle daily for 14 consecutive days. Seven days after the initiation of drug treatment, Aβ peptide (1-42 or 42-1) and a marker for cell proliferation were administered. After additional 7 days (at 14 days after initiation of the study) the animals were subjected to memory assessment (Morris Water Maze), and then euthanized. Immunohistochemical procedures were performed on hippocampal tissues.

For the second (ii): the experimental duration was 5 weeks (35 days: Scheme 1C), allowing for the quantitative evaluation of the survival of and the phenotypic appraisal of Aβ peptide-induced proliferated cells (Scheme 1, C). Mice were treated similarly to those described above, but were administered drug or vehicle daily for 5 weeks. After 5 weeks, subgroups were subjected to memory tasks and then euthanized for immunohistochemical evaluation.

Mice received either i.p. vehicle (1% carboxy methyl cellulose solution (Fluka, Cat # 21901) prepared in sterile saline) or i.p. 3,6′-dithiothalidomide, prepared as a suspension in the vehicle at a dose of 56 mg/kg. At the time of Aβ peptide administration mice were anesthetized with ketamine (100 mg/kg, i.p.) and xylazine (10 mg/kg, i.p.) and positioned in a stereotactic apparatus (Kopf Instruments, Tujunga, CA, USA). Aβ_1-42 and the reverse peptide Aβ_42-1 (American Peptide, Sunnyvale, CA) were reconstituted in phosphate-buffered saline (pH 7.4) and aggregated by incubation at 37 °C for 7 days prior to administration, as described previously (Choi and Bosetti 2009). Aβ_1-42 and Aβ_42-1 (400 pmol) were injected intracerebroventricularly (i.c.v.) into the lateral ventricle using a 10 μl syringe (World Precision Instruments, Sarasota FL) and syringe pump (Stoelting, Wood Dale, IL) at a rate of 1μl/min. Selection of the dose of Aβ_1-42 and Aβ_42-1 was based on previous studies (Yan et al. 2001; Jhoo et al. 2004; Prediger et al. 2007). The coordinates for the stereotaxic infusion were -2.5 mm dorsal/ventral, -1 mm lateral, and -0.5 mm anterior/posterior from the bregma (George Paxinos 2001). A marker for cell proliferation, 5-bromo-2′-deoxyuridine (BrdU, 90 mg/kg, 10 mg/ml in 0.9% saline; Sigma Aldrich, St. Louis, MO) was administered i.p. at the time of Aβ peptide administration, and daily for seven days thereafter. Animals were allowed to recover from surgery and returned to their home cages until the time of their experimental endpoint.

Immunohistochemistry and immunoflourescence

Immunohistochemistry and immunofluorescence triple labeling for BrdU were performed on free-floating 40 μm sagittal sections that were pretreated by denaturing DNA, as described previously (van Praag et al. 1999a). Antibodies used for BrdU immunohistochemistry were mouse anti-BrdU (1:100, DAKO, Denmark), for triple labeling immunofluorescence assessments: rat anti-BrdU (1:200; Accurate Chemical & Scientific, Westbury NY), rabbit anti-S-100β (1:200; Abcam, Cambridge, MA), mouse anti-NeuN (1:1500; Chemicon, Billerica, MA) were used to detect cells with a neuronal phenotype. Mouse anti-synaptophysin (SYN) (1:2500; Millipore Corp - Bioscience Division, Billerica, MA) and goat anti-doublecortin (DCX) (1:200, Santa Cruz, CA) were used to detect morphological development of new neurons. Rabbit anti-CD11b (1:200; Chemicon, Temecula, CA) was used to detect activated microglia cells. For immunohistochemistry, the peroxidase method (ABC system, with biotinylated donkey anti-mouse IgG antibodies and diaminobenzidine as chromogen; Vector laboratories) was used. Images were detected with a light microscope Olympus CKX41, using X40 and X60 objective (Olympus). For immunofluorescence studies the fluorescent antibodies 488-Alexa Fluor anti-mouse IgG (1:500), 594-Alexa Fluor anti-rat IgG (1:500), and 405-Alexa Fluor anti-rabbit IgG (1:500), 488-Alexa Fluor anti-goat IgG (1:200) (Invitrogen, Carlsbad, CA) were used. Fluorescent signals were detected and processed using a Zeiss Axiovert 200 microscope equipped with confocal laser system LSM 510 META. The images were acquired using a 63x oil immersion objective (Zeiss), and were assembled using Adobe Photoshop CS. The images that show the development of new neurons in the hippocampus were obtained by confocal Z-stack 3D acquisition and processed with LSM image examiner software.

Cell counting

BrdU-labeled cells were counted in the subgranular zone (SGZ), which was defined as a two-nucleus-wide band below the apparent border between the granule cell layer (GCL) and the hilus, as previously described (Kempermann et al. 2003). We made determinations in one of every six sections, and covered the entire area of dentate gyrus (DG). The numbers of BrdU-labeled cells in each mouse were calculated by summing the number of labeled nuclei of all slices analyzed. To determine the percentage of neuronal differentiation of newborn cells, one in every six sections were analyzed for the entire area of the DG. For each animal, fifty BrdU-positive cells were randomly selected and analyzed for co-expression of BrdU with NeuN a marker of a neuronal phenotype, or with S100β a marker for a glial phenotype, or neither NeuN nor S100β (van Praag et al. 1999b). The number of activated microglia per section was quantified by counting the number of CD11b-stained cells within 0.3 mm² area of the hippocampus. For each measurement, two blinded independent investigators counted 3-4 brains per group, 3 sections per brain.

Morris Water Maze Assessment

Spatial learning and memory were assessed using the Morris Water Maze. Briefly, the experimental apparatus consisted of a circular plastic pool (diameter, 97 cm; height, 60 cm) filled with water (23 ± 2°C) that was colored with nontoxic white paint to obscure the location of a submerged platform. A target platform (10 × 10 cm) was submerged 1 cm below the water surface and placed at the midpoint of one quadrant. This platform was located in a fixed position, equidistant from the center and the wall of the pool, and was kept constant for each mouse throughout training. Visual cues were placed around the tank to orient the mice. The water maze test was performed in two sets of animals one and four weeks following Aβ_1-42 administration. The acquisition training session was performed 4 days before the test session and consisted of 5 trials for 4 days, during which the animals were allowed to search for the platform for 60 sec. In the event that a mouse failed to locate the platform in the allotted time, it was manually guided to the location. Mice were allowed to remain on the platform for 30 sec for the first day, and then returned to their home cages. On subsequent days, the animals were allowed to remain on the platform for 15 sec.

The test session was performed 24 h after the training sessions, on the last day prior to euthanizing the animals. In the test session, the platform was removed from the pool and each mouse was allowed to swim for 60 sec. This test session also consisted of 5 trials, in which the average time spent in the correct quadrant, where the platform was located on the training session, was calculated.

Gene expression

Fresh frozen mouse hippocampus were processed for RNA extraction using the Qiagen RNeasy Lipid Tissue Mini kit (Qiagen, Valencia, CA), following the manufacturer’s instructions. Retro-Transcription (RT) was done using the Moloney murine leukemia virus-reverse transcriptase (MMLV-RT) (Invitrogen). 2.5 mg of total RNA were mixed with 2.2 μl of 0.2 ng/μl random hexamer (Invitrogen), 10 μl of 56 buffer (Invitrogen), 10 μl of 2 mM dNTPs, 1 μl of 1 mM DTT (Invitrogen), 0.4 μl of 33 U/μl RNasin (Promega), 2 μl MMLV-RT (200 U/ μl), in a final volume of 50 μl. The reaction mix was incubated at 37 °C for 2 h and then the enzyme heat inactivated at 95 °C for 10 min. Quantitative real-time polymerase chain reaction (RT-PCR) was performed on TNF-α (TNF-α: Mm00443258_m1), Tumor necrosis factor receptor I (Tnfrsf1b: Mm00441889_m1); Tumor necrosis factor receptor II (Tnfrsf1a: Mm00441883) and phosphoglicerate kinase 1 (pgk1: Mm01225301_m1). PCR reactions were performed using the Applied Biosystems 7500 system Real-time PCR system (Foster City, CA, USA) with Taqman probes following the manufacturer’s instructions. Data were analyzed using the comparative threshold cycle (DD Ct) method (Livak and Schmittgen 2001). Results were normalized with Pgk1 as the endogenous control, and expressed as fold difference from the Aβ_42-1-injected mice, as previously reported (Toscano et al. 2007).

Statistical analysis

All data are expressed as means ± SEM. Statistical significance was assessed with a one-way ANOVA followed by a Bonferroni’s test using GraphPad Prism. Significance was taken at p < 0.05.

Results

In all experiments, we also examined the group of mice injected with the control Aβ_42-1 peptide and treated with 3,6′-thiothalidomide.When we compared Aβ_42-1-injected mice treated with 3,6′-thiothalidomide and Aβ_42-1-injected mice treated with vehicle, there were no significant changes in microglia activation, numbers of BrdU-proliferative cells and BrdU-cells that differentiated in neurons, memory impairment and memory recovery (data not shown). Since 3,6′-dithiothalidomide treatment alone did not cause any effects in control mice, this group was not included in the graphs.

3,6′-dithiothalidomide treatment decreases microglia activation in response to Aβ_1-42 – induced neuroinflammation

Mice were treated with 3,6′-dithiothalidomide or vehicle for 14 days, 7 days before and 7 days after injection of Aβ_1-42 or Aβ_42-1. Aβ_1-42 injection caused a robust inflammatory response, characterized by the presence of activated microglia cells within hippocampus in mice injected with vehicle (Fig. 1a, b). Intense CD11b-immunoreactive microglia with enhanced staining intensity, retracted processes, and amoeboid appearance were observed in the hippocampus of these mice (Fig. 1a). In the hippocampus of Aβ_1-42-injected mice treated with 3,6′-dithiothalidomide and in Aβ_42-1-injected mice, only a few CD11b-immunoreactive microglia cells were observed (Fig. 1a, b), which retained a resting morphology with small cell bodies, thin, and ramified processes.

Figure 1.

a. Representative photomicrographs of CD11b immunoreactivity in the hippocampus subfield A) Aβ_42-1 vehicle-injected mice (control), B) Vehicle-treated Aβ_1-42 injected-mice, C) 3,6′-dithiothalidomide-treated Aβ_1-42 -injected mice. b. Quantification of CD11b-positive cells in the hippocampus. Mean ± SEM (n = 4 per group). ***p < 0.001 compared with vehicle-treated - Aβ_42-1 injected mice; ^###p < 0.001 compared with 3,6′-dithiothalidomide-treated Aβ_1-42 -injected mice. Scale bars = 100 μm.

3,6′-dithiothalidomide treatment attenuates the effect of Aβ_1-42 injection on hippocampal progenitor cell proliferation

To analyze the effects of treatment with 3,6′-dithiothalidomide on Aβ_1-42 inflammation-induced proliferation of hippocampal progenitor cells, mice were treated for a total of 14 days with 3,6′-dithiothalidomide, 7 days after initiation of drug treatment the appropriate Aβ peptides were administered. Seven days later, a subset of animals was euthanized and the numbers of BrdU-labeled cells were counted in the DG of hippocampus (Scheme 1 B, Fig. 2a). Aβ_1-42-induced brain inflammation caused an increase (64%) in the number of BrdU positive proliferating cells, as compared with mice injected with the control peptide (Aβ_42-1). Treatment with 3,6′-dithiothalidomide substantially attenuated this effect. Comparing the numbers of BrdU-cells from the drug treated animals with those observed in vehicle treated animals there was a 36% reduction in proliferated cells counts induced by Aβ_1-42 insult (427.6 ± 32.59 for 3,6′-dithiothalidomide treated animals vs. 547.6 ± 29.91 for non drug treated animals, p < 0.01, Fig. 2b). The distribution of the BrdU-labeled proliferative cells was found mainly in clusters at the border between the granule cell layer and the hilus of the DG, with no differences in the distributions being evident among the different treatments.

Figure 2.

a. Effects of 3,6′-dithiothalidomide treatment on hippocampal BrdU proliferative cells after Aβ_1-42 administration are presented. Representative photomicrographs of BrdU immunohistochemistry in the DG, 7 days after i.c.v. injection of either the control peptide Aβ_42-1 (A), or the neurotoxic peptide Aβ_1-42 (B, C) are shown. (D) Inset showing high-magnification images of BrdU immunostaining cells. Scale bars = 100 μm (A, B, C); 50 μm (D). b. A quantitative assessment of BrdU-labeled cells as described above, are shown. Data are mean ± SEM (n=5). Data were analyzed using a 1-way ANOVA followed by Bonferroni’s post-hoc test. *p < 0.05; ***p < 0.001 vs. Aβ_42-1 injected mice; *p < 0.05 vs. Aβ_42-1 injected mice; ^##p < 0.01 vs. β_1-42 injected mice treated with vehicle.

3,6′-dithiothalidomide treatment represses the effect of Aβ_1-42 –induced inflammation on BrdU-cells survival at 4 weeks

The effects of 3,6′-dithiothalidomide on the survival of BrdU cells induced by Aβ_1-42 mediated neuroinflammation were assessed (Scheme 1, C). We analyzed the number of BrdU-cells in the granule cell layer and subgranular zone of the DG of mice treated with vehicle or 3,6′-dithiothalidomide 4 weeks after either Aβ_1-42 or control peptide administration. Aβ_1-42 significantly increased the number of BrdU cells survived compared with the control peptide group (128 ± 1.69 for Aβ_1-42 versus 91 ± 2.14 for Aβ_42-1, a 40% increase in cell number, p < 0.001). In mice injected with Aβ_1-42, the treatment with 3,6′-dithiothalidomide repressed and normalized the number of surviving BrdU-cells to control peptide levels (91 ± 2.14 for Aβ_42-1 vs. 93.8 ± 1.7 for drug + Aβ_1-42; Fig. 3).

Figure 3.

Effects of 3,6′-dithiothalidomide treatment on hippocampal derived BrdU-cell survival at 4 weeks after peptide administration are shown. Presented are quantified data for BrdU-cells in mice administered with the neurotoxic peptide Aβ_1-42 treated with vehicle or 3,6′-dithiothalidomide and mice administered with control Aβ_42-1 peptide. Data are mean ± SEM (n=5). Data were analyzed using a 1-way ANOVA followed by Bonferroni’s post-hoc test. ***p < 0.001 vs. Aβ_42-1 injected mice; ^###p<0.001 vs. Aβ_1-42 injected mice treated with vehicle.

Aβ_1-42 –induced increase in neurogenesis was attenuated by 3,6′-dithiothalidomide treatment

To determine the fate of hippocampal derived proliferating cells, we examined the phenotype of newly generated BrdU-cells 4 weeks after Aβ_1-42-induced inflammation by co-immunolabeling for BrdU along with neuronal (NeuN) and glial (S100β) markers (Fig. 4a). The percentage of BrdU positive cells that co-labeled for NeuN was significantly increased in Aβ_1-42-injected mice compared with control mice injected with the reverse peptide (p < 0.001, Fig.4b, Table 1). In animals treated with 3,6′-dithiothalidomide and Aβ_1-42, there was a significant elevation in neurons when compared to control animals, which was less than that induced by Aβ_1-42 plus vehicle (p < 0.01, Fig. 4b, Table 1). Aβ_42-1-injected mice showed a significantly higher fraction of BrdU labeled cells that proved neither neuronal nor glial. Indicating that these cells either remained undifferentiated or had differentiated into cells with an alternative phenotype. This was not the case for Aβ_1-42 injected mice treated with either vehicle or 3,6′-dithiothalidomide (p < 0.001; Table 1). Interestingly, there was no significant difference between the groups with regard to percentage of newborn cells differentiated into glia (Table 1).

Figure 4.

a. Effects of 3,6′-dithiothalidomide treatment on the differentiation of hippocampal neuronal progenitor cells, 4 weeks after Aβ administration are presented. Representative confocal images of the DG of mice administered with the control peptide Aβ_42-1 (A), with the neurotoxic peptide Aβ_1-42 treated with vehicle (B) or 3,6′-dithiothalidomide (C) are shown. Hippocampal sections were triple-labeled for BrdU (red), NeuN indicating neuronal phenotype (green), and S100β selective for glial phenotype (blue), assessments were made with immunofluorescence microscopy, scale bars = 20 μm. b. The quantification of the percentage of BrdU-cells that differentiated into neurons is described. Data are mean ± SEM (n=5). Data were analyzed using a 1-way ANOVA followed by Bonferroni’s post-hoc test. p < 0.05; ***p < 0.001 vs. Aβ_42-1 injected mice; **p < 0.01 vs. Aβ_42-1 injected mice; ^##p<0.01 vs. Aβ_1-42 injected mice treated with vehicle.

Table 1. Phenotype of BrdU-positive cells.

Phenotype of surviving cells was determined by immunofluorescent triple labeling for BrdU, NeuN (neurons), S100β (glia). The percentage of BrdU-positive cells double labeled for either NeuN or s100β or neither marker is presented. All data are means ± S.E.M. and were analyzed using 1-way ANOVA followed by a Bonferroni’s post-hoc test).

	Aβ42-1 + Vehicle	Aβ1-42 + Vehicle	Aβ1-42 + 3,6′-Dithiothalidomide
Neuron (%)	70.69 (2.16)	96.01 (3.13)***	85.40 (2.54)** ##
Glia (%)	9.73 (0.55)	3.99 (2)	9.74 (2.32)
Other (%)	19.58 (2.11)	0 (0)***	4.86 (1.6)*** #

^*p < 0.05;

^***p < 0.001;

^**p < 0.01 vs. control group (Aβ_42-1 + Vehicle) .

^#p < 0.05,

^##p < 0.01 compared with Aβ_1-42 + Vehicle.

New hippocampal neurons show normal morphological development 4 weeks after Aβ_1-42 injection

To determine whether Aβ_1-42-induced newly generated granule neurons develop normal morphology we performed doublecortin (DCX)-synaptophysin (SYN) double staining immunofluorescence, as previously described (van Praag et al. 2002; Tronel et al. 2010). SYN (red), which is expressed by synaptic vesicle membranes, was detected around the newly generated DCX-positive (green) granule cells (Fig. 5), suggesting that the new immature neurons are physically integrated into the surrounding hippocampal tissue. Co-labeled cells observed in the DG region showed no morphological differences irrespective of treatment group (Fig. 5).

Figure 5.

Representative images of morphological development of new neurons are presented at 4 weeks after administration of either Aβ peptide. The degree of neuronal integration appears to be similar between the groups. Confocal images of the DG of mice that receive i.c.v. control Aβ_42-1 (A) and the neurotoxic Aβ_1-42 peptide with vehicle (B) or 3,6′-dithiothalidomide (C) are shown. Sections were double-labeled for SYN (red) and DCX (green). Images were obtained with immunofluorescence microscopy, scale bar = 10 μm.

Aβ_1-42-induced memory deficits are abolished by 3,6′-dithiothalidomide treatment

To explore the effects of 3,6′-dithiothalidomide on memory tasks after Aβ_1-42 inflammation, mice were subjected to Morris Water Maze assessment at 7 days or 4 weeks after Aβ administration. Training sessions were performed 4 days before the test session, and consisted of 5 trials for 3 days. The escape latency decreased over the course of the acquisition training in all mice analyzed 7 days after Aβ injection (Fig. 6a). The time to find the platform was significantly increased in mice injected with Aβ_1-42 peptide treated with vehicle compared to mice injected with the reverse peptide (p < 0.01; Fig. 6a). Following 3 days of training, mice were tested using an average of 5 trials. Aβ_1-42-induced inflammation significantly impaired memory acquisition compared with control Aβ_42-1-injected mice (17.09 ± 3.54 vs. 5.48 ± 0.67; p < 0.01; Fig. 6b). Treatment with 3,6′-dithiothalidomide prevented memory deficits compared to the respective vehicle-treated mice (7.09 ± 1.05 vs. 17.09 ± 3.54; p < 0.05; Fig. 6b). Four weeks after Aβ injection mice had fully recovered their memory function and there were no differences between the groups (data not shown).

Figure 6.

Effect of Aβ_1-42 administration and Aβ_1-42 plus 3,6′-dithiothalidomide treatment on mouse memory performance in the Morris Water Maze test, 7 days after Aβ peptide injection are shown. a. The escape latency during the training and memory assessment for mice that received i.c.v. control peptide and the neurotoxic Aβ_1-42 peptide treated with vehicle or 3,6′-dithiothalidomide are presented. b. Quantified probe trial data of mice that received i.c.v. control peptide, or Aβ_1-42 treated with vehicle or 3,6′-dithiothalidomide 7 days after Aβ peptide administration are shown. Data are mean ± SEM (n=5). Data were analyzed using a 1-way ANOVA followed by Bonferroni’s post-hoc test. Significance was set at p < 0.05. **p < 0.01 vs. Aβ_42-1 peptide-injected mice; ^#p < 0.05 vs. Aβ_1-42 peptide-injected mice.

Discussion

Recent studies have shown that neuroinflammation-associated changes in the brain such as microglia activation, induction of NF-κB transcription factor, and release of inflammatory mediators, including TNF-α, can contribute to impaired neurogenesis (Monje et al. 2003; Russo et al. 2011b; Russo et al. 2011a). In this study, we investigated whether TNF-α is involved in the neurogenic signaling of hippocampal progenitor cells following Aβ_1-42-induced inflammation (Akiyama et al. 2000; Choi and Bosetti 2009; Ferretti and Cuello 2011) by using the potent TNF-α synthesis inhibitor, 3,6′-dithiothalidomide, a small lipophilic experimental drug that readily enters the brain (Tweedie et al. 2007; Tweedie et al. 2009). We report that the treatment with 3,6′-dithiothalidomide normalizes microglia activation, proliferation and survival of progenitor cells, differentiation of new neurons, and cognitive functions to the control conditions, thus repressing the effects of Aβ_1-42 on neuronal tissue and on the neurogenic hippocampal niche.

TNF-α participates in the regulation of turnover of neural stem/progenitor cells under physiological conditions (Iosif et al. 2006), and also has been implicated in the pathogenesis of certain neurodegenerative and neurological disorders like AD, Parkinson’s disease, stroke and head trauma (Hallenbeck 2002; Hirsch et al. 2003; Li et al. 2004). TNF-α regulates several cellular processes, including inflammation, cell differentiation, cell death and survival through activation of two TNF receptors: TNFR1 or TNFR2 (Wajant et al. 2003). These receptors have been proposed to facilitate distinct TNF-α-mediated effects in the brain: TNFR1, with its intracellular “death domain”, contributes to neuronal death and damage, and primarily responds to soluble TNF-α (Grell et al. 1998). In contrast, TNFR2 is associated with cell survival and neuroprotection (Fontaine et al. 2002; Yang et al. 2002; Wajant et al. 2003; Marchetti et al. 2004), and is primarily activated via membrane-bound TNF-α (Grell et al. 1995). The pro-survival or pro-death roles of TNF-α likely depend on which TNF receptor is activated. Under pathological conditions, soluble and membrane-bound TNF-α levels may be differentially elevated and diverse types of neurons may have different expression ratios of the two TNF-α receptors (Yang et al. 2002). Furthermore, the net effect of TNF-α is dependent on several factors, such as the site, degree, and duration of neuroinflammation, the level and cellular origin of the cytokine, the type of target cells, the expression level of the two receptors and their affinity to TNF-α (Fontaine et al. 2002; Hallenbeck 2002; Heldmann et al. 2005). Due to the diverse bioactivities of TNF-α, it is unclear under which conditions TNF-α promotes beneficial or deleterious effects on neuronal tissue and on the hippocampal neurogenic niche.

The cellular synthesis of TNF-α, similar to other inflammatory cytokines, is closely regulated at the post-transcriptional level of mRNA stability through its 3′-untranslated region (3′-UTR) (Moreira et al. 1993), which allows for rapid alterations in TNF-α production in response to exogenous and endogenous-induced changes in the brain microenvironment. The presence of adenylate-uridylate-rich elements (AREs) within the 3-UTR of TNF-α mRNA permits post-transcriptional repression that, through interaction with specific RNA-binding proteins, can either promote its stability and thereby increase its synthesis or, alternatively, target it for rapid degradation or inhibition of translation, to lower its rate of generation (Patil et al. 2008; Khera et al. 2010). In this regard, thalidomide has been shown to increase translational blockade and thereby, reduce the rate of TNF-α protein synthesis (Sampaio et al. 1991; Moreira et al. 1993). 3,6′-dithiothalidomide, like thalidomide, regulates TNF-α mRNA stability via its 3′-UTR (Zhu et al. 2003), but with the isosteric replacement of carbonyl groups by thiocarbonyls has incremental increases in TNF-α inhibitory activity of up to 30-fold, without toxicity (Tweedie et al. 2009). A recent study showed how, using the same concentration, 3,6′-dithiothalidomide but not thalidomide lowered TNF-α protein level following LPS-challenged RAW 264.7 cells (Tweedie et al. 2009). Thus, the potent action of 3,6′-dithiothalidomide allows the administration of lower and better tolerated doses during anti-inflammatory treatment (Baratz et al. 2011; Belarbi et al. 2012; Tweedie et al. 2012, in press).

Neuronal progenitor cells express both TNFR1 and TNFR2, suggesting that TNF-α can modulate the formation of new hippocampal neurons under pathological conditions by acting on these two receptors with differential effects on proliferation and survival (Iosif et al. 2006). TNF-α is also released by neuronal progenitor cells (Heldmann et al. 2005), and can act directly on the neuronal progenitor cells via either autocrine mechanisms or by release from neighboring microglia cells (Heldmann et al. 2005).

TNF-α, mainly via TNFR2 signaling, has been shown to promote neurogenesis in various models of injury, such as ischemia, demyelination and status epilepticus. Specifically, immunotherapy with an antibody to TNF-α attenuates the increase in neurogenesis after middle cerebral artery occlusion (Heldmann et al., 2005). Furthermore, studies in TNFR2 knock-out mice subjected to cuprizone-induced demyelination indicated that TNF-α via TNFR2 is critical to oligodendrocyte regeneration by promoting proliferation of oligodendrocyte progenitors (Arnett et al. 2001). TNFR2 knock-out mice show reduced neurogenesis also after status epilepticus, which is associated with inflammation and elevated TNF-α level (Iosif et al. 2006). In this study, we showed that following Aβ_1-42 injection and 3,6′-dithiothalidomide treatment TNF receptors mRNA are not modulated (S2), although TNF-α mRNA level is drastically up-regulated also after drug treatment (S1). Since 3,6′-dithiothalidomide inhibits TNF-α protein synthesis (Tweedie et al. 2009; Baratz et al. 2011; Belarbi et al. 2012; Tweedie et al. 2012, in press) by a mechanism related to the destabilizing of its mRNA (Greig et al. 2004); one probable reason is that microglial cells respond with a robust increase in transcription of TNF-α mRNA to compensate for any reduction in TNF-α protein synthesis.

Several studies reported that an increase in neurogenesis is associated with cognitive improvement (Stone et al. 2011; Marlatt et al. 2012); conversely, impairment in neurogenesis is associated with cognitive deficits (Kim et al. 2011; Mishra et al. 2012). Taking into account the effect of inflammation on neurogenesis and memory tasks, in this study we showed that Aβ_1-42 induced neuroinflammation is accompanied by enhanced neurogenesis and cognitive impairment.

Altered neurogenesis has been shown in various AD transgenic mouse models, with reporting of both decreased and increased neurogenesis (Lazarov and Marr 2010; Marlatt and Lucassen 2010). Possible explanations for these apparently conflicting data could be ascribed to neurodegenerative stage-dependent effects as well as strain, genetic background and specific transgene mutations, which could differentially affect the different stages of proliferation, survival and differentiation of newly formed cells (Verret et al. 2007). In this regard, several studies have demonstrated that hippocampal neurogenesis in vivo is significantly increased at early stages of neurodegeneration, and although neurogenesis and the differentiation of newly generated neurons into a mature phenotype are still significantly increased at late stages of neurodegeneration, these neurons are impaired in comparison with the early stage neurons (Jin et al. 2004b). In addition, it has been reported that Aβ levels and/or conformation may affect neurogenesis, for instance, oligomeric Aβ₄₂ has been shown to enhance neuronal differentiation of embryonic and postnatal NSC in vitro (Lopez-Toledano and Shelanski 2004). Furthermore, studies from human postmortem brain of AD patients showed increased hippocampal neurogenesis (Jin et al., 2004). Our data suggest that increased neurogenesis occurs perhaps as an endogenous compensatory response to neuroinflammation and at early neurodegenerative events following Aβ_1-42 injection, where these events are likely mediated by TNFR2 signaling actions (Heldmann et al. 2005; Iosif et al. 2006).

In this study, we found that treatment with 3,6′-dithiothalidomide attenuates the memory disruption, and at the same time attenuates the compensatory response and the number of new neurons generated by hippocampal neurogenic niche after Aβ_1-42 injection (Scheme 2). Thus, it is possible that TNF-α plays a dual role and is involved in mediating at the same time neurodegeneration (memory disruption) and regeneration (increase of neurogenesis), through TNFR1 and TNFR2 signaling respectively, and that there is a cooperative action of both receptors during an inflammatory response (Fontaine et al. 2002). TNF-α also has a homeostatic role in limiting and controlling the extent and duration of an inflammatory response (Hallenbeck 2002; Heldmann et al. 2005), probably through modification of the TNF-α core signaling framework and by generating feedback responses that suppress inflammation (Hallenbeck 2002).

Scheme 2.

Mechanism linking hippocampal neurogenesis and memory tasks in response to Aβ_1-42-induced neuroinflammation and after 3,6′-dithiothalidomide treatment. 3,6′-dithiothalidomide treatment leads to a reduction of TNF-α protein level and consequently, attenuates the neuroinflammatory response and neuronal damage after Aβ_1-42 injection. The hippocampal neurogenic niche, in turn, responds with an attenuated compensatory increase in nurogenesis.

Understanding which factors regulate hippocampal neurogenesis and its effects on cognition and memory could have important therapeutic implications to help gain insight on the cellular mechanisms underlying AD and to development of new therapeutic strategies. A recent study from our groups showed that chronic 3,6′-dithiothalidomide administration to an elderly symptomatic cohort of 3×Tg AD mice reduced multiple hallmark features of AD (Tweedie et al. 2012, in press). The data presented in this manuscript indicate that while treatment with 3,6′-dithiothalidomide attenuates neurogenesis caused by Aβ_1-42–induced inflammation, importantly it prevents the subsequent memory impairments most likely due to reduced neuronal cell damage and an attenuated neuroinflammatory response related to lower TNF-α protein levels (Scheme 2), suggesting that the anti-inflammatory effect is a critical component in the improvement of memory function. Here, we can hypothesize that the new neurons generated in a brain with less neuroinflammation could have a higher probability to integrate and synaptically participate to the hippocampal circuitry, since neuroinflammation and its propagation limit the integration of new neurons (Jin et al. 2004a).

In conclusion, our data indicate that TNF-α synthesis inhibition by 3,6′-dithiothalidomide treatment prevents memory impairments induced by a Aβ_1-42-neuroinflammation and suggest that this drug should be further investigated as a therapeutic in AD as well as other animal models of neurodegenerative disease with an inflammatory component.

Abbreviations

SVZ

subventricular zone

SGL

subgranular zone

DG

dentate gyrus

GCL

granule cell layer

AD

Alzheimer’s disease

Aß

β-amyloid

ADDLs

Aß–derived diffusible ligands

TNF-α

tumor necrosis factor-α

i.c.v.

intracerebroventricular injection

BrdU

5-bromo-2′-deoxyuridine

DCX

doublecortin

SYN

synaptophysin

NF-κB

nuclear factor-kappaB

TNFR

tumor necrosis factor receptor

3′-UTR

3′-untranslated region

AREs

adenylate-uridylate-rich elements

NSC

neuronal stem cells

TGF-β

transforming growth factor-β

BDNF

brain-derived neurotrophic factor

NGF

nerve growth factor

FGF

fibroblast growth factor